diff options
Diffstat (limited to 'contrib/llvm/tools/llvm-objdump')
| -rw-r--r-- | contrib/llvm/tools/llvm-objdump/COFFDump.cpp | 8 | ||||
| -rw-r--r-- | contrib/llvm/tools/llvm-objdump/MachODump.cpp | 4587 | ||||
| -rw-r--r-- | contrib/llvm/tools/llvm-objdump/llvm-objdump.cpp | 616 | ||||
| -rw-r--r-- | contrib/llvm/tools/llvm-objdump/llvm-objdump.h | 24 |
4 files changed, 4807 insertions, 428 deletions
diff --git a/contrib/llvm/tools/llvm-objdump/COFFDump.cpp b/contrib/llvm/tools/llvm-objdump/COFFDump.cpp index 4a20b91..976a921 100644 --- a/contrib/llvm/tools/llvm-objdump/COFFDump.cpp +++ b/contrib/llvm/tools/llvm-objdump/COFFDump.cpp @@ -167,7 +167,7 @@ resolveSectionAndAddress(const COFFObjectFile *Obj, const SymbolRef &Sym, if (std::error_code EC = Sym.getSection(iter)) return EC; ResolvedSection = Obj->getCOFFSection(*iter); - return object_error::success; + return std::error_code(); } // Given a vector of relocations for a section and an offset into this section @@ -182,7 +182,7 @@ static std::error_code resolveSymbol(const std::vector<RelocationRef> &Rels, return EC; if (Ofs == Offset) { Sym = *I->getSymbol(); - return object_error::success; + return std::error_code(); } } return object_error::parse_failed; @@ -204,7 +204,7 @@ getSectionContents(const COFFObjectFile *Obj, return EC; if (std::error_code EC = Obj->getSectionContents(Section, Contents)) return EC; - return object_error::success; + return std::error_code(); } // Given a vector of relocations for a section and an offset into this section @@ -217,7 +217,7 @@ static std::error_code resolveSymbolName(const std::vector<RelocationRef> &Rels, return EC; if (std::error_code EC = Sym.getName(Name)) return EC; - return object_error::success; + return std::error_code(); } static void printCOFFSymbolAddress(llvm::raw_ostream &Out, diff --git a/contrib/llvm/tools/llvm-objdump/MachODump.cpp b/contrib/llvm/tools/llvm-objdump/MachODump.cpp index 03fad5f..bf7451e 100644 --- a/contrib/llvm/tools/llvm-objdump/MachODump.cpp +++ b/contrib/llvm/tools/llvm-objdump/MachODump.cpp @@ -18,6 +18,7 @@ #include "llvm/ADT/Triple.h" #include "llvm/Config/config.h" #include "llvm/DebugInfo/DIContext.h" +#include "llvm/DebugInfo/DWARF/DWARFContext.h" #include "llvm/MC/MCAsmInfo.h" #include "llvm/MC/MCContext.h" #include "llvm/MC/MCDisassembler.h" @@ -36,6 +37,7 @@ #include "llvm/Support/Format.h" #include "llvm/Support/FormattedStream.h" #include "llvm/Support/GraphWriter.h" +#include "llvm/Support/LEB128.h" #include "llvm/Support/MachO.h" #include "llvm/Support/MemoryBuffer.h" #include "llvm/Support/TargetRegistry.h" @@ -62,13 +64,79 @@ static cl::opt<std::string> DSYMFile("dsym", static cl::opt<bool> FullLeadingAddr("full-leading-addr", cl::desc("Print full leading address")); -static cl::opt<bool> - PrintImmHex("print-imm-hex", - cl::desc("Use hex format for immediate values")); +static cl::opt<bool> NoLeadingAddr("no-leading-addr", + cl::desc("Print no leading address")); + +cl::opt<bool> llvm::UniversalHeaders("universal-headers", + cl::desc("Print Mach-O universal headers " + "(requires -macho)")); + +cl::opt<bool> + llvm::ArchiveHeaders("archive-headers", + cl::desc("Print archive headers for Mach-O archives " + "(requires -macho)")); + +cl::opt<bool> + ArchiveMemberOffsets("archive-member-offsets", + cl::desc("Print the offset to each archive member for " + "Mach-O archives (requires -macho and " + "-archive-headers)")); + +cl::opt<bool> + llvm::IndirectSymbols("indirect-symbols", + cl::desc("Print indirect symbol table for Mach-O " + "objects (requires -macho)")); + +cl::opt<bool> + llvm::DataInCode("data-in-code", + cl::desc("Print the data in code table for Mach-O objects " + "(requires -macho)")); + +cl::opt<bool> + llvm::LinkOptHints("link-opt-hints", + cl::desc("Print the linker optimization hints for " + "Mach-O objects (requires -macho)")); + +cl::list<std::string> + llvm::DumpSections("section", + cl::desc("Prints the specified segment,section for " + "Mach-O objects (requires -macho)")); + +cl::opt<bool> llvm::Raw("raw", + cl::desc("Have -section dump the raw binary contents")); + +cl::opt<bool> + llvm::InfoPlist("info-plist", + cl::desc("Print the info plist section as strings for " + "Mach-O objects (requires -macho)")); + +cl::opt<bool> + llvm::DylibsUsed("dylibs-used", + cl::desc("Print the shared libraries used for linked " + "Mach-O files (requires -macho)")); + +cl::opt<bool> + llvm::DylibId("dylib-id", + cl::desc("Print the shared library's id for the dylib Mach-O " + "file (requires -macho)")); + +cl::opt<bool> + llvm::NonVerbose("non-verbose", + cl::desc("Print the info for Mach-O objects in " + "non-verbose or numeric form (requires -macho)")); cl::opt<bool> - llvm::UniversalHeaders("universal-headers", - cl::desc("Print Mach-O universal headers")); + llvm::ObjcMetaData("objc-meta-data", + cl::desc("Print the Objective-C runtime meta data for " + "Mach-O files (requires -macho)")); + +cl::opt<std::string> llvm::DisSymName( + "dis-symname", + cl::desc("disassemble just this symbol's instructions (requires -macho")); + +static cl::opt<bool> NoSymbolicOperands( + "no-symbolic-operands", + cl::desc("do not symbolic operands when disassembling (requires -macho)")); static cl::list<std::string> ArchFlags("arch", cl::desc("architecture(s) from a Mach-O file to dump"), @@ -147,7 +215,7 @@ static bool compareDiceTableEntries(const DiceTableEntry &i, return j.first >= i.first && j.first < i.first + Length; } -static uint64_t DumpDataInCode(const char *bytes, uint64_t Length, +static uint64_t DumpDataInCode(const uint8_t *bytes, uint64_t Length, unsigned short Kind) { uint32_t Value, Size = 1; @@ -156,19 +224,19 @@ static uint64_t DumpDataInCode(const char *bytes, uint64_t Length, case MachO::DICE_KIND_DATA: if (Length >= 4) { if (!NoShowRawInsn) - DumpBytes(StringRef(bytes, 4)); + dumpBytes(ArrayRef<uint8_t>(bytes, 4), outs()); Value = bytes[3] << 24 | bytes[2] << 16 | bytes[1] << 8 | bytes[0]; outs() << "\t.long " << Value; Size = 4; } else if (Length >= 2) { if (!NoShowRawInsn) - DumpBytes(StringRef(bytes, 2)); + dumpBytes(ArrayRef<uint8_t>(bytes, 2), outs()); Value = bytes[1] << 8 | bytes[0]; outs() << "\t.short " << Value; Size = 2; } else { if (!NoShowRawInsn) - DumpBytes(StringRef(bytes, 2)); + dumpBytes(ArrayRef<uint8_t>(bytes, 2), outs()); Value = bytes[0]; outs() << "\t.byte " << Value; Size = 1; @@ -180,14 +248,14 @@ static uint64_t DumpDataInCode(const char *bytes, uint64_t Length, break; case MachO::DICE_KIND_JUMP_TABLE8: if (!NoShowRawInsn) - DumpBytes(StringRef(bytes, 1)); + dumpBytes(ArrayRef<uint8_t>(bytes, 1), outs()); Value = bytes[0]; outs() << "\t.byte " << format("%3u", Value) << "\t@ KIND_JUMP_TABLE8\n"; Size = 1; break; case MachO::DICE_KIND_JUMP_TABLE16: if (!NoShowRawInsn) - DumpBytes(StringRef(bytes, 2)); + dumpBytes(ArrayRef<uint8_t>(bytes, 2), outs()); Value = bytes[1] << 8 | bytes[0]; outs() << "\t.short " << format("%5u", Value & 0xffff) << "\t@ KIND_JUMP_TABLE16\n"; @@ -196,7 +264,7 @@ static uint64_t DumpDataInCode(const char *bytes, uint64_t Length, case MachO::DICE_KIND_JUMP_TABLE32: case MachO::DICE_KIND_ABS_JUMP_TABLE32: if (!NoShowRawInsn) - DumpBytes(StringRef(bytes, 4)); + dumpBytes(ArrayRef<uint8_t>(bytes, 4), outs()); Value = bytes[3] << 24 | bytes[2] << 16 | bytes[1] << 8 | bytes[0]; outs() << "\t.long " << Value; if (Kind == MachO::DICE_KIND_JUMP_TABLE32) @@ -209,8 +277,7 @@ static uint64_t DumpDataInCode(const char *bytes, uint64_t Length, return Size; } -static void getSectionsAndSymbols(const MachO::mach_header Header, - MachOObjectFile *MachOObj, +static void getSectionsAndSymbols(MachOObjectFile *MachOObj, std::vector<SectionRef> &Sections, std::vector<SymbolRef> &Symbols, SmallVectorImpl<uint64_t> &FoundFns, @@ -228,10 +295,8 @@ static void getSectionsAndSymbols(const MachO::mach_header Header, Sections.push_back(Section); } - MachOObjectFile::LoadCommandInfo Command = - MachOObj->getFirstLoadCommandInfo(); bool BaseSegmentAddressSet = false; - for (unsigned i = 0;; ++i) { + for (const auto &Command : MachOObj->load_commands()) { if (Command.C.cmd == MachO::LC_FUNCTION_STARTS) { // We found a function starts segment, parse the addresses for later // consumption. @@ -247,11 +312,831 @@ static void getSectionsAndSymbols(const MachO::mach_header Header, BaseSegmentAddress = SLC.vmaddr; } } + } +} + +static void PrintIndirectSymbolTable(MachOObjectFile *O, bool verbose, + uint32_t n, uint32_t count, + uint32_t stride, uint64_t addr) { + MachO::dysymtab_command Dysymtab = O->getDysymtabLoadCommand(); + uint32_t nindirectsyms = Dysymtab.nindirectsyms; + if (n > nindirectsyms) + outs() << " (entries start past the end of the indirect symbol " + "table) (reserved1 field greater than the table size)"; + else if (n + count > nindirectsyms) + outs() << " (entries extends past the end of the indirect symbol " + "table)"; + outs() << "\n"; + uint32_t cputype = O->getHeader().cputype; + if (cputype & MachO::CPU_ARCH_ABI64) + outs() << "address index"; + else + outs() << "address index"; + if (verbose) + outs() << " name\n"; + else + outs() << "\n"; + for (uint32_t j = 0; j < count && n + j < nindirectsyms; j++) { + if (cputype & MachO::CPU_ARCH_ABI64) + outs() << format("0x%016" PRIx64, addr + j * stride) << " "; + else + outs() << format("0x%08" PRIx32, addr + j * stride) << " "; + MachO::dysymtab_command Dysymtab = O->getDysymtabLoadCommand(); + uint32_t indirect_symbol = O->getIndirectSymbolTableEntry(Dysymtab, n + j); + if (indirect_symbol == MachO::INDIRECT_SYMBOL_LOCAL) { + outs() << "LOCAL\n"; + continue; + } + if (indirect_symbol == + (MachO::INDIRECT_SYMBOL_LOCAL | MachO::INDIRECT_SYMBOL_ABS)) { + outs() << "LOCAL ABSOLUTE\n"; + continue; + } + if (indirect_symbol == MachO::INDIRECT_SYMBOL_ABS) { + outs() << "ABSOLUTE\n"; + continue; + } + outs() << format("%5u ", indirect_symbol); + if (verbose) { + MachO::symtab_command Symtab = O->getSymtabLoadCommand(); + if (indirect_symbol < Symtab.nsyms) { + symbol_iterator Sym = O->getSymbolByIndex(indirect_symbol); + SymbolRef Symbol = *Sym; + StringRef SymName; + Symbol.getName(SymName); + outs() << SymName; + } else { + outs() << "?"; + } + } + outs() << "\n"; + } +} + +static void PrintIndirectSymbols(MachOObjectFile *O, bool verbose) { + for (const auto &Load : O->load_commands()) { + if (Load.C.cmd == MachO::LC_SEGMENT_64) { + MachO::segment_command_64 Seg = O->getSegment64LoadCommand(Load); + for (unsigned J = 0; J < Seg.nsects; ++J) { + MachO::section_64 Sec = O->getSection64(Load, J); + uint32_t section_type = Sec.flags & MachO::SECTION_TYPE; + if (section_type == MachO::S_NON_LAZY_SYMBOL_POINTERS || + section_type == MachO::S_LAZY_SYMBOL_POINTERS || + section_type == MachO::S_LAZY_DYLIB_SYMBOL_POINTERS || + section_type == MachO::S_THREAD_LOCAL_VARIABLE_POINTERS || + section_type == MachO::S_SYMBOL_STUBS) { + uint32_t stride; + if (section_type == MachO::S_SYMBOL_STUBS) + stride = Sec.reserved2; + else + stride = 8; + if (stride == 0) { + outs() << "Can't print indirect symbols for (" << Sec.segname << "," + << Sec.sectname << ") " + << "(size of stubs in reserved2 field is zero)\n"; + continue; + } + uint32_t count = Sec.size / stride; + outs() << "Indirect symbols for (" << Sec.segname << "," + << Sec.sectname << ") " << count << " entries"; + uint32_t n = Sec.reserved1; + PrintIndirectSymbolTable(O, verbose, n, count, stride, Sec.addr); + } + } + } else if (Load.C.cmd == MachO::LC_SEGMENT) { + MachO::segment_command Seg = O->getSegmentLoadCommand(Load); + for (unsigned J = 0; J < Seg.nsects; ++J) { + MachO::section Sec = O->getSection(Load, J); + uint32_t section_type = Sec.flags & MachO::SECTION_TYPE; + if (section_type == MachO::S_NON_LAZY_SYMBOL_POINTERS || + section_type == MachO::S_LAZY_SYMBOL_POINTERS || + section_type == MachO::S_LAZY_DYLIB_SYMBOL_POINTERS || + section_type == MachO::S_THREAD_LOCAL_VARIABLE_POINTERS || + section_type == MachO::S_SYMBOL_STUBS) { + uint32_t stride; + if (section_type == MachO::S_SYMBOL_STUBS) + stride = Sec.reserved2; + else + stride = 4; + if (stride == 0) { + outs() << "Can't print indirect symbols for (" << Sec.segname << "," + << Sec.sectname << ") " + << "(size of stubs in reserved2 field is zero)\n"; + continue; + } + uint32_t count = Sec.size / stride; + outs() << "Indirect symbols for (" << Sec.segname << "," + << Sec.sectname << ") " << count << " entries"; + uint32_t n = Sec.reserved1; + PrintIndirectSymbolTable(O, verbose, n, count, stride, Sec.addr); + } + } + } + } +} + +static void PrintDataInCodeTable(MachOObjectFile *O, bool verbose) { + MachO::linkedit_data_command DIC = O->getDataInCodeLoadCommand(); + uint32_t nentries = DIC.datasize / sizeof(struct MachO::data_in_code_entry); + outs() << "Data in code table (" << nentries << " entries)\n"; + outs() << "offset length kind\n"; + for (dice_iterator DI = O->begin_dices(), DE = O->end_dices(); DI != DE; + ++DI) { + uint32_t Offset; + DI->getOffset(Offset); + outs() << format("0x%08" PRIx32, Offset) << " "; + uint16_t Length; + DI->getLength(Length); + outs() << format("%6u", Length) << " "; + uint16_t Kind; + DI->getKind(Kind); + if (verbose) { + switch (Kind) { + case MachO::DICE_KIND_DATA: + outs() << "DATA"; + break; + case MachO::DICE_KIND_JUMP_TABLE8: + outs() << "JUMP_TABLE8"; + break; + case MachO::DICE_KIND_JUMP_TABLE16: + outs() << "JUMP_TABLE16"; + break; + case MachO::DICE_KIND_JUMP_TABLE32: + outs() << "JUMP_TABLE32"; + break; + case MachO::DICE_KIND_ABS_JUMP_TABLE32: + outs() << "ABS_JUMP_TABLE32"; + break; + default: + outs() << format("0x%04" PRIx32, Kind); + break; + } + } else + outs() << format("0x%04" PRIx32, Kind); + outs() << "\n"; + } +} - if (i == Header.ncmds - 1) +static void PrintLinkOptHints(MachOObjectFile *O) { + MachO::linkedit_data_command LohLC = O->getLinkOptHintsLoadCommand(); + const char *loh = O->getData().substr(LohLC.dataoff, 1).data(); + uint32_t nloh = LohLC.datasize; + outs() << "Linker optimiztion hints (" << nloh << " total bytes)\n"; + for (uint32_t i = 0; i < nloh;) { + unsigned n; + uint64_t identifier = decodeULEB128((const uint8_t *)(loh + i), &n); + i += n; + outs() << " identifier " << identifier << " "; + if (i >= nloh) + return; + switch (identifier) { + case 1: + outs() << "AdrpAdrp\n"; + break; + case 2: + outs() << "AdrpLdr\n"; + break; + case 3: + outs() << "AdrpAddLdr\n"; break; + case 4: + outs() << "AdrpLdrGotLdr\n"; + break; + case 5: + outs() << "AdrpAddStr\n"; + break; + case 6: + outs() << "AdrpLdrGotStr\n"; + break; + case 7: + outs() << "AdrpAdd\n"; + break; + case 8: + outs() << "AdrpLdrGot\n"; + break; + default: + outs() << "Unknown identifier value\n"; + break; + } + uint64_t narguments = decodeULEB128((const uint8_t *)(loh + i), &n); + i += n; + outs() << " narguments " << narguments << "\n"; + if (i >= nloh) + return; + + for (uint32_t j = 0; j < narguments; j++) { + uint64_t value = decodeULEB128((const uint8_t *)(loh + i), &n); + i += n; + outs() << "\tvalue " << format("0x%" PRIx64, value) << "\n"; + if (i >= nloh) + return; + } + } +} + +static void PrintDylibs(MachOObjectFile *O, bool JustId) { + unsigned Index = 0; + for (const auto &Load : O->load_commands()) { + if ((JustId && Load.C.cmd == MachO::LC_ID_DYLIB) || + (!JustId && (Load.C.cmd == MachO::LC_ID_DYLIB || + Load.C.cmd == MachO::LC_LOAD_DYLIB || + Load.C.cmd == MachO::LC_LOAD_WEAK_DYLIB || + Load.C.cmd == MachO::LC_REEXPORT_DYLIB || + Load.C.cmd == MachO::LC_LAZY_LOAD_DYLIB || + Load.C.cmd == MachO::LC_LOAD_UPWARD_DYLIB))) { + MachO::dylib_command dl = O->getDylibIDLoadCommand(Load); + if (dl.dylib.name < dl.cmdsize) { + const char *p = (const char *)(Load.Ptr) + dl.dylib.name; + if (JustId) + outs() << p << "\n"; + else { + outs() << "\t" << p; + outs() << " (compatibility version " + << ((dl.dylib.compatibility_version >> 16) & 0xffff) << "." + << ((dl.dylib.compatibility_version >> 8) & 0xff) << "." + << (dl.dylib.compatibility_version & 0xff) << ","; + outs() << " current version " + << ((dl.dylib.current_version >> 16) & 0xffff) << "." + << ((dl.dylib.current_version >> 8) & 0xff) << "." + << (dl.dylib.current_version & 0xff) << ")\n"; + } + } else { + outs() << "\tBad offset (" << dl.dylib.name << ") for name of "; + if (Load.C.cmd == MachO::LC_ID_DYLIB) + outs() << "LC_ID_DYLIB "; + else if (Load.C.cmd == MachO::LC_LOAD_DYLIB) + outs() << "LC_LOAD_DYLIB "; + else if (Load.C.cmd == MachO::LC_LOAD_WEAK_DYLIB) + outs() << "LC_LOAD_WEAK_DYLIB "; + else if (Load.C.cmd == MachO::LC_LAZY_LOAD_DYLIB) + outs() << "LC_LAZY_LOAD_DYLIB "; + else if (Load.C.cmd == MachO::LC_REEXPORT_DYLIB) + outs() << "LC_REEXPORT_DYLIB "; + else if (Load.C.cmd == MachO::LC_LOAD_UPWARD_DYLIB) + outs() << "LC_LOAD_UPWARD_DYLIB "; + else + outs() << "LC_??? "; + outs() << "command " << Index++ << "\n"; + } + } + } +} + +typedef DenseMap<uint64_t, StringRef> SymbolAddressMap; + +static void CreateSymbolAddressMap(MachOObjectFile *O, + SymbolAddressMap *AddrMap) { + // Create a map of symbol addresses to symbol names. + for (const SymbolRef &Symbol : O->symbols()) { + SymbolRef::Type ST; + Symbol.getType(ST); + if (ST == SymbolRef::ST_Function || ST == SymbolRef::ST_Data || + ST == SymbolRef::ST_Other) { + uint64_t Address; + Symbol.getAddress(Address); + StringRef SymName; + Symbol.getName(SymName); + if (!SymName.startswith(".objc")) + (*AddrMap)[Address] = SymName; + } + } +} + +// GuessSymbolName is passed the address of what might be a symbol and a +// pointer to the SymbolAddressMap. It returns the name of a symbol +// with that address or nullptr if no symbol is found with that address. +static const char *GuessSymbolName(uint64_t value, SymbolAddressMap *AddrMap) { + const char *SymbolName = nullptr; + // A DenseMap can't lookup up some values. + if (value != 0xffffffffffffffffULL && value != 0xfffffffffffffffeULL) { + StringRef name = AddrMap->lookup(value); + if (!name.empty()) + SymbolName = name.data(); + } + return SymbolName; +} + +static void DumpCstringChar(const char c) { + char p[2]; + p[0] = c; + p[1] = '\0'; + outs().write_escaped(p); +} + +static void DumpCstringSection(MachOObjectFile *O, const char *sect, + uint32_t sect_size, uint64_t sect_addr, + bool print_addresses) { + for (uint32_t i = 0; i < sect_size; i++) { + if (print_addresses) { + if (O->is64Bit()) + outs() << format("%016" PRIx64, sect_addr + i) << " "; + else + outs() << format("%08" PRIx64, sect_addr + i) << " "; + } + for (; i < sect_size && sect[i] != '\0'; i++) + DumpCstringChar(sect[i]); + if (i < sect_size && sect[i] == '\0') + outs() << "\n"; + } +} + +static void DumpLiteral4(uint32_t l, float f) { + outs() << format("0x%08" PRIx32, l); + if ((l & 0x7f800000) != 0x7f800000) + outs() << format(" (%.16e)\n", f); + else { + if (l == 0x7f800000) + outs() << " (+Infinity)\n"; + else if (l == 0xff800000) + outs() << " (-Infinity)\n"; + else if ((l & 0x00400000) == 0x00400000) + outs() << " (non-signaling Not-a-Number)\n"; + else + outs() << " (signaling Not-a-Number)\n"; + } +} + +static void DumpLiteral4Section(MachOObjectFile *O, const char *sect, + uint32_t sect_size, uint64_t sect_addr, + bool print_addresses) { + for (uint32_t i = 0; i < sect_size; i += sizeof(float)) { + if (print_addresses) { + if (O->is64Bit()) + outs() << format("%016" PRIx64, sect_addr + i) << " "; + else + outs() << format("%08" PRIx64, sect_addr + i) << " "; + } + float f; + memcpy(&f, sect + i, sizeof(float)); + if (O->isLittleEndian() != sys::IsLittleEndianHost) + sys::swapByteOrder(f); + uint32_t l; + memcpy(&l, sect + i, sizeof(uint32_t)); + if (O->isLittleEndian() != sys::IsLittleEndianHost) + sys::swapByteOrder(l); + DumpLiteral4(l, f); + } +} + +static void DumpLiteral8(MachOObjectFile *O, uint32_t l0, uint32_t l1, + double d) { + outs() << format("0x%08" PRIx32, l0) << " " << format("0x%08" PRIx32, l1); + uint32_t Hi, Lo; + if (O->isLittleEndian()) { + Hi = l1; + Lo = l0; + } else { + Hi = l0; + Lo = l1; + } + // Hi is the high word, so this is equivalent to if(isfinite(d)) + if ((Hi & 0x7ff00000) != 0x7ff00000) + outs() << format(" (%.16e)\n", d); + else { + if (Hi == 0x7ff00000 && Lo == 0) + outs() << " (+Infinity)\n"; + else if (Hi == 0xfff00000 && Lo == 0) + outs() << " (-Infinity)\n"; + else if ((Hi & 0x00080000) == 0x00080000) + outs() << " (non-signaling Not-a-Number)\n"; else - Command = MachOObj->getNextLoadCommandInfo(Command); + outs() << " (signaling Not-a-Number)\n"; + } +} + +static void DumpLiteral8Section(MachOObjectFile *O, const char *sect, + uint32_t sect_size, uint64_t sect_addr, + bool print_addresses) { + for (uint32_t i = 0; i < sect_size; i += sizeof(double)) { + if (print_addresses) { + if (O->is64Bit()) + outs() << format("%016" PRIx64, sect_addr + i) << " "; + else + outs() << format("%08" PRIx64, sect_addr + i) << " "; + } + double d; + memcpy(&d, sect + i, sizeof(double)); + if (O->isLittleEndian() != sys::IsLittleEndianHost) + sys::swapByteOrder(d); + uint32_t l0, l1; + memcpy(&l0, sect + i, sizeof(uint32_t)); + memcpy(&l1, sect + i + sizeof(uint32_t), sizeof(uint32_t)); + if (O->isLittleEndian() != sys::IsLittleEndianHost) { + sys::swapByteOrder(l0); + sys::swapByteOrder(l1); + } + DumpLiteral8(O, l0, l1, d); + } +} + +static void DumpLiteral16(uint32_t l0, uint32_t l1, uint32_t l2, uint32_t l3) { + outs() << format("0x%08" PRIx32, l0) << " "; + outs() << format("0x%08" PRIx32, l1) << " "; + outs() << format("0x%08" PRIx32, l2) << " "; + outs() << format("0x%08" PRIx32, l3) << "\n"; +} + +static void DumpLiteral16Section(MachOObjectFile *O, const char *sect, + uint32_t sect_size, uint64_t sect_addr, + bool print_addresses) { + for (uint32_t i = 0; i < sect_size; i += 16) { + if (print_addresses) { + if (O->is64Bit()) + outs() << format("%016" PRIx64, sect_addr + i) << " "; + else + outs() << format("%08" PRIx64, sect_addr + i) << " "; + } + uint32_t l0, l1, l2, l3; + memcpy(&l0, sect + i, sizeof(uint32_t)); + memcpy(&l1, sect + i + sizeof(uint32_t), sizeof(uint32_t)); + memcpy(&l2, sect + i + 2 * sizeof(uint32_t), sizeof(uint32_t)); + memcpy(&l3, sect + i + 3 * sizeof(uint32_t), sizeof(uint32_t)); + if (O->isLittleEndian() != sys::IsLittleEndianHost) { + sys::swapByteOrder(l0); + sys::swapByteOrder(l1); + sys::swapByteOrder(l2); + sys::swapByteOrder(l3); + } + DumpLiteral16(l0, l1, l2, l3); + } +} + +static void DumpLiteralPointerSection(MachOObjectFile *O, + const SectionRef &Section, + const char *sect, uint32_t sect_size, + uint64_t sect_addr, + bool print_addresses) { + // Collect the literal sections in this Mach-O file. + std::vector<SectionRef> LiteralSections; + for (const SectionRef &Section : O->sections()) { + DataRefImpl Ref = Section.getRawDataRefImpl(); + uint32_t section_type; + if (O->is64Bit()) { + const MachO::section_64 Sec = O->getSection64(Ref); + section_type = Sec.flags & MachO::SECTION_TYPE; + } else { + const MachO::section Sec = O->getSection(Ref); + section_type = Sec.flags & MachO::SECTION_TYPE; + } + if (section_type == MachO::S_CSTRING_LITERALS || + section_type == MachO::S_4BYTE_LITERALS || + section_type == MachO::S_8BYTE_LITERALS || + section_type == MachO::S_16BYTE_LITERALS) + LiteralSections.push_back(Section); + } + + // Set the size of the literal pointer. + uint32_t lp_size = O->is64Bit() ? 