X-Git-Url: http://demsky.eecs.uci.edu/git/?a=blobdiff_plain;f=lib%2FObject%2FMachOObjectFile.cpp;h=c6815ea27d24cfd02f8289a5a872cfc10c6cf2b5;hb=12af22e8cc217827cf4f118b0f5e4ebbda9925ae;hp=23c2d6d8bda07d17a19f3ad061ae23e9740ad51d;hpb=a40b3522c82f5684a8a2c59a631b8c445288f6e1;p=oota-llvm.git diff --git a/lib/Object/MachOObjectFile.cpp b/lib/Object/MachOObjectFile.cpp index 23c2d6d8bda..c6815ea27d2 100644 --- a/lib/Object/MachOObjectFile.cpp +++ b/lib/Object/MachOObjectFile.cpp @@ -13,10 +13,14 @@ //===----------------------------------------------------------------------===// #include "llvm/Object/MachO.h" +#include "llvm/ADT/STLExtras.h" +#include "llvm/ADT/StringSwitch.h" #include "llvm/ADT/Triple.h" #include "llvm/Support/DataExtractor.h" #include "llvm/Support/Format.h" #include "llvm/Support/Host.h" +#include "llvm/Support/LEB128.h" +#include "llvm/Support/MachO.h" #include "llvm/Support/MemoryBuffer.h" #include "llvm/Support/raw_ostream.h" #include @@ -26,206 +30,19 @@ using namespace llvm; using namespace object; -namespace llvm { -namespace object { - -struct nlist_base { - uint32_t n_strx; - uint8_t n_type; - uint8_t n_sect; - uint16_t n_desc; -}; - -struct section_base { - char sectname[16]; - char segname[16]; -}; - -template -static void SwapValue(T &Value) { - Value = sys::SwapByteOrder(Value); +namespace { + struct section_base { + char sectname[16]; + char segname[16]; + }; } template -static void SwapStruct(T &Value); - -template<> -void SwapStruct(MachO::any_relocation_info &H) { - SwapValue(H.r_word0); - SwapValue(H.r_word1); -} - -template<> -void SwapStruct(MachO::load_command &L) { - SwapValue(L.cmd); - SwapValue(L.cmdsize); -} - -template<> -void SwapStruct(nlist_base &S) { - SwapValue(S.n_strx); - SwapValue(S.n_desc); -} - -template<> -void SwapStruct(MachO::section &S) { - SwapValue(S.addr); - SwapValue(S.size); - SwapValue(S.offset); - SwapValue(S.align); - SwapValue(S.reloff); - SwapValue(S.nreloc); - SwapValue(S.flags); - SwapValue(S.reserved1); - SwapValue(S.reserved2); -} - -template<> -void SwapStruct(MachO::section_64 &S) { - SwapValue(S.addr); - SwapValue(S.size); - SwapValue(S.offset); - SwapValue(S.align); - SwapValue(S.reloff); - SwapValue(S.nreloc); - SwapValue(S.flags); - SwapValue(S.reserved1); - SwapValue(S.reserved2); - SwapValue(S.reserved3); -} - -template<> -void SwapStruct(MachO::nlist &S) { - SwapValue(S.n_strx); - SwapValue(S.n_desc); - SwapValue(S.n_value); -} - -template<> -void SwapStruct(MachO::nlist_64 &S) { - SwapValue(S.n_strx); - SwapValue(S.n_desc); - SwapValue(S.n_value); -} - -template<> -void SwapStruct(MachO::mach_header &H) { - SwapValue(H.magic); - SwapValue(H.cputype); - SwapValue(H.cpusubtype); - SwapValue(H.filetype); - SwapValue(H.ncmds); - SwapValue(H.sizeofcmds); - SwapValue(H.flags); -} - -template<> -void SwapStruct(MachO::mach_header_64 &H) { - SwapValue(H.magic); - SwapValue(H.cputype); - SwapValue(H.cpusubtype); - SwapValue(H.filetype); - SwapValue(H.ncmds); - SwapValue(H.sizeofcmds); - SwapValue(H.flags); - SwapValue(H.reserved); -} - -template<> -void SwapStruct(MachO::symtab_command &C) { - SwapValue(C.cmd); - SwapValue(C.cmdsize); - SwapValue(C.symoff); - SwapValue(C.nsyms); - SwapValue(C.stroff); - SwapValue(C.strsize); -} - -template<> -void SwapStruct(MachO::dysymtab_command &C) { - SwapValue(C.cmd); - SwapValue(C.cmdsize); - SwapValue(C.ilocalsym); - SwapValue(C.nlocalsym); - SwapValue(C.iextdefsym); - SwapValue(C.nextdefsym); - SwapValue(C.iundefsym); - SwapValue(C.nundefsym); - SwapValue(C.tocoff); - SwapValue(C.ntoc); - SwapValue(C.modtaboff); - SwapValue(C.nmodtab); - SwapValue(C.extrefsymoff); - SwapValue(C.nextrefsyms); - SwapValue(C.indirectsymoff); - SwapValue(C.nindirectsyms); - SwapValue(C.extreloff); - SwapValue(C.nextrel); - SwapValue(C.locreloff); - SwapValue(C.nlocrel); -} - -template<> -void SwapStruct(MachO::linkedit_data_command &C) { - SwapValue(C.cmd); - SwapValue(C.cmdsize); - SwapValue(C.dataoff); - SwapValue(C.datasize); -} - -template<> -void SwapStruct(MachO::segment_command &C) { - SwapValue(C.cmd); - SwapValue(C.cmdsize); - SwapValue(C.vmaddr); - SwapValue(C.vmsize); - SwapValue(C.fileoff); - SwapValue(C.filesize); - SwapValue(C.maxprot); - SwapValue(C.initprot); - SwapValue(C.nsects); - SwapValue(C.flags); -} - -template<> -void SwapStruct(MachO::segment_command_64 &C) { - SwapValue(C.cmd); - SwapValue(C.cmdsize); - SwapValue(C.vmaddr); - SwapValue(C.vmsize); - SwapValue(C.fileoff); - SwapValue(C.filesize); - SwapValue(C.maxprot); - SwapValue(C.initprot); - SwapValue(C.nsects); - SwapValue(C.flags); -} - -template<> -void SwapStruct(uint32_t &C) { - SwapValue(C); -} - -template<> -void SwapStruct(MachO::linker_options_command &C) { - SwapValue(C.cmd); - SwapValue(C.cmdsize); - SwapValue(C.count); -} - -template<> -void SwapStruct(MachO::data_in_code_entry &C) { - SwapValue(C.offset); - SwapValue(C.length); - SwapValue(C.kind); -} - -template -T getStruct(const MachOObjectFile *O, const char *P) { +static T getStruct(const MachOObjectFile *O, const char *P) { T Cmd; memcpy(&Cmd, P, sizeof(T)); if (O->isLittleEndian() != sys::IsLittleEndianHost) - SwapStruct(Cmd); + MachO::swapStruct(Cmd); return Cmd; } @@ -259,10 +76,10 @@ static const char *getPtr(const MachOObjectFile *O, size_t Offset) { return O->getData().substr(Offset, 1).data(); } -static nlist_base +static MachO::nlist_base getSymbolTableEntryBase(const MachOObjectFile *O, DataRefImpl DRI) { const char *P = reinterpret_cast(DRI.p); - return getStruct(O, P); + return getStruct(O, P); } static StringRef parseSegmentOrSectionName(const char *P) { @@ -296,16 +113,16 @@ static void printRelocationTargetName(const MachOObjectFile *O, if (IsScattered) { uint32_t Val = O->getPlainRelocationSymbolNum(RE); - for (symbol_iterator SI = O->symbol_begin(), SE = O->symbol_end(); - SI != SE; ++SI) { - error_code ec; + for (const SymbolRef &Symbol : O->symbols()) { + std::error_code ec; uint64_t Addr; StringRef Name; - if ((ec = SI->getAddress(Addr))) + if ((ec = Symbol.getAddress(Addr))) report_fatal_error(ec.message()); - if (Addr != Val) continue; - if ((ec = SI->getName(Name))) + if (Addr != Val) + continue; + if ((ec = Symbol.getName(Name))) report_fatal_error(ec.message()); fmt << Name; return; @@ -313,16 +130,16 @@ static void printRelocationTargetName(const MachOObjectFile *O, // If we couldn't find a symbol