8 : 4; + + // Collect the external relocation symbols for the the literal pointers. + std::vector<std::pair<uint64_t, SymbolRef>> Relocs; + for (const RelocationRef &Reloc : Section.relocations()) { + DataRefImpl Rel; + MachO::any_relocation_info RE; + bool isExtern = false; + Rel = Reloc.getRawDataRefImpl(); + RE = O->getRelocation(Rel); + isExtern = O->getPlainRelocationExternal(RE); + if (isExtern) { + uint64_t RelocOffset; + Reloc.getOffset(RelocOffset); + symbol_iterator RelocSym = Reloc.getSymbol(); + Relocs.push_back(std::make_pair(RelocOffset, *RelocSym)); + } + } + array_pod_sort(Relocs.begin(), Relocs.end()); + + // Dump each literal pointer. + for (uint32_t i = 0; i < sect_size; i += lp_size) { + if (print_addresses) { + if (O->is64Bit()) + outs() << format("%016" PRIx64, sect_addr + i) << " "; + else + outs() << format("%08" PRIx64, sect_addr + i) << " "; + } + uint64_t lp; + if (O->is64Bit()) { + memcpy(&lp, sect + i, sizeof(uint64_t)); + if (O->isLittleEndian() != sys::IsLittleEndianHost) + sys::swapByteOrder(lp); + } else { + uint32_t li; + memcpy(&li, sect + i, sizeof(uint32_t)); + if (O->isLittleEndian() != sys::IsLittleEndianHost) + sys::swapByteOrder(li); + lp = li; + } + + // First look for an external relocation entry for this literal pointer. + auto Reloc = std::find_if( + Relocs.begin(), Relocs.end(), + [&](const std::pair<uint64_t, SymbolRef> &P) { return P.first == i; }); + if (Reloc != Relocs.end()) { + symbol_iterator RelocSym = Reloc->second; + StringRef SymName; + RelocSym->getName(SymName); + outs() << "external relocation entry for symbol:" << SymName << "\n"; + continue; + } + + // For local references see what the section the literal pointer points to. + auto Sect = std::find_if(LiteralSections.begin(), LiteralSections.end(), + [&](const SectionRef &R) { + return lp >= R.getAddress() && + lp < R.getAddress() + R.getSize(); + }); + if (Sect == LiteralSections.end()) { + outs() << format("0x%" PRIx64, lp) << " (not in a literal section)\n"; + continue; + } + + uint64_t SectAddress = Sect->getAddress(); + uint64_t SectSize = Sect->getSize(); + + StringRef SectName; + Sect->getName(SectName); + DataRefImpl Ref = Sect->getRawDataRefImpl(); + StringRef SegmentName = O->getSectionFinalSegmentName(Ref); + outs() << SegmentName << ":" << SectName << ":"; + + uint32_t section_type; + if (O->is64Bit()) { + const MachO::section_64 Sec = O->getSection64(Ref); + section_type = Sec.flags & MachO::SECTION_TYPE; + } else { + const MachO::section Sec = O->getSection(Ref); + section_type = Sec.flags & MachO::SECTION_TYPE; + } + + StringRef BytesStr; + Sect->getContents(BytesStr); + const char *Contents = reinterpret_cast<const char *>(BytesStr.data()); + + switch (section_type) { + case MachO::S_CSTRING_LITERALS: + for (uint64_t i = lp - SectAddress; i < SectSize && Contents[i] != '\0'; + i++) { + DumpCstringChar(Contents[i]); + } + outs() << "\n"; + break; + case MachO::S_4BYTE_LITERALS: + float f; + memcpy(&f, Contents + (lp - SectAddress), sizeof(float)); + uint32_t l; + memcpy(&l, Contents + (lp - SectAddress), sizeof(uint32_t)); + if (O->isLittleEndian() != sys::IsLittleEndianHost) { + sys::swapByteOrder(f); + sys::swapByteOrder(l); + } + DumpLiteral4(l, f); + break; + case MachO::S_8BYTE_LITERALS: { + double d; + memcpy(&d, Contents + (lp - SectAddress), sizeof(double)); + uint32_t l0, l1; + memcpy(&l0, Contents + (lp - SectAddress), sizeof(uint32_t)); + memcpy(&l1, Contents + (lp - SectAddress) + sizeof(uint32_t), + sizeof(uint32_t)); + if (O->isLittleEndian() != sys::IsLittleEndianHost) { + sys::swapByteOrder(f); + sys::swapByteOrder(l0); + sys::swapByteOrder(l1); + } + DumpLiteral8(O, l0, l1, d); + break; + } + case MachO::S_16BYTE_LITERALS: { + uint32_t l0, l1, l2, l3; + memcpy(&l0, Contents + (lp - SectAddress), sizeof(uint32_t)); + memcpy(&l1, Contents + (lp - SectAddress) + sizeof(uint32_t), + sizeof(uint32_t)); + memcpy(&l2, Contents + (lp - SectAddress) + 2 * sizeof(uint32_t), + sizeof(uint32_t)); + memcpy(&l3, Contents + (lp - SectAddress) + 3 * sizeof(uint32_t), + sizeof(uint32_t)); + if (O->isLittleEndian() != sys::IsLittleEndianHost) { + sys::swapByteOrder(l0); + sys::swapByteOrder(l1); + sys::swapByteOrder(l2); + sys::swapByteOrder(l3); + } + DumpLiteral16(l0, l1, l2, l3); + break; + } + } + } +} + +static void DumpInitTermPointerSection(MachOObjectFile *O, const char *sect, + uint32_t sect_size, uint64_t sect_addr, + SymbolAddressMap *AddrMap, + bool verbose) { + uint32_t stride; + if (O->is64Bit()) + stride = sizeof(uint64_t); + else + stride = sizeof(uint32_t); + for (uint32_t i = 0; i < sect_size; i += stride) { + const char *SymbolName = nullptr; + if (O->is64Bit()) { + outs() << format("0x%016" PRIx64, sect_addr + i * stride) << " "; + uint64_t pointer_value; + memcpy(&pointer_value, sect + i, stride); + if (O->isLittleEndian() != sys::IsLittleEndianHost) + sys::swapByteOrder(pointer_value); + outs() << format("0x%016" PRIx64, pointer_value); + if (verbose) + SymbolName = GuessSymbolName(pointer_value, AddrMap); + } else { + outs() << format("0x%08" PRIx64, sect_addr + i * stride) << " "; + uint32_t pointer_value; + memcpy(&pointer_value, sect + i, stride); + if (O->isLittleEndian() != sys::IsLittleEndianHost) + sys::swapByteOrder(pointer_value); + outs() << format("0x%08" PRIx32, pointer_value); + if (verbose) + SymbolName = GuessSymbolName(pointer_value, AddrMap); + } + if (SymbolName) + outs() << " " << SymbolName; + outs() << "\n"; + } +} + +static void DumpRawSectionContents(MachOObjectFile *O, const char *sect, + uint32_t size, uint64_t addr) { + uint32_t cputype = O->getHeader().cputype; + if (cputype == MachO::CPU_TYPE_I386 || cputype == MachO::CPU_TYPE_X86_64) { + uint32_t j; + for (uint32_t i = 0; i < size; i += j, addr += j) { + if (O->is64Bit()) + outs() << format("%016" PRIx64, addr) << "\t"; + else + outs() << format("%08" PRIx64, addr) << "\t"; + for (j = 0; j < 16 && i + j < size; j++) { + uint8_t byte_word = *(sect + i + j); + outs() << format("%02" PRIx32, (uint32_t)byte_word) << " "; + } + outs() << "\n"; + } + } else { + uint32_t j; + for (uint32_t i = 0; i < size; i += j, addr += j) { + if (O->is64Bit()) + outs() << format("%016" PRIx64, addr) << "\t"; + else + outs() << format("%08" PRIx64, sect) << "\t"; + for (j = 0; j < 4 * sizeof(int32_t) && i + j < size; + j += sizeof(int32_t)) { + if (i + j + sizeof(int32_t) < size) { + uint32_t long_word; + memcpy(&long_word, sect + i + j, sizeof(int32_t)); + if (O->isLittleEndian() != sys::IsLittleEndianHost) + sys::swapByteOrder(long_word); + outs() << format("%08" PRIx32, long_word) << " "; + } else { + for (uint32_t k = 0; i + j + k < size; k++) { + uint8_t byte_word = *(sect + i + j); + outs() << format("%02" PRIx32, (uint32_t)byte_word) << " "; + } + } + } + outs() << "\n"; + } + } +} + +static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF, + StringRef DisSegName, StringRef DisSectName); +static void DumpProtocolSection(MachOObjectFile *O, const char *sect, + uint32_t size, uint32_t addr); + +static void DumpSectionContents(StringRef Filename, MachOObjectFile *O, + bool verbose) { + SymbolAddressMap AddrMap; + if (verbose) + CreateSymbolAddressMap(O, &AddrMap); + + for (unsigned i = 0; i < DumpSections.size(); ++i) { + StringRef DumpSection = DumpSections[i]; + std::pair<StringRef, StringRef> DumpSegSectName; + DumpSegSectName = DumpSection.split(','); + StringRef DumpSegName, DumpSectName; + if (DumpSegSectName.second.size()) { + DumpSegName = DumpSegSectName.first; + DumpSectName = DumpSegSectName.second; + } else { + DumpSegName = ""; + DumpSectName = DumpSegSectName.first; + } + for (const SectionRef &Section : O->sections()) { + StringRef SectName; + Section.getName(SectName); + DataRefImpl Ref = Section.getRawDataRefImpl(); + StringRef SegName = O->getSectionFinalSegmentName(Ref); + if ((DumpSegName.empty() || SegName == DumpSegName) && + (SectName == DumpSectName)) { + + uint32_t section_flags; + if (O->is64Bit()) { + const MachO::section_64 Sec = O->getSection64(Ref); + section_flags = Sec.flags; + + } else { + const MachO::section Sec = O->getSection(Ref); + section_flags = Sec.flags; + } + uint32_t section_type = section_flags & MachO::SECTION_TYPE; + + StringRef BytesStr; + Section.getContents(BytesStr); + const char *sect = reinterpret_cast<const char *>(BytesStr.data()); + uint32_t sect_size = BytesStr.size(); + uint64_t sect_addr = Section.getAddress(); + + if (Raw) { + outs().write(BytesStr.data(), BytesStr.size()); + continue; + } + + outs() << "Contents of (" << SegName << "," << SectName + << ") section\n"; + + if (verbose) { + if ((section_flags & MachO::S_ATTR_PURE_INSTRUCTIONS) || + (section_flags & MachO::S_ATTR_SOME_INSTRUCTIONS)) { + DisassembleMachO(Filename, O, SegName, SectName); + continue; + } + if (SegName == "__TEXT" && SectName == "__info_plist") { + outs() << sect; + continue; + } + if (SegName == "__OBJC" && SectName == "__protocol") { + DumpProtocolSection(O, sect, sect_size, sect_addr); + continue; + } + switch (section_type) { + case MachO::S_REGULAR: + DumpRawSectionContents(O, sect, sect_size, sect_addr); + break; + case MachO::S_ZEROFILL: + outs() << "zerofill section and has no contents in the file\n"; + break; + case MachO::S_CSTRING_LITERALS: + DumpCstringSection(O, sect, sect_size, sect_addr, !NoLeadingAddr); + break; + case MachO::S_4BYTE_LITERALS: + DumpLiteral4Section(O, sect, sect_size, sect_addr, !NoLeadingAddr); + break; + case MachO::S_8BYTE_LITERALS: + DumpLiteral8Section(O, sect, sect_size, sect_addr, !NoLeadingAddr); + break; + case MachO::S_16BYTE_LITERALS: + DumpLiteral16Section(O, sect, sect_size, sect_addr, !NoLeadingAddr); + break; + case MachO::S_LITERAL_POINTERS: + DumpLiteralPointerSection(O, Section, sect, sect_size, sect_addr, + !NoLeadingAddr); + break; + case MachO::S_MOD_INIT_FUNC_POINTERS: + case MachO::S_MOD_TERM_FUNC_POINTERS: + DumpInitTermPointerSection(O, sect, sect_size, sect_addr, &AddrMap, + verbose); + break; + default: + outs() << "Unknown section type (" + << format("0x%08" PRIx32, section_type) << ")\n"; + DumpRawSectionContents(O, sect, sect_size, sect_addr); + break; + } + } else { + if (section_type == MachO::S_ZEROFILL) + outs() << "zerofill section and has no contents in the file\n"; + else + DumpRawSectionContents(O, sect, sect_size, sect_addr); + } + } + } + } +} + +static void DumpInfoPlistSectionContents(StringRef Filename, + MachOObjectFile *O) { + for (const SectionRef &Section : O->sections()) { + StringRef SectName; + Section.getName(SectName); + DataRefImpl Ref = Section.getRawDataRefImpl(); + StringRef SegName = O->getSectionFinalSegmentName(Ref); + if (SegName == "__TEXT" && SectName == "__info_plist") { + outs() << "Contents of (" << SegName << "," << SectName << ") section\n"; + StringRef BytesStr; + Section.getContents(BytesStr); + const char *sect = reinterpret_cast<const char *>(BytesStr.data()); + outs() << sect; + return; + } } } @@ -289,7 +1174,7 @@ static bool checkMachOAndArchFlags(ObjectFile *O, StringRef Filename) { return true; } -static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF); +static void printObjcMetaData(MachOObjectFile *O, bool verbose); // ProcessMachO() is passed a single opened Mach-O file, which may be an // archive member and or in a slice of a universal file. It prints the @@ -300,9 +1185,11 @@ static void ProcessMachO(StringRef Filename, MachOObjectFile *MachOOF, StringRef ArchitectureName = StringRef()) { // If we are doing some processing here on the Mach-O file print the header // info. And don't print it otherwise like in the case of printing the - // UniversalHeaders. + // UniversalHeaders or ArchiveHeaders. if (Disassemble || PrivateHeaders || ExportsTrie || Rebase || Bind || - LazyBind || WeakBind) { + LazyBind || WeakBind || IndirectSymbols || DataInCode || LinkOptHints || + DylibsUsed || DylibId || ObjcMetaData || + (DumpSections.size() != 0 && !Raw)) { outs() << Filename; if (!ArchiveMemberName.empty()) outs() << '(' << ArchiveMemberName << ')'; @@ -312,23 +1199,35 @@ static void ProcessMachO(StringRef Filename, MachOObjectFile *MachOOF, } if (Disassemble) - DisassembleMachO(Filename, MachOOF); - // TODO: These should/could be printed in Darwin's otool(1) or nm(1) style - // for -macho. Or just used a new option that maps to the otool(1) - // option like -r, -l, etc. Or just the normal llvm-objdump option - // but now for this slice so that the -arch options can be used. - // if (Relocations) - // PrintRelocations(MachOOF); - // if (SectionHeaders) - // PrintSectionHeaders(MachOOF); - // if (SectionContents) - // PrintSectionContents(MachOOF); - // if (SymbolTable) - // PrintSymbolTable(MachOOF); - // if (UnwindInfo) - // PrintUnwindInfo(MachOOF); + DisassembleMachO(Filename, MachOOF, "__TEXT", "__text"); + if (IndirectSymbols) + PrintIndirectSymbols(MachOOF, !NonVerbose); + if (DataInCode) + PrintDataInCodeTable(MachOOF, !NonVerbose); + if (LinkOptHints) + PrintLinkOptHints(MachOOF); + if (Relocations) + PrintRelocations(MachOOF); + if (SectionHeaders) + PrintSectionHeaders(MachOOF); + if (SectionContents) + PrintSectionContents(MachOOF); + if (DumpSections.size() != 0) + DumpSectionContents(Filename, MachOOF, !NonVerbose); + if (InfoPlist) + DumpInfoPlistSectionContents(Filename, MachOOF); + if (DylibsUsed) + PrintDylibs(MachOOF, false); + if (DylibId) + PrintDylibs(MachOOF, true); + if (SymbolTable) + PrintSymbolTable(MachOOF); + if (UnwindInfo) + printMachOUnwindInfo(MachOOF); if (PrivateHeaders) printMachOFileHeader(MachOOF); + if (ObjcMetaData) + printObjcMetaData(MachOOF, !NonVerbose); if (ExportsTrie) printExportsTrie(MachOOF); if (Rebase) @@ -514,6 +1413,106 @@ static void printMachOUniversalHeaders(const object::MachOUniversalBinary *UB, } } +static void printArchiveChild(Archive::Child &C, bool verbose, + bool print_offset) { + if (print_offset) + outs() << C.getChildOffset() << "\t"; + sys::fs::perms Mode = C.getAccessMode(); + if (verbose) { + // FIXME: this first dash, "-", is for (Mode & S_IFMT) == S_IFREG. + // But there is nothing in sys::fs::perms for S_IFMT or S_IFREG. + outs() << "-"; + if (Mode & sys::fs::owner_read) + outs() << "r"; + else + outs() << "-"; + if (Mode & sys::fs::owner_write) + outs() << "w"; + else + outs() << "-"; + if (Mode & sys::fs::owner_exe) + outs() << "x"; + else + outs() << "-"; + if (Mode & sys::fs::group_read) + outs() << "r"; + else + outs() << "-"; + if (Mode & sys::fs::group_write) + outs() << "w"; + else + outs() << "-"; + if (Mode & sys::fs::group_exe) + outs() << "x"; + else + outs() << "-"; + if (Mode & sys::fs::others_read) + outs() << "r"; + else + outs() << "-"; + if (Mode & sys::fs::others_write) + outs() << "w"; + else + outs() << "-"; + if (Mode & sys::fs::others_exe) + outs() << "x"; + else + outs() << "-"; + } else { + outs() << format("0%o ", Mode); + } + + unsigned UID = C.getUID(); + outs() << format("%3d/", UID); + unsigned GID = C.getGID(); + outs() << format("%-3d ", GID); + uint64_t Size = C.getRawSize(); + outs() << format("%5" PRId64, Size) << " "; + + StringRef RawLastModified = C.getRawLastModified(); + if (verbose) { + unsigned Seconds; + if (RawLastModified.getAsInteger(10, Seconds)) + outs() << "(date: \"%s\" contains non-decimal chars) " << RawLastModified; + else { + // Since cime(3) returns a 26 character string of the form: + // "Sun Sep 16 01:03:52 1973\n\0" + // just print 24 characters. + time_t t = Seconds; + outs() << format("%.24s ", ctime(&t)); + } + } else { + outs() << RawLastModified << " "; + } + + if (verbose) { + ErrorOr<StringRef> NameOrErr = C.getName(); + if (NameOrErr.getError()) { + StringRef RawName = C.getRawName(); + outs() << RawName << "\n"; + } else { + StringRef Name = NameOrErr.get(); + outs() << Name << "\n"; + } + } else { + StringRef RawName = C.getRawName(); + outs() << RawName << "\n"; + } +} + +static void printArchiveHeaders(Archive *A, bool verbose, bool print_offset) { + if (A->hasSymbolTable()) { + Archive::child_iterator S = A->getSymbolTableChild(); + Archive::Child C = *S; + printArchiveChild(C, verbose, print_offset); + } + for (Archive::child_iterator I = A->child_begin(), E = A->child_end(); I != E; + ++I) { + Archive::Child C = *I; + printArchiveChild(C, verbose, print_offset); + } +} + // ParseInputMachO() parses the named Mach-O file in Filename and handles the // -arch flags selecting just those slices as specified by them and also parses // archive files. Then for each individual Mach-O file ProcessMachO() is @@ -542,6 +1541,8 @@ void llvm::ParseInputMachO(StringRef Filename) { if (Archive *A = dyn_cast<Archive>(&Bin)) { outs() << "Archive : " << Filename << "\n"; + if (ArchiveHeaders) + printArchiveHeaders(A, !NonVerbose, ArchiveMemberOffsets); for (Archive::child_iterator I = A->child_begin(), E = A->child_end(); I != E; ++I) { ErrorOr<std::unique_ptr<Binary>> ChildOrErr = I->getAsBinary(); @@ -557,7 +1558,7 @@ void llvm::ParseInputMachO(StringRef Filename) { } if (UniversalHeaders) { if (MachOUniversalBinary *UB = dyn_cast<MachOUniversalBinary>(&Bin)) - printMachOUniversalHeaders(UB, true); + printMachOUniversalHeaders(UB, !NonVerbose); } if (MachOUniversalBinary *UB = dyn_cast<MachOUniversalBinary>(&Bin)) { // If we have a list of architecture flags specified dump only those. @@ -587,6 +1588,8 @@ void llvm::ParseInputMachO(StringRef Filename) { if (!ArchitectureName.empty()) outs() << " (architecture " << ArchitectureName << ")"; outs() << "\n"; + if (ArchiveHeaders) + printArchiveHeaders(A.get(), !NonVerbose, ArchiveMemberOffsets); for (Archive::child_iterator AI = A->child_begin(), AE = A->child_end(); AI != AE; ++AI) { @@ -627,6 +1630,8 @@ void llvm::ParseInputMachO(StringRef Filename) { I->getAsArchive()) { std::unique_ptr<Archive> &A = *AOrErr; outs() << "Archive : " << Filename << "\n"; + if (ArchiveHeaders) + printArchiveHeaders(A.get(), !NonVerbose, ArchiveMemberOffsets); for (Archive::child_iterator AI = A->child_begin(), AE = A->child_end(); AI != AE; ++AI) { @@ -662,6 +1667,8 @@ void llvm::ParseInputMachO(StringRef Filename) { if (!ArchitectureName.empty()) outs() << " (architecture " << ArchitectureName << ")"; outs() << "\n"; + if (ArchiveHeaders) + printArchiveHeaders(A.get(), !NonVerbose, ArchiveMemberOffsets); for (Archive::child_iterator AI = A->child_begin(), AE = A->child_end(); AI != AE; ++AI) { ErrorOr<std::unique_ptr<Binary>> ChildOrErr = AI->getAsBinary(); @@ -691,7 +1698,6 @@ void llvm::ParseInputMachO(StringRef Filename) { << "Unrecognized file type.\n"; } -typedef DenseMap<uint64_t, StringRef> SymbolAddressMap; typedef std::pair<uint64_t, const char *> BindInfoEntry; typedef std::vector<BindInfoEntry> BindTable; typedef BindTable::iterator bind_table_iterator; @@ -712,21 +1718,6 @@ struct DisassembleInfo { BindTable *bindtable; }; -// GuessSymbolName is passed the address of what might be a symbol and a -// pointer to the DisassembleInfo struct. It returns the name of a symbol -// with that address or nullptr if no symbol is found with that address. -static const char *GuessSymbolName(uint64_t value, - struct DisassembleInfo *info) { - const char *SymbolName = nullptr; - // A DenseMap can't lookup up some values. - if (value != 0xffffffffffffffffULL && value != 0xfffffffffffffffeULL) { - StringRef name = info->AddrMap->lookup(value); - if (!name.empty()) - SymbolName = name.data(); - } - return SymbolName; -} - // SymbolizerGetOpInfo() is the operand information call back function. // This is called to get the symbolic information for operand(s) of an // instruction when it is being done. This routine does this from @@ -742,8 +1733,8 @@ static const char *GuessSymbolName(uint64_t value, // names and addends of the symbolic expression to add for the operand. The // value of TagType is currently 1 (for the LLVMOpInfo1 struct). If symbolic // information is returned then this function returns 1 else it returns 0. -int SymbolizerGetOpInfo(void *DisInfo, uint64_t Pc, uint64_t Offset, - uint64_t Size, int TagType, void *TagBuf) { +static int SymbolizerGetOpInfo(void *DisInfo, uint64_t Pc, uint64_t Offset, + uint64_t Size, int TagType, void *TagBuf) { struct DisassembleInfo *info = (struct DisassembleInfo *)DisInfo; struct LLVMOpInfo1 *op_info = (struct LLVMOpInfo1 *)TagBuf; uint64_t value = op_info->Value; @@ -755,7 +1746,7 @@ int SymbolizerGetOpInfo(void *DisInfo, uint64_t Pc, uint64_t Offset, // If the TagType is not the value 1 which it code knows about or if no // verbose symbolic information is wanted then just return 0, indicating no // information is being returned. - if (TagType != 1 || info->verbose == false) + if (TagType != 1 || !info->verbose) return 0; unsigned int Arch = info->O->getArch(); @@ -817,8 +1808,8 @@ int SymbolizerGetOpInfo(void *DisInfo, uint64_t Pc, uint64_t Offset, } if (reloc_found && (r_type == MachO::GENERIC_RELOC_SECTDIFF || r_type == MachO::GENERIC_RELOC_LOCAL_SECTDIFF)) { - const char *add = GuessSymbolName(r_value, info); - const char *sub = GuessSymbolName(pair_r_value, info); + const char *add = GuessSymbolName(r_value, info->AddrMap); + const char *sub = GuessSymbolName(pair_r_value, info->AddrMap); uint32_t offset = value - (r_value - pair_r_value); op_info->AddSymbol.Present = 1; if (add != nullptr) @@ -838,7 +1829,8 @@ int SymbolizerGetOpInfo(void *DisInfo, uint64_t Pc, uint64_t Offset, // (if any) for an entry that matches this segment offset. // uint32_t seg_offset = (Pc + Offset); return 0; - } else if (Arch == Triple::x86_64) { + } + if (Arch == Triple::x86_64) { if (Size != 1 && Size != 2 && Size != 4 && Size != 0) return 0; // First search the section's relocation entries (if any) for an entry @@ -904,89 +1896,77 @@ int SymbolizerGetOpInfo(void *DisInfo, uint64_t Pc, uint64_t Offset, // (if any) for an entry that matches this segment offset. // uint64_t seg_offset = (Pc + Offset); return 0; - } else if (Arch == Triple::arm) { + } + if (Arch == Triple::arm) { if (Offset != 0 || (Size != 4 && Size != 2)) return 0; // First search the section's relocation entries (if any) for an entry // for this section offset. uint32_t sect_addr = info->S.getAddress(); uint32_t sect_offset = (Pc + Offset) - sect_addr; - bool reloc_found = false; DataRefImpl Rel; MachO::any_relocation_info RE; bool isExtern = false; SymbolRef Symbol; bool r_scattered = false; uint32_t r_value, pair_r_value, r_type, r_length, other_half; - for (const RelocationRef &Reloc : info->S.relocations()) { - uint64_t RelocOffset; - Reloc.getOffset(RelocOffset); - if (RelocOffset == sect_offset) { - Rel = Reloc.getRawDataRefImpl(); - RE = info->O->getRelocation(Rel); - r_length = info->O->getAnyRelocationLength(RE); - r_scattered = info->O->isRelocationScattered(RE); - if (r_scattered) { - r_value = info->O->getScatteredRelocationValue(RE); - r_type = info->O->getScatteredRelocationType(RE); - } else { - r_type = info->O->getAnyRelocationType(RE); - isExtern = info->O->getPlainRelocationExternal(RE); - if (isExtern) { - symbol_iterator RelocSym = Reloc.getSymbol(); - Symbol = *RelocSym; - } - } - if (r_type == MachO::ARM_RELOC_HALF || - r_type == MachO::ARM_RELOC_SECTDIFF || - r_type == MachO::ARM_RELOC_LOCAL_SECTDIFF || - r_type == MachO::ARM_RELOC_HALF_SECTDIFF) { - DataRefImpl RelNext = Rel; - info->O->moveRelocationNext(RelNext); - MachO::any_relocation_info RENext; - RENext = info->O->getRelocation(RelNext); - other_half = info->O->getAnyRelocationAddress(RENext) & 0xffff; - if (info->O->isRelocationScattered(RENext)) - pair_r_value = info->O->getScatteredRelocationValue(RENext); - } - reloc_found = true; - break; + auto Reloc = + std::find_if(info->S.relocations().begin(), info->S.relocations().end(), + [&](const RelocationRef &Reloc) { + uint64_t RelocOffset; + Reloc.getOffset(RelocOffset); + return RelocOffset == sect_offset; + }); + + if (Reloc == info->S.relocations().end()) + return 0; + + Rel = Reloc->getRawDataRefImpl(); + RE = info->O->getRelocation(Rel); + r_length = info->O->getAnyRelocationLength(RE); + r_scattered = info->O->isRelocationScattered(RE); + if (r_scattered) { + r_value = info->O->getScatteredRelocationValue(RE); + r_type = info->O->getScatteredRelocationType(RE); + } else { + r_type = info->O->getAnyRelocationType(RE); + isExtern = info->O->getPlainRelocationExternal(RE); + if (isExtern) { + symbol_iterator RelocSym = Reloc->getSymbol(); + Symbol = *RelocSym; } } - if (reloc_found && isExtern) { + if (r_type == MachO::ARM_RELOC_HALF || + r_type == MachO::ARM_RELOC_SECTDIFF || + r_type == MachO::ARM_RELOC_LOCAL_SECTDIFF || + r_type == MachO::ARM_RELOC_HALF_SECTDIFF) { + DataRefImpl RelNext = Rel; + info->O->moveRelocationNext(RelNext); + MachO::any_relocation_info RENext; + RENext = info->O->getRelocation(RelNext); + other_half = info->O->getAnyRelocationAddress(RENext) & 0xffff; + if (info->O->isRelocationScattered(RENext)) + pair_r_value = info->O->getScatteredRelocationValue(RENext); + } + + if (isExtern) { StringRef SymName; Symbol.getName(SymName); const char *name = SymName.data(); op_info->AddSymbol.Present = 1; op_info->AddSymbol.Name = name; - if (value != 0) { - switch (r_type) { - case MachO::ARM_RELOC_HALF: - if ((r_length & 0x1) == 1) { - op_info->Value = value << 16 | other_half; - op_info->VariantKind = LLVMDisassembler_VariantKind_ARM_HI16; - } else { - op_info->Value = other_half << 16 | value; - op_info->VariantKind = LLVMDisassembler_VariantKind_ARM_LO16; - } - break; - default: - break; - } - } else { - switch (r_type) { - case MachO::ARM_RELOC_HALF: - if ((r_length & 0x1) == 1) { - op_info->Value = value << 16 | other_half; - op_info->VariantKind = LLVMDisassembler_VariantKind_ARM_HI16; - } else { - op_info->Value = other_half << 16 | value; - op_info->VariantKind = LLVMDisassembler_VariantKind_ARM_LO16; - } - break; - default: - break; + switch (r_type) { + case MachO::ARM_RELOC_HALF: + if ((r_length & 0x1) == 1) { + op_info->Value = value << 16 | other_half; + op_info->VariantKind = LLVMDisassembler_VariantKind_ARM_HI16; + } else { + op_info->Value = other_half << 16 | value; + op_info->VariantKind = LLVMDisassembler_VariantKind_ARM_LO16; } + break; + default: + break; } return 1; } @@ -994,33 +1974,31 @@ int SymbolizerGetOpInfo(void *DisInfo, uint64_t Pc, uint64_t Offset, // return 0 so the code in tryAddingSymbolicOperand() can use the // SymbolLookUp call back with the branch target address to look up the // symbol and possiblity add an annotation for a symbol stub. - if (reloc_found && isExtern == 0 && (r_type == MachO::ARM_RELOC_BR24 || - r_type == MachO::ARM_THUMB_RELOC_BR22)) + if (isExtern == 0 && (r_type == MachO::ARM_RELOC_BR24 || + r_type == MachO::ARM_THUMB_RELOC_BR22)) return 0; uint32_t offset = 0; - if (reloc_found) { - if (r_type == MachO::ARM_RELOC_HALF || - r_type == MachO::ARM_RELOC_HALF_SECTDIFF) { - if ((r_length & 0x1) == 1) - value = value << 16 | other_half; - else - value = other_half << 16 | value; - } - if (r_scattered && (r_type != MachO::ARM_RELOC_HALF && - r_type != MachO::ARM_RELOC_HALF_SECTDIFF)) { - offset = value - r_value; - value = r_value; - } + if (r_type == MachO::ARM_RELOC_HALF || + r_type == MachO::ARM_RELOC_HALF_SECTDIFF) { + if ((r_length & 0x1) == 1) + value = value << 16 | other_half; + else + value = other_half << 16 | value; + } + if (r_scattered && (r_type != MachO::ARM_RELOC_HALF && + r_type != MachO::ARM_RELOC_HALF_SECTDIFF)) { + offset = value - r_value; + value = r_value; } - if (reloc_found && r_type == MachO::ARM_RELOC_HALF_SECTDIFF) { + if (r_type == MachO::ARM_RELOC_HALF_SECTDIFF) { if ((r_length & 0x1) == 1) op_info->VariantKind = LLVMDisassembler_VariantKind_ARM_HI16; else op_info->VariantKind = LLVMDisassembler_VariantKind_ARM_LO16; - const char *add = GuessSymbolName(r_value, info); - const char *sub = GuessSymbolName(pair_r_value, info); + const char *add = GuessSymbolName(r_value, info->AddrMap); + const char *sub = GuessSymbolName(pair_r_value, info->AddrMap); int32_t offset = value - (r_value - pair_r_value); op_info->AddSymbol.Present = 1; if (add != nullptr) @@ -1036,118 +2014,102 @@ int SymbolizerGetOpInfo(void *DisInfo, uint64_t Pc, uint64_t Offset, return 1; } - if (reloc_found == false) - return 0; - op_info->AddSymbol.Present = 1; op_info->Value = offset; - if (reloc_found) { - if (r_type == MachO::ARM_RELOC_HALF) { - if ((r_length & 0x1) == 1) - op_info->VariantKind = LLVMDisassembler_VariantKind_ARM_HI16; - else - op_info->VariantKind = LLVMDisassembler_VariantKind_ARM_LO16; - } + if (r_type == MachO::ARM_RELOC_HALF) { + if ((r_length & 0x1) == 1) + op_info->VariantKind = LLVMDisassembler_VariantKind_ARM_HI16; + else + op_info->VariantKind = LLVMDisassembler_VariantKind_ARM_LO16; } - const char *add = GuessSymbolName(value, info); + const char *add = GuessSymbolName(value, info->AddrMap); if (add != nullptr) { op_info->AddSymbol.Name = add; return 1; } op_info->AddSymbol.Value = value; return 1; - } else if (Arch == Triple::aarch64) { + } + if (Arch == Triple::aarch64) { if (Offset != 0 || Size != 4) return 0; // First search the section's relocation entries (if any) for an entry // for this section offset. uint64_t sect_addr = info->S.getAddress(); uint64_t sect_offset = (Pc + Offset) - sect_addr; - bool reloc_found = false; - DataRefImpl Rel; - MachO::any_relocation_info RE; - bool isExtern = false; - SymbolRef Symbol; - uint32_t r_type = 0; - for (const RelocationRef &Reloc : info->S.relocations()) { - uint64_t RelocOffset; - Reloc.getOffset(RelocOffset); - if (RelocOffset == sect_offset) { - Rel = Reloc.getRawDataRefImpl(); - RE = info->O->getRelocation(Rel); - r_type = info->O->getAnyRelocationType(RE); - if (r_type == MachO::ARM64_RELOC_ADDEND) { - DataRefImpl RelNext = Rel; - info->O->moveRelocationNext(RelNext); - MachO::any_relocation_info RENext = info->O->getRelocation(RelNext); - if (value == 0) { - value = info->O->getPlainRelocationSymbolNum(RENext); - op_info->Value = value; - } - } - // NOTE: Scattered relocations don't exist on arm64. - isExtern = info->O->getPlainRelocationExternal(RE); - if (isExtern) { - symbol_iterator RelocSym = Reloc.getSymbol(); - Symbol = *RelocSym; - } - reloc_found = true; - break; + auto Reloc = + std::find_if(info->S.relocations().begin(), info->S.relocations().end(), + [&](const RelocationRef &Reloc) { + uint64_t RelocOffset; + Reloc.getOffset(RelocOffset); + return RelocOffset == sect_offset; + }); + + if (Reloc == info->S.relocations().end()) + return 0; + + DataRefImpl Rel = Reloc->getRawDataRefImpl(); + MachO::any_relocation_info RE = info->O->getRelocation(Rel); + uint32_t r_type = info->O->getAnyRelocationType(RE); + if (r_type == MachO::ARM64_RELOC_ADDEND) { + DataRefImpl RelNext = Rel; + info->O->moveRelocationNext(RelNext); + MachO::any_relocation_info RENext = info->O->getRelocation(RelNext); + if (value == 0) { + value = info->O->getPlainRelocationSymbolNum(RENext); + op_info->Value = value; } } - if (reloc_found && isExtern) { - StringRef SymName; - Symbol.getName(SymName); - const char *name = SymName.data(); - op_info->AddSymbol.Present = 1; - op_info->AddSymbol.Name = name; + // NOTE: Scattered relocations don't exist on arm64. + if (!info->O->getPlainRelocationExternal(RE)) + return 0; + StringRef SymName; + Reloc->getSymbol()->getName(SymName); + const char *name = SymName.data(); + op_info->AddSymbol.Present = 1; + op_info->AddSymbol.Name = name; - switch (r_type) { - case MachO::ARM64_RELOC_PAGE21: - /* @page */ - op_info->VariantKind = LLVMDisassembler_VariantKind_ARM64_PAGE; - break; - case MachO::ARM64_RELOC_PAGEOFF12: - /* @pageoff */ - op_info->VariantKind = LLVMDisassembler_VariantKind_ARM64_PAGEOFF; - break; - case MachO::ARM64_RELOC_GOT_LOAD_PAGE21: - /* @gotpage */ - op_info->VariantKind = LLVMDisassembler_VariantKind_ARM64_GOTPAGE; - break; - case MachO::ARM64_RELOC_GOT_LOAD_PAGEOFF12: - /* @gotpageoff */ - op_info->VariantKind = LLVMDisassembler_VariantKind_ARM64_GOTPAGEOFF; - break; - case MachO::ARM64_RELOC_TLVP_LOAD_PAGE21: - /* @tvlppage is not implemented in llvm-mc */ - op_info->VariantKind = LLVMDisassembler_VariantKind_ARM64_TLVP; - break; - case MachO::ARM64_RELOC_TLVP_LOAD_PAGEOFF12: - /* @tvlppageoff is not implemented in llvm-mc */ - op_info->VariantKind = LLVMDisassembler_VariantKind_ARM64_TLVOFF; - break; - default: - case MachO::ARM64_RELOC_BRANCH26: - op_info->VariantKind = LLVMDisassembler_VariantKind_None; - break; - } - return 1; + switch (r_type) { + case MachO::ARM64_RELOC_PAGE21: + /* @page */ + op_info->VariantKind = LLVMDisassembler_VariantKind_ARM64_PAGE; + break; + case MachO::ARM64_RELOC_PAGEOFF12: + /* @pageoff */ + op_info->VariantKind = LLVMDisassembler_VariantKind_ARM64_PAGEOFF; + break; + case MachO::ARM64_RELOC_GOT_LOAD_PAGE21: + /* @gotpage */ + op_info->VariantKind = LLVMDisassembler_VariantKind_ARM64_GOTPAGE; + break; + case MachO::ARM64_RELOC_GOT_LOAD_PAGEOFF12: + /* @gotpageoff */ + op_info->VariantKind = LLVMDisassembler_VariantKind_ARM64_GOTPAGEOFF; + break; + case MachO::ARM64_RELOC_TLVP_LOAD_PAGE21: + /* @tvlppage is not implemented in llvm-mc */ + op_info->VariantKind = LLVMDisassembler_VariantKind_ARM64_TLVP; + break; + case MachO::ARM64_RELOC_TLVP_LOAD_PAGEOFF12: + /* @tvlppageoff is not implemented in llvm-mc */ + op_info->VariantKind = LLVMDisassembler_VariantKind_ARM64_TLVOFF; + break; + default: + case MachO::ARM64_RELOC_BRANCH26: + op_info->VariantKind = LLVMDisassembler_VariantKind_None; + break; } - return 0; - } else { - return 0; + return 1; } + return 0; } // GuessCstringPointer is passed the address of what might be a pointer to a // literal string in a cstring section. If that address is in a cstring section // it returns a pointer to that string. Else it returns nullptr. -const char *GuessCstringPointer(uint64_t ReferenceValue, - struct DisassembleInfo *info) { - uint32_t LoadCommandCount = info->O->getHeader().ncmds; - MachOObjectFile::LoadCommandInfo Load = info->O->getFirstLoadCommandInfo(); - for (unsigned I = 0;; ++I) { +static const char *GuessCstringPointer(uint64_t ReferenceValue, + struct DisassembleInfo *info) { + for (const auto &Load : info->O->load_commands()) { if (Load.C.cmd == MachO::LC_SEGMENT_64) { MachO::segment_command_64 Seg = info->O->getSegment64LoadCommand(Load); for (unsigned J = 0; J < Seg.nsects; ++J) { @@ -1191,10 +2153,6 @@ const char *GuessCstringPointer(uint64_t ReferenceValue, } } } - if (I == LoadCommandCount - 1) - break; - else - Load = info->O->getNextLoadCommandInfo(Load); } return nullptr; } @@ -1205,11 +2163,9 @@ const char *GuessCstringPointer(uint64_t ReferenceValue, // symbol name being referenced by the stub or pointer. static const char *GuessIndirectSymbol(uint64_t ReferenceValue, struct DisassembleInfo *info) { - uint32_t LoadCommandCount = info->O->getHeader().ncmds; - MachOObjectFile::LoadCommandInfo Load = info->O->getFirstLoadCommandInfo(); MachO::dysymtab_command Dysymtab = info->O->getDysymtabLoadCommand(); MachO::symtab_command Symtab = info->O->getSymtabLoadCommand(); - for (unsigned I = 0;; ++I) { + for (const auto &Load : info->O->load_commands()) { if (Load.C.cmd == MachO::LC_SEGMENT_64) { MachO::segment_command_64 Seg = info->O->getSegment64LoadCommand(Load); for (unsigned J = 0; J < Seg.nsects; ++J) { @@ -1279,10 +2235,6 @@ static const char *GuessIndirectSymbol(uint64_t ReferenceValue, } } } - if (I == LoadCommandCount - 1) - break; - else - Load = info->O->getNextLoadCommandInfo(Load); } return nullptr; } @@ -1369,9 +2321,7 @@ static uint64_t GuessPointerPointer(uint64_t ReferenceValue, selref = false; msgref = false; cfstring = false; - uint32_t LoadCommandCount = info->O->getHeader().ncmds; - MachOObjectFile::LoadCommandInfo Load = info->O->getFirstLoadCommandInfo(); - for (unsigned I = 0;; ++I) { + for (const auto &Load : info->O->load_commands()) { if (Load.C.cmd == MachO::LC_SEGMENT_64) { MachO::segment_command_64 Seg = info->O->getSegment64LoadCommand(Load); for (unsigned J = 0; J < Seg.nsects; ++J) { @@ -1416,10 +2366,6 @@ static uint64_t GuessPointerPointer(uint64_t ReferenceValue, } } // TODO: Look for LC_SEGMENT for 32-bit Mach-O files. - if (I == LoadCommandCount - 1) - break; - else - Load = info->O->getNextLoadCommandInfo(Load); } return 0; } @@ -1429,14 +2375,24 @@ static uint64_t GuessPointerPointer(uint64_t ReferenceValue, // offset into the section, number of bytes left in the section past the offset // and which section is was being referenced. If the Address is not in a // section nullptr is returned. -const char *get_pointer_64(uint64_t Address, uint32_t &offset, uint32_t &left, - SectionRef &S, DisassembleInfo *info) { +static const char *get_pointer_64(uint64_t Address, uint32_t &offset, + uint32_t &left, SectionRef &S, + DisassembleInfo *info, + bool objc_only = false) { offset = 0; left = 0; S = SectionRef(); for (unsigned SectIdx = 0; SectIdx != info->Sections->size(); SectIdx++) { uint64_t SectAddress = ((*(info->Sections))[SectIdx]).getAddress(); uint64_t SectSize = ((*(info->Sections))[SectIdx]).getSize(); + if (objc_only) { + StringRef SectName; + ((*(info->Sections))[SectIdx]).getName(SectName); + DataRefImpl Ref = ((*(info->Sections))[SectIdx]).getRawDataRefImpl(); + StringRef SegName = info->O->getSectionFinalSegmentName(Ref); + if (SegName != "__OBJC" && SectName != "__cstring") + continue; + } if (Address >= SectAddress && Address < SectAddress + SectSize) { S = (*(info->Sections))[SectIdx]; offset = Address - SectAddress; @@ -1449,13 +2405,24 @@ const char *get_pointer_64(uint64_t Address, uint32_t &offset, uint32_t &left, return nullptr; } +static const char *get_pointer_32(uint32_t Address, uint32_t &offset, + uint32_t &left, SectionRef &S, + DisassembleInfo *info, + bool objc_only = false) { + return get_pointer_64(Address, offset, left, S, info, objc_only); +} + // get_symbol_64() returns the name of a symbol (or nullptr) and the address of // the symbol indirectly through n_value. Based on the relocation information // for the specified section offset in the specified section reference. -const char *get_symbol_64(uint32_t sect_offset, SectionRef S, - DisassembleInfo *info, uint64_t &n_value) { +// If no relocation information is found and a non-zero ReferenceValue for the +// symbol is passed, look up that address in the info's AddrMap. +static const char * +get_symbol_64(uint32_t sect_offset, SectionRef S, DisassembleInfo *info, + uint64_t &n_value, + uint64_t ReferenceValue = UnknownAddressOrSize) { n_value = 0; - if (info->verbose == false) + if (!info->verbose) return nullptr; // See if there is an external relocation entry at the sect_offset. @@ -1487,6 +2454,8 @@ const char *get_symbol_64(uint32_t sect_offset, SectionRef S, const char *SymbolName = nullptr; if (reloc_found && isExtern) { Symbol.getAddress(n_value); + if (n_value == UnknownAddressOrSize) + n_value = 0; StringRef name; Symbol.getName(name); if (!name.empty()) { @@ -1504,17 +2473,21 @@ const char *get_symbol_64(uint32_t sect_offset, SectionRef S, // // NOTE: need add passing the database_offset to this routine. - // TODO: We did not find an external relocation entry so look up the - // ReferenceValue as an address of a symbol and if found return that symbol's - // name. - // - // NOTE: need add passing the ReferenceValue to this routine. Then that code - // would simply be this: - // SymbolName = GuessSymbolName(ReferenceValue, info); + // We did not find an external relocation entry so look up the ReferenceValue + // as an address of a symbol and if found return that symbol's name. + if (ReferenceValue != UnknownAddressOrSize) + SymbolName = GuessSymbolName(ReferenceValue, info->AddrMap); return SymbolName; } +static const char *get_symbol_32(uint32_t sect_offset, SectionRef S, + DisassembleInfo *info, + uint32_t ReferenceValue) { + uint64_t n_value64; + return get_symbol_64(sect_offset, S, info, n_value64, ReferenceValue); +} + // These are structs in the Objective-C meta data and read to produce the // comments for disassembly. While these are part of the ABI they are no // public defintions. So the are here not in include/llvm/Support/MachO.h . @@ -1536,6 +2509,14 @@ struct class64_t { uint64_t data; // class_ro64_t * (64-bit pointer) }; +struct class32_t { + uint32_t isa; /* class32_t * (32-bit pointer) */ + uint32_t superclass; /* class32_t * (32-bit pointer) */ + uint32_t cache; /* Cache (32-bit pointer) */ + uint32_t vtable; /* IMP * (32-bit pointer) */ + uint32_t data; /* class_ro32_t * (32-bit pointer) */ +}; + struct class_ro64_t { uint32_t flags; uint32_t instanceStart; @@ -1550,6 +2531,276 @@ struct class_ro64_t { uint64_t baseProperties; // const struct objc_property_list (64-bit pointer) }; +struct class_ro32_t { + uint32_t flags; + uint32_t instanceStart; + uint32_t instanceSize; + uint32_t ivarLayout; /* const uint8_t * (32-bit pointer) */ + uint32_t name; /* const char * (32-bit pointer) */ + uint32_t baseMethods; /* const method_list_t * (32-bit pointer) */ + uint32_t baseProtocols; /* const protocol_list_t * (32-bit pointer) */ + uint32_t ivars; /* const ivar_list_t * (32-bit pointer) */ + uint32_t weakIvarLayout; /* const uint8_t * (32-bit pointer) */ + uint32_t baseProperties; /* const struct objc_property_list * + (32-bit pointer) */ +}; + +/* Values for class_ro{64,32}_t->flags */ +#define RO_META (1 << 0) +#define RO_ROOT (1 << 1) +#define RO_HAS_CXX_STRUCTORS (1 << 2) + +struct method_list64_t { + uint32_t entsize; + uint32_t count; + /* struct method64_t first; These structures follow inline */ +}; + +struct method_list32_t { + uint32_t entsize; + uint32_t count; + /* struct method32_t first; These structures follow inline */ +}; + +struct method64_t { + uint64_t name; /* SEL (64-bit pointer) */ + uint64_t types; /* const char * (64-bit pointer) */ + uint64_t imp; /* IMP (64-bit pointer) */ +}; + +struct method32_t { + uint32_t name; /* SEL (32-bit pointer) */ + uint32_t types; /* const char * (32-bit pointer) */ + uint32_t imp; /* IMP (32-bit pointer) */ +}; + +struct protocol_list64_t { + uint64_t count; /* uintptr_t (a 64-bit value) */ + /* struct protocol64_t * list[0]; These pointers follow inline */ +}; + +struct protocol_list32_t { + uint32_t count; /* uintptr_t (a 32-bit value) */ + /* struct protocol32_t * list[0]; These pointers follow inline */ +}; + +struct protocol64_t { + uint64_t isa; /* id * (64-bit pointer) */ + uint64_t name; /* const char * (64-bit pointer) */ + uint64_t protocols; /* struct protocol_list64_t * + (64-bit pointer) */ + uint64_t instanceMethods; /* method_list_t * (64-bit pointer) */ + uint64_t classMethods; /* method_list_t * (64-bit pointer) */ + uint64_t optionalInstanceMethods; /* method_list_t * (64-bit pointer) */ + uint64_t optionalClassMethods; /* method_list_t * (64-bit pointer) */ + uint64_t instanceProperties; /* struct objc_property_list * + (64-bit pointer) */ +}; + +struct protocol32_t { + uint32_t isa; /* id * (32-bit pointer) */ + uint32_t name; /* const char * (32-bit pointer) */ + uint32_t protocols; /* struct protocol_list_t * + (32-bit pointer) */ + uint32_t instanceMethods; /* method_list_t * (32-bit pointer) */ + uint32_t classMethods; /* method_list_t * (32-bit pointer) */ + uint32_t optionalInstanceMethods; /* method_list_t * (32-bit pointer) */ + uint32_t optionalClassMethods; /* method_list_t * (32-bit pointer) */ + uint32_t instanceProperties; /* struct objc_property_list * + (32-bit pointer) */ +}; + +struct ivar_list64_t { + uint32_t entsize; + uint32_t count; + /* struct ivar64_t first; These structures follow inline */ +}; + +struct ivar_list32_t { + uint32_t entsize; + uint32_t count; + /* struct ivar32_t first; These structures follow inline */ +}; + +struct ivar64_t { + uint64_t offset; /* uintptr_t * (64-bit pointer) */ + uint64_t name; /* const char * (64-bit pointer) */ + uint64_t type; /* const char * (64-bit pointer) */ + uint32_t alignment; + uint32_t size; +}; + +struct ivar32_t { + uint32_t offset; /* uintptr_t * (32-bit pointer) */ + uint32_t name; /* const char * (32-bit pointer) */ + uint32_t type; /* const char * (32-bit pointer) */ + uint32_t alignment; + uint32_t size; +}; + +struct objc_property_list64 { + uint32_t entsize; + uint32_t count; + /* struct objc_property64 first; These structures follow inline */ +}; + +struct objc_property_list32 { + uint32_t entsize; + uint32_t count; + /* struct objc_property32 first; These structures follow inline */ +}; + +struct objc_property64 { + uint64_t name; /* const char * (64-bit pointer) */ + uint64_t attributes; /* const char * (64-bit pointer) */ +}; + +struct objc_property32 { + uint32_t name; /* const char * (32-bit pointer) */ + uint32_t attributes; /* const char * (32-bit pointer) */ +}; + +struct category64_t { + uint64_t name; /* const char * (64-bit pointer) */ + uint64_t cls; /* struct class_t * (64-bit pointer) */ + uint64_t instanceMethods; /* struct method_list_t * (64-bit pointer) */ + uint64_t classMethods; /* struct method_list_t * (64-bit pointer) */ + uint64_t protocols; /* struct protocol_list_t * (64-bit pointer) */ + uint64_t instanceProperties; /* struct objc_property_list * + (64-bit pointer) */ +}; + +struct category32_t { + uint32_t name; /* const char * (32-bit pointer) */ + uint32_t cls; /* struct class_t * (32-bit pointer) */ + uint32_t instanceMethods; /* struct method_list_t * (32-bit pointer) */ + uint32_t classMethods; /* struct method_list_t * (32-bit pointer) */ + uint32_t protocols; /* struct protocol_list_t * (32-bit pointer) */ + uint32_t instanceProperties; /* struct objc_property_list * + (32-bit pointer) */ +}; + +struct objc_image_info64 { + uint32_t version; + uint32_t flags; +}; +struct objc_image_info32 { + uint32_t version; + uint32_t flags; +}; +struct imageInfo_t { + uint32_t version; + uint32_t flags; +}; +/* masks for objc_image_info.flags */ +#define OBJC_IMAGE_IS_REPLACEMENT (1 << 0) +#define OBJC_IMAGE_SUPPORTS_GC (1 << 1) + +struct message_ref64 { + uint64_t imp; /* IMP (64-bit pointer) */ + uint64_t sel; /* SEL (64-bit pointer) */ +}; + +struct message_ref32 { + uint32_t imp; /* IMP (32-bit pointer) */ + uint32_t sel; /* SEL (32-bit pointer) */ +}; + +// Objective-C 1 (32-bit only) meta data structs. + +struct objc_module_t { + uint32_t version; + uint32_t size; + uint32_t name; /* char * (32-bit pointer) */ + uint32_t symtab; /* struct objc_symtab * (32-bit pointer) */ +}; + +struct objc_symtab_t { + uint32_t sel_ref_cnt; + uint32_t refs; /* SEL * (32-bit pointer) */ + uint16_t cls_def_cnt; + uint16_t cat_def_cnt; + // uint32_t defs[1]; /* void * (32-bit pointer) variable size */ +}; + +struct objc_class_t { + uint32_t isa; /* struct objc_class * (32-bit pointer) */ + uint32_t super_class; /* struct objc_class * (32-bit pointer) */ + uint32_t name; /* const char * (32-bit pointer) */ + int32_t version; + int32_t info; + int32_t instance_size; + uint32_t ivars; /* struct objc_ivar_list * (32-bit pointer) */ + uint32_t methodLists; /* struct objc_method_list ** (32-bit pointer) */ + uint32_t cache; /* struct objc_cache * (32-bit pointer) */ + uint32_t protocols; /* struct objc_protocol_list * (32-bit pointer) */ +}; + +#define CLS_GETINFO(cls, infomask) ((cls)->info & (infomask)) +// class is not a metaclass +#define CLS_CLASS 0x1 +// class is a metaclass +#define CLS_META 0x2 + +struct objc_category_t { + uint32_t category_name; /* char * (32-bit pointer) */ + uint32_t class_name; /* char * (32-bit pointer) */ + uint32_t instance_methods; /* struct objc_method_list * (32-bit pointer) */ + uint32_t class_methods; /* struct objc_method_list * (32-bit pointer) */ + uint32_t protocols; /* struct objc_protocol_list * (32-bit ptr) */ +}; + +struct objc_ivar_t { + uint32_t ivar_name; /* char * (32-bit pointer) */ + uint32_t ivar_type; /* char * (32-bit pointer) */ + int32_t ivar_offset; +}; + +struct objc_ivar_list_t { + int32_t ivar_count; + // struct objc_ivar_t ivar_list[1]; /* variable length structure */ +}; + +struct objc_method_list_t { + uint32_t obsolete; /* struct objc_method_list * (32-bit pointer) */ + int32_t method_count; + // struct objc_method_t method_list[1]; /* variable length structure */ +}; + +struct objc_method_t { + uint32_t method_name; /* SEL, aka struct objc_selector * (32-bit pointer) */ + uint32_t method_types; /* char * (32-bit pointer) */ + uint32_t method_imp; /* IMP, aka function pointer, (*IMP)(id, SEL, ...) + (32-bit pointer) */ +}; + +struct objc_protocol_list_t { + uint32_t next; /* struct objc_protocol_list * (32-bit pointer) */ + int32_t count; + // uint32_t list[1]; /* Protocol *, aka struct objc_protocol_t * + // (32-bit pointer) */ +}; + +struct objc_protocol_t { + uint32_t isa; /* struct objc_class * (32-bit pointer) */ + uint32_t protocol_name; /* char * (32-bit pointer) */ + uint32_t protocol_list; /* struct objc_protocol_list * (32-bit pointer) */ + uint32_t instance_methods; /* struct objc_method_description_list * + (32-bit pointer) */ + uint32_t class_methods; /* struct objc_method_description_list * + (32-bit pointer) */ +}; + +struct objc_method_description_list_t { + int32_t count; + // struct objc_method_description_t list[1]; +}; + +struct objc_method_description_t { + uint32_t name; /* SEL, aka struct objc_selector * (32-bit pointer) */ + uint32_t types; /* char * (32-bit pointer) */ +}; + inline void swapStruct(struct cfstring64_t &cfs) { sys::swapByteOrder(cfs.isa); sys::swapByteOrder(cfs.flags); @@ -1565,6 +2816,14 @@ inline void swapStruct(struct class64_t &c) { sys::swapByteOrder(c.data); } +inline void swapStruct(struct class32_t &c) { + sys::swapByteOrder(c.isa); + sys::swapByteOrder(c.superclass); + sys::swapByteOrder(c.cache); + sys::swapByteOrder(c.vtable); + sys::swapByteOrder(c.data); +} + inline void swapStruct(struct class_ro64_t &cro) { sys::swapByteOrder(cro.flags); sys::swapByteOrder(cro.instanceStart); @@ -1579,6 +2838,238 @@ inline void swapStruct(struct class_ro64_t &cro) { sys::swapByteOrder(cro.baseProperties); } +inline void swapStruct(struct class_ro32_t &cro) { + sys::swapByteOrder(cro.flags); + sys::swapByteOrder(cro.instanceStart); + sys::swapByteOrder(cro.instanceSize); + sys::swapByteOrder(cro.ivarLayout); + sys::swapByteOrder(cro.name); + sys::swapByteOrder(cro.baseMethods); + sys::swapByteOrder(cro.baseProtocols); + sys::swapByteOrder(cro.ivars); + sys::swapByteOrder(cro.weakIvarLayout); + sys::swapByteOrder(cro.baseProperties); +} + +inline void swapStruct(struct method_list64_t &ml) { + sys::swapByteOrder(ml.entsize); + sys::swapByteOrder(ml.count); +} + +inline void swapStruct(struct method_list32_t &ml) { + sys::swapByteOrder(ml.entsize); + sys::swapByteOrder(ml.count); +} + +inline void swapStruct(struct method64_t &m) { + sys::swapByteOrder(m.name); + sys::swapByteOrder(m.types); + sys::swapByteOrder(m.imp); +} + +inline void swapStruct(struct method32_t &m) { + sys::swapByteOrder(m.name); + sys::swapByteOrder(m.types); + sys::swapByteOrder(m.imp); +} + +inline void swapStruct(struct protocol_list64_t &pl) { + sys::swapByteOrder(pl.count); +} + +inline void swapStruct(struct protocol_list32_t &pl) { + sys::swapByteOrder(pl.count); +} + +inline void