that this relocation refers to, try // to find a section beginning instead. - for (section_iterator SI = O->section_begin(), SE = O->section_end(); - SI != SE; ++SI) { - error_code ec; + for (const SectionRef &Section : O->sections()) { + std::error_code ec; uint64_t Addr; StringRef Name; - if ((ec = SI->getAddress(Addr))) + if ((ec = Section.getAddress(Addr))) report_fatal_error(ec.message()); - if (Addr != Val) continue; - if ((ec = SI->getName(Name))) + if (Addr != Val) + continue; + if ((ec = Section.getName(Name))) report_fatal_error(ec.message()); fmt << Name; return; @@ -407,11 +224,11 @@ static uint32_t getSectionFlags(const MachOObjectFile *O, return Sect.flags; } -MachOObjectFile::MachOObjectFile(MemoryBuffer *Object, bool IsLittleEndian, - bool Is64bits, error_code &EC, - bool BufferOwned) - : ObjectFile(getMachOType(IsLittleEndian, Is64bits), Object, BufferOwned), - SymtabLoadCmd(NULL), DysymtabLoadCmd(NULL), DataInCodeLoadCmd(NULL) { +MachOObjectFile::MachOObjectFile(MemoryBufferRef Object, bool IsLittleEndian, + bool Is64bits, std::error_code &EC) + : ObjectFile(getMachOType(IsLittleEndian, Is64bits), Object), + SymtabLoadCmd(nullptr), DysymtabLoadCmd(nullptr), + DataInCodeLoadCmd(nullptr), DyldInfoLoadCmd(nullptr) { uint32_t LoadCommandCount = this->getHeader().ncmds; MachO::LoadCommandType SegmentLoadType = is64Bit() ? MachO::LC_SEGMENT_64 : MachO::LC_SEGMENT; @@ -427,12 +244,22 @@ MachOObjectFile::MachOObjectFile(MemoryBuffer *Object, bool IsLittleEndian, } else if (Load.C.cmd == MachO::LC_DATA_IN_CODE) { assert(!DataInCodeLoadCmd && "Multiple data in code tables"); DataInCodeLoadCmd = Load.Ptr; + } else if (Load.C.cmd == MachO::LC_DYLD_INFO || + Load.C.cmd == MachO::LC_DYLD_INFO_ONLY) { + assert(!DyldInfoLoadCmd && "Multiple dyldinfo load commands"); + DyldInfoLoadCmd = Load.Ptr; } else if (Load.C.cmd == SegmentLoadType) { uint32_t NumSections = getSegmentLoadCommandNumSections(this, Load); for (unsigned J = 0; J < NumSections; ++J) { const char *Sec = getSectionPtr(this, Load, J); Sections.push_back(Sec); } + } else if (Load.C.cmd == MachO::LC_LOAD_DYLIB || + Load.C.cmd == MachO::LC_LOAD_WEAK_DYLIB || + Load.C.cmd == MachO::LC_LAZY_LOAD_DYLIB || + Load.C.cmd == MachO::LC_REEXPORT_DYLIB || + Load.C.cmd == MachO::LC_LOAD_UPWARD_DYLIB) { + Libraries.push_back(Load.Ptr); } if (I == LoadCommandCount - 1) @@ -449,55 +276,64 @@ void MachOObjectFile::moveSymbolNext(DataRefImpl &Symb) const { Symb.p += SymbolTableEntrySize; } -error_code MachOObjectFile::getSymbolName(DataRefImpl Symb, - StringRef &Res) const { +std::error_code MachOObjectFile::getSymbolName(DataRefImpl Symb, + StringRef &Res) const { StringRef StringTable = getStringTableData(); - nlist_base Entry = getSymbolTableEntryBase(this, Symb); + MachO::nlist_base Entry = getSymbolTableEntryBase(this, Symb); const char *Start = &StringTable.data()[Entry.n_strx]; Res = StringRef(Start); return object_error::success; } -error_code MachOObjectFile::getSymbolAddress(DataRefImpl Symb, - uint64_t &Res) const { +// getIndirectName() returns the name of the alias'ed symbol who's string table +// index is in the n_value field. +std::error_code MachOObjectFile::getIndirectName(DataRefImpl Symb, + StringRef &Res) const { + StringRef StringTable = getStringTableData(); + uint64_t NValue; if (is64Bit()) { MachO::nlist_64 Entry = getSymbol64TableEntry(Symb); - Res = Entry.n_value; + NValue = Entry.n_value; + if ((Entry.n_type & MachO::N_TYPE) != MachO::N_INDR) + return object_error::parse_failed; } else { MachO::nlist Entry = getSymbolTableEntry(Symb); - Res = Entry.n_value; + NValue = Entry.n_value; + if ((Entry.n_type & MachO::N_TYPE) != MachO::N_INDR) + return object_error::parse_failed; } + if (NValue >= StringTable.size()) + return object_error::parse_failed; + const char *Start = &StringTable.data()[NValue]; + Res = StringRef(Start); return object_error::success; } -error_code -MachOObjectFile::getSymbolFileOffset(DataRefImpl Symb, - uint64_t &Res) const { - nlist_base Entry = getSymbolTableEntryBase(this, Symb); - getSymbolAddress(Symb, Res); - if (Entry.n_sect) { - uint64_t Delta; - DataRefImpl SecRel; - SecRel.d.a = Entry.n_sect-1; - if (is64Bit()) { - MachO::section_64 Sec = getSection64(SecRel); - Delta = Sec.offset - Sec.addr; - } else { - MachO::section Sec = getSection(SecRel); - Delta = Sec.offset - Sec.addr; - } - - Res += Delta; +std::error_code MachOObjectFile::getSymbolAddress(DataRefImpl Symb, + uint64_t &Res) const { + if (is64Bit()) { + MachO::nlist_64 Entry = getSymbol64TableEntry(Symb); + if ((Entry.n_type & MachO::N_TYPE) == MachO::N_UNDF && + Entry.n_value == 0) + Res = UnknownAddressOrSize; + else + Res = Entry.n_value; + } else { + MachO::nlist Entry = getSymbolTableEntry(Symb); + if ((Entry.n_type & MachO::N_TYPE) == MachO::N_UNDF && + Entry.n_value == 0) + Res = UnknownAddressOrSize; + else + Res = Entry.n_value; } - return object_error::success; } -error_code MachOObjectFile::getSymbolAlignment(DataRefImpl DRI, - uint32_t &Result) const { +std::error_code MachOObjectFile::getSymbolAlignment(DataRefImpl DRI, + uint32_t &Result) const { uint32_t flags = getSymbolFlags(DRI); if (flags & SymbolRef::SF_Common) { - nlist_base Entry = getSymbolTableEntryBase(this, DRI); + MachO::nlist_base Entry = getSymbolTableEntryBase(this, DRI); Result = 1 << MachO::GET_COMM_ALIGN(Entry.n_desc); } else { Result = 0; @@ -505,15 +341,19 @@ error_code MachOObjectFile::getSymbolAlignment(DataRefImpl DRI, return object_error::success; } -error_code MachOObjectFile::getSymbolSize(DataRefImpl DRI, - uint64_t &Result) const { +std::error_code MachOObjectFile::getSymbolSize(DataRefImpl DRI, + uint64_t &Result) const { uint64_t BeginOffset; uint64_t EndOffset = 0; uint8_t SectionIndex; - nlist_base Entry = getSymbolTableEntryBase(this, DRI); + MachO::nlist_base Entry = getSymbolTableEntryBase(this, DRI); uint64_t Value; getSymbolAddress(DRI, Value); + if (Value == UnknownAddressOrSize) { + Result = UnknownAddressOrSize; + return object_error::success; + } BeginOffset = Value; @@ -528,10 +368,12 @@ error_code MachOObjectFile::getSymbolSize(DataRefImpl DRI, } // Unfortunately symbols are unsorted so we need to touch all // symbols from load command - for (symbol_iterator I = symbol_begin(), E = symbol_end(); I != E; ++I) { - DataRefImpl DRI = I->getRawDataRefImpl(); + for (const SymbolRef &Symbol : symbols()) { + DataRefImpl DRI = Symbol.getRawDataRefImpl(); Entry = getSymbolTableEntryBase(this, DRI); getSymbolAddress(DRI, Value); + if (Value == UnknownAddressOrSize) + continue; if (Entry.n_sect == SectionIndex && Value > BeginOffset) if (!EndOffset || Value < EndOffset) EndOffset = Value; @@ -548,9 +390,9 @@ error_code MachOObjectFile::getSymbolSize(DataRefImpl DRI, return object_error::success; } -error_code MachOObjectFile::getSymbolType(DataRefImpl Symb, - SymbolRef::Type &Res) const { - nlist_base Entry = getSymbolTableEntryBase(this, Symb); +std::error_code MachOObjectFile::getSymbolType(DataRefImpl Symb, + SymbolRef::Type &Res) const { + MachO::nlist_base Entry = getSymbolTableEntryBase(this, Symb); uint8_t n_type = Entry.n_type; Res = SymbolRef::ST_Other; @@ -573,7 +415,7 @@ error_code MachOObjectFile::getSymbolType(DataRefImpl Symb, } uint32_t MachOObjectFile::getSymbolFlags(DataRefImpl DRI) const { - nlist_base Entry = getSymbolTableEntryBase(this, DRI); + MachO::nlist_base Entry = getSymbolTableEntryBase(this, DRI); uint8_t MachOType = Entry.n_type; uint16_t MachOFlags = Entry.n_desc; @@ -583,6 +425,9 @@ uint32_t MachOObjectFile::getSymbolFlags(DataRefImpl DRI) const { if ((MachOType & MachO::N_TYPE) == MachO::N_UNDF) Result |= SymbolRef::SF_Undefined; + if ((MachOType & MachO::N_TYPE) == MachO::N_INDR) + Result |= SymbolRef::SF_Indirect; + if (MachOType & MachO::N_STAB) Result |= SymbolRef::SF_FormatSpecific; @@ -591,7 +436,7 @@ uint32_t MachOObjectFile::getSymbolFlags(DataRefImpl DRI) const { if ((MachOType & MachO::N_TYPE) == MachO::N_UNDF) { uint64_t Value; getSymbolAddress(DRI, Value); - if (Value) + if (Value && Value != UnknownAddressOrSize) Result |= SymbolRef::SF_Common; } } @@ -599,16 +444,18 @@ uint32_t MachOObjectFile::getSymbolFlags(DataRefImpl DRI) const { if (MachOFlags & (MachO::N_WEAK_REF | MachO::N_WEAK_DEF)) Result |= SymbolRef::SF_Weak; + if (MachOFlags & (MachO::N_ARM_THUMB_DEF)) + Result |= SymbolRef::SF_Thumb; + if ((MachOType & MachO::N_TYPE) == MachO::N_ABS) Result |= SymbolRef::SF_Absolute; return Result; } -error_code -MachOObjectFile::getSymbolSection(DataRefImpl Symb, - section_iterator &Res) const { - nlist_base Entry = getSymbolTableEntryBase(this, Symb); +std::error_code MachOObjectFile::getSymbolSection(DataRefImpl Symb, + section_iterator &Res) const { + MachO::nlist_base Entry = getSymbolTableEntryBase(this, Symb); uint8_t index = Entry.n_sect; if (index == 0) { @@ -622,24 +469,19 @@ MachOObjectFile::getSymbolSection(DataRefImpl Symb, return object_error::success; } -error_code MachOObjectFile::getSymbolValue(DataRefImpl Symb, - uint64_t &Val) const { - report_fatal_error("getSymbolValue unimplemented in MachOObjectFile"); -} - void MachOObjectFile::moveSectionNext(DataRefImpl &Sec) const { Sec.d.a++; } -error_code -MachOObjectFile::getSectionName(DataRefImpl Sec, StringRef &Result) const { +std::error_code MachOObjectFile::getSectionName(DataRefImpl Sec, + StringRef &Result) const { ArrayRef Raw = getSectionRawName(Sec); Result = parseSegmentOrSectionName(Raw.data()); return object_error::success; } -error_code -MachOObjectFile::getSectionAddress(DataRefImpl Sec, uint64_t &Res) const { +std::error_code MachOObjectFile::getSectionAddress(DataRefImpl Sec, + uint64_t &Res) const { if (is64Bit()) { MachO::section_64 Sect = getSection64(Sec); Res = Sect.addr; @@ -650,8 +492,8 @@ MachOObjectFile::getSectionAddress(DataRefImpl Sec, uint64_t &Res) const { return object_error::success; } -error_code -MachOObjectFile::getSectionSize(DataRefImpl Sec, uint64_t &Res) const { +std::error_code MachOObjectFile::getSectionSize(DataRefImpl Sec, + uint64_t &Res) const { if (is64Bit()) { MachO::section_64 Sect = getSection64(Sec); Res = Sect.size; @@ -663,8 +505,8 @@ MachOObjectFile::getSectionSize(DataRefImpl Sec, uint64_t &Res) const { return object_error::success; } -error_code -MachOObjectFile::getSectionContents(DataRefImpl Sec, StringRef &Res) const { +std::error_code MachOObjectFile::getSectionContents(DataRefImpl Sec, + StringRef &Res) const { uint32_t Offset; uint64_t Size; @@ -682,8 +524,8 @@ MachOObjectFile::getSectionContents(DataRefImpl Sec, StringRef &Res) const { return object_error::success; } -error_code -MachOObjectFile::getSectionAlignment(DataRefImpl Sec, uint64_t &Res) const { +std::error_code MachOObjectFile::getSectionAlignment(DataRefImpl Sec, + uint64_t &Res) const { uint32_t Align; if (is64Bit()) { MachO::section_64 Sect = getSection64(Sec); @@ -697,26 +539,34 @@ MachOObjectFile::getSectionAlignment(DataRefImpl Sec, uint64_t &Res) const { return object_error::success; } -error_code -MachOObjectFile::isSectionText(DataRefImpl Sec, bool &Res) const { +std::error_code MachOObjectFile::isSectionText(DataRefImpl Sec, + bool &Res) const { uint32_t Flags = getSectionFlags(this, Sec); Res = Flags & MachO::S_ATTR_PURE_INSTRUCTIONS; return object_error::success; } -error_code MachOObjectFile::isSectionData(DataRefImpl DRI, bool &Result) const { - // FIXME: Unimplemented. - Result = false; +std::error_code MachOObjectFile::isSectionData(DataRefImpl Sec, + bool &Result) const { + uint32_t Flags = getSectionFlags(this, Sec); + unsigned SectionType = Flags & MachO::SECTION_TYPE; + Result = !(Flags & MachO::S_ATTR_PURE_INSTRUCTIONS) && + !(SectionType == MachO::S_ZEROFILL || + SectionType == MachO::S_GB_ZEROFILL); return object_error::success; } -error_code MachOObjectFile::isSectionBSS(DataRefImpl DRI, bool &Result) const { - // FIXME: Unimplemented. - Result = false; +std::error_code MachOObjectFile::isSectionBSS(DataRefImpl Sec, + bool &Result) const { + uint32_t Flags = getSectionFlags(this, Sec); + unsigned SectionType = Flags & MachO::SECTION_TYPE; + Result = !