swapStruct(struct protocol64_t &p) { + sys::swapByteOrder(p.isa); + sys::swapByteOrder(p.name); + sys::swapByteOrder(p.protocols); + sys::swapByteOrder(p.instanceMethods); + sys::swapByteOrder(p.classMethods); + sys::swapByteOrder(p.optionalInstanceMethods); + sys::swapByteOrder(p.optionalClassMethods); + sys::swapByteOrder(p.instanceProperties); +} + +inline void swapStruct(struct protocol32_t &p) { + sys::swapByteOrder(p.isa); + sys::swapByteOrder(p.name); + sys::swapByteOrder(p.protocols); + sys::swapByteOrder(p.instanceMethods); + sys::swapByteOrder(p.classMethods); + sys::swapByteOrder(p.optionalInstanceMethods); + sys::swapByteOrder(p.optionalClassMethods); + sys::swapByteOrder(p.instanceProperties); +} + +inline void swapStruct(struct ivar_list64_t &il) { + sys::swapByteOrder(il.entsize); + sys::swapByteOrder(il.count); +} + +inline void swapStruct(struct ivar_list32_t &il) { + sys::swapByteOrder(il.entsize); + sys::swapByteOrder(il.count); +} + +inline void swapStruct(struct ivar64_t &i) { + sys::swapByteOrder(i.offset); + sys::swapByteOrder(i.name); + sys::swapByteOrder(i.type); + sys::swapByteOrder(i.alignment); + sys::swapByteOrder(i.size); +} + +inline void swapStruct(struct ivar32_t &i) { + sys::swapByteOrder(i.offset); + sys::swapByteOrder(i.name); + sys::swapByteOrder(i.type); + sys::swapByteOrder(i.alignment); + sys::swapByteOrder(i.size); +} + +inline void swapStruct(struct objc_property_list64 &pl) { + sys::swapByteOrder(pl.entsize); + sys::swapByteOrder(pl.count); +} + +inline void swapStruct(struct objc_property_list32 &pl) { + sys::swapByteOrder(pl.entsize); + sys::swapByteOrder(pl.count); +} + +inline void swapStruct(struct objc_property64 &op) { + sys::swapByteOrder(op.name); + sys::swapByteOrder(op.attributes); +} + +inline void swapStruct(struct objc_property32 &op) { + sys::swapByteOrder(op.name); + sys::swapByteOrder(op.attributes); +} + +inline void swapStruct(struct category64_t &c) { + sys::swapByteOrder(c.name); + sys::swapByteOrder(c.cls); + sys::swapByteOrder(c.instanceMethods); + sys::swapByteOrder(c.classMethods); + sys::swapByteOrder(c.protocols); + sys::swapByteOrder(c.instanceProperties); +} + +inline void swapStruct(struct category32_t &c) { + sys::swapByteOrder(c.name); + sys::swapByteOrder(c.cls); + sys::swapByteOrder(c.instanceMethods); + sys::swapByteOrder(c.classMethods); + sys::swapByteOrder(c.protocols); + sys::swapByteOrder(c.instanceProperties); +} + +inline void swapStruct(struct objc_image_info64 &o) { + sys::swapByteOrder(o.version); + sys::swapByteOrder(o.flags); +} + +inline void swapStruct(struct objc_image_info32 &o) { + sys::swapByteOrder(o.version); + sys::swapByteOrder(o.flags); +} + +inline void swapStruct(struct imageInfo_t &o) { + sys::swapByteOrder(o.version); + sys::swapByteOrder(o.flags); +} + +inline void swapStruct(struct message_ref64 &mr) { + sys::swapByteOrder(mr.imp); + sys::swapByteOrder(mr.sel); +} + +inline void swapStruct(struct message_ref32 &mr) { + sys::swapByteOrder(mr.imp); + sys::swapByteOrder(mr.sel); +} + +inline void swapStruct(struct objc_module_t &module) { + sys::swapByteOrder(module.version); + sys::swapByteOrder(module.size); + sys::swapByteOrder(module.name); + sys::swapByteOrder(module.symtab); +} + +inline void swapStruct(struct objc_symtab_t &symtab) { + sys::swapByteOrder(symtab.sel_ref_cnt); + sys::swapByteOrder(symtab.refs); + sys::swapByteOrder(symtab.cls_def_cnt); + sys::swapByteOrder(symtab.cat_def_cnt); +} + +inline void swapStruct(struct objc_class_t &objc_class) { + sys::swapByteOrder(objc_class.isa); + sys::swapByteOrder(objc_class.super_class); + sys::swapByteOrder(objc_class.name); + sys::swapByteOrder(objc_class.version); + sys::swapByteOrder(objc_class.info); + sys::swapByteOrder(objc_class.instance_size); + sys::swapByteOrder(objc_class.ivars); + sys::swapByteOrder(objc_class.methodLists); + sys::swapByteOrder(objc_class.cache); + sys::swapByteOrder(objc_class.protocols); +} + +inline void swapStruct(struct objc_category_t &objc_category) { + sys::swapByteOrder(objc_category.category_name); + sys::swapByteOrder(objc_category.class_name); + sys::swapByteOrder(objc_category.instance_methods); + sys::swapByteOrder(objc_category.class_methods); + sys::swapByteOrder(objc_category.protocols); +} + +inline void swapStruct(struct objc_ivar_list_t &objc_ivar_list) { + sys::swapByteOrder(objc_ivar_list.ivar_count); +} + +inline void swapStruct(struct objc_ivar_t &objc_ivar) { + sys::swapByteOrder(objc_ivar.ivar_name); + sys::swapByteOrder(objc_ivar.ivar_type); + sys::swapByteOrder(objc_ivar.ivar_offset); +} + +inline void swapStruct(struct objc_method_list_t &method_list) { + sys::swapByteOrder(method_list.obsolete); + sys::swapByteOrder(method_list.method_count); +} + +inline void swapStruct(struct objc_method_t &method) { + sys::swapByteOrder(method.method_name); + sys::swapByteOrder(method.method_types); + sys::swapByteOrder(method.method_imp); +} + +inline void swapStruct(struct objc_protocol_list_t &protocol_list) { + sys::swapByteOrder(protocol_list.next); + sys::swapByteOrder(protocol_list.count); +} + +inline void swapStruct(struct objc_protocol_t &protocol) { + sys::swapByteOrder(protocol.isa); + sys::swapByteOrder(protocol.protocol_name); + sys::swapByteOrder(protocol.protocol_list); + sys::swapByteOrder(protocol.instance_methods); + sys::swapByteOrder(protocol.class_methods); +} + +inline void swapStruct(struct objc_method_description_list_t &mdl) { + sys::swapByteOrder(mdl.count); +} + +inline void swapStruct(struct objc_method_description_t &md) { + sys::swapByteOrder(md.name); + sys::swapByteOrder(md.types); +} + static const char *get_dyld_bind_info_symbolname(uint64_t ReferenceValue, struct DisassembleInfo *info); @@ -1587,9 +3078,9 @@ static const char *get_dyld_bind_info_symbolname(uint64_t ReferenceValue, // address of the pointer, so when the pointer is zero as it can be in an .o // file, that is used to look for an external relocation entry with a symbol // name. -const char *get_objc2_64bit_class_name(uint64_t pointer_value, - uint64_t ReferenceValue, - struct DisassembleInfo *info) { +static const char *get_objc2_64bit_class_name(uint64_t pointer_value, + uint64_t ReferenceValue, + struct DisassembleInfo *info) { const char *r; uint32_t offset, left; SectionRef S; @@ -1638,8 +3129,8 @@ const char *get_objc2_64bit_class_name(uint64_t pointer_value, // get_objc2_64bit_cfstring_name is used for disassembly and is passed a // pointer to a cfstring and returns its name or nullptr. -const char *get_objc2_64bit_cfstring_name(uint64_t ReferenceValue, - struct DisassembleInfo *info) { +static const char *get_objc2_64bit_cfstring_name(uint64_t ReferenceValue, + struct DisassembleInfo *info) { const char *r, *name; uint32_t offset, left; SectionRef S; @@ -1672,8 +3163,8 @@ const char *get_objc2_64bit_cfstring_name(uint64_t ReferenceValue, // who's symbol's n_value is the real pointer to the selector name. If that is // the case the real pointer to the selector name is returned else 0 is // returned -uint64_t get_objc2_64bit_selref(uint64_t ReferenceValue, - struct DisassembleInfo *info) { +static uint64_t get_objc2_64bit_selref(uint64_t ReferenceValue, + struct DisassembleInfo *info) { uint32_t offset, left; SectionRef S; @@ -1687,6 +3178,2411 @@ uint64_t get_objc2_64bit_selref(uint64_t ReferenceValue, return n_value; } +static const SectionRef get_section(MachOObjectFile *O, const char *segname, + const char *sectname) { + for (const SectionRef &Section : O->sections()) { + StringRef SectName; + Section.getName(SectName); + DataRefImpl Ref = Section.getRawDataRefImpl(); + StringRef SegName = O->getSectionFinalSegmentName(Ref); + if (SegName == segname && SectName == sectname) + return Section; + } + return SectionRef(); +} + +static void +walk_pointer_list_64(const char *listname, const SectionRef S, + MachOObjectFile *O, struct DisassembleInfo *info, + void (*func)(uint64_t, struct DisassembleInfo *info)) { + if (S == SectionRef()) + return; + + StringRef SectName; + S.getName(SectName); + DataRefImpl Ref = S.getRawDataRefImpl(); + StringRef SegName = O->getSectionFinalSegmentName(Ref); + outs() << "Contents of (" << SegName << "," << SectName << ") section\n"; + + StringRef BytesStr; + S.getContents(BytesStr); + const char *Contents = reinterpret_cast<const char *>(BytesStr.data()); + + for (uint32_t i = 0; i < S.getSize(); i += sizeof(uint64_t)) { + uint32_t left = S.getSize() - i; + uint32_t size = left < sizeof(uint64_t) ? left : sizeof(uint64_t); + uint64_t p = 0; + memcpy(&p, Contents + i, size); + if (i + sizeof(uint64_t) > S.getSize()) + outs() << listname << " list pointer extends past end of (" << SegName + << "," << SectName << ") section\n"; + outs() << format("%016" PRIx64, S.getAddress() + i) << " "; + + if (O->isLittleEndian() != sys::IsLittleEndianHost) + sys::swapByteOrder(p); + + uint64_t n_value = 0; + const char *name = get_symbol_64(i, S, info, n_value, p); + if (name == nullptr) + name = get_dyld_bind_info_symbolname(S.getAddress() + i, info); + + if (n_value != 0) { + outs() << format("0x%" PRIx64, n_value); + if (p != 0) + outs() << " + " << format("0x%" PRIx64, p); + } else + outs() << format("0x%" PRIx64, p); + if (name != nullptr) + outs() << " " << name; + outs() << "\n"; + + p += n_value; + if (func) + func(p, info); + } +} + +static void +walk_pointer_list_32(const char *listname, const SectionRef S, + MachOObjectFile *O, struct DisassembleInfo *info, + void (*func)(uint32_t, struct DisassembleInfo *info)) { + if (S == SectionRef()) + return; + + StringRef SectName; + S.getName(SectName); + DataRefImpl Ref = S.getRawDataRefImpl(); + StringRef SegName = O->getSectionFinalSegmentName(Ref); + outs() << "Contents of (" << SegName << "," << SectName << ") section\n"; + + StringRef BytesStr; + S.getContents(BytesStr); + const char *Contents = reinterpret_cast<const char *>(BytesStr.data()); + + for (uint32_t i = 0; i < S.getSize(); i += sizeof(uint32_t)) { + uint32_t left = S.getSize() - i; + uint32_t size = left < sizeof(uint32_t) ? left : sizeof(uint32_t); + uint32_t p = 0; + memcpy(&p, Contents + i, size); + if (i + sizeof(uint32_t) > S.getSize()) + outs() << listname << " list pointer extends past end of (" << SegName + << "," << SectName << ") section\n"; + uint32_t Address = S.getAddress() + i; + outs() << format("%08" PRIx32, Address) << " "; + + if (O->isLittleEndian() != sys::IsLittleEndianHost) + sys::swapByteOrder(p); + outs() << format("0x%" PRIx32, p); + + const char *name = get_symbol_32(i, S, info, p); + if (name != nullptr) + outs() << " " << name; + outs() << "\n"; + + if (func) + func(p, info); + } +} + +static void print_layout_map(const char *layout_map, uint32_t left) { + outs() << " layout map: "; + do { + outs() << format("0x%02" PRIx32, (*layout_map) & 0xff) << " "; + left--; + layout_map++; + } while (*layout_map != '\0' && left != 0); + outs() << "\n"; +} + +static void print_layout_map64(uint64_t p, struct DisassembleInfo *info) { + uint32_t offset, left; + SectionRef S; + const char *layout_map; + + if (p == 0) + return; + layout_map = get_pointer_64(p, offset, left, S, info); + print_layout_map(layout_map, left); +} + +static void print_layout_map32(uint32_t p, struct DisassembleInfo *info) { + uint32_t offset, left; + SectionRef S; + const char *layout_map; + + if (p == 0) + return; + layout_map = get_pointer_32(p, offset, left, S, info); + print_layout_map(layout_map, left); +} + +static void print_method_list64_t(uint64_t p, struct DisassembleInfo *info, + const char *indent) { + struct method_list64_t ml; + struct method64_t m; + const char *r; + uint32_t offset, xoffset, left, i; + SectionRef S, xS; + const char *name, *sym_name; + uint64_t n_value; + + r = get_pointer_64(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&ml, '\0', sizeof(struct method_list64_t)); + if (left < sizeof(struct method_list64_t)) { + memcpy(&ml, r, left); + outs() << " (method_list_t entends past the end of the section)\n"; + } else + memcpy(&ml, r, sizeof(struct method_list64_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(ml); + outs() << indent << "\t\t entsize " << ml.entsize << "\n"; + outs() << indent << "\t\t count " << ml.count << "\n"; + + p += sizeof(struct method_list64_t); + offset += sizeof(struct method_list64_t); + for (i = 0; i < ml.count; i++) { + r = get_pointer_64(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&m, '\0', sizeof(struct method64_t)); + if (left < sizeof(struct method64_t)) { + memcpy(&ml, r, left); + outs() << indent << " (method_t entends past the end of the section)\n"; + } else + memcpy(&m, r, sizeof(struct method64_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(m); + + outs() << indent << "\t\t name "; + sym_name = get_symbol_64(offset + offsetof(struct method64_t, name), S, + info, n_value, m.name); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (m.name != 0) + outs() << " + " << format("0x%" PRIx64, m.name); + } else + outs() << format("0x%" PRIx64, m.name); + name = get_pointer_64(m.name + n_value, xoffset, left, xS, info); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + + outs() << indent << "\t\t types "; + sym_name = get_symbol_64(offset + offsetof(struct method64_t, types), S, + info, n_value, m.types); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (m.types != 0) + outs() << " + " << format("0x%" PRIx64, m.types); + } else + outs() << format("0x%" PRIx64, m.types); + name = get_pointer_64(m.types + n_value, xoffset, left, xS, info); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + + outs() << indent << "\t\t imp "; + name = get_symbol_64(offset + offsetof(struct method64_t, imp), S, info, + n_value, m.imp); + if (info->verbose && name == nullptr) { + if (n_value != 0) { + outs() << format("0x%" PRIx64, n_value) << " "; + if (m.imp != 0) + outs() << "+ " << format("0x%" PRIx64, m.imp) << " "; + } else + outs() << format("0x%" PRIx64, m.imp) << " "; + } + if (name != nullptr) + outs() << name; + outs() << "\n"; + + p += sizeof(struct method64_t); + offset += sizeof(struct method64_t); + } +} + +static void print_method_list32_t(uint64_t p, struct DisassembleInfo *info, + const char *indent) { + struct method_list32_t ml; + struct method32_t m; + const char *r, *name; + uint32_t offset, xoffset, left, i; + SectionRef S, xS; + + r = get_pointer_32(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&ml, '\0', sizeof(struct method_list32_t)); + if (left < sizeof(struct method_list32_t)) { + memcpy(&ml, r, left); + outs() << " (method_list_t entends past the end of the section)\n"; + } else + memcpy(&ml, r, sizeof(struct method_list32_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(ml); + outs() << indent << "\t\t entsize " << ml.entsize << "\n"; + outs() << indent << "\t\t count " << ml.count << "\n"; + + p += sizeof(struct method_list32_t); + offset += sizeof(struct method_list32_t); + for (i = 0; i < ml.count; i++) { + r = get_pointer_32(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&m, '\0', sizeof(struct method32_t)); + if (left < sizeof(struct method32_t)) { + memcpy(&ml, r, left); + outs() << indent << " (method_t entends past the end of the section)\n"; + } else + memcpy(&m, r, sizeof(struct method32_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(m); + + outs() << indent << "\t\t name " << format("0x%" PRIx32, m.name); + name = get_pointer_32(m.name, xoffset, left, xS, info); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + + outs() << indent << "\t\t types " << format("0x%" PRIx32, m.types); + name = get_pointer_32(m.types, xoffset, left, xS, info); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + + outs() << indent << "\t\t imp " << format("0x%" PRIx32, m.imp); + name = get_symbol_32(offset + offsetof(struct method32_t, imp), S, info, + m.imp); + if (name != nullptr) + outs() << " " << name; + outs() << "\n"; + + p += sizeof(struct method32_t); + offset += sizeof(struct method32_t); + } +} + +static bool print_method_list(uint32_t p, struct DisassembleInfo *info) { + uint32_t offset, left, xleft; + SectionRef S; + struct objc_method_list_t method_list; + struct objc_method_t method; + const char *r, *methods, *name, *SymbolName; + int32_t i; + + r = get_pointer_32(p, offset, left, S, info, true); + if (r == nullptr) + return true; + + outs() << "\n"; + if (left > sizeof(struct objc_method_list_t)) { + memcpy(&method_list, r, sizeof(struct objc_method_list_t)); + } else { + outs() << "\t\t objc_method_list extends past end of the section\n"; + memset(&method_list, '\0', sizeof(struct objc_method_list_t)); + memcpy(&method_list, r, left); + } + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(method_list); + + outs() << "\t\t obsolete " + << format("0x%08" PRIx32, method_list.obsolete) << "\n"; + outs() << "\t\t method_count " << method_list.method_count << "\n"; + + methods = r + sizeof(struct objc_method_list_t); + for (i = 0; i < method_list.method_count; i++) { + if ((i + 1) * sizeof(struct objc_method_t) > left) { + outs() << "\t\t remaining method's extend past the of the section\n"; + break; + } + memcpy(&method, methods + i * sizeof(struct objc_method_t), + sizeof(struct objc_method_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(method); + + outs() << "\t\t method_name " + << format("0x%08" PRIx32, method.method_name); + if (info->verbose) { + name = get_pointer_32(method.method_name, offset, xleft, S, info, true); + if (name != nullptr) + outs() << format(" %.*s", xleft, name); + else + outs() << " (not in an __OBJC section)"; + } + outs() << "\n"; + + outs() << "\t\t method_types " + << format("0x%08" PRIx32, method.method_types); + if (info->verbose) { + name = get_pointer_32(method.method_types, offset, xleft, S, info, true); + if (name != nullptr) + outs() << format(" %.*s", xleft, name); + else + outs() << " (not in an __OBJC section)"; + } + outs() << "\n"; + + outs() << "\t\t method_imp " + << format("0x%08" PRIx32, method.method_imp) << " "; + if (info->verbose) { + SymbolName = GuessSymbolName(method.method_imp, info->AddrMap); + if (SymbolName != nullptr) + outs() << SymbolName; + } + outs() << "\n"; + } + return false; +} + +static void print_protocol_list64_t(uint64_t p, struct DisassembleInfo *info) { + struct protocol_list64_t pl; + uint64_t q, n_value; + struct protocol64_t pc; + const char *r; + uint32_t offset, xoffset, left, i; + SectionRef S, xS; + const char *name, *sym_name; + + r = get_pointer_64(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&pl, '\0', sizeof(struct protocol_list64_t)); + if (left < sizeof(struct protocol_list64_t)) { + memcpy(&pl, r, left); + outs() << " (protocol_list_t entends past the end of the section)\n"; + } else + memcpy(&pl, r, sizeof(struct protocol_list64_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(pl); + outs() << " count " << pl.count << "\n"; + + p += sizeof(struct protocol_list64_t); + offset += sizeof(struct protocol_list64_t); + for (i = 0; i < pl.count; i++) { + r = get_pointer_64(p, offset, left, S, info); + if (r == nullptr) + return; + q = 0; + if (left < sizeof(uint64_t)) { + memcpy(&q, r, left); + outs() << " (protocol_t * entends past the end of the section)\n"; + } else + memcpy(&q, r, sizeof(uint64_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + sys::swapByteOrder(q); + + outs() << "\t\t list[" << i << "] "; + sym_name = get_symbol_64(offset, S, info, n_value, q); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (q != 0) + outs() << " + " << format("0x%" PRIx64, q); + } else + outs() << format("0x%" PRIx64, q); + outs() << " (struct protocol_t *)\n"; + + r = get_pointer_64(q + n_value, offset, left, S, info); + if (r == nullptr) + return; + memset(&pc, '\0', sizeof(struct protocol64_t)); + if (left < sizeof(struct protocol64_t)) { + memcpy(&pc, r, left); + outs() << " (protocol_t entends past the end of the section)\n"; + } else + memcpy(&pc, r, sizeof(struct protocol64_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(pc); + + outs() << "\t\t\t isa " << format("0x%" PRIx64, pc.isa) << "\n"; + + outs() << "\t\t\t name "; + sym_name = get_symbol_64(offset + offsetof(struct protocol64_t, name), S, + info, n_value, pc.name); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (pc.name != 0) + outs() << " + " << format("0x%" PRIx64, pc.name); + } else + outs() << format("0x%" PRIx64, pc.name); + name = get_pointer_64(pc.name + n_value, xoffset, left, xS, info); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + + outs() << "\t\t\tprotocols " << format("0x%" PRIx64, pc.protocols) << "\n"; + + outs() << "\t\t instanceMethods "; + sym_name = + get_symbol_64(offset + offsetof(struct protocol64_t, instanceMethods), + S, info, n_value, pc.instanceMethods); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (pc.instanceMethods != 0) + outs() << " + " << format("0x%" PRIx64, pc.instanceMethods); + } else + outs() << format("0x%" PRIx64, pc.instanceMethods); + outs() << " (struct method_list_t *)\n"; + if (pc.instanceMethods + n_value != 0) + print_method_list64_t(pc.instanceMethods + n_value, info, "\t"); + + outs() << "\t\t classMethods "; + sym_name = + get_symbol_64(offset + offsetof(struct protocol64_t, classMethods), S, + info, n_value, pc.classMethods); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (pc.classMethods != 0) + outs() << " + " << format("0x%" PRIx64, pc.classMethods); + } else + outs() << format("0x%" PRIx64, pc.classMethods); + outs() << " (struct method_list_t *)\n"; + if (pc.classMethods + n_value != 0) + print_method_list64_t(pc.classMethods + n_value, info, "\t"); + + outs() << "\t optionalInstanceMethods " + << format("0x%" PRIx64, pc.optionalInstanceMethods) << "\n"; + outs() << "\t optionalClassMethods " + << format("0x%" PRIx64, pc.optionalClassMethods) << "\n"; + outs() << "\t instanceProperties " + << format("0x%" PRIx64, pc.instanceProperties) << "\n"; + + p += sizeof(uint64_t); + offset += sizeof(uint64_t); + } +} + +static void print_protocol_list32_t(uint32_t p, struct DisassembleInfo *info) { + struct protocol_list32_t pl; + uint32_t q; + struct protocol32_t pc; + const char *r; + uint32_t offset, xoffset, left, i; + SectionRef S, xS; + const char *name; + + r = get_pointer_32(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&pl, '\0', sizeof(struct protocol_list32_t)); + if (left < sizeof(struct protocol_list32_t)) { + memcpy(&pl, r, left); + outs() << " (protocol_list_t entends past the end of the section)\n"; + } else + memcpy(&pl, r, sizeof(struct protocol_list32_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(pl); + outs() << " count " << pl.count << "\n"; + + p += sizeof(struct protocol_list32_t); + offset += sizeof(struct protocol_list32_t); + for (i = 0; i < pl.count; i++) { + r = get_pointer_32(p, offset, left, S, info); + if (r == nullptr) + return; + q = 0; + if (left < sizeof(uint32_t)) { + memcpy(&q, r, left); + outs() << " (protocol_t * entends past the end of the section)\n"; + } else + memcpy(&q, r, sizeof(uint32_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + sys::swapByteOrder(q); + outs() << "\t\t list[" << i << "] " << format("0x%" PRIx32, q) + << " (struct protocol_t *)\n"; + r = get_pointer_32(q, offset, left, S, info); + if (r == nullptr) + return; + memset(&pc, '\0', sizeof(struct protocol32_t)); + if (left < sizeof(struct protocol32_t)) { + memcpy(&pc, r, left); + outs() << " (protocol_t entends past the end of the section)\n"; + } else + memcpy(&pc, r, sizeof(struct protocol32_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(pc); + outs() << "\t\t\t isa " << format("0x%" PRIx32, pc.isa) << "\n"; + outs() << "\t\t\t name " << format("0x%" PRIx32, pc.name); + name = get_pointer_32(pc.name, xoffset, left, xS, info); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + outs() << "\t\t\tprotocols " << format("0x%" PRIx32, pc.protocols) << "\n"; + outs() << "\t\t instanceMethods " + << format("0x%" PRIx32, pc.instanceMethods) + << " (struct method_list_t *)\n"; + if (pc.instanceMethods != 0) + print_method_list32_t(pc.instanceMethods, info, "\t"); + outs() << "\t\t classMethods " << format("0x%" PRIx32, pc.classMethods) + << " (struct method_list_t *)\n"; + if (pc.classMethods != 0) + print_method_list32_t(pc.classMethods, info, "\t"); + outs() << "\t optionalInstanceMethods " + << format("0x%" PRIx32, pc.optionalInstanceMethods) << "\n"; + outs() << "\t optionalClassMethods " + << format("0x%" PRIx32, pc.optionalClassMethods) << "\n"; + outs() << "\t instanceProperties " + << format("0x%" PRIx32, pc.instanceProperties) << "\n"; + p += sizeof(uint32_t); + offset += sizeof(uint32_t); + } +} + +static void print_indent(uint32_t indent) { + for (uint32_t i = 0; i < indent;) { + if (indent - i >= 8) { + outs() << "\t"; + i += 8; + } else { + for (uint32_t j = i; j < indent; j++) + outs() << " "; + return; + } + } +} + +static bool print_method_description_list(uint32_t p, uint32_t indent, + struct DisassembleInfo *info) { + uint32_t offset, left, xleft; + SectionRef S; + struct objc_method_description_list_t mdl; + struct objc_method_description_t md; + const char *r, *list, *name; + int32_t i; + + r = get_pointer_32(p, offset, left, S, info, true); + if (r == nullptr) + return true; + + outs() << "\n"; + if (left > sizeof(struct objc_method_description_list_t)) { + memcpy(&mdl, r, sizeof(struct objc_method_description_list_t)); + } else { + print_indent(indent); + outs() << " objc_method_description_list extends past end of the section\n"; + memset(&mdl, '\0', sizeof(struct objc_method_description_list_t)); + memcpy(&mdl, r, left); + } + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(mdl); + + print_indent(indent); + outs() << " count " << mdl.count << "\n"; + + list = r + sizeof(struct objc_method_description_list_t); + for (i = 0; i < mdl.count; i++) { + if ((i + 1) * sizeof(struct objc_method_description_t) > left) { + print_indent(indent); + outs() << " remaining list entries extend past the of the section\n"; + break; + } + print_indent(indent); + outs() << " list[" << i << "]\n"; + memcpy(&md, list + i * sizeof(struct objc_method_description_t), + sizeof(struct objc_method_description_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(md); + + print_indent(indent); + outs() << " name " << format("0x%08" PRIx32, md.name); + if (info->verbose) { + name = get_pointer_32(md.name, offset, xleft, S, info, true); + if (name != nullptr) + outs() << format(" %.*s", xleft, name); + else + outs() << " (not in an __OBJC section)"; + } + outs() << "\n"; + + print_indent(indent); + outs() << " types " << format("0x%08" PRIx32, md.types); + if (info->verbose) { + name = get_pointer_32(md.types, offset, xleft, S, info, true); + if (name != nullptr) + outs() << format(" %.*s", xleft, name); + else + outs() << " (not in an __OBJC section)"; + } + outs() << "\n"; + } + return false; +} + +static bool print_protocol_list(uint32_t p, uint32_t indent, + struct DisassembleInfo *info); + +static bool print_protocol(uint32_t p, uint32_t indent, + struct DisassembleInfo *info) { + uint32_t offset, left; + SectionRef S; + struct objc_protocol_t protocol; + const char *r, *name; + + r = get_pointer_32(p, offset, left, S, info, true); + if (r == nullptr) + return true; + + outs() << "\n"; + if (left >= sizeof(struct objc_protocol_t)) { + memcpy(&protocol, r, sizeof(struct objc_protocol_t)); + } else { + print_indent(indent); + outs() << " Protocol extends past end of the section\n"; + memset(&protocol, '\0', sizeof(struct objc_protocol_t)); + memcpy(&protocol, r, left); + } + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(protocol); + + print_indent(indent); + outs() << " isa " << format("0x%08" PRIx32, protocol.isa) + << "\n"; + + print_indent(indent); + outs() << " protocol_name " + << format("0x%08" PRIx32, protocol.protocol_name); + if (info->verbose) { + name = get_pointer_32(protocol.protocol_name, offset, left, S, info, true); + if (name != nullptr) + outs() << format(" %.*s", left, name); + else + outs() << " (not in an __OBJC section)"; + } + outs() << "\n"; + + print_indent(indent); + outs() << " protocol_list " + << format("0x%08" PRIx32, protocol.protocol_list); + if (print_protocol_list(protocol.protocol_list, indent + 4, info)) + outs() << " (not in an __OBJC section)\n"; + + print_indent(indent); + outs() << " instance_methods " + << format("0x%08" PRIx32, protocol.instance_methods); + if (print_method_description_list(protocol.instance_methods, indent, info)) + outs() << " (not in an __OBJC section)\n"; + + print_indent(indent); + outs() << " class_methods " + << format("0x%08" PRIx32, protocol.class_methods); + if (print_method_description_list(protocol.class_methods, indent, info)) + outs() << " (not in an __OBJC section)\n"; + + return false; +} + +static bool print_protocol_list(uint32_t p, uint32_t indent, + struct DisassembleInfo *info) { + uint32_t offset, left, l; + SectionRef S; + struct objc_protocol_list_t protocol_list; + const char *r, *list; + int32_t i; + + r = get_pointer_32(p, offset, left, S, info, true); + if (r == nullptr) + return true; + + outs() << "\n"; + if (left > sizeof(struct objc_protocol_list_t)) { + memcpy(&protocol_list, r, sizeof(struct objc_protocol_list_t)); + } else { + outs() << "\t\t objc_protocol_list_t extends past end of the section\n"; + memset(&protocol_list, '\0', sizeof(struct objc_protocol_list_t)); + memcpy(&protocol_list, r, left); + } + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(protocol_list); + + print_indent(indent); + outs() << " next " << format("0x%08" PRIx32, protocol_list.next) + << "\n"; + print_indent(indent); + outs() << " count " << protocol_list.count << "\n"; + + list = r + sizeof(struct objc_protocol_list_t); + for (i = 0; i < protocol_list.count; i++) { + if ((i + 1) * sizeof(uint32_t) > left) { + outs() << "\t\t remaining list entries extend past the of the section\n"; + break; + } + memcpy(&l, list + i * sizeof(uint32_t), sizeof(uint32_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + sys::swapByteOrder(l); + + print_indent(indent); + outs() << " list[" << i << "] " << format("0x%08" PRIx32, l); + if (print_protocol(l, indent, info)) + outs() << "(not in an __OBJC section)\n"; + } + return false; +} + +static void print_ivar_list64_t(uint64_t p, struct DisassembleInfo *info) { + struct ivar_list64_t il; + struct ivar64_t i; + const char *r; + uint32_t offset, xoffset, left, j; + SectionRef S, xS; + const char *name, *sym_name, *ivar_offset_p; + uint64_t ivar_offset, n_value; + + r = get_pointer_64(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&il, '\0', sizeof(struct ivar_list64_t)); + if (left < sizeof(struct ivar_list64_t)) { + memcpy(&il, r, left); + outs() << " (ivar_list_t entends past the end of the section)\n"; + } else + memcpy(&il, r, sizeof(struct ivar_list64_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(il); + outs() << " entsize " << il.entsize << "\n"; + outs() << " count " << il.count << "\n"; + + p += sizeof(struct ivar_list64_t); + offset += sizeof(struct ivar_list64_t); + for (j = 0; j < il.count; j++) { + r = get_pointer_64(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&i, '\0', sizeof(struct ivar64_t)); + if (left < sizeof(struct ivar64_t)) { + memcpy(&i, r, left); + outs() << " (ivar_t entends past the end of the section)\n"; + } else + memcpy(&i, r, sizeof(struct ivar64_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(i); + + outs() << "\t\t\t offset "; + sym_name = get_symbol_64(offset + offsetof(struct ivar64_t, offset), S, + info, n_value, i.offset); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (i.offset != 0) + outs() << " + " << format("0x%" PRIx64, i.offset); + } else + outs() << format("0x%" PRIx64, i.offset); + ivar_offset_p = get_pointer_64(i.offset + n_value, xoffset, left, xS, info); + if (ivar_offset_p != nullptr && left >= sizeof(*ivar_offset_p)) { + memcpy(&ivar_offset, ivar_offset_p, sizeof(ivar_offset)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + sys::swapByteOrder(ivar_offset); + outs() << " " << ivar_offset << "\n"; + } else + outs() << "\n"; + + outs() << "\t\t\t name "; + sym_name = get_symbol_64(offset + offsetof(struct ivar64_t, name), S, info, + n_value, i.name); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (i.name != 0) + outs() << " + " << format("0x%" PRIx64, i.name); + } else + outs() << format("0x%" PRIx64, i.name); + name = get_pointer_64(i.name + n_value, xoffset, left, xS, info); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + + outs() << "\t\t\t type "; + sym_name = get_symbol_64(offset + offsetof(struct ivar64_t, type), S, info, + n_value, i.name); + name = get_pointer_64(i.type + n_value, xoffset, left, xS, info); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (i.type != 0) + outs() << " + " << format("0x%" PRIx64, i.type); + } else + outs() << format("0x%" PRIx64, i.type); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + + outs() << "\t\t\talignment " << i.alignment << "\n"; + outs() << "\t\t\t size " << i.size << "\n"; + + p += sizeof(struct ivar64_t); + offset += sizeof(struct ivar64_t); + } +} + +static void print_ivar_list32_t(uint32_t p, struct DisassembleInfo *info) { + struct ivar_list32_t il; + struct ivar32_t i; + const char *r; + uint32_t offset, xoffset, left, j; + SectionRef S, xS; + const char *name, *ivar_offset_p; + uint32_t ivar_offset; + + r = get_pointer_32(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&il, '\0', sizeof(struct ivar_list32_t)); + if (left < sizeof(struct ivar_list32_t)) { + memcpy(&il, r, left); + outs() << " (ivar_list_t entends past the end of the section)\n"; + } else + memcpy(&il, r, sizeof(struct ivar_list32_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(il); + outs() << " entsize " << il.entsize << "\n"; + outs() << " count " << il.count << "\n"; + + p += sizeof(struct ivar_list32_t); + offset += sizeof(struct ivar_list32_t); + for (j = 0; j < il.count; j++) { + r = get_pointer_32(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&i, '\0', sizeof(struct ivar32_t)); + if (left < sizeof(struct ivar32_t)) { + memcpy(&i, r, left); + outs() << " (ivar_t entends past the end of the section)\n"; + } else + memcpy(&i, r, sizeof(struct ivar32_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(i); + + outs() << "\t\t\t offset " << format("0x%" PRIx32, i.offset); + ivar_offset_p = get_pointer_32(i.offset, xoffset, left, xS, info); + if (ivar_offset_p != nullptr && left >= sizeof(*ivar_offset_p)) { + memcpy(&ivar_offset, ivar_offset_p, sizeof(ivar_offset)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + sys::swapByteOrder(ivar_offset); + outs() << " " << ivar_offset << "\n"; + } else + outs() << "\n"; + + outs() << "\t\t\t name " << format("0x%" PRIx32, i.name); + name = get_pointer_32(i.name, xoffset, left, xS, info); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + + outs() << "\t\t\t type " << format("0x%" PRIx32, i.type); + name = get_pointer_32(i.type, xoffset, left, xS, info); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + + outs() << "\t\t\talignment " << i.alignment << "\n"; + outs() << "\t\t\t size " << i.size << "\n"; + + p += sizeof(struct ivar32_t); + offset += sizeof(struct ivar32_t); + } +} + +static void print_objc_property_list64(uint64_t p, + struct DisassembleInfo *info) { + struct objc_property_list64 opl; + struct objc_property64 op; + const char *r; + uint32_t offset, xoffset, left, j; + SectionRef S, xS; + const char *name, *sym_name; + uint64_t n_value; + + r = get_pointer_64(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&opl, '\0', sizeof(struct objc_property_list64)); + if (left < sizeof(struct objc_property_list64)) { + memcpy(&opl, r, left); + outs() << " (objc_property_list entends past the end of the section)\n"; + } else + memcpy(&opl, r, sizeof(struct objc_property_list64)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(opl); + outs() << " entsize " << opl.entsize << "\n"; + outs() << " count " << opl.count << "\n"; + + p += sizeof(struct objc_property_list64); + offset += sizeof(struct objc_property_list64); + for (j = 0; j < opl.count; j++) { + r = get_pointer_64(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&op, '\0', sizeof(struct objc_property64)); + if (left < sizeof(struct objc_property64)) { + memcpy(&op, r, left); + outs() << " (objc_property entends past the end of the section)\n"; + } else + memcpy(&op, r, sizeof(struct objc_property64)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(op); + + outs() << "\t\t\t name "; + sym_name = get_symbol_64(offset + offsetof(struct objc_property64, name), S, + info, n_value, op.name); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (op.name != 0) + outs() << " + " << format("0x%" PRIx64, op.name); + } else + outs() << format("0x%" PRIx64, op.name); + name = get_pointer_64(op.name + n_value, xoffset, left, xS, info); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + + outs() << "\t\t\tattributes "; + sym_name = + get_symbol_64(offset + offsetof(struct objc_property64, attributes), S, + info, n_value, op.attributes); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (op.attributes != 0) + outs() << " + " << format("0x%" PRIx64, op.attributes); + } else + outs() << format("0x%" PRIx64, op.attributes); + name = get_pointer_64(op.attributes + n_value, xoffset, left, xS, info); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + + p += sizeof(struct objc_property64); + offset += sizeof(struct objc_property64); + } +} + +static void print_objc_property_list32(uint32_t p, + struct DisassembleInfo *info) { + struct objc_property_list32 opl; + struct objc_property32 op; + const char *r; + uint32_t offset, xoffset, left, j; + SectionRef S, xS; + const char *name; + + r = get_pointer_32(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&opl, '\0', sizeof(struct objc_property_list32)); + if (left < sizeof(struct objc_property_list32)) { + memcpy(&opl, r, left); + outs() << " (objc_property_list entends past the end of the section)\n"; + } else + memcpy(&opl, r, sizeof(struct objc_property_list32)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(opl); + outs() << " entsize " << opl.entsize << "\n"; + outs() << " count " << opl.count << "\n"; + + p += sizeof(struct objc_property_list32); + offset += sizeof(struct objc_property_list32); + for (j = 0; j < opl.count; j++) { + r = get_pointer_32(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&op, '\0', sizeof(struct objc_property32)); + if (left < sizeof(struct objc_property32)) { + memcpy(&op, r, left); + outs() << " (objc_property entends past the end of the section)\n"; + } else + memcpy(&op, r, sizeof(struct objc_property32)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(op); + + outs() << "\t\t\t name " << format("0x%" PRIx32, op.name); + name = get_pointer_32(op.name, xoffset, left, xS, info); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + + outs() << "\t\t\tattributes " << format("0x%" PRIx32, op.attributes); + name = get_pointer_32(op.attributes, xoffset, left, xS, info); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + + p += sizeof(struct objc_property32); + offset += sizeof(struct objc_property32); + } +} + +static void print_class_ro64_t(uint64_t p, struct DisassembleInfo *info, + bool &is_meta_class) { + struct class_ro64_t cro; + const char *r; + uint32_t offset, xoffset, left; + SectionRef S, xS; + const char *name, *sym_name; + uint64_t n_value; + + r = get_pointer_64(p, offset, left, S, info); + if (r == nullptr || left < sizeof(struct class_ro64_t)) + return; + memset(&cro, '\0', sizeof(struct class_ro64_t)); + if (left < sizeof(struct class_ro64_t)) { + memcpy(&cro, r, left); + outs() << " (class_ro_t entends past the end of the section)\n"; + } else + memcpy(&cro, r, sizeof(struct class_ro64_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(cro); + outs() << " flags " << format("0x%" PRIx32, cro.flags); + if (cro.flags & RO_META) + outs() << " RO_META"; + if (cro.flags & RO_ROOT) + outs() << " RO_ROOT"; + if (cro.flags & RO_HAS_CXX_STRUCTORS) + outs() << " RO_HAS_CXX_STRUCTORS"; + outs() << "\n"; + outs() << " instanceStart " << cro.instanceStart << "\n"; + outs() << " instanceSize " << cro.instanceSize << "\n"; + outs() << " reserved " << format("0x%" PRIx32, cro.reserved) + << "\n"; + outs() << " ivarLayout " << format("0x%" PRIx64, cro.ivarLayout) + << "\n"; + print_layout_map64(cro.ivarLayout, info); + + outs() << " name "; + sym_name = get_symbol_64(offset + offsetof(struct class_ro64_t, name), S, + info, n_value, cro.name); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (cro.name != 0) + outs() << " + " << format("0x%" PRIx64, cro.name); + } else + outs() << format("0x%" PRIx64, cro.name); + name = get_pointer_64(cro.name + n_value, xoffset, left, xS, info); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + + outs() << " baseMethods "; + sym_name = get_symbol_64(offset + offsetof(struct class_ro64_t, baseMethods), + S, info, n_value, cro.baseMethods); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (cro.baseMethods != 0) + outs() << " + " << format("0x%" PRIx64, cro.baseMethods); + } else + outs() << format("0x%" PRIx64, cro.baseMethods); + outs() << " (struct method_list_t *)\n"; + if (cro.baseMethods + n_value != 0) + print_method_list64_t(cro.baseMethods + n_value, info, ""); + + outs() << " baseProtocols "; + sym_name = + get_symbol_64(offset + offsetof(struct class_ro64_t, baseProtocols), S, + info, n_value, cro.baseProtocols); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (cro.baseProtocols != 0) + outs() << " + " << format("0x%" PRIx64, cro.baseProtocols); + } else + outs() << format("0x%" PRIx64, cro.baseProtocols); + outs() << "\n"; + if (cro.baseProtocols + n_value != 0) + print_protocol_list64_t(cro.baseProtocols + n_value, info); + + outs() << " ivars "; + sym_name = get_symbol_64(offset + offsetof(struct class_ro64_t, ivars), S, + info, n_value, cro.ivars); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (cro.ivars != 0) + outs() << " + " << format("0x%" PRIx64, cro.ivars); + } else + outs() << format("0x%" PRIx64, cro.ivars); + outs() << "\n"; + if (cro.ivars + n_value != 0) + print_ivar_list64_t(cro.ivars + n_value, info); + + outs() << " weakIvarLayout "; + sym_name = + get_symbol_64(offset + offsetof(struct class_ro64_t, weakIvarLayout), S, + info, n_value, cro.weakIvarLayout); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (cro.weakIvarLayout != 0) + outs() << " + " << format("0x%" PRIx64, cro.weakIvarLayout); + } else + outs() << format("0x%" PRIx64, cro.weakIvarLayout); + outs() << "\n"; + print_layout_map64(cro.weakIvarLayout + n_value, info); + + outs() << " baseProperties "; + sym_name = + get_symbol_64(offset + offsetof(struct class_ro64_t, baseProperties), S, + info, n_value, cro.baseProperties); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (cro.baseProperties != 0) + outs() << " + " << format("0x%" PRIx64, cro.baseProperties); + } else + outs() << format("0x%" PRIx64, cro.baseProperties); + outs() << "\n"; + if (cro.baseProperties + n_value != 0) + print_objc_property_list64(cro.baseProperties + n_value, info); + + is_meta_class = (cro.flags & RO_META) ? true : false; +} + +static void print_class_ro32_t(uint32_t p, struct DisassembleInfo *info, + bool &is_meta_class) { + struct class_ro32_t cro; + const char *r; + uint32_t offset, xoffset, left; + SectionRef S, xS; + const char *name; + + r = get_pointer_32(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&cro, '\0', sizeof(struct class_ro32_t)); + if (left < sizeof(struct class_ro32_t)) { + memcpy(&cro, r, left); + outs() << " (class_ro_t entends past the end of the section)\n"; + } else + memcpy(&cro, r, sizeof(struct class_ro32_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(cro); + outs() << " flags " << format("0x%" PRIx32, cro.flags); + if (cro.flags & RO_META) + outs() << " RO_META"; + if (cro.flags & RO_ROOT) + outs() << " RO_ROOT"; + if (cro.flags & RO_HAS_CXX_STRUCTORS) + outs() << " RO_HAS_CXX_STRUCTORS"; + outs() << "\n"; + outs() << " instanceStart " << cro.instanceStart << "\n"; + outs() << " instanceSize " << cro.instanceSize << "\n"; + outs() << " ivarLayout " << format("0x%" PRIx32, cro.ivarLayout) + << "\n"; + print_layout_map32(cro.ivarLayout, info); + + outs() << " name " << format("0x%" PRIx32, cro.name); + name = get_pointer_32(cro.name, xoffset, left, xS, info); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + + outs() << " baseMethods " + << format("0x%" PRIx32, cro.baseMethods) + << " (struct method_list_t *)\n"; + if (cro.baseMethods != 0) + print_method_list32_t(cro.baseMethods, info, ""); + + outs() << " baseProtocols " + << format("0x%" PRIx32, cro.baseProtocols) << "\n"; + if (cro.baseProtocols != 0) + print_protocol_list32_t(cro.baseProtocols, info); + outs() << " ivars " << format("0x%" PRIx32, cro.ivars) + << "\n"; + if (cro.ivars != 0) + print_ivar_list32_t(cro.ivars, info); + outs() << " weakIvarLayout " + << format("0x%" PRIx32, cro.weakIvarLayout) << "\n"; + print_layout_map32(cro.weakIvarLayout, info); + outs() << " baseProperties " + << format("0x%" PRIx32, cro.baseProperties) << "\n"; + if (cro.baseProperties != 0) + print_objc_property_list32(cro.baseProperties, info); + is_meta_class = (cro.flags & RO_META) ? true : false; +} + +static void print_class64_t(uint64_t p, struct DisassembleInfo *info) { + struct class64_t c; + const char *r; + uint32_t offset, left; + SectionRef S; + const char *name; + uint64_t isa_n_value, n_value; + + r = get_pointer_64(p, offset, left, S, info); + if (r == nullptr || left < sizeof(struct class64_t)) + return; + memset(&c, '\0', sizeof(struct class64_t)); + if (left < sizeof(struct class64_t)) { + memcpy(&c, r, left); + outs() << " (class_t entends past the end of the section)\n"; + } else + memcpy(&c, r, sizeof(struct class64_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(c); + + outs() << " isa " << format("0x%" PRIx64, c.isa); + name = get_symbol_64(offset + offsetof(struct class64_t, isa), S, info, + isa_n_value, c.isa); + if (name != nullptr) + outs() << " " << name; + outs() << "\n"; + + outs() << " superclass " << format("0x%" PRIx64, c.superclass); + name = get_symbol_64(offset + offsetof(struct class64_t, superclass), S, info, + n_value, c.superclass); + if (name != nullptr) + outs() << " " << name; + outs() << "\n"; + + outs() << " cache " << format("0x%" PRIx64, c.cache); + name = get_symbol_64(offset + offsetof(struct class64_t, cache), S, info, + n_value, c.cache); + if (name != nullptr) + outs() << " " << name; + outs() << "\n"; + + outs() << " vtable " << format("0x%" PRIx64, c.vtable); + name = get_symbol_64(offset + offsetof(struct class64_t, vtable), S, info, + n_value, c.vtable); + if (name != nullptr) + outs() << " " << name; + outs() << "\n"; + + name = get_symbol_64(offset + offsetof(struct class64_t, data), S, info, + n_value, c.data); + outs() << " data "; + if (n_value != 0) { + if (info->verbose && name != nullptr) + outs() << name; + else + outs() << format("0x%" PRIx64, n_value); + if (c.data != 0) + outs() << " + " << format("0x%" PRIx64, c.data); + } else + outs() << format("0x%" PRIx64, c.data); + outs() << " (struct class_ro_t *)"; + + // This is a Swift class if some of the low bits of the pointer are set. + if ((c.data + n_value) & 0x7) + outs() << " Swift class"; + outs() << "\n"; + bool is_meta_class; + print_class_ro64_t((c.data + n_value) & ~0x7, info, is_meta_class); + + if (is_meta_class == false) { + outs() << "Meta Class\n"; + print_class64_t(c.isa + isa_n_value, info); + } +} + +static void print_class32_t(uint32_t p, struct DisassembleInfo *info) { + struct class32_t c; + const char *r; + uint32_t offset, left; + SectionRef S; + const char *name; + + r = get_pointer_32(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&c, '\0', sizeof(struct class32_t)); + if (left < sizeof(struct class32_t)) { + memcpy(&c, r, left); + outs() << " (class_t entends past the end of the section)\n"; + } else + memcpy(&c, r, sizeof(struct class32_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(c); + + outs() << " isa " << format("0x%" PRIx32, c.isa); + name = + get_symbol_32(offset + offsetof(struct class32_t, isa), S, info, c.isa); + if (name != nullptr) + outs() << " " << name; + outs() << "\n"; + + outs() << " superclass " << format("0x%" PRIx32, c.superclass); + name = get_symbol_32(offset + offsetof(struct class32_t, superclass), S, info, + c.superclass); + if (name != nullptr) + outs() << " " << name; + outs() << "\n"; + + outs() << " cache " << format("0x%" PRIx32, c.cache); + name = get_symbol_32(offset + offsetof(struct class32_t, cache), S, info, + c.cache); + if (name != nullptr) + outs() << " " << name; + outs() << "\n"; + + outs() << " vtable " << format("0x%" PRIx32, c.vtable); + name = get_symbol_32(offset + offsetof(struct class32_t, vtable), S, info, + c.vtable); + if (name != nullptr) + outs() << " " << name; + outs() << "\n"; + + name = + get_symbol_32(offset + offsetof(struct class32_t, data), S, info, c.data); + outs() << " data " << format("0x%" PRIx32, c.data) + << " (struct class_ro_t *)"; + + // This is a Swift class if some of the low bits of the pointer are set. + if (c.data & 0x3) + outs() << " Swift class"; + outs() << "\n"; + bool is_meta_class; + print_class_ro32_t(c.data & ~0x3, info, is_meta_class); + + if (is_meta_class == false) { + outs() << "Meta Class\n"; + print_class32_t(c.isa, info); + } +} + +static void print_objc_class_t(struct objc_class_t *objc_class, + struct DisassembleInfo *info) { + uint32_t offset, left, xleft; + const char *name, *p, *ivar_list; + SectionRef S; + int32_t i; + struct objc_ivar_list_t objc_ivar_list; + struct objc_ivar_t ivar; + + outs() << "\t\t isa " << format("0x%08" PRIx32, objc_class->isa); + if (info->verbose && CLS_GETINFO(objc_class, CLS_META)) { + name = get_pointer_32(objc_class->isa, offset, left, S, info, true); + if (name != nullptr) + outs() << format(" %.