(Flags & MachO::S_ATTR_PURE_INSTRUCTIONS) && + (SectionType == MachO::S_ZEROFILL || + SectionType == MachO::S_GB_ZEROFILL); return object_error::success; } -error_code +std::error_code MachOObjectFile::isSectionRequiredForExecution(DataRefImpl Sec, bool &Result) const { // FIXME: Unimplemented. @@ -724,15 +574,15 @@ MachOObjectFile::isSectionRequiredForExecution(DataRefImpl Sec, return object_error::success; } -error_code MachOObjectFile::isSectionVirtual(DataRefImpl Sec, - bool &Result) const { +std::error_code MachOObjectFile::isSectionVirtual(DataRefImpl Sec, + bool &Result) const { // FIXME: Unimplemented. Result = false; return object_error::success; } -error_code -MachOObjectFile::isSectionZeroInit(DataRefImpl Sec, bool &Res) const { +std::error_code MachOObjectFile::isSectionZeroInit(DataRefImpl Sec, + bool &Res) const { uint32_t Flags = getSectionFlags(this, Sec); unsigned SectionType = Flags & MachO::SECTION_TYPE; Res = SectionType == MachO::S_ZEROFILL || @@ -740,8 +590,8 @@ MachOObjectFile::isSectionZeroInit(DataRefImpl Sec, bool &Res) const { return object_error::success; } -error_code MachOObjectFile::isSectionReadOnlyData(DataRefImpl Sec, - bool &Result) const { +std::error_code MachOObjectFile::isSectionReadOnlyData(DataRefImpl Sec, + bool &Result) const { // Consider using the code from isSectionText to look for __const sections. // Alternately, emit S_ATTR_PURE_INSTRUCTIONS and/or S_ATTR_SOME_INSTRUCTIONS // to use section attributes to distinguish code from data. @@ -751,9 +601,9 @@ error_code MachOObjectFile::isSectionReadOnlyData(DataRefImpl Sec, return object_error::success; } -error_code -MachOObjectFile::sectionContainsSymbol(DataRefImpl Sec, DataRefImpl Symb, - bool &Result) const { +std::error_code MachOObjectFile::sectionContainsSymbol(DataRefImpl Sec, + DataRefImpl Symb, + bool &Result) const { SymbolRef::Type ST; this->getSymbolType(Symb, ST); if (ST == SymbolRef::ST_Unknown) { @@ -774,55 +624,50 @@ MachOObjectFile::sectionContainsSymbol(DataRefImpl Sec, DataRefImpl Symb, } relocation_iterator MachOObjectFile::section_rel_begin(DataRefImpl Sec) const { - uint32_t Offset; - if (is64Bit()) { - MachO::section_64 Sect = getSection64(Sec); - Offset = Sect.reloff; - } else { - MachO::section Sect = getSection(Sec); - Offset = Sect.reloff; - } - DataRefImpl Ret; - Ret.p = reinterpret_cast(getPtr(this, Offset)); + Ret.d.a = Sec.d.a; + Ret.d.b = 0; return relocation_iterator(RelocationRef(Ret, this)); } relocation_iterator MachOObjectFile::section_rel_end(DataRefImpl Sec) const { - uint32_t Offset; uint32_t Num; if (is64Bit()) { MachO::section_64 Sect = getSection64(Sec); - Offset = Sect.reloff; Num = Sect.nreloc; } else { MachO::section Sect = getSection(Sec); - Offset = Sect.reloff; Num = Sect.nreloc; } - const MachO::any_relocation_info *P = - reinterpret_cast(getPtr(this, Offset)); - DataRefImpl Ret; - Ret.p = reinterpret_cast(P + Num); + Ret.d.a = Sec.d.a; + Ret.d.b = Num; return relocation_iterator(RelocationRef(Ret, this)); } void MachOObjectFile::moveRelocationNext(DataRefImpl &Rel) const { - const MachO::any_relocation_info *P = - reinterpret_cast(Rel.p); - Rel.p = reinterpret_cast(P + 1); + ++Rel.d.b; } -error_code -MachOObjectFile::getRelocationAddress(DataRefImpl Rel, uint64_t &Res) const { - report_fatal_error("getRelocationAddress not implemented in MachOObjectFile"); +std::error_code MachOObjectFile::getRelocationAddress(DataRefImpl Rel, + uint64_t &Res) const { + uint64_t Offset; + getRelocationOffset(Rel, Offset); + + DataRefImpl Sec; + Sec.d.a = Rel.d.a; + uint64_t SecAddress; + getSectionAddress(Sec, SecAddress); + Res = SecAddress + Offset; + return object_error::success; } -error_code MachOObjectFile::getRelocationOffset(DataRefImpl Rel, - uint64_t &Res) const { +std::error_code MachOObjectFile::getRelocationOffset(DataRefImpl Rel, + uint64_t &Res) const { + assert(getHeader().filetype == MachO::MH_OBJECT && + "Only implemented for MH_OBJECT"); MachO::any_relocation_info RE = getRelocation(Rel); Res = getAnyRelocationAddress(RE); return object_error::success; @@ -831,6 +676,9 @@ error_code MachOObjectFile::getRelocationOffset(DataRefImpl Rel, symbol_iterator MachOObjectFile::getRelocationSymbol(DataRefImpl Rel) const { MachO::any_relocation_info RE = getRelocation(Rel); + if (isRelocationScattered(RE)) + return symbol_end(); + uint32_t SymbolIdx = getPlainRelocationSymbolNum(RE); bool isExtern = getPlainRelocationExternal(RE); if (!isExtern) @@ -846,14 +694,14 @@ MachOObjectFile::getRelocationSymbol(DataRefImpl Rel) const { return symbol_iterator(SymbolRef(Sym, this)); } -error_code MachOObjectFile::getRelocationType(DataRefImpl Rel, - uint64_t &Res) const { +std::error_code MachOObjectFile::getRelocationType(DataRefImpl Rel, + uint64_t &Res) const { MachO::any_relocation_info RE = getRelocation(Rel); Res = getAnyRelocationType(RE); return object_error::success; } -error_code +std::error_code MachOObjectFile::getRelocationTypeName(DataRefImpl Rel, SmallVectorImpl &Result) const { StringRef res; @@ -916,6 +764,22 @@ MachOObjectFile::getRelocationTypeName(DataRefImpl Rel, res = Table[RType]; break; } + case Triple::aarch64: { + static const char *const Table[] = { + "ARM64_RELOC_UNSIGNED", "ARM64_RELOC_SUBTRACTOR", + "ARM64_RELOC_BRANCH26", "ARM64_RELOC_PAGE21", + "ARM64_RELOC_PAGEOFF12", "ARM64_RELOC_GOT_LOAD_PAGE21", + "ARM64_RELOC_GOT_LOAD_PAGEOFF12", "ARM64_RELOC_POINTER_TO_GOT", + "ARM64_RELOC_TLVP_LOAD_PAGE21", "ARM64_RELOC_TLVP_LOAD_PAGEOFF12", + "ARM64_RELOC_ADDEND" + }; + + if (RType >= array_lengthof(Table)) + res = "Unknown"; + else + res = Table[RType]; + break; + } case Triple::ppc: { static const char *const Table[] = { "PPC_RELOC_VANILLA", @@ -949,7 +813,7 @@ MachOObjectFile::getRelocationTypeName(DataRefImpl Rel, return object_error::success; } -error_code +std::error_code MachOObjectFile::getRelocationValueString(DataRefImpl Rel, SmallVectorImpl &Result) const { MachO::any_relocation_info RE = getRelocation(Rel); @@ -978,7 +842,7 @@ MachOObjectFile::getRelocationValueString(DataRefImpl Rel, } case MachO::X86_64_RELOC_SUBTRACTOR: { DataRefImpl RelNext = Rel; - RelNext.d.a++; + moveRelocationNext(RelNext); MachO::any_relocation_info RENext = getRelocation(RelNext); // X86_64_RELOC_SUBTRACTOR must be followed by a relocation of type @@ -1026,7 +890,7 @@ MachOObjectFile::getRelocationValueString(DataRefImpl Rel, return object_error::success; case MachO::GENERIC_RELOC_SECTDIFF: { DataRefImpl RelNext = Rel; - RelNext.d.a++; + moveRelocationNext(RelNext); MachO::any_relocation_info RENext = getRelocation(RelNext); // X86 sect diff's must be followed by a relocation of type @@ -1048,7 +912,7 @@ MachOObjectFile::getRelocationValueString(DataRefImpl Rel, switch (Type) { case MachO::GENERIC_RELOC_LOCAL_SECTDIFF: { DataRefImpl RelNext = Rel; - RelNext.d.a++; + moveRelocationNext(RelNext); MachO::any_relocation_info RENext = getRelocation(RelNext); // X86 sect diff's must be followed by a relocation of type @@ -1087,7 +951,7 @@ MachOObjectFile::getRelocationValueString(DataRefImpl