*s", left, name); + else + outs() << " (not in an __OBJC section)"; + } + outs() << "\n"; + + outs() << "\t super_class " + << format("0x%08" PRIx32, objc_class->super_class); + if (info->verbose) { + name = get_pointer_32(objc_class->super_class, offset, left, S, info, true); + if (name != nullptr) + outs() << format(" %.*s", left, name); + else + outs() << " (not in an __OBJC section)"; + } + outs() << "\n"; + + outs() << "\t\t name " << format("0x%08" PRIx32, objc_class->name); + if (info->verbose) { + name = get_pointer_32(objc_class->name, offset, left, S, info, true); + if (name != nullptr) + outs() << format(" %.*s", left, name); + else + outs() << " (not in an __OBJC section)"; + } + outs() << "\n"; + + outs() << "\t\t version " << format("0x%08" PRIx32, objc_class->version) + << "\n"; + + outs() << "\t\t info " << format("0x%08" PRIx32, objc_class->info); + if (info->verbose) { + if (CLS_GETINFO(objc_class, CLS_CLASS)) + outs() << " CLS_CLASS"; + else if (CLS_GETINFO(objc_class, CLS_META)) + outs() << " CLS_META"; + } + outs() << "\n"; + + outs() << "\t instance_size " + << format("0x%08" PRIx32, objc_class->instance_size) << "\n"; + + p = get_pointer_32(objc_class->ivars, offset, left, S, info, true); + outs() << "\t\t ivars " << format("0x%08" PRIx32, objc_class->ivars); + if (p != nullptr) { + if (left > sizeof(struct objc_ivar_list_t)) { + outs() << "\n"; + memcpy(&objc_ivar_list, p, sizeof(struct objc_ivar_list_t)); + } else { + outs() << " (entends past the end of the section)\n"; + memset(&objc_ivar_list, '\0', sizeof(struct objc_ivar_list_t)); + memcpy(&objc_ivar_list, p, left); + } + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(objc_ivar_list); + outs() << "\t\t ivar_count " << objc_ivar_list.ivar_count << "\n"; + ivar_list = p + sizeof(struct objc_ivar_list_t); + for (i = 0; i < objc_ivar_list.ivar_count; i++) { + if ((i + 1) * sizeof(struct objc_ivar_t) > left) { + outs() << "\t\t remaining ivar's extend past the of the section\n"; + break; + } + memcpy(&ivar, ivar_list + i * sizeof(struct objc_ivar_t), + sizeof(struct objc_ivar_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(ivar); + + outs() << "\t\t\tivar_name " << format("0x%08" PRIx32, ivar.ivar_name); + if (info->verbose) { + name = get_pointer_32(ivar.ivar_name, offset, xleft, S, info, true); + if (name != nullptr) + outs() << format(" %.*s", xleft, name); + else + outs() << " (not in an __OBJC section)"; + } + outs() << "\n"; + + outs() << "\t\t\tivar_type " << format("0x%08" PRIx32, ivar.ivar_type); + if (info->verbose) { + name = get_pointer_32(ivar.ivar_type, offset, xleft, S, info, true); + if (name != nullptr) + outs() << format(" %.*s", xleft, name); + else + outs() << " (not in an __OBJC section)"; + } + outs() << "\n"; + + outs() << "\t\t ivar_offset " + << format("0x%08" PRIx32, ivar.ivar_offset) << "\n"; + } + } else { + outs() << " (not in an __OBJC section)\n"; + } + + outs() << "\t\t methods " << format("0x%08" PRIx32, objc_class->methodLists); + if (print_method_list(objc_class->methodLists, info)) + outs() << " (not in an __OBJC section)\n"; + + outs() << "\t\t cache " << format("0x%08" PRIx32, objc_class->cache) + << "\n"; + + outs() << "\t\tprotocols " << format("0x%08" PRIx32, objc_class->protocols); + if (print_protocol_list(objc_class->protocols, 16, info)) + outs() << " (not in an __OBJC section)\n"; +} + +static void print_objc_objc_category_t(struct objc_category_t *objc_category, + struct DisassembleInfo *info) { + uint32_t offset, left; + const char *name; + SectionRef S; + + outs() << "\t category name " + << format("0x%08" PRIx32, objc_category->category_name); + if (info->verbose) { + name = get_pointer_32(objc_category->category_name, offset, left, S, info, + true); + if (name != nullptr) + outs() << format(" %.*s", left, name); + else + outs() << " (not in an __OBJC section)"; + } + outs() << "\n"; + + outs() << "\t\t class name " + << format("0x%08" PRIx32, objc_category->class_name); + if (info->verbose) { + name = + get_pointer_32(objc_category->class_name, offset, left, S, info, true); + if (name != nullptr) + outs() << format(" %.*s", left, name); + else + outs() << " (not in an __OBJC section)"; + } + outs() << "\n"; + + outs() << "\t instance methods " + << format("0x%08" PRIx32, objc_category->instance_methods); + if (print_method_list(objc_category->instance_methods, info)) + outs() << " (not in an __OBJC section)\n"; + + outs() << "\t class methods " + << format("0x%08" PRIx32, objc_category->class_methods); + if (print_method_list(objc_category->class_methods, info)) + outs() << " (not in an __OBJC section)\n"; +} + +static void print_category64_t(uint64_t p, struct DisassembleInfo *info) { + struct category64_t c; + const char *r; + uint32_t offset, xoffset, left; + SectionRef S, xS; + const char *name, *sym_name; + uint64_t n_value; + + r = get_pointer_64(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&c, '\0', sizeof(struct category64_t)); + if (left < sizeof(struct category64_t)) { + memcpy(&c, r, left); + outs() << " (category_t entends past the end of the section)\n"; + } else + memcpy(&c, r, sizeof(struct category64_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(c); + + outs() << " name "; + sym_name = get_symbol_64(offset + offsetof(struct category64_t, name), S, + info, n_value, c.name); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (c.name != 0) + outs() << " + " << format("0x%" PRIx64, c.name); + } else + outs() << format("0x%" PRIx64, c.name); + name = get_pointer_64(c.name + n_value, xoffset, left, xS, info); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + + outs() << " cls "; + sym_name = get_symbol_64(offset + offsetof(struct category64_t, cls), S, info, + n_value, c.cls); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (c.cls != 0) + outs() << " + " << format("0x%" PRIx64, c.cls); + } else + outs() << format("0x%" PRIx64, c.cls); + outs() << "\n"; + if (c.cls + n_value != 0) + print_class64_t(c.cls + n_value, info); + + outs() << " instanceMethods "; + sym_name = + get_symbol_64(offset + offsetof(struct category64_t, instanceMethods), S, + info, n_value, c.instanceMethods); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (c.instanceMethods != 0) + outs() << " + " << format("0x%" PRIx64, c.instanceMethods); + } else + outs() << format("0x%" PRIx64, c.instanceMethods); + outs() << "\n"; + if (c.instanceMethods + n_value != 0) + print_method_list64_t(c.instanceMethods + n_value, info, ""); + + outs() << " classMethods "; + sym_name = get_symbol_64(offset + offsetof(struct category64_t, classMethods), + S, info, n_value, c.classMethods); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (c.classMethods != 0) + outs() << " + " << format("0x%" PRIx64, c.classMethods); + } else + outs() << format("0x%" PRIx64, c.classMethods); + outs() << "\n"; + if (c.classMethods + n_value != 0) + print_method_list64_t(c.classMethods + n_value, info, ""); + + outs() << " protocols "; + sym_name = get_symbol_64(offset + offsetof(struct category64_t, protocols), S, + info, n_value, c.protocols); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (c.protocols != 0) + outs() << " + " << format("0x%" PRIx64, c.protocols); + } else + outs() << format("0x%" PRIx64, c.protocols); + outs() << "\n"; + if (c.protocols + n_value != 0) + print_protocol_list64_t(c.protocols + n_value, info); + + outs() << "instanceProperties "; + sym_name = + get_symbol_64(offset + offsetof(struct category64_t, instanceProperties), + S, info, n_value, c.instanceProperties); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (c.instanceProperties != 0) + outs() << " + " << format("0x%" PRIx64, c.instanceProperties); + } else + outs() << format("0x%" PRIx64, c.instanceProperties); + outs() << "\n"; + if (c.instanceProperties + n_value != 0) + print_objc_property_list64(c.instanceProperties + n_value, info); +} + +static void print_category32_t(uint32_t p, struct DisassembleInfo *info) { + struct category32_t c; + const char *r; + uint32_t offset, left; + SectionRef S, xS; + const char *name; + + r = get_pointer_32(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&c, '\0', sizeof(struct category32_t)); + if (left < sizeof(struct category32_t)) { + memcpy(&c, r, left); + outs() << " (category_t entends past the end of the section)\n"; + } else + memcpy(&c, r, sizeof(struct category32_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(c); + + outs() << " name " << format("0x%" PRIx32, c.name); + name = get_symbol_32(offset + offsetof(struct category32_t, name), S, info, + c.name); + if (name != NULL) + outs() << " " << name; + outs() << "\n"; + + outs() << " cls " << format("0x%" PRIx32, c.cls) << "\n"; + if (c.cls != 0) + print_class32_t(c.cls, info); + outs() << " instanceMethods " << format("0x%" PRIx32, c.instanceMethods) + << "\n"; + if (c.instanceMethods != 0) + print_method_list32_t(c.instanceMethods, info, ""); + outs() << " classMethods " << format("0x%" PRIx32, c.classMethods) + << "\n"; + if (c.classMethods != 0) + print_method_list32_t(c.classMethods, info, ""); + outs() << " protocols " << format("0x%" PRIx32, c.protocols) << "\n"; + if (c.protocols != 0) + print_protocol_list32_t(c.protocols, info); + outs() << "instanceProperties " << format("0x%" PRIx32, c.instanceProperties) + << "\n"; + if (c.instanceProperties != 0) + print_objc_property_list32(c.instanceProperties, info); +} + +static void print_message_refs64(SectionRef S, struct DisassembleInfo *info) { + uint32_t i, left, offset, xoffset; + uint64_t p, n_value; + struct message_ref64 mr; + const char *name, *sym_name; + const char *r; + SectionRef xS; + + if (S == SectionRef()) + return; + + StringRef SectName; + S.getName(SectName); + DataRefImpl Ref = S.getRawDataRefImpl(); + StringRef SegName = info->O->getSectionFinalSegmentName(Ref); + outs() << "Contents of (" << SegName << "," << SectName << ") section\n"; + offset = 0; + for (i = 0; i < S.getSize(); i += sizeof(struct message_ref64)) { + p = S.getAddress() + i; + r = get_pointer_64(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&mr, '\0', sizeof(struct message_ref64)); + if (left < sizeof(struct message_ref64)) { + memcpy(&mr, r, left); + outs() << " (message_ref entends past the end of the section)\n"; + } else + memcpy(&mr, r, sizeof(struct message_ref64)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(mr); + + outs() << " imp "; + name = get_symbol_64(offset + offsetof(struct message_ref64, imp), S, info, + n_value, mr.imp); + if (n_value != 0) { + outs() << format("0x%" PRIx64, n_value) << " "; + if (mr.imp != 0) + outs() << "+ " << format("0x%" PRIx64, mr.imp) << " "; + } else + outs() << format("0x%" PRIx64, mr.imp) << " "; + if (name != nullptr) + outs() << " " << name; + outs() << "\n"; + + outs() << " sel "; + sym_name = get_symbol_64(offset + offsetof(struct message_ref64, sel), S, + info, n_value, mr.sel); + if (n_value != 0) { + if (info->verbose && sym_name != nullptr) + outs() << sym_name; + else + outs() << format("0x%" PRIx64, n_value); + if (mr.sel != 0) + outs() << " + " << format("0x%" PRIx64, mr.sel); + } else + outs() << format("0x%" PRIx64, mr.sel); + name = get_pointer_64(mr.sel + n_value, xoffset, left, xS, info); + if (name != nullptr) + outs() << format(" %.*s", left, name); + outs() << "\n"; + + offset += sizeof(struct message_ref64); + } +} + +static void print_message_refs32(SectionRef S, struct DisassembleInfo *info) { + uint32_t i, left, offset, xoffset, p; + struct message_ref32 mr; + const char *name, *r; + SectionRef xS; + + if (S == SectionRef()) + return; + + StringRef SectName; + S.getName(SectName); + DataRefImpl Ref = S.getRawDataRefImpl(); + StringRef SegName = info->O->getSectionFinalSegmentName(Ref); + outs() << "Contents of (" << SegName << "," << SectName << ") section\n"; + offset = 0; + for (i = 0; i < S.getSize(); i += sizeof(struct message_ref64)) { + p = S.getAddress() + i; + r = get_pointer_32(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&mr, '\0', sizeof(struct message_ref32)); + if (left < sizeof(struct message_ref32)) { + memcpy(&mr, r, left); + outs() << " (message_ref entends past the end of the section)\n"; + } else + memcpy(&mr, r, sizeof(struct message_ref32)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(mr); + + outs() << " imp " << format("0x%" PRIx32, mr.imp); + name = get_symbol_32(offset + offsetof(struct message_ref32, imp), S, info, + mr.imp); + if (name != nullptr) + outs() << " " << name; + outs() << "\n"; + + outs() << " sel " << format("0x%" PRIx32, mr.sel); + name = get_pointer_32(mr.sel, xoffset, left, xS, info); + if (name != nullptr) + outs() << " " << name; + outs() << "\n"; + + offset += sizeof(struct message_ref32); + } +} + +static void print_image_info64(SectionRef S, struct DisassembleInfo *info) { + uint32_t left, offset, swift_version; + uint64_t p; + struct objc_image_info64 o; + const char *r; + + StringRef SectName; + S.getName(SectName); + DataRefImpl Ref = S.getRawDataRefImpl(); + StringRef SegName = info->O->getSectionFinalSegmentName(Ref); + outs() << "Contents of (" << SegName << "," << SectName << ") section\n"; + p = S.getAddress(); + r = get_pointer_64(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&o, '\0', sizeof(struct objc_image_info64)); + if (left < sizeof(struct objc_image_info64)) { + memcpy(&o, r, left); + outs() << " (objc_image_info entends past the end of the section)\n"; + } else + memcpy(&o, r, sizeof(struct objc_image_info64)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(o); + outs() << " version " << o.version << "\n"; + outs() << " flags " << format("0x%" PRIx32, o.flags); + if (o.flags & OBJC_IMAGE_IS_REPLACEMENT) + outs() << " OBJC_IMAGE_IS_REPLACEMENT"; + if (o.flags & OBJC_IMAGE_SUPPORTS_GC) + outs() << " OBJC_IMAGE_SUPPORTS_GC"; + swift_version = (o.flags >> 8) & 0xff; + if (swift_version != 0) { + if (swift_version == 1) + outs() << " Swift 1.0"; + else if (swift_version == 2) + outs() << " Swift 1.1"; + else + outs() << " unknown future Swift version (" << swift_version << ")"; + } + outs() << "\n"; +} + +static void print_image_info32(SectionRef S, struct DisassembleInfo *info) { + uint32_t left, offset, swift_version, p; + struct objc_image_info32 o; + const char *r; + + StringRef SectName; + S.getName(SectName); + DataRefImpl Ref = S.getRawDataRefImpl(); + StringRef SegName = info->O->getSectionFinalSegmentName(Ref); + outs() << "Contents of (" << SegName << "," << SectName << ") section\n"; + p = S.getAddress(); + r = get_pointer_32(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&o, '\0', sizeof(struct objc_image_info32)); + if (left < sizeof(struct objc_image_info32)) { + memcpy(&o, r, left); + outs() << " (objc_image_info entends past the end of the section)\n"; + } else + memcpy(&o, r, sizeof(struct objc_image_info32)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(o); + outs() << " version " << o.version << "\n"; + outs() << " flags " << format("0x%" PRIx32, o.flags); + if (o.flags & OBJC_IMAGE_IS_REPLACEMENT) + outs() << " OBJC_IMAGE_IS_REPLACEMENT"; + if (o.flags & OBJC_IMAGE_SUPPORTS_GC) + outs() << " OBJC_IMAGE_SUPPORTS_GC"; + swift_version = (o.flags >> 8) & 0xff; + if (swift_version != 0) { + if (swift_version == 1) + outs() << " Swift 1.0"; + else if (swift_version == 2) + outs() << " Swift 1.1"; + else + outs() << " unknown future Swift version (" << swift_version << ")"; + } + outs() << "\n"; +} + +static void print_image_info(SectionRef S, struct DisassembleInfo *info) { + uint32_t left, offset, p; + struct imageInfo_t o; + const char *r; + + StringRef SectName; + S.getName(SectName); + DataRefImpl Ref = S.getRawDataRefImpl(); + StringRef SegName = info->O->getSectionFinalSegmentName(Ref); + outs() << "Contents of (" << SegName << "," << SectName << ") section\n"; + p = S.getAddress(); + r = get_pointer_32(p, offset, left, S, info); + if (r == nullptr) + return; + memset(&o, '\0', sizeof(struct imageInfo_t)); + if (left < sizeof(struct imageInfo_t)) { + memcpy(&o, r, left); + outs() << " (imageInfo entends past the end of the section)\n"; + } else + memcpy(&o, r, sizeof(struct imageInfo_t)); + if (info->O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(o); + outs() << " version " << o.version << "\n"; + outs() << " flags " << format("0x%" PRIx32, o.flags); + if (o.flags & 0x1) + outs() << " F&C"; + if (o.flags & 0x2) + outs() << " GC"; + if (o.flags & 0x4) + outs() << " GC-only"; + else + outs() << " RR"; + outs() << "\n"; +} + +static void printObjc2_64bit_MetaData(MachOObjectFile *O, bool verbose) { + SymbolAddressMap AddrMap; + if (verbose) + CreateSymbolAddressMap(O, &AddrMap); + + std::vector<SectionRef> Sections; + for (const SectionRef &Section : O->sections()) { + StringRef SectName; + Section.getName(SectName); + Sections.push_back(Section); + } + + struct DisassembleInfo info; + // Set up the block of info used by the Symbolizer call backs. + info.verbose = verbose; + info.O = O; + info.AddrMap = &AddrMap; + info.Sections = &Sections; + info.class_name = nullptr; + info.selector_name = nullptr; + info.method = nullptr; + info.demangled_name = nullptr; + info.bindtable = nullptr; + info.adrp_addr = 0; + info.adrp_inst = 0; + + const SectionRef CL = get_section(O, "__OBJC2", "__class_list"); + if (CL != SectionRef()) { + info.S = CL; + walk_pointer_list_64("class", CL, O, &info, print_class64_t); + } else { + const SectionRef CL = get_section(O, "__DATA", "__objc_classlist"); + info.S = CL; + walk_pointer_list_64("class", CL, O, &info, print_class64_t); + } + + const SectionRef CR = get_section(O, "__OBJC2", "__class_refs"); + if (CR != SectionRef()) { + info.S = CR; + walk_pointer_list_64("class refs", CR, O, &info, nullptr); + } else { + const SectionRef CR = get_section(O, "__DATA", "__objc_classrefs"); + info.S = CR; + walk_pointer_list_64("class refs", CR, O, &info, nullptr); + } + + const SectionRef SR = get_section(O, "__OBJC2", "__super_refs"); + if (SR != SectionRef()) { + info.S = SR; + walk_pointer_list_64("super refs", SR, O, &info, nullptr); + } else { + const SectionRef SR = get_section(O, "__DATA", "__objc_superrefs"); + info.S = SR; + walk_pointer_list_64("super refs", SR, O, &info, nullptr); + } + + const SectionRef CA = get_section(O, "__OBJC2", "__category_list"); + if (CA != SectionRef()) { + info.S = CA; + walk_pointer_list_64("category", CA, O, &info, print_category64_t); + } else { + const SectionRef CA = get_section(O, "__DATA", "__objc_catlist"); + info.S = CA; + walk_pointer_list_64("category", CA, O, &info, print_category64_t); + } + + const SectionRef PL = get_section(O, "__OBJC2", "__protocol_list"); + if (PL != SectionRef()) { + info.S = PL; + walk_pointer_list_64("protocol", PL, O, &info, nullptr); + } else { + const SectionRef PL = get_section(O, "__DATA", "__objc_protolist"); + info.S = PL; + walk_pointer_list_64("protocol", PL, O, &info, nullptr); + } + + const SectionRef MR = get_section(O, "__OBJC2", "__message_refs"); + if (MR != SectionRef()) { + info.S = MR; + print_message_refs64(MR, &info); + } else { + const SectionRef MR = get_section(O, "__DATA", "__objc_msgrefs"); + info.S = MR; + print_message_refs64(MR, &info); + } + + const SectionRef II = get_section(O, "__OBJC2", "__image_info"); + if (II != SectionRef()) { + info.S = II; + print_image_info64(II, &info); + } else { + const SectionRef II = get_section(O, "__DATA", "__objc_imageinfo"); + info.S = II; + print_image_info64(II, &info); + } + + if (info.bindtable != nullptr) + delete info.bindtable; +} + +static void printObjc2_32bit_MetaData(MachOObjectFile *O, bool verbose) { + SymbolAddressMap AddrMap; + if (verbose) + CreateSymbolAddressMap(O, &AddrMap); + + std::vector<SectionRef> Sections; + for (const SectionRef &Section : O->sections()) { + StringRef SectName; + Section.getName(SectName); + Sections.push_back(Section); + } + + struct DisassembleInfo info; + // Set up the block of info used by the Symbolizer call backs. + info.verbose = verbose; + info.O = O; + info.AddrMap = &AddrMap; + info.Sections = &Sections; + info.class_name = nullptr; + info.selector_name = nullptr; + info.method = nullptr; + info.demangled_name = nullptr; + info.bindtable = nullptr; + info.adrp_addr = 0; + info.adrp_inst = 0; + + const SectionRef CL = get_section(O, "__OBJC2", "__class_list"); + if (CL != SectionRef()) { + info.S = CL; + walk_pointer_list_32("class", CL, O, &info, print_class32_t); + } else { + const SectionRef CL = get_section(O, "__DATA", "__objc_classlist"); + info.S = CL; + walk_pointer_list_32("class", CL, O, &info, print_class32_t); + } + + const SectionRef CR = get_section(O, "__OBJC2", "__class_refs"); + if (CR != SectionRef()) { + info.S = CR; + walk_pointer_list_32("class refs", CR, O, &info, nullptr); + } else { + const SectionRef CR = get_section(O, "__DATA", "__objc_classrefs"); + info.S = CR; + walk_pointer_list_32("class refs", CR, O, &info, nullptr); + } + + const SectionRef SR = get_section(O, "__OBJC2", "__super_refs"); + if (SR != SectionRef()) { + info.S = SR; + walk_pointer_list_32("super refs", SR, O, &info, nullptr); + } else { + const SectionRef SR = get_section(O, "__DATA", "__objc_superrefs"); + info.S = SR; + walk_pointer_list_32("super refs", SR, O, &info, nullptr); + } + + const SectionRef CA = get_section(O, "__OBJC2", "__category_list"); + if (CA != SectionRef()) { + info.S = CA; + walk_pointer_list_32("category", CA, O, &info, print_category32_t); + } else { + const SectionRef CA = get_section(O, "__DATA", "__objc_catlist"); + info.S = CA; + walk_pointer_list_32("category", CA, O, &info, print_category32_t); + } + + const SectionRef PL = get_section(O, "__OBJC2", "__protocol_list"); + if (PL != SectionRef()) { + info.S = PL; + walk_pointer_list_32("protocol", PL, O, &info, nullptr); + } else { + const SectionRef PL = get_section(O, "__DATA", "__objc_protolist"); + info.S = PL; + walk_pointer_list_32("protocol", PL, O, &info, nullptr); + } + + const SectionRef MR = get_section(O, "__OBJC2", "__message_refs"); + if (MR != SectionRef()) { + info.S = MR; + print_message_refs32(MR, &info); + } else { + const SectionRef MR = get_section(O, "__DATA", "__objc_msgrefs"); + info.S = MR; + print_message_refs32(MR, &info); + } + + const SectionRef II = get_section(O, "__OBJC2", "__image_info"); + if (II != SectionRef()) { + info.S = II; + print_image_info32(II, &info); + } else { + const SectionRef II = get_section(O, "__DATA", "__objc_imageinfo"); + info.S = II; + print_image_info32(II, &info); + } +} + +static bool printObjc1_32bit_MetaData(MachOObjectFile *O, bool verbose) { + uint32_t i, j, p, offset, xoffset, left, defs_left, def; + const char *r, *name, *defs; + struct objc_module_t module; + SectionRef S, xS; + struct objc_symtab_t symtab; + struct objc_class_t objc_class; + struct objc_category_t objc_category; + + outs() << "Objective-C segment\n"; + S = get_section(O, "__OBJC", "__module_info"); + if (S == SectionRef()) + return false; + + SymbolAddressMap AddrMap; + if (verbose) + CreateSymbolAddressMap(O, &AddrMap); + + std::vector<SectionRef> Sections; + for (const SectionRef &Section : O->sections()) { + StringRef SectName; + Section.getName(SectName); + Sections.push_back(Section); + } + + struct DisassembleInfo info; + // Set up the block of info used by the Symbolizer call backs. + info.verbose = verbose; + info.O = O; + info.AddrMap = &AddrMap; + info.Sections = &Sections; + info.class_name = nullptr; + info.selector_name = nullptr; + info.method = nullptr; + info.demangled_name = nullptr; + info.bindtable = nullptr; + info.adrp_addr = 0; + info.adrp_inst = 0; + + for (i = 0; i < S.getSize(); i += sizeof(struct objc_module_t)) { + p = S.getAddress() + i; + r = get_pointer_32(p, offset, left, S, &info, true); + if (r == nullptr) + return true; + memset(&module, '\0', sizeof(struct objc_module_t)); + if (left < sizeof(struct objc_module_t)) { + memcpy(&module, r, left); + outs() << " (module extends past end of __module_info section)\n"; + } else + memcpy(&module, r, sizeof(struct objc_module_t)); + if (O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(module); + + outs() << "Module " << format("0x%" PRIx32, p) << "\n"; + outs() << " version " << module.version << "\n"; + outs() << " size " << module.size << "\n"; + outs() << " name "; + name = get_pointer_32(module.name, xoffset, left, xS, &info, true); + if (name != nullptr) + outs() << format("%.*s", left, name); + else + outs() << format("0x%08" PRIx32, module.name) + << "(not in an __OBJC section)"; + outs() << "\n"; + + r = get_pointer_32(module.symtab, xoffset, left, xS, &info, true); + if (module.symtab == 0 || r == nullptr) { + outs() << " symtab " << format("0x%08" PRIx32, module.symtab) + << " (not in an __OBJC section)\n"; + continue; + } + outs() << " symtab " << format("0x%08" PRIx32, module.symtab) << "\n"; + memset(&symtab, '\0', sizeof(struct objc_symtab_t)); + defs_left = 0; + defs = nullptr; + if (left < sizeof(struct objc_symtab_t)) { + memcpy(&symtab, r, left); + outs() << "\tsymtab extends past end of an __OBJC section)\n"; + } else { + memcpy(&symtab, r, sizeof(struct objc_symtab_t)); + if (left > sizeof(struct objc_symtab_t)) { + defs_left = left - sizeof(struct objc_symtab_t); + defs = r + sizeof(struct objc_symtab_t); + } + } + if (O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(symtab); + + outs() << "\tsel_ref_cnt " << symtab.sel_ref_cnt << "\n"; + r = get_pointer_32(symtab.refs, xoffset, left, xS, &info, true); + outs() << "\trefs " << format("0x%08" PRIx32, symtab.refs); + if (r == nullptr) + outs() << " (not in an __OBJC section)"; + outs() << "\n"; + outs() << "\tcls_def_cnt " << symtab.cls_def_cnt << "\n"; + outs() << "\tcat_def_cnt " << symtab.cat_def_cnt << "\n"; + if (symtab.cls_def_cnt > 0) + outs() << "\tClass Definitions\n"; + for (j = 0; j < symtab.cls_def_cnt; j++) { + if ((j + 1) * sizeof(uint32_t) > defs_left) { + outs() << "\t(remaining class defs entries entends past the end of the " + << "section)\n"; + break; + } + memcpy(&def, defs + j * sizeof(uint32_t), sizeof(uint32_t)); + if (O->isLittleEndian() != sys::IsLittleEndianHost) + sys::swapByteOrder(def); + + r = get_pointer_32(def, xoffset, left, xS, &info, true); + outs() << "\tdefs[" << j << "] " << format("0x%08" PRIx32, def); + if (r != nullptr) { + if (left > sizeof(struct objc_class_t)) { + outs() << "\n"; + memcpy(&objc_class, r, sizeof(struct objc_class_t)); + } else { + outs() << " (entends past the end of the section)\n"; + memset(&objc_class, '\0', sizeof(struct objc_class_t)); + memcpy(&objc_class, r, left); + } + if (O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(objc_class); + print_objc_class_t(&objc_class, &info); + } else { + outs() << "(not in an __OBJC section)\n"; + } + + if (CLS_GETINFO(&objc_class, CLS_CLASS)) { + outs() << "\tMeta Class"; + r = get_pointer_32(objc_class.isa, xoffset, left, xS, &info, true); + if (r != nullptr) { + if (left > sizeof(struct objc_class_t)) { + outs() << "\n"; + memcpy(&objc_class, r, sizeof(struct objc_class_t)); + } else { + outs() << " (entends past the end of the section)\n"; + memset(&objc_class, '\0', sizeof(struct objc_class_t)); + memcpy(&objc_class, r, left); + } + if (O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(objc_class); + print_objc_class_t(&objc_class, &info); + } else { + outs() << "(not in an __OBJC section)\n"; + } + } + } + if (symtab.cat_def_cnt > 0) + outs() << "\tCategory Definitions\n"; + for (j = 0; j < symtab.cat_def_cnt; j++) { + if ((j + symtab.cls_def_cnt + 1) * sizeof(uint32_t) > defs_left) { + outs() << "\t(remaining category defs entries entends past the end of " + << "the section)\n"; + break; + } + memcpy(&def, defs + (j + symtab.cls_def_cnt) * sizeof(uint32_t), + sizeof(uint32_t)); + if (O->isLittleEndian() != sys::IsLittleEndianHost) + sys::swapByteOrder(def); + + r = get_pointer_32(def, xoffset, left, xS, &info, true); + outs() << "\tdefs[" << j + symtab.cls_def_cnt << "] " + << format("0x%08" PRIx32, def); + if (r != nullptr) { + if (left > sizeof(struct objc_category_t)) { + outs() << "\n"; + memcpy(&objc_category, r, sizeof(struct objc_category_t)); + } else { + outs() << " (entends past the end of the section)\n"; + memset(&objc_category, '\0', sizeof(struct objc_category_t)); + memcpy(&objc_category, r, left); + } + if (O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(objc_category); + print_objc_objc_category_t(&objc_category, &info); + } else { + outs() << "(not in an __OBJC section)\n"; + } + } + } + const SectionRef II = get_section(O, "__OBJC", "__image_info"); + if (II != SectionRef()) + print_image_info(II, &info); + + return true; +} + +static void DumpProtocolSection(MachOObjectFile *O, const char *sect, + uint32_t size, uint32_t addr) { + SymbolAddressMap AddrMap; + CreateSymbolAddressMap(O, &AddrMap); + + std::vector<SectionRef> Sections; + for (const SectionRef &Section : O->sections()) { + StringRef SectName; + Section.getName(SectName); + Sections.push_back(Section); + } + + struct DisassembleInfo info; + // Set up the block of info used by the Symbolizer call backs. + info.verbose = true; + info.O = O; + info.AddrMap = &AddrMap; + info.Sections = &Sections; + info.class_name = nullptr; + info.selector_name = nullptr; + info.method = nullptr; + info.demangled_name = nullptr; + info.bindtable = nullptr; + info.adrp_addr = 0; + info.adrp_inst = 0; + + const char *p; + struct objc_protocol_t protocol; + uint32_t left, paddr; + for (p = sect; p < sect + size; p += sizeof(struct objc_protocol_t)) { + memset(&protocol, '\0', sizeof(struct objc_protocol_t)); + left = size - (p - sect); + if (left < sizeof(struct objc_protocol_t)) { + outs() << "Protocol extends past end of __protocol section\n"; + memcpy(&protocol, p, left); + } else + memcpy(&protocol, p, sizeof(struct objc_protocol_t)); + if (O->isLittleEndian() != sys::IsLittleEndianHost) + swapStruct(protocol); + paddr = addr + (p - sect); + outs() << "Protocol " << format("0x%" PRIx32, paddr); + if (print_protocol(paddr, 0, &info)) + outs() << "(not in an __OBJC section)\n"; + } +} + +static void printObjcMetaData(MachOObjectFile *O, bool verbose) { + if (O->is64Bit()) + printObjc2_64bit_MetaData(O, verbose); + else { + MachO::mach_header H; + H = O->getHeader(); + if (H.cputype == MachO::CPU_TYPE_ARM) + printObjc2_32bit_MetaData(O, verbose); + else { + // This is the 32-bit non-arm cputype case. Which is normally + // the first Objective-C ABI. But it may be the case of a + // binary for the iOS simulator which is the second Objective-C + // ABI. In that case printObjc1_32bit_MetaData() will determine that + // and return false. + if (printObjc1_32bit_MetaData(O, verbose) == false) + printObjc2_32bit_MetaData(O, verbose); + } + } +} + // GuessLiteralPointer returns a string which for the item in the Mach-O file // for the address passed in as ReferenceValue for printing as a comment with // the instruction and also returns the corresponding type of that item @@ -1705,9 +5601,10 @@ uint64_t get_objc2_64bit_selref(uint64_t ReferenceValue, // // If there is no item in the Mach-O file for the address passed in as // ReferenceValue nullptr is returned and ReferenceType is unchanged. -const char *GuessLiteralPointer(uint64_t ReferenceValue, uint64_t ReferencePC, - uint64_t *ReferenceType, - struct DisassembleInfo *info) { +static const char *GuessLiteralPointer(uint64_t ReferenceValue, + uint64_t ReferencePC, + uint64_t *ReferenceType, + struct DisassembleInfo *info) { // First see if there is an external relocation entry at the ReferencePC. uint64_t sect_addr = info->S.getAddress(); uint64_t sect_offset = ReferencePC - sect_addr; @@ -1749,7 +5646,7 @@ const char *GuessLiteralPointer(uint64_t ReferenceValue, uint64_t ReferencePC, bool classref, selref, msgref, cfstring; uint64_t pointer_value = GuessPointerPointer(ReferenceValue, info, classref, selref, msgref, cfstring); - if (classref == true && pointer_value == 0) { + if (classref && pointer_value == 0) { // Note the ReferenceValue is a pointer into the __objc_classrefs section. // And the pointer_value in that section is typically zero as it will be // set by dyld as part of the "bind information". @@ -1765,7 +5662,7 @@ const char *GuessLiteralPointer(uint64_t ReferenceValue, uint64_t ReferencePC, } } - if (classref == true) { + if (classref) { *ReferenceType = LLVMDisassembler_ReferenceType_Out_Objc_Class_Ref; const char *name = get_objc2_64bit_class_name(pointer_value, ReferenceValue, info); @@ -1776,13 +5673,13 @@ const char *GuessLiteralPointer(uint64_t ReferenceValue, uint64_t ReferencePC, return name; } - if (cfstring == true) { + if (cfstring) { *ReferenceType = LLVMDisassembler_ReferenceType_Out_Objc_CFString_Ref; const char *name = get_objc2_64bit_cfstring_name(ReferenceValue, info); return name; } - if (selref == true && pointer_value == 0) + if (selref && pointer_value == 0) pointer_value = get_objc2_64bit_selref(ReferenceValue, info); if (pointer_value != 0) @@ -1790,10 +5687,10 @@ const char *GuessLiteralPointer(uint64_t ReferenceValue, uint64_t ReferencePC, const char *name = GuessCstringPointer(ReferenceValue, info); if (name) { - if (pointer_value != 0 && selref == true) { + if (pointer_value != 0 && selref) { *ReferenceType = LLVMDisassembler_ReferenceType_Out_Objc_Selector_Ref; info->selector_name = name; - } else if (pointer_value != 0 && msgref == true) { + } else if (pointer_value != 0 && msgref) { info->class_name = nullptr; *ReferenceType = LLVMDisassembler_ReferenceType_Out_Objc_Message_Ref; info->selector_name = name; @@ -1841,19 +5738,20 @@ const char *GuessLiteralPointer(uint64_t ReferenceValue, uint64_t ReferencePC, // SymbolValue is checked to be an address of literal pointer, symbol pointer, // or an Objective-C meta data reference. If so the output ReferenceType is // set to correspond to that as well as setting the ReferenceName. -const char *SymbolizerSymbolLookUp(void *DisInfo, uint64_t ReferenceValue, - uint64_t *ReferenceType, - uint64_t ReferencePC, - const char **ReferenceName) { +static const char *SymbolizerSymbolLookUp(void *DisInfo, + uint64_t ReferenceValue, + uint64_t *ReferenceType, + uint64_t ReferencePC, + const char **ReferenceName) { struct DisassembleInfo *info = (struct DisassembleInfo *)DisInfo; // If no verbose symbolic information is wanted then just return nullptr. - if (info->verbose == false) { + if (!info->verbose) { *ReferenceName = nullptr; *ReferenceType = LLVMDisassembler_ReferenceType_InOut_None; return nullptr; } - const char *SymbolName = GuessSymbolName(ReferenceValue, info); + const char *SymbolName = GuessSymbolName(ReferenceValue, info->AddrMap); if (*ReferenceType == LLVMDisassembler_ReferenceType_In_Branch) { *ReferenceName = GuessIndirectSymbol(ReferenceValue, info); @@ -2017,7 +5915,8 @@ static void emitComments(raw_svector_ostream &CommentStream, CommentStream.resync(); } -static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF) { +static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF, + StringRef DisSegName, StringRef DisSectName) { const char *McpuDefault = nullptr; const Target *ThumbTarget = nullptr; const Target *TheTarget = GetTarget(MachOOF, &McpuDefault, &ThumbTarget); @@ -2059,12 +5958,12 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF) { if (RelInfo) { Symbolizer.reset(TheTarget->createMCSymbolizer( TripleName, SymbolizerGetOpInfo, SymbolizerSymbolLookUp, - &SymbolizerInfo, &Ctx, RelInfo.release())); + &SymbolizerInfo, &Ctx, std::move(RelInfo))); DisAsm->setSymbolizer(std::move(Symbolizer)); } int AsmPrinterVariant = AsmInfo->getAssemblerDialect(); std::unique_ptr<MCInstPrinter> IP(TheTarget->createMCInstPrinter( - AsmPrinterVariant, *AsmInfo, *InstrInfo, *MRI, *STI)); + Triple(TripleName), AsmPrinterVariant, *AsmInfo, *InstrInfo, *MRI)); // Set the display preference for hex vs. decimal immediates. IP->setPrintImmHex(PrintImmHex); // Comment stream and backing vector. @@ -2107,13 +6006,13 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF) { if (ThumbRelInfo) { ThumbSymbolizer.reset(ThumbTarget->createMCSymbolizer( ThumbTripleName, SymbolizerGetOpInfo, SymbolizerSymbolLookUp, - &ThumbSymbolizerInfo, PtrThumbCtx, ThumbRelInfo.release())); + &ThumbSymbolizerInfo, PtrThumbCtx, std::move(ThumbRelInfo))); ThumbDisAsm->setSymbolizer(std::move(ThumbSymbolizer)); } int ThumbAsmPrinterVariant = ThumbAsmInfo->getAssemblerDialect(); ThumbIP.reset(ThumbTarget->createMCInstPrinter( - ThumbAsmPrinterVariant, *ThumbAsmInfo, *ThumbInstrInfo, *ThumbMRI, - *ThumbSTI)); + Triple(ThumbTripleName), ThumbAsmPrinterVariant, *ThumbAsmInfo, + *ThumbInstrInfo, *ThumbMRI)); // Set the display preference for hex vs. decimal immediates. ThumbIP->setPrintImmHex(PrintImmHex); } @@ -2135,7 +6034,7 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF) { SmallVector<uint64_t, 8> FoundFns; uint64_t BaseSegmentAddress; - getSectionsAndSymbols(Header, MachOOF, Sections, Symbols, FoundFns, + getSectionsAndSymbols(MachOOF, Sections, Symbols, FoundFns, BaseSegmentAddress); // Sort the symbols by address, just in case they didn't come in that way. @@ -2182,31 +6081,21 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF) { } // Setup the DIContext - diContext.reset(DIContext::getDWARFContext(*DbgObj)); + diContext.reset(new DWARFContextInMemory(*DbgObj)); } - // TODO: For now this only disassembles the (__TEXT,__text) section (see the - // checks in the code below at the top of this loop). It should allow a - // darwin otool(1) like -s option to disassemble any named segment & section - // that is marked as containing instructions with the attributes - // S_ATTR_PURE_INSTRUCTIONS or S_ATTR_SOME_INSTRUCTIONS in the flags field of - // the section structure. - outs() << "(__TEXT,__text) section\n"; + if (DumpSections.size() == 0) + outs() << "(" << DisSegName << "," << DisSectName << ") section\n"; for (unsigned SectIdx = 0; SectIdx != Sections.size(); SectIdx++) { - - bool SectIsText = Sections[SectIdx].isText(); - if (SectIsText == false) - continue; - StringRef SectName; - if (Sections[SectIdx].getName(SectName) || SectName != "__text") - continue; // Skip non-text sections + if (Sections[SectIdx].getName(SectName) || SectName != DisSectName) + continue; DataRefImpl DR = Sections[SectIdx].getRawDataRefImpl(); StringRef SegmentName = MachOOF->getSectionFinalSegmentName(DR); - if (SegmentName != "__TEXT") + if (SegmentName != DisSegName) continue; StringRef BytesStr; @@ -2234,6 +6123,7 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF) { // Create a map of symbol addresses to symbol names for use by // the SymbolizerSymbolLookUp() routine. SymbolAddressMap AddrMap; + bool DisSymNameFound = false; for (const SymbolRef &Symbol : MachOOF->symbols()) { SymbolRef::Type ST; Symbol.getType(ST); @@ -2244,10 +6134,16 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF) { StringRef SymName; Symbol.getName(SymName); AddrMap[Address] = SymName; + if (!DisSymName.empty() && DisSymName == SymName) + DisSymNameFound = true; } } + if (!DisSymName.empty() && !DisSymNameFound) { + outs() << "Can't find -dis-symname: " << DisSymName << "\n"; + return; + } // Set up the block of info used by the Symbolizer call backs. - SymbolizerInfo.verbose = true; + SymbolizerInfo.verbose = !NoSymbolicOperands; SymbolizerInfo.O = MachOOF; SymbolizerInfo.S = Sections[SectIdx]; SymbolizerInfo.AddrMap = &AddrMap; @@ -2260,7 +6156,7 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF) { SymbolizerInfo.adrp_addr = 0; SymbolizerInfo.adrp_inst = 0; // Same for the ThumbSymbolizer - ThumbSymbolizerInfo.verbose = true; + ThumbSymbolizerInfo.verbose = !NoSymbolicOperands; ThumbSymbolizerInfo.O = MachOOF; ThumbSymbolizerInfo.S = Sections[SectIdx]; ThumbSymbolizerInfo.AddrMap = &AddrMap; @@ -2288,6 +6184,10 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF) { if (!containsSym) continue; + // If we are only disassembling one symbol see if this is that symbol. + if (!DisSymName.empty() && DisSymName != SymName) + continue; + // Start at the address of the symbol relative to the section's address. uint64_t Start = 0; uint64_t SectionAddress = Sections[SectIdx].getAddress(); @@ -2328,13 +6228,15 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF) { MCInst Inst; uint64_t PC = SectAddress + Index; - if (FullLeadingAddr) { - if (MachOOF->is64Bit()) - outs() << format("%016" PRIx64, PC); - else - outs() << format("%08" PRIx64, PC); - } else { - outs() << format("%8" PRIx64 ":", PC); + if (!NoLeadingAddr) { + if (FullLeadingAddr) { + if (MachOOF->is64Bit()) + outs() << format("%016" PRIx64, PC); + else + outs() << format("%08" PRIx64, PC); + } else { + outs() << format("%8" PRIx64 ":", PC); + } } if (!NoShowRawInsn) outs() << "\t"; @@ -2351,9 +6253,7 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF) { DTI->second.getLength(Length); uint16_t Kind; DTI->second.getKind(Kind); - Size = DumpDataInCode(reinterpret_cast<const char *>(Bytes.data()) + - Index, - Length, Kind); + Size = DumpDataInCode(Bytes.data() + Index, Length, Kind); if ((Kind == MachO::DICE_KIND_JUMP_TABLE8) && (PC == (DTI->first + Length - 1)) && (Length & 1)) Size++; @@ -2372,16 +6272,15 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF) { DebugOut, Annotations); if (gotInst) { if (!NoShowRawInsn) { - DumpBytes(StringRef( - reinterpret_cast<const char *>(Bytes.data()) + Index, Size)); + dumpBytes(ArrayRef<uint8_t>(Bytes.data() + Index, Size), outs()); } formatted_raw_ostream FormattedOS(outs()); Annotations.flush(); StringRef AnnotationsStr = Annotations.str(); if (isThumb) - ThumbIP->printInst(&Inst, FormattedOS, AnnotationsStr); + ThumbIP->printInst(&Inst, FormattedOS, AnnotationsStr, *ThumbSTI); else - IP->printInst(&Inst, FormattedOS, AnnotationsStr); + IP->printInst(&Inst, FormattedOS, AnnotationsStr, *STI); emitComments(CommentStream, CommentsToEmit, FormattedOS, *AsmInfo); // Print debug info. @@ -2426,21 +6325,21 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF) { uint64_t PC = SectAddress + Index; if (DisAsm->getInstruction(Inst, InstSize, Bytes.slice(Index), PC, DebugOut, nulls())) { - if (FullLeadingAddr) { - if (MachOOF->is64Bit()) - outs() << format("%016" PRIx64, PC); - else - outs() << format("%08" PRIx64, PC); - } else { - outs() << format("%8" PRIx64 ":", PC); + if (!NoLeadingAddr) { + if (FullLeadingAddr) { + if (MachOOF->is64Bit()) + outs() << format("%016" PRIx64, PC); + else + outs() << format("%08" PRIx64, PC); + } else { + outs() << format("%8" PRIx64 ":", PC); + } } if (!NoShowRawInsn) { outs() << "\t"; - DumpBytes( - StringRef(reinterpret_cast<const char *>(Bytes.data()) + Index, - InstSize)); + dumpBytes(ArrayRef<uint8_t>(Bytes.data() + Index, InstSize), outs()); } - IP->printInst(&Inst, outs(), ""); + IP->printInst(&Inst, outs(), "", *STI); outs() << "\n"; } else { unsigned int Arch = MachOOF->getArch(); @@ -2544,7 +6443,7 @@ static void findUnwindRelocNameAddend(const MachOObjectFile *Obj, } auto RE = Obj->getRelocation(Reloc.getRawDataRefImpl()); - SectionRef RelocSection = Obj->getRelocationSection(RE); + SectionRef RelocSection = Obj->getAnyRelocationSection(RE); uint64_t SectionAddr = RelocSection.getAddress(); @@ -4427,15 +8326,12 @@ static void PrintLinkEditDataCommand(MachO::linkedit_data_command ld, outs() << "\n"; } -static void PrintLoadCommands(const MachOObjectFile *Obj, uint32_t ncmds, - uint32_t filetype, uint32_t cputype, - bool verbose) { - if (ncmds == 0) - return; +static void PrintLoadCommands(const MachOObjectFile *Obj, uint32_t filetype, + uint32_t cputype, bool verbose) { StringRef Buf = Obj->getData(); - MachOObjectFile::LoadCommandInfo Command = Obj->getFirstLoadCommandInfo(); - for (unsigned i = 0;; ++i) { - outs() << "Load command " << i << "\n"; + unsigned Index = 0; + for (const auto &Command : Obj->load_commands()) { + outs() << "Load command " << Index++ << "\n"; if (Command.C.cmd == MachO::LC_SEGMENT) { MachO::segment_command SLC = Obj->getSegmentLoadCommand(Command); const char *sg_segname = SLC.segname; @@ -4554,14 +8450,10 @@ static void PrintLoadCommands(const MachOObjectFile *Obj, uint32_t ncmds, // TODO: get and print the raw bytes of the load command. } // TODO: print all the other kinds of load commands. - if (i == ncmds - 1) - break; - else - Command = Obj->getNextLoadCommandInfo(Command); } } -static void getAndPrintMachHeader(const MachOObjectFile *Obj, uint32_t &ncmds, +static void getAndPrintMachHeader(const MachOObjectFile *Obj, uint32_t &filetype, uint32_t &cputype, bool verbose) { if (Obj->is64Bit()) { @@ -4569,7 +8461,6 @@ static void getAndPrintMachHeader(const MachOObjectFile *Obj, uint32_t &ncmds, H_64 = Obj->getHeader64(); PrintMachHeader(H_64.magic, H_64.cputype, H_64.cpusubtype, H_64.filetype, H_64.ncmds, H_64.sizeofcmds, H_64.flags, verbose); - ncmds = H_64.ncmds; filetype = H_64.filetype; cputype = H_64.cputype; } else { @@ -4577,7 +8468,6 @@ static void getAndPrintMachHeader(const MachOObjectFile *Obj, uint32_t &ncmds, H = Obj->getHeader(); PrintMachHeader(H.magic, H.cputype, H.cpusubtype, H.filetype, H.ncmds, H.sizeofcmds, H.flags, verbose); - ncmds = H.ncmds; filetype = H.filetype; cputype = H.cputype; } @@ -4585,11 +8475,10 @@ static void getAndPrintMachHeader(const MachOObjectFile *Obj, uint32_t &ncmds, void llvm::printMachOFileHeader(const object::ObjectFile *Obj) { const MachOObjectFile *file = dyn_cast<const MachOObjectFile>(Obj); - uint32_t ncmds = 0; uint32_t filetype = 0; uint32_t cputype = 0; - getAndPrintMachHeader(file, ncmds, filetype, cputype, true); - PrintLoadCommands(file, ncmds, filetype, cputype, true); + getAndPrintMachHeader(file, filetype, cputype, !NonVerbose); + PrintLoadCommands(file, filetype, cputype, !NonVerbose); } //===----------------------------------------------------------------------===// diff --git a/contrib/llvm/tools/llvm-objdump/llvm-objdump.cpp b/contrib/llvm/tools/llvm-objdump/llvm-objdump.cpp index aff6272..1152a15 100644 --- a/contrib/llvm/tools/llvm-objdump/llvm-objdump.cpp +++ b/contrib/llvm/tools/llvm-objdump/llvm-objdump.cpp @@ -32,12 +32,14 @@ #include "llvm/MC/MCRelocationInfo.h" #include "llvm/MC/MCSubtargetInfo.h" #include "llvm/Object/Archive.h" +#include "llvm/Object/ELFObjectFile.h" #include "llvm/Object/COFF.h" #include "llvm/Object/MachO.h" #include "llvm/Object/ObjectFile.h" #include "llvm/Support/Casting.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/Debug.h" +#include "llvm/Support/Errc.h" #include "llvm/Support/FileSystem.h" #include "llvm/Support/Format.h" #include "llvm/Support/GraphWriter.h" @@ -68,14 +70,14 @@ static cl::alias Disassembled("d", cl::desc("Alias for --disassemble"), cl::aliasopt(Disassemble)); -static cl::opt<bool> -Relocations("r", cl::desc("Display the relocation entries in the file")); +cl::opt<bool> +llvm::Relocations("r", cl::desc("Display the relocation entries in the file")); -static cl::opt<bool> -SectionContents("s", cl::desc("Display the content of each section")); +cl::opt<bool> +llvm::SectionContents("s", cl::desc("Display the content of each section")); -static cl::opt<bool> -SymbolTable("t", cl::desc("Display the symbol table")); +cl::opt<bool> +llvm::SymbolTable("t", cl::desc("Display the symbol table")); cl::opt<bool> llvm::ExportsTrie("exports-trie", cl::desc("Display mach-o exported symbols")); @@ -111,9 +113,9 @@ cl::opt<std::string> llvm::ArchName("arch-name", cl::desc("Target arch to disassemble for, " "see -version for available targets")); -static cl::opt<bool> -SectionHeaders("section-headers", cl::desc("Display summaries of the headers " - "for each section.")); +cl::opt<bool> +llvm::SectionHeaders("section-headers", cl::desc("Display summaries of the " + "headers for each section.")); static cl::alias SectionHeadersShort("headers", cl::desc("Alias for --section-headers"), cl::aliasopt(SectionHeaders)); @@ -132,8 +134,8 @@ llvm::NoShowRawInsn("no-show-raw-insn", cl::desc("When disassembling " "instructions, do not print " "the instruction bytes.")); -static cl::opt<bool> -UnwindInfo("unwind-info", cl::desc("Display unwind information")); +cl::opt<bool> +llvm::UnwindInfo("unwind-info", cl::desc("Display unwind information")); static cl::alias UnwindInfoShort("u", cl::desc("Alias for --unwind-info"), @@ -147,6 +149,10 @@ static cl::alias PrivateHeadersShort("p", cl::desc("Alias for --private-headers"), cl::aliasopt(PrivateHeaders)); +cl::opt<bool> + llvm::PrintImmHex("print-imm-hex", + cl::desc("Use hex format for immediate values")); + static StringRef ToolName; static int ReturnValue = EXIT_SUCCESS; @@ -160,6 +166,12 @@ bool llvm::error(std::error_code EC) { return true; } +static void report_error(StringRef File, std::error_code EC) { + assert(EC); + errs() << ToolName << ": '" << File << "': " << EC.message() << ".