Rel, printRelocationTargetName(this, RE, fmt); DataRefImpl RelNext = Rel; - RelNext.d.a++; + moveRelocationNext(RelNext); MachO::any_relocation_info RENext = getRelocation(RelNext); // ARM half relocs must be followed by a relocation of type @@ -1125,8 +989,8 @@ MachOObjectFile::getRelocationValueString(DataRefImpl Rel, return object_error::success; } -error_code -MachOObjectFile::getRelocationHidden(DataRefImpl Rel, bool &Result) const { +std::error_code MachOObjectFile::getRelocationHidden(DataRefImpl Rel, + bool &Result) const { unsigned Arch = getArch(); uint64_t Type; getRelocationType(Rel, Type); @@ -1153,30 +1017,197 @@ MachOObjectFile::getRelocationHidden(DataRefImpl Rel, bool &Result) const { return object_error::success; } -error_code MachOObjectFile::getLibraryNext(DataRefImpl LibData, - LibraryRef &Res) const { - report_fatal_error("Needed libraries unimplemented in MachOObjectFile"); -} +// +// guessLibraryShortName() is passed a name of a dynamic library and returns a +// guess on what the short name is. Then name is returned as a substring of the +// StringRef Name passed in. The name of the dynamic library is recognized as +// a framework if it has one of the two following forms: +// Foo.framework/Versions/A/Foo +// Foo.framework/Foo +// Where A and Foo can be any string. And may contain a trailing suffix +// starting with an underbar. If the Name is recognized as a framework then +// isFramework is set to true else it is set to false. If the Name has a +// suffix then Suffix is set to the substring in Name that contains the suffix +// else it is set to a NULL StringRef. +// +// The Name of the dynamic library is recognized as a library name if it has +// one of the two following forms: +// libFoo.A.dylib +// libFoo.dylib +// The library may have a suffix trailing the name Foo of the form: +// libFoo_profile.A.dylib +// libFoo_profile.dylib +// +// The Name of the dynamic library is also recognized as a library name if it +// has the following form: +// Foo.qtx +// +// If the Name of the dynamic library is none of the forms above then a NULL +// StringRef is returned. +// +StringRef MachOObjectFile::guessLibraryShortName(StringRef Name, + bool &isFramework, + StringRef &Suffix) { + StringRef Foo, F, DotFramework, V, Dylib, Lib, Dot, Qtx; + size_t a, b, c, d, Idx; + + isFramework = false; + Suffix = StringRef(); + + // Pull off the last component and make Foo point to it + a = Name.rfind('/'); + if (a == Name.npos || a == 0) + goto guess_library; + Foo = Name.slice(a+1, Name.npos); + + // Look for a suffix starting with a '_' + Idx = Foo.rfind('_'); + if (Idx != Foo.npos && Foo.size() >= 2) { + Suffix = Foo.slice(Idx, Foo.npos); + Foo = Foo.slice(0, Idx); + } + + // First look for the form Foo.framework/Foo + b = Name.rfind('/', a); + if (b == Name.npos) + Idx = 0; + else + Idx = b+1; + F = Name.slice(Idx, Idx + Foo.size()); + DotFramework = Name.slice(Idx + Foo.size(), + Idx + Foo.size() + sizeof(".framework/")-1); + if (F == Foo && DotFramework == ".framework/") { + isFramework = true; + return Foo; + } + + // Next look for the form Foo.framework/Versions/A/Foo + if (b == Name.npos) + goto guess_library; + c = Name.rfind('/', b); + if (c == Name.npos || c == 0) + goto guess_library; + V = Name.slice(c+1, Name.npos); + if (!V.startswith("Versions/")) + goto guess_library; + d = Name.rfind('/', c); + if (d == Name.npos) + Idx = 0; + else + Idx = d+1; + F = Name.slice(Idx, Idx + Foo.size()); + DotFramework = Name.slice(Idx + Foo.size(), + Idx + Foo.size() + sizeof(".framework/")-1); + if (F == Foo && DotFramework == ".framework/") { + isFramework = true; + return Foo; + } + +guess_library: + // pull off the suffix after the "." and make a point to it + a = Name.rfind('.'); + if (a == Name.npos || a == 0) + return StringRef(); + Dylib = Name.slice(a, Name.npos); + if (Dylib != ".dylib") + goto guess_qtx; + + // First pull off the version letter for the form Foo.A.dylib if any. + if (a >= 3) { + Dot = Name.slice(a-2, a-1); + if (Dot == ".") + a = a - 2; + } -error_code MachOObjectFile::getLibraryPath(DataRefImpl LibData, - StringRef &Res) const { - report_fatal_error("Needed libraries unimplemented in MachOObjectFile"); + b = Name.rfind('/', a); + if (b == Name.npos) + b = 0; + else + b = b+1; + // ignore any suffix after an underbar like Foo_profile.A.dylib + Idx = Name.find('_', b); + if (Idx != Name.npos && Idx != b) { + Lib = Name.slice(b, Idx); + Suffix = Name.slice(Idx, a); + } + else + Lib = Name.slice(b, a); + // There are incorrect library names of the form: + // libATS.A_profile.dylib so check for these. + if (Lib.size() >= 3) { + Dot = Lib.slice(Lib.size()-2, Lib.size()-1); + if (Dot == ".") + Lib = Lib.slice(0, Lib.size()-2); + } + return Lib; + +guess_qtx: + Qtx = Name.slice(a, Name.npos); + if (Qtx != ".qtx") + return StringRef(); + b = Name.rfind('/', a); + if (b == Name.npos) + Lib = Name.slice(0, a); + else + Lib = Name.slice(b+1, a); + // There are library names of the form: QT.A.qtx so check for these. + if (Lib.size() >= 3) { + Dot = Lib.slice(Lib.size()-2, Lib.size()-1); + if (Dot == ".") + Lib = Lib.slice(0, Lib.size()-2); + } + return Lib; } -symbol_iterator MachOObjectFile::symbol_begin() const { - DataRefImpl DRI; - if (!SymtabLoadCmd) - return symbol_iterator(SymbolRef(DRI, this)); +// getLibraryShortNameByIndex() is used to get the short name of the library +// for an undefined symbol in a linked Mach-O binary that was linked with the +// normal two-level namespace default (that is MH_TWOLEVEL in the header). +// It is passed the index (0 - based) of the library as translated from +// GET_LIBRARY_ORDINAL (1 - based). +std::error_code MachOObjectFile::getLibraryShortNameByIndex(unsigned Index, + StringRef &Res) const { + if (Index >= Libraries.size()) + return object_error::parse_failed; - MachO::symtab_command Symtab = getSymtabLoadCommand(); - DRI.p = reinterpret_cast(getPtr(this, Symtab.symoff)); - return symbol_iterator(SymbolRef(DRI, this)); + MachO::dylib_command D = + getStruct(this, Libraries[Index]); + if (D.dylib.name >= D.cmdsize) + return object_error::parse_failed; + + // If the cache of LibrariesShortNames is not built up do that first for + // all the Libraries. + if (LibrariesShortNames.size() == 0) { + for (unsigned i = 0; i < Libraries.size(); i++) { + MachO::dylib_command D = + getStruct(this, Libraries[i]); + if (D.dylib.name >= D.cmdsize) { + LibrariesShortNames.push_back(StringRef()); + continue; + } + const char *P = (const char *)(Libraries[i]) + D.dylib.name; + StringRef Name = StringRef(P); + StringRef Suffix; + bool isFramework; + StringRef shortName = guessLibraryShortName(Name, isFramework, Suffix); + if (shortName == StringRef()) + LibrariesShortNames.push_back(Name); + else + LibrariesShortNames.push_back(shortName); + } + } + + Res = LibrariesShortNames[Index]; + return object_error::success; +} + +basic_symbol_iterator MachOObjectFile::symbol_begin_impl() const { + return getSymbolByIndex(0); } -symbol_iterator MachOObjectFile::symbol_end() const { +basic_symbol_iterator MachOObjectFile::symbol_end_impl() const { DataRefImpl DRI; if (!SymtabLoadCmd) - return symbol_iterator(SymbolRef(DRI, this)); + return basic_symbol_iterator(SymbolRef(DRI, this)); MachO::symtab_command Symtab = getSymtabLoadCommand(); unsigned SymbolTableEntrySize = is64Bit() ? @@ -1185,7 +1216,21 @@ symbol_iterator MachOObjectFile::symbol_end() const { unsigned Offset = Symtab.symoff + Symtab.nsyms * SymbolTableEntrySize; DRI.p = reinterpret_cast(getPtr(this, Offset)); - return symbol_iterator(SymbolRef(DRI, this)); + return basic_symbol_iterator(SymbolRef(DRI, this)); +} + +basic_symbol_iterator MachOObjectFile::getSymbolByIndex(unsigned Index) const { + DataRefImpl DRI; + if (!SymtabLoadCmd) + return basic_symbol_iterator(SymbolRef(DRI, this)); + + MachO::symtab_command Symtab = getSymtabLoadCommand(); + assert(Index < Symtab.nsyms && "Requested symbol index is out of range."); + unsigned SymbolTableEntrySize = + is64Bit() ? sizeof(MachO::nlist_64) : sizeof(MachO::nlist); + DRI.p = reinterpret_cast(getPtr(this, Symtab.symoff)); + DRI.p += Index * SymbolTableEntrySize; + return basic_symbol_iterator(SymbolRef(DRI, this)); } section_iterator MachOObjectFile::section_begin() const { @@ -1199,16 +1244,6 @@ section_iterator MachOObjectFile::section_end() const { return section_iterator(SectionRef(DRI, this)); } -library_iterator MachOObjectFile::needed_library_begin() const { - // TODO: implement - report_fatal_error("Needed libraries unimplemented in MachOObjectFile"); -} - -library_iterator MachOObjectFile::needed_library_end() const { - // TODO: implement - report_fatal_error("Needed libraries unimplemented in MachOObjectFile"); -} - uint8_t MachOObjectFile::getBytesInAddress() const { return is64Bit() ? 8 : 4; } @@ -1238,6 +1273,8 @@ StringRef MachOObjectFile::getFileFormatName() const { switch (CPUType) { case llvm::MachO::CPU_TYPE_X86_64: return "Mach-O 64-bit x86-64"; + case llvm::MachO::CPU_TYPE_ARM64: + return "Mach-O arm64"; case llvm::MachO::CPU_TYPE_POWERPC64: return "Mach-O 64-bit ppc64"; default: @@ -1253,6 +1290,8 @@ Triple::ArchType MachOObjectFile::getArch(uint32_t CPUType) { return Triple::x86_64; case llvm::MachO::CPU_TYPE_ARM: return Triple::arm; + case llvm::MachO::CPU_TYPE_ARM64: + return Triple::aarch64; case llvm::MachO::CPU_TYPE_POWERPC: return Triple::ppc; case llvm::MachO::CPU_TYPE_POWERPC64: @@ -1262,13 +1301,181 @@ Triple::ArchType MachOObjectFile::getArch(uint32_t CPUType) { } } +Triple MachOObjectFile::getArch(uint32_t CPUType, uint32_t CPUSubType, + const char **McpuDefault) { + if (McpuDefault) + *McpuDefault = nullptr; + + switch (CPUType) { + case MachO::CPU_TYPE_I386: + switch (CPUSubType & ~MachO::CPU_SUBTYPE_MASK) { + case MachO::CPU_SUBTYPE_I386_ALL: + return Triple("i386-apple-darwin"); + default: + return Triple(); + } + case MachO::CPU_TYPE_X86_64: + switch (CPUSubType & ~MachO::CPU_SUBTYPE_MASK) { + case MachO::CPU_SUBTYPE_X86_64_ALL: + return Triple("x86_64-apple-darwin"); + case MachO::CPU_SUBTYPE_X86_64_H: + return Triple("x86_64h-apple-darwin"); + default: + return Triple(); + } + case MachO::CPU_TYPE_ARM: + switch (CPUSubType & ~MachO::CPU_SUBTYPE_MASK) { + case MachO::CPU_SUBTYPE_ARM_V4T: + return Triple("armv4t-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_V5TEJ: + return Triple("armv5e-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_XSCALE: + return Triple("xscale-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_V6: + return Triple("armv6-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_V6M: + if (McpuDefault) + *McpuDefault = "cortex-m0"; + return Triple("armv6m-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_V7: + return Triple("armv7-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_V7EM: + if (McpuDefault) + *McpuDefault = "cortex-m4"; + return Triple("armv7em-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_V7K: + return Triple("armv7k-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_V7M: + if (McpuDefault) + *McpuDefault = "cortex-m3"; + return Triple("armv7m-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_V7S: + return Triple("armv7s-apple-darwin"); + default: + return Triple(); + } + case MachO::CPU_TYPE_ARM64: + switch (CPUSubType & ~MachO::CPU_SUBTYPE_MASK) { + case MachO::CPU_SUBTYPE_ARM64_ALL: + return Triple("arm64-apple-darwin"); + default: + return Triple(); + } + case MachO::CPU_TYPE_POWERPC: + switch (CPUSubType & ~MachO::CPU_SUBTYPE_MASK) { + case MachO::CPU_SUBTYPE_POWERPC_ALL: + return Triple("ppc-apple-darwin"); + default: + return Triple(); + } + case MachO::CPU_TYPE_POWERPC64: + switch (CPUSubType & ~MachO::CPU_SUBTYPE_MASK) { + case MachO::CPU_SUBTYPE_POWERPC_ALL: + return Triple("ppc64-apple-darwin"); + default: + return Triple(); + } + default: + return Triple(); + } +} + +Triple MachOObjectFile::getThumbArch(uint32_t CPUType, uint32_t CPUSubType, + const char **McpuDefault) { + if (McpuDefault) + *McpuDefault = nullptr; + + switch (CPUType) { + case MachO::CPU_TYPE_ARM: + switch (CPUSubType & ~MachO::CPU_SUBTYPE_MASK) { + case MachO::CPU_SUBTYPE_ARM_V4T: + return Triple("thumbv4t-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_V5TEJ: + return Triple("thumbv5e-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_XSCALE: + return Triple("xscale-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_V6: + return Triple("thumbv6-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_V6M: + if (McpuDefault) + *McpuDefault = "cortex-m0"; + return Triple("thumbv6m-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_V7: + return Triple("thumbv7-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_V7EM: + if (McpuDefault) + *McpuDefault = "cortex-m4"; + return Triple("thumbv7em-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_V7K: + return Triple("thumbv7k-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_V7M: + if (McpuDefault) + *McpuDefault = "cortex-m3"; + return Triple("thumbv7m-apple-darwin"); + case MachO::CPU_SUBTYPE_ARM_V7S: + return Triple("thumbv7s-apple-darwin"); + default: + return Triple(); + } + default: + return Triple(); + } +} + +Triple MachOObjectFile::getArch(uint32_t CPUType, uint32_t CPUSubType, + const char **McpuDefault, + Triple *ThumbTriple) { + Triple T = MachOObjectFile::getArch(CPUType, CPUSubType, McpuDefault); + *ThumbTriple = MachOObjectFile::getThumbArch(CPUType, CPUSubType, + McpuDefault); + return T; +} + +Triple MachOObjectFile::getHostArch() { + return Triple(sys::getDefaultTargetTriple()); +} + +bool MachOObjectFile::isValidArch(StringRef ArchFlag) { + return StringSwitch(ArchFlag) + .Case("i386", true) + .Case("x86_64", true) + .Case("x86_64h", true) + .Case("armv4t", true) + .Case("arm", true) + .Case("armv5e", true) + .Case("armv6", true) + .Case("armv6m", true) + .Case("armv7em", true) + .Case("armv7k", true) + .Case("armv7m", true) + .Case("armv7s", true) + .Case("arm64", true) + .Case("ppc", true) + .Case("ppc64", true) + .Default(false); +} + unsigned MachOObjectFile::getArch() const { return getArch(getCPUType(this)); } -StringRef MachOObjectFile::getLoadName() const { - // TODO: Implement - report_fatal_error("get_load_name() unimplemented in MachOObjectFile"); +Triple MachOObjectFile::getArch(const char **McpuDefault, + Triple *ThumbTriple) const { + Triple T; + if (is64Bit()) { + MachO::mach_header_64 H_64; + H_64 = getHeader64(); + T = MachOObjectFile::getArch(H_64.cputype, H_64.cpusubtype, McpuDefault); + *ThumbTriple = MachOObjectFile::getThumbArch(H_64.cputype, H_64.cpusubtype, + McpuDefault); + } else { + MachO::mach_header H; + H = getHeader(); + T = MachOObjectFile::getArch(H.cputype, H.cpusubtype, McpuDefault); + *ThumbTriple = MachOObjectFile::getThumbArch(H.cputype, H.cpusubtype, + McpuDefault); + } + return T; } relocation_iterator MachOObjectFile::section_rel_begin(unsigned Index) const { @@ -1304,6 +1511,183 @@ dice_iterator MachOObjectFile::end_dices() const { return dice_iterator(DiceRef(DRI, this)); } +ExportEntry::ExportEntry(ArrayRef T) + : Trie(T), Malformed(false), Done(false) { } + +void ExportEntry::moveToFirst() { + pushNode(0); + pushDownUntilBottom(); +} + +void ExportEntry::moveToEnd() { + Stack.clear(); + Done = true; +} + +bool ExportEntry::operator==(const ExportEntry &Other) const { + // Common case, one at end, other iterating from begin. + if (Done || Other.Done) + return (Done == Other.Done); + // Not equal if different stack sizes. + if (Stack.size() != Other.Stack.size()) + return false; + // Not equal if different cumulative strings. + if (!CumulativeString.str().equals(Other.CumulativeString.str())) + return false; + // Equal if all nodes in both stacks match. + for (unsigned i=0; i < Stack.size(); ++i) { + if (Stack[i].Start != Other.Stack[i].Start) + return false; + } + return true; +} + +uint64_t ExportEntry::readULEB128(const uint8_t *&Ptr) { + unsigned Count; + uint64_t Result = decodeULEB128(Ptr, &Count); + Ptr += Count; + if (Ptr > Trie.end()) { + Ptr = Trie.end(); + Malformed = true; + } + return Result; +} + +StringRef ExportEntry::name() const { + return CumulativeString.str(); +} + +uint64_t ExportEntry::flags() const { + return Stack.back().Flags; +} + +uint64_t ExportEntry::address() const { + return Stack.back().Address; +} + +uint64_t ExportEntry::other() const { + return Stack.back().Other; +} + +StringRef ExportEntry::otherName() const { + const char* ImportName = Stack.back().ImportName; + if (ImportName) + return StringRef(ImportName); + return StringRef(); +} + +uint32_t ExportEntry::nodeOffset() const { + return Stack.back().Start - Trie.begin(); +} + +ExportEntry::NodeState::NodeState(const uint8_t *Ptr) + : Start(Ptr), Current(Ptr), Flags(0), Address(0), Other(0), + ImportName(nullptr), ChildCount(0), NextChildIndex(0), + ParentStringLength(0), IsExportNode(false) { +} + +void ExportEntry::pushNode(uint64_t offset) { + const uint8_t *Ptr = Trie.begin() + offset; + NodeState State(Ptr); + uint64_t ExportInfoSize = readULEB128(State.Current); + State.IsExportNode = (ExportInfoSize != 0); + const uint8_t* Children = State.Current + ExportInfoSize; + if (State.IsExportNode) { + State.Flags = readULEB128(State.Current); + if (State.Flags & MachO::EXPORT_SYMBOL_FLAGS_REEXPORT) { + State.Address = 0; + State.Other = readULEB128(State.Current); // dylib ordinal + State.ImportName = reinterpret_cast(State.Current); + } else { + State.Address = readULEB128(State.Current); + if (State.Flags & MachO::EXPORT_SYMBOL_FLAGS_STUB_AND_RESOLVER) + State.Other = readULEB128(State.Current); + } + } + State.ChildCount = *Children; + State.Current = Children + 1; + State.NextChildIndex = 0; + State.ParentStringLength = CumulativeString.size(); + Stack.push_back(State); +} + +void ExportEntry::pushDownUntilBottom() { + while (Stack.back().NextChildIndex < Stack.back().ChildCount) { + NodeState &Top = Stack.back(); + CumulativeString.resize(Top.ParentStringLength); + for (;*Top.Current != 0; Top.Current++) { + char C = *Top.Current; + CumulativeString.push_back(C); + } + Top.Current += 1; + uint64_t childNodeIndex = readULEB128(Top.Current); + Top.NextChildIndex += 1; + pushNode(childNodeIndex); + } + if (!Stack.back().IsExportNode) { + Malformed = true; + moveToEnd(); + } +} + +// We have a trie data structure and need a way to walk it that is compatible +// with the C++ iterator model. The solution is a non-recursive depth first +// traversal where the iterator contains a stack of parent nodes along with a +// string that is the accumulation of all edge strings along the parent chain +// to this point. +// +// There is one “export” node for each exported symbol. But because some +// symbols may be a prefix of another symbol (e.g. _dup and _dup2), an export +// node may have child nodes too. +// +// The algorithm for moveNext() is to keep moving down the leftmost unvisited +// child until hitting a node with no children (which is an export node or +// else the trie is malformed). On the way down, each node is pushed on the +// stack ivar. If there is no more ways down, it pops up one and