\n"; + ReturnValue = EXIT_FAILURE; +} + static const Target *getTarget(const ObjectFile *Obj = nullptr) { // Figure out the target triple. llvm::Triple TheTriple("unknown-unknown-unknown"); @@ -194,32 +206,6 @@ static const Target *getTarget(const ObjectFile *Obj = nullptr) { return TheTarget; } -void llvm::DumpBytes(StringRef bytes) { - static const char hex_rep[] = "0123456789abcdef"; - // FIXME: The real way to do this is to figure out the longest instruction - // and align to that size before printing. I'll fix this when I get - // around to outputting relocations. - // 15 is the longest x86 instruction - // 3 is for the hex rep of a byte + a space. - // 1 is for the null terminator. - enum { OutputSize = (15 * 3) + 1 }; - char output[OutputSize]; - - assert(bytes.size() <= 15 - && "DumpBytes only supports instructions of up to 15 bytes"); - memset(output, ' ', sizeof(output)); - unsigned index = 0; - for (StringRef::iterator i = bytes.begin(), - e = bytes.end(); i != e; ++i) { - output[index] = hex_rep[(*i & 0xF0) >> 4]; - output[index + 1] = hex_rep[*i & 0xF]; - index += 3; - } - - output[sizeof(output) - 1] = 0; - outs() << output; -} - bool llvm::RelocAddressLess(RelocationRef a, RelocationRef b) { uint64_t a_addr, b_addr; if (error(a.getOffset(a_addr))) return false; @@ -227,6 +213,474 @@ bool llvm::RelocAddressLess(RelocationRef a, RelocationRef b) { return a_addr < b_addr; } +namespace { +class PrettyPrinter { +public: + virtual ~PrettyPrinter(){} + virtual void printInst(MCInstPrinter &IP, const MCInst *MI, + ArrayRef<uint8_t> Bytes, uint64_t Address, + raw_ostream &OS, StringRef Annot, + MCSubtargetInfo const &STI) { + outs() << format("%8" PRIx64 ":", Address); + if (!NoShowRawInsn) { + outs() << "\t"; + dumpBytes(Bytes, outs()); + } + IP.printInst(MI, outs(), "", STI); + } +}; +PrettyPrinter PrettyPrinterInst; +class HexagonPrettyPrinter : public PrettyPrinter { +public: + void printLead(ArrayRef<uint8_t> Bytes, uint64_t Address, + raw_ostream &OS) { + uint32_t opcode = + (Bytes[3] << 24) | (Bytes[2] << 16) | (Bytes[1] << 8) | Bytes[0]; + OS << format("%8" PRIx64 ":", Address); + if (!NoShowRawInsn) { + OS << "\t"; + dumpBytes(Bytes.slice(0, 4), OS); + OS << format("%08" PRIx32, opcode); + } + } + void printInst(MCInstPrinter &IP, const MCInst *MI, + ArrayRef<uint8_t> Bytes, uint64_t Address, + raw_ostream &OS, StringRef Annot, + MCSubtargetInfo const &STI) override { + std::string Buffer; + { + raw_string_ostream TempStream(Buffer); + IP.printInst(MI, TempStream, "", STI); + } + StringRef Contents(Buffer); + // Split off bundle attributes + auto PacketBundle = Contents.rsplit('\n'); + // Split off first instruction from the rest + auto HeadTail = PacketBundle.first.split('\n'); + auto Preamble = " { "; + auto Separator = ""; + while(!HeadTail.first.empty()) { + OS << Separator; + Separator = "\n"; + printLead(Bytes, Address, OS); + OS << Preamble; + Preamble = " "; + StringRef Inst; + auto Duplex = HeadTail.first.split('\v'); + if(!Duplex.second.empty()){ + OS << Duplex.first; + OS << "; "; + Inst = Duplex.second; + } + else + Inst = HeadTail.first; + OS << Inst; + Bytes = Bytes.slice(4); + Address += 4; + HeadTail = HeadTail.second.split('\n'); + } + OS << " } " << PacketBundle.second; + } +}; +HexagonPrettyPrinter HexagonPrettyPrinterInst; +PrettyPrinter &selectPrettyPrinter(Triple const &Triple) { + switch(Triple.getArch()) { + default: + return PrettyPrinterInst; + case Triple::hexagon: + return HexagonPrettyPrinterInst; + } +} +} + +template <class ELFT> +static const typename ELFObjectFile<ELFT>::Elf_Rel * +getRel(const ELFFile<ELFT> &EF, DataRefImpl Rel) { + typedef typename ELFObjectFile<ELFT>::Elf_Rel Elf_Rel; + return EF.template getEntry<Elf_Rel>(Rel.d.a, Rel.d.b); +} + +template <class ELFT> +static const typename ELFObjectFile<ELFT>::Elf_Rela * +getRela(const ELFFile<ELFT> &EF, DataRefImpl Rela) { + typedef typename ELFObjectFile<ELFT>::Elf_Rela Elf_Rela; + return EF.template getEntry<Elf_Rela>(Rela.d.a, Rela.d.b); +} + +template <class ELFT> +static std::error_code getRelocationValueString(const ELFObjectFile<ELFT> *Obj, + DataRefImpl Rel, + SmallVectorImpl<char> &Result) { + typedef typename ELFObjectFile<ELFT>::Elf_Sym Elf_Sym; + typedef typename ELFObjectFile<ELFT>::Elf_Shdr Elf_Shdr; + const ELFFile<ELFT> &EF = *Obj->getELFFile(); + + const Elf_Shdr *sec = EF.getSection(Rel.d.a); + uint8_t type; + StringRef res; + int64_t addend = 0; + uint16_t symbol_index = 0; + switch (sec->sh_type) { + default: + return object_error::parse_failed; + case ELF::SHT_REL: { + type = getRel(EF, Rel)->getType(EF.isMips64EL()); + symbol_index = getRel(EF, Rel)->getSymbol(EF.isMips64EL()); + // TODO: Read implicit addend from section data. + break; + } + case ELF::SHT_RELA: { + type = getRela(EF, Rel)->getType(EF.isMips64EL()); + symbol_index = getRela(EF, Rel)->getSymbol(EF.isMips64EL()); + addend = getRela(EF, Rel)->r_addend; + break; + } + } + const Elf_Sym *symb = + EF.template getEntry<Elf_Sym>(sec->sh_link, symbol_index); + StringRef Target; + const Elf_Shdr *SymSec = EF.getSection(symb); + if (symb->getType() == ELF::STT_SECTION) { + ErrorOr<StringRef> SecName = EF.getSectionName(SymSec); + if (std::error_code EC = SecName.getError()) + return EC; + Target = *SecName; + } else { + ErrorOr<StringRef> SymName = + EF.getSymbolName(EF.getSection(sec->sh_link), symb); + if (!SymName) + return SymName.getError(); + Target = *SymName; + } + switch (EF.getHeader()->e_machine) { + case ELF::EM_X86_64: + switch (type) { + case ELF::R_X86_64_PC8: + case ELF::R_X86_64_PC16: + case ELF::R_X86_64_PC32: { + std::string fmtbuf; + raw_string_ostream fmt(fmtbuf); + fmt << Target << (addend < 0 ? "" : "+") << addend << "-P"; + fmt.flush(); + Result.append(fmtbuf.begin(), fmtbuf.end()); + } break; + case ELF::R_X86_64_8: + case ELF::R_X86_64_16: + case ELF::R_X86_64_32: + case ELF::R_X86_64_32S: + case ELF::R_X86_64_64: { + std::string fmtbuf; + raw_string_ostream fmt(fmtbuf); + fmt << Target << (addend < 0 ? "" : "+") << addend; + fmt.flush(); + Result.append(fmtbuf.begin(), fmtbuf.end()); + } break; + default: + res = "Unknown"; + } + break; + case ELF::EM_AARCH64: { + std::string fmtbuf; + raw_string_ostream fmt(fmtbuf); + fmt << Target; + if (addend != 0) + fmt << (addend < 0 ? "" : "+") << addend; + fmt.flush(); + Result.append(fmtbuf.begin(), fmtbuf.end()); + break; + } + case ELF::EM_386: + case ELF::EM_ARM: + case ELF::EM_HEXAGON: + case ELF::EM_MIPS: + res = Target; + break; + default: + res = "Unknown"; + } + if (Result.empty()) + Result.append(res.begin(), res.end()); + return std::error_code(); +} + +static std::error_code getRelocationValueString(const ELFObjectFileBase *Obj, + const RelocationRef &RelRef, + SmallVectorImpl<char> &Result) { + DataRefImpl Rel = RelRef.getRawDataRefImpl(); + if (auto *ELF32LE = dyn_cast<ELF32LEObjectFile>(Obj)) + return getRelocationValueString(ELF32LE, Rel, Result); + if (auto *ELF64LE = dyn_cast<ELF64LEObjectFile>(Obj)) + return getRelocationValueString(ELF64LE, Rel, Result); + if (auto *ELF32BE = dyn_cast<ELF32BEObjectFile>(Obj)) + return getRelocationValueString(ELF32BE, Rel, Result); + auto *ELF64BE = cast<ELF64BEObjectFile>(Obj); + return getRelocationValueString(ELF64BE, Rel, Result); +} + +static std::error_code getRelocationValueString(const COFFObjectFile *Obj, + const RelocationRef &Rel, + SmallVectorImpl<char> &Result) { + symbol_iterator SymI = Rel.getSymbol(); + StringRef SymName; + if (std::error_code EC = SymI->getName(SymName)) + return EC; + Result.append(SymName.begin(), SymName.end()); + return std::error_code(); +} + +static void printRelocationTargetName(const MachOObjectFile *O, + const MachO::any_relocation_info &RE, + raw_string_ostream &fmt) { + bool IsScattered = O->isRelocationScattered(RE); + + // Target of a scattered relocation is an address. In the interest of + // generating pretty output, scan through the symbol table looking for a + // symbol that aligns with that address. If we find one, print it. + // Otherwise, we just print the hex address of the target. + if (IsScattered) { + uint32_t Val = O->getPlainRelocationSymbolNum(RE); + + for (const SymbolRef &Symbol : O->symbols()) { + std::error_code ec; + uint64_t Addr; + StringRef Name; + + if ((ec = Symbol.getAddress(Addr))) + report_fatal_error(ec.message()); + if (Addr != Val) + continue; + if ((ec = Symbol.getName(Name))) + report_fatal_error(ec.message()); + fmt << Name; + return; + } + + // If we couldn't find a symbol that this relocation refers to, try + // to find a section beginning instead. + for (const SectionRef &Section : O->sections()) { + std::error_code ec; + + StringRef Name; + uint64_t Addr = Section.getAddress(); + if (Addr != Val) + continue; + if ((ec = Section.getName(Name))) + report_fatal_error(ec.message()); + fmt << Name; + return; + } + + fmt << format("0x%x", Val); + return; + } + + StringRef S; + bool isExtern = O->getPlainRelocationExternal(RE); + uint64_t Val = O->getPlainRelocationSymbolNum(RE); + + if (isExtern) { + symbol_iterator SI = O->symbol_begin(); + advance(SI, Val); + SI->getName(S); + } else { + section_iterator SI = O->section_begin(); + // Adjust for the fact that sections are 1-indexed. + advance(SI, Val - 1); + SI->getName(S); + } + + fmt << S; +} + +static std::error_code getRelocationValueString(const MachOObjectFile *Obj, + const RelocationRef &RelRef, + SmallVectorImpl<char> &Result) { + DataRefImpl Rel = RelRef.getRawDataRefImpl(); + MachO::any_relocation_info RE = Obj->getRelocation(Rel); + + unsigned Arch = Obj->getArch(); + + std::string fmtbuf; + raw_string_ostream fmt(fmtbuf); + unsigned Type = Obj->getAnyRelocationType(RE); + bool IsPCRel = Obj->getAnyRelocationPCRel(RE); + + // Determine any addends that should be displayed with the relocation. + // These require decoding the relocation type, which is triple-specific. + + // X86_64 has entirely custom relocation types. + if (Arch == Triple::x86_64) { + bool isPCRel = Obj->getAnyRelocationPCRel(RE); + + switch (Type) { + case MachO::X86_64_RELOC_GOT_LOAD: + case MachO::X86_64_RELOC_GOT: { + printRelocationTargetName(Obj, RE, fmt); + fmt << "@GOT"; + if (isPCRel) + fmt << "PCREL"; + break; + } + case MachO::X86_64_RELOC_SUBTRACTOR: { + DataRefImpl RelNext = Rel; + Obj->moveRelocationNext(RelNext); + MachO::any_relocation_info RENext = Obj->getRelocation(RelNext); + + // X86_64_RELOC_SUBTRACTOR must be followed by a relocation of type + // X86_64_RELOC_UNSIGNED. + // NOTE: Scattered relocations don't exist on x86_64. + unsigned RType = Obj->getAnyRelocationType(RENext); + if (RType != MachO::X86_64_RELOC_UNSIGNED) + report_fatal_error("Expected X86_64_RELOC_UNSIGNED after " + "X86_64_RELOC_SUBTRACTOR."); + + // The X86_64_RELOC_UNSIGNED contains the minuend symbol; + // X86_64_RELOC_SUBTRACTOR contains the subtrahend. + printRelocationTargetName(Obj, RENext, fmt); + fmt << "-"; + printRelocationTargetName(Obj, RE, fmt); + break; + } + case MachO::X86_64_RELOC_TLV: + printRelocationTargetName(Obj, RE, fmt); + fmt << "@TLV"; + if (isPCRel) + fmt << "P"; + break; + case MachO::X86_64_RELOC_SIGNED_1: + printRelocationTargetName(Obj, RE, fmt); + fmt << "-1"; + break; + case MachO::X86_64_RELOC_SIGNED_2: + printRelocationTargetName(Obj, RE, fmt); + fmt << "-2"; + break; + case MachO::X86_64_RELOC_SIGNED_4: + printRelocationTargetName(Obj, RE, fmt); + fmt << "-4"; + break; + default: + printRelocationTargetName(Obj, RE, fmt); + break; + } + // X86 and ARM share some relocation types in common. + } else if (Arch == Triple::x86 || Arch == Triple::arm || + Arch == Triple::ppc) { + // Generic relocation types... + switch (Type) { + case MachO::GENERIC_RELOC_PAIR: // prints no info + return std::error_code(); + case MachO::GENERIC_RELOC_SECTDIFF: { + DataRefImpl RelNext = Rel; + Obj->moveRelocationNext(RelNext); + MachO::any_relocation_info RENext = Obj->getRelocation(RelNext); + + // X86 sect diff's must be followed by a relocation of type + // GENERIC_RELOC_PAIR. + unsigned RType = Obj->getAnyRelocationType(RENext); + + if (RType != MachO::GENERIC_RELOC_PAIR) + report_fatal_error("Expected GENERIC_RELOC_PAIR after " + "GENERIC_RELOC_SECTDIFF."); + + printRelocationTargetName(Obj, RE, fmt); + fmt << "-"; + printRelocationTargetName(Obj, RENext, fmt); + break; + } + } + + if (Arch == Triple::x86 || Arch == Triple::ppc) { + switch (Type) { + case MachO::GENERIC_RELOC_LOCAL_SECTDIFF: { + DataRefImpl RelNext = Rel; + Obj->moveRelocationNext(RelNext); + MachO::any_relocation_info RENext = Obj->getRelocation(RelNext); + + // X86 sect diff's must be followed by a relocation of type + // GENERIC_RELOC_PAIR. + unsigned RType = Obj->getAnyRelocationType(RENext); + if (RType != MachO::GENERIC_RELOC_PAIR) + report_fatal_error("Expected GENERIC_RELOC_PAIR after " + "GENERIC_RELOC_LOCAL_SECTDIFF."); + + printRelocationTargetName(Obj, RE, fmt); + fmt << "-"; + printRelocationTargetName(Obj, RENext, fmt); + break; + } + case MachO::GENERIC_RELOC_TLV: { + printRelocationTargetName(Obj, RE, fmt); + fmt << "@TLV"; + if (IsPCRel) + fmt << "P"; + break; + } + default: + printRelocationTargetName(Obj, RE, fmt); + } + } else { // ARM-specific relocations + switch (Type) { + case MachO::ARM_RELOC_HALF: + case MachO::ARM_RELOC_HALF_SECTDIFF: { + // Half relocations steal a bit from the length field to encode + // whether this is an upper16 or a lower16 relocation. + bool isUpper = Obj->getAnyRelocationLength(RE) >> 1; + + if (isUpper) + fmt << ":upper16:("; + else + fmt << ":lower16:("; + printRelocationTargetName(Obj, RE, fmt); + + DataRefImpl RelNext = Rel; + Obj->moveRelocationNext(RelNext); + MachO::any_relocation_info RENext = Obj->getRelocation(RelNext); + + // ARM half relocs must be followed by a relocation of type + // ARM_RELOC_PAIR. + unsigned RType = Obj->getAnyRelocationType(RENext); + if (RType != MachO::ARM_RELOC_PAIR) + report_fatal_error("Expected ARM_RELOC_PAIR after " + "ARM_RELOC_HALF"); + + // NOTE: The half of the target virtual address is stashed in the + // address field of the secondary relocation, but we can't reverse + // engineer the constant offset from it without decoding the movw/movt + // instruction to find the other half in its immediate field. + + // ARM_RELOC_HALF_SECTDIFF encodes the second section in the + // symbol/section pointer of the follow-on relocation. + if (Type == MachO::ARM_RELOC_HALF_SECTDIFF) { + fmt << "-"; + printRelocationTargetName(Obj, RENext, fmt); + } + + fmt << ")"; + break; + } + default: { printRelocationTargetName(Obj, RE, fmt); } + } + } + } else + printRelocationTargetName(Obj, RE, fmt); + + fmt.flush(); + Result.append(fmtbuf.begin(), fmtbuf.end()); + return std::error_code(); +} + +static std::error_code getRelocationValueString(const RelocationRef &Rel, + SmallVectorImpl<char> &Result) { + const ObjectFile *Obj = Rel.getObjectFile(); + if (auto *ELF = dyn_cast<ELFObjectFileBase>(Obj)) + return getRelocationValueString(ELF, Rel, Result); + if (auto *COFF = dyn_cast<COFFObjectFile>(Obj)) + return getRelocationValueString(COFF, Rel, Result); + auto *MachO = cast<MachOObjectFile>(Obj); + return getRelocationValueString(MachO, Rel, Result); +} + static void DisassembleObject(const ObjectFile *Obj, bool InlineRelocs) { const Target *TheTarget = getTarget(Obj); // getTarget() will have already issued a diagnostic if necessary, so @@ -287,12 +741,14 @@ static void DisassembleObject(const ObjectFile *Obj, bool InlineRelocs) { int AsmPrinterVariant = AsmInfo->getAssemblerDialect(); std::unique_ptr<MCInstPrinter> IP(TheTarget->createMCInstPrinter( - AsmPrinterVariant, *AsmInfo, *MII, *MRI, *STI)); + Triple(TripleName), AsmPrinterVariant, *AsmInfo, *MII, *MRI)); if (!IP) { errs() << "error: no instruction printer for target " << TripleName << '\n'; return; } + IP->setPrintImmHex(PrintImmHex); + PrettyPrinter &PIP = selectPrettyPrinter(Triple(TripleName)); StringRef Fmt = Obj->getBytesInAddress() > 4 ? "\t\t%016" PRIx64 ": " : "\t\t\t%08" PRIx64 ": "; @@ -365,11 +821,9 @@ static void DisassembleObject(const ObjectFile *Obj, bool InlineRelocs) { outs() << SegmentName << ","; outs() << name << ':'; - // If the section has no symbols just insert a dummy one and disassemble - // the whole section. - if (Symbols.empty()) - Symbols.push_back(std::make_pair(0, name)); - + // If the section has no symbol at the start, just insert a dummy one. + if (Symbols.empty() || Symbols[0].first != 0) + Symbols.insert(Symbols.begin(), std::make_pair(0, name)); SmallString<40> Comments; raw_svector_ostream CommentStream(Comments); @@ -409,13 +863,9 @@ static void DisassembleObject(const ObjectFile *Obj, bool InlineRelocs) { if (DisAsm->getInstruction(Inst, Size, Bytes.slice(Index), SectionAddr + Index, DebugOut, CommentStream)) { - outs() << format("%8" PRIx64 ":", SectionAddr + Index); - if (!NoShowRawInsn) { - outs() << "\t"; - DumpBytes(StringRef( - reinterpret_cast<const char *>(Bytes.data()) + Index, Size)); - } - IP->printInst(&Inst, outs(), ""); + PIP.printInst(*IP, &Inst, + Bytes.slice(Index, Size), + SectionAddr + Index, outs(), "", *STI); outs() << CommentStream.str(); Comments.clear(); outs() << "\n"; @@ -440,8 +890,8 @@ static void DisassembleObject(const ObjectFile *Obj, bool InlineRelocs) { // Stop when rel_cur's address is past the current instruction. if (addr >= Index + Size) break; if (error(rel_cur->getTypeName(name))) goto skip_print_rel; - if (error(rel_cur->getValueString(val))) goto skip_print_rel; - + if (error(getRelocationValueString(*rel_cur, val))) + goto skip_print_rel; outs() << format(Fmt.data(), SectionAddr + addr) << name << "\t" << val << "\n"; @@ -453,7 +903,7 @@ static void DisassembleObject(const ObjectFile *Obj, bool InlineRelocs) { } } -static void PrintRelocations(const ObjectFile *Obj) { +void llvm::PrintRelocations(const ObjectFile *Obj) { StringRef Fmt = Obj->getBytesInAddress() > 4 ? "%016" PRIx64 : "%08" PRIx64; // Regular objdump doesn't print relocations in non-relocatable object @@ -481,7 +931,7 @@ static void PrintRelocations(const ObjectFile *Obj) { continue; if (error(Reloc.getOffset(address))) continue; - if (error(Reloc.getValueString(valuestr))) + if (error(getRelocationValueString(Reloc, valuestr))) continue; outs() << format(Fmt.data(), address) << " " << relocname << " " << valuestr << "\n"; @@ -490,7 +940,7 @@ static void PrintRelocations(const ObjectFile *Obj) { } } -static void PrintSectionHeaders(const ObjectFile *Obj) { +void llvm::PrintSectionHeaders(const ObjectFile *Obj) { outs() << "Sections:\n" "Idx Name Size Address Type\n"; unsigned i = 0; @@ -511,7 +961,7 @@ static void PrintSectionHeaders(const ObjectFile *Obj) { } } -static void PrintSectionContents(const ObjectFile *Obj) { +void llvm::PrintSectionContents(const ObjectFile *Obj) { std::error_code EC; for (const SectionRef &Section : Obj->sections()) { StringRef Name; @@ -614,7 +1064,7 @@ static void PrintCOFFSymbolTable(const COFFObjectFile *coff) { } } -static void PrintSymbolTable(const ObjectFile *o) { +void llvm::PrintSymbolTable(const ObjectFile *o) { outs() << "SYMBOL TABLE:\n"; if (const COFFObjectFile *coff = dyn_cast<const COFFObjectFile>(o)) { @@ -622,27 +1072,35 @@ static void PrintSymbolTable(const ObjectFile *o) { return; } for (const SymbolRef &Symbol : o->symbols()) { - StringRef Name; uint64_t Address; SymbolRef::Type Type; - uint64_t Size; uint32_t Flags = Symbol.getFlags(); section_iterator Section = o->section_end(); - if (error(Symbol.getName(Name))) - continue; if (error(Symbol.getAddress(Address))) continue; if (error(Symbol.getType(Type))) continue; - if (error(Symbol.getSize(Size))) - continue; + uint64_t Size = Symbol.getSize(); if (error(Symbol.getSection(Section))) continue; + StringRef Name; + if (Type == SymbolRef::ST_Debug && Section != o->section_end()) { + Section->getName(Name); + } else if (error(Symbol.getName(Name))) { + continue; + } bool Global = Flags & SymbolRef::SF_Global; bool Weak = Flags & SymbolRef::SF_Weak; bool Absolute = Flags & SymbolRef::SF_Absolute; + bool Common = Flags & SymbolRef::SF_Common; + bool Hidden = Flags & SymbolRef::SF_Hidden; + if (Common) { + uint32_t Alignment = Symbol.getAlignment(); + Address = Size; + Size = Alignment; + } if (Address == UnknownAddressOrSize) Address = 0; if (Size == UnknownAddressOrSize) @@ -672,6 +1130,8 @@ static void PrintSymbolTable(const ObjectFile *o) { << ' '; if (Absolute) { outs() << "*ABS*"; + } else if (Common) { + outs() << "*COM*"; } else if (Section == o->section_end()) { outs() << "*UND*"; } else { @@ -687,8 +1147,11 @@ static void PrintSymbolTable(const ObjectFile *o) { outs() << SectionName; } outs() << '\t' - << format("%08" PRIx64 " ", Size) - << Name + << format("%08" PRIx64 " ", Size); + if (Hidden) { + outs() << ".hidden "; + } + outs() << Name << '\n'; } } @@ -812,15 +1275,13 @@ static void DumpArchive(const Archive *a) { if (std::error_code EC = ChildOrErr.getError()) { // Ignore non-object files. if (EC != object_error::invalid_file_type) - errs() << ToolName << ": '" << a->getFileName() << "': " << EC.message() - << ".\n"; + report_error(a->getFileName(), EC); continue; } if (ObjectFile *o = dyn_cast<ObjectFile>(&*ChildOrErr.get())) DumpObject(o); else - errs() << ToolName << ": '" << a->getFileName() << "': " - << "Unrecognized file type.\n"; + report_error(a->getFileName(), object_error::invalid_file_type); } } @@ -828,7 +1289,7 @@ static void DumpArchive(const Archive *a) { static void DumpInput(StringRef file) { // If file isn't stdin, check that it exists. if (file != "-" && !sys::fs::exists(file)) { - errs() << ToolName << ": '" << file << "': " << "No such file\n"; + report_error(file, errc::no_such_file_or_directory); return; } @@ -843,7 +1304,7 @@ static void DumpInput(StringRef file) { // Attempt to open the binary. ErrorOr<OwningBinary<Binary>> BinaryOrErr = createBinary(file); if (std::error_code EC = BinaryOrErr.getError()) { - errs() << ToolName << ": '" << file << "': " << EC.message() << ".\n"; + report_error(file, EC); return; } Binary &Binary = *BinaryOrErr.get().getBinary(); @@ -853,7 +1314,7 @@ static void DumpInput(StringRef file) { else if (ObjectFile *o = dyn_cast<ObjectFile>(&Binary)) DumpObject(o); else - errs() << ToolName << ": '" << file << "': " << "Unrecognized file type.\n"; + report_error(file, object_error::invalid_file_type); } int main(int argc, char **argv) { @@ -892,7 +1353,16 @@ int main(int argc, char **argv) { && !Bind && !LazyBind && !WeakBind - && !(UniversalHeaders && MachOOpt)) { + && !(UniversalHeaders && MachOOpt) + && !(ArchiveHeaders && MachOOpt) + && !(IndirectSymbols && MachOOpt) + && !(DataInCode && MachOOpt) + && !(LinkOptHints && MachOOpt) + && !(InfoPlist && MachOOpt) + && !(DylibsUsed && MachOOpt) + && !(DylibId && MachOOpt) + && !(ObjcMetaData && MachOOpt) + && !(DumpSections.size() != 0 && MachOOpt)) { cl::PrintHelpMessage(); return 2; } diff --git a/contrib/llvm/tools/llvm-objdump/llvm-objdump.h b/contrib/llvm/tools/llvm-objdump/llvm-objdump.h index f829dd1..b4d34f4 100644 --- a/contrib/llvm/tools/llvm-objdump/llvm-objdump.h +++ b/contrib/llvm/tools/llvm-objdump/llvm-objdump.h @@ -1,4 +1,3 @@ -//===-- llvm-objdump.h ----------------------------------------------------===// // // The LLVM Compiler Infrastructure // @@ -26,6 +25,8 @@ extern cl::opt<std::string> TripleName; extern cl::opt<std::string> ArchName; extern cl::opt<std::string> MCPU; extern cl::list<std::string> MAttrs; +extern cl::list<std::string> DumpSections; +extern cl::opt<bool> Raw; extern cl::opt<bool> Disassemble; extern cl::opt<bool> NoShowRawInsn; extern cl::opt<bool> PrivateHeaders; @@ -35,11 +36,26 @@ extern cl::opt<bool> Bind; extern cl::opt<bool> LazyBind; extern cl::opt<bool> WeakBind; extern cl::opt<bool> UniversalHeaders; +extern cl::opt<bool> ArchiveHeaders; +extern cl::opt<bool> IndirectSymbols; +extern cl::opt<bool> DataInCode; +extern cl::opt<bool> LinkOptHints; +extern cl::opt<bool> InfoPlist; +extern cl::opt<bool> DylibsUsed; +extern cl::opt<bool> DylibId; +extern cl::opt<bool> ObjcMetaData; +extern cl::opt<std::string> DisSymName; +extern cl::opt<bool> NonVerbose; +extern cl::opt<bool> Relocations; +extern cl::opt<bool> SectionHeaders; +extern cl::opt<bool> SectionContents; +extern cl::opt<bool> SymbolTable; +extern cl::opt<bool> UnwindInfo; +extern cl::opt<bool> PrintImmHex; // Various helper functions. bool error(std::error_code ec); bool RelocAddressLess(object::RelocationRef a, object::RelocationRef b); -void DumpBytes(StringRef bytes); void ParseInputMachO(StringRef Filename); void printCOFFUnwindInfo(const object::COFFObjectFile* o); void printMachOUnwindInfo(const object::MachOObjectFile* o); @@ -56,6 +72,10 @@ void printRebaseTable(const object::ObjectFile *o); void printBindTable(const object::ObjectFile *o); void printLazyBindTable(const object::ObjectFile *o); void printWeakBindTable(const object::ObjectFile *o); +void PrintRelocations(const object::ObjectFile *o); +void PrintSectionHeaders(const object::ObjectFile *o); +void PrintSectionContents(const object::ObjectFile *o); +void PrintSymbolTable(const object::ObjectFile *o); } // end namespace llvm |