tries to go +// down a sibling path until a childless node is reached. +void ExportEntry::moveNext() { + if (Stack.empty() || !Stack.back().IsExportNode) { + Malformed = true; + moveToEnd(); + return; + } + + Stack.pop_back(); + while (!Stack.empty()) { + NodeState &Top = Stack.back(); + if (Top.NextChildIndex < Top.ChildCount) { + pushDownUntilBottom(); + // Now at the next export node. + return; + } else { + if (Top.IsExportNode) { + // This node has no children but is itself an export node. + CumulativeString.resize(Top.ParentStringLength); + return; + } + Stack.pop_back(); + } + } + Done = true; +} + +iterator_range +MachOObjectFile::exports(ArrayRef Trie) { + ExportEntry Start(Trie); + Start.moveToFirst(); + + ExportEntry Finish(Trie); + Finish.moveToEnd(); + + return iterator_range(export_iterator(Start), + export_iterator(Finish)); +} + +iterator_range MachOObjectFile::exports() const { + return exports(getDyldInfoExportsTrie()); +} + + StringRef MachOObjectFile::getSectionFinalSegmentName(DataRefImpl Sec) const { ArrayRef Raw = getSectionRawFinalSegmentName(Sec); @@ -1314,14 +1698,14 @@ ArrayRef MachOObjectFile::getSectionRawName(DataRefImpl Sec) const { const section_base *Base = reinterpret_cast(Sections[Sec.d.a]); - return ArrayRef(Base->sectname); + return makeArrayRef(Base->sectname); } ArrayRef MachOObjectFile::getSectionRawFinalSegmentName(DataRefImpl Sec) const { const section_base *Base = reinterpret_cast(Sections[Sec.d.a]); - return ArrayRef(Base->segname); + return makeArrayRef(Base->segname); } bool @@ -1467,10 +1851,59 @@ MachOObjectFile::getLinkerOptionsLoadCommand(const LoadCommandInfo &L) const { return getStruct(this, L.Ptr); } +MachO::version_min_command +MachOObjectFile::getVersionMinLoadCommand(const LoadCommandInfo &L) const { + return getStruct(this, L.Ptr); +} + +MachO::dylib_command +MachOObjectFile::getDylibIDLoadCommand(const LoadCommandInfo &L) const { + return getStruct(this, L.Ptr); +} + +MachO::dyld_info_command +MachOObjectFile::getDyldInfoLoadCommand(const LoadCommandInfo &L) const { + return getStruct(this, L.Ptr); +} + +MachO::dylinker_command +MachOObjectFile::getDylinkerCommand(const LoadCommandInfo &L) const { + return getStruct(this, L.Ptr); +} + +MachO::uuid_command +MachOObjectFile::getUuidCommand(const LoadCommandInfo &L) const { + return getStruct(this, L.Ptr); +} + +MachO::source_version_command +MachOObjectFile::getSourceVersionCommand(const LoadCommandInfo &L) const { + return getStruct(this, L.Ptr); +} + +MachO::entry_point_command +MachOObjectFile::getEntryPointCommand(const LoadCommandInfo &L) const { + return getStruct(this, L.Ptr); +} + + MachO::any_relocation_info MachOObjectFile::getRelocation(DataRefImpl Rel) const { - const char *P = reinterpret_cast(Rel.p); - return getStruct(this, P); + DataRefImpl Sec; + Sec.d.a = Rel.d.a; + uint32_t Offset; + if (is64Bit()) { + MachO::section_64 Sect = getSection64(Sec); + Offset = Sect.reloff; + } else { + MachO::section Sect = getSection(Sec); + Offset = Sect.reloff; + } + + auto P = reinterpret_cast( + getPtr(this, Offset)) + Rel.d.b; + return getStruct( + this, reinterpret_cast(P)); } MachO::data_in_code_entry @@ -1523,6 +1956,62 @@ MachOObjectFile::getDataInCodeLoadCommand() const { return Cmd; } +ArrayRef MachOObjectFile::getDyldInfoRebaseOpcodes() const { + if (!DyldInfoLoadCmd) + return ArrayRef(); + + MachO::dyld_info_command DyldInfo + = getStruct(this, DyldInfoLoadCmd); + const uint8_t *Ptr = reinterpret_cast( + getPtr(this, DyldInfo.rebase_off)); + return ArrayRef(Ptr, DyldInfo.rebase_size); +} + +ArrayRef MachOObjectFile::getDyldInfoBindOpcodes() const { + if (!DyldInfoLoadCmd) + return ArrayRef(); + + MachO::dyld_info_command DyldInfo + = getStruct(this, DyldInfoLoadCmd); + const uint8_t *Ptr = reinterpret_cast( + getPtr(this, DyldInfo.bind_off)); + return ArrayRef(Ptr, DyldInfo.bind_size); +} + +ArrayRef MachOObjectFile::getDyldInfoWeakBindOpcodes() const { + if (!DyldInfoLoadCmd) + return ArrayRef(); + + MachO::dyld_info_command DyldInfo + = getStruct(this, DyldInfoLoadCmd); + const uint8_t *Ptr = reinterpret_cast( + getPtr(this, DyldInfo.weak_bind_off)); + return ArrayRef(Ptr, DyldInfo.weak_bind_size); +} + +ArrayRef MachOObjectFile::getDyldInfoLazyBindOpcodes() const { + if (!DyldInfoLoadCmd) + return ArrayRef(); + + MachO::dyld_info_command DyldInfo + = getStruct(this, DyldInfoLoadCmd); + const uint8_t *Ptr = reinterpret_cast( + getPtr(this, DyldInfo.lazy_bind_off)); + return ArrayRef(Ptr, DyldInfo.lazy_bind_size); +} + +ArrayRef MachOObjectFile::getDyldInfoExportsTrie() const { + if (!DyldInfoLoadCmd) + return ArrayRef(); + + MachO::dyld_info_command DyldInfo + = getStruct(this, DyldInfoLoadCmd); + const uint8_t *Ptr = reinterpret_cast( + getPtr(this, DyldInfo.export_off)); + return ArrayRef(Ptr, DyldInfo.export_size); +} + + StringRef MachOObjectFile::getStringTableData() const { MachO::symtab_command S = getSymtabLoadCommand(); return getData().substr(S.stroff, S.strsize); @@ -1545,28 +2034,28 @@ void MachOObjectFile::ReadULEB128s(uint64_t Index, } } -ErrorOr ObjectFile::createMachOObjectFile(MemoryBuffer *Buffer, - bool BufferOwned) { - StringRef Magic = Buffer->getBuffer().slice(0, 4); - error_code EC; - OwningPtr Ret; +bool MachOObjectFile::isRelocatableObject() const { + return getHeader().filetype == MachO::MH_OBJECT; +} + +ErrorOr> +ObjectFile::createMachOObjectFile(MemoryBufferRef Buffer) { + StringRef Magic = Buffer.getBuffer().slice(0, 4); + std::error_code EC; + std::unique_ptr Ret; if (Magic == "\xFE\xED\xFA\xCE") - Ret.reset(new MachOObjectFile(Buffer, false, false, EC, BufferOwned)); + Ret.reset(new MachOObjectFile(Buffer, false, false, EC)); else if (Magic == "\xCE\xFA\xED\xFE") - Ret.reset(new MachOObjectFile(Buffer, true, false, EC, BufferOwned)); + Ret.reset(new MachOObjectFile(Buffer, true, false, EC)); else if (Magic == "\xFE\xED\xFA\xCF") - Ret.reset(new MachOObjectFile(Buffer, false, true, EC, BufferOwned)); + Ret.reset(new MachOObjectFile(Buffer, false, true, EC)); else if (Magic == "\xCF\xFA\xED\xFE") - Ret.reset(new MachOObjectFile(Buffer, true, true, EC, BufferOwned)); - else { - delete Buffer; + Ret.reset(new MachOObjectFile(Buffer, true, true, EC)); + else return object_error::parse_failed; - } if (EC) return EC; - return Ret.take(); + return std::move(Ret); } -} // end namespace object -} // end namespace llvm