tools/llvm-objcopy/MachO/MachOReader.cpp

   1 //===- MachOReader.cpp ------------------------------------------*- C++ -*-===//
   2 //
   3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
   4 // See https://llvm.org/LICENSE.txt for license information.
   5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8
   9 #include "MachOReader.h"
  10 #include "../llvm-objcopy.h"
  11 #include "Object.h"
  12 #include "llvm/BinaryFormat/MachO.h"
  13 #include "llvm/Object/MachO.h"
  14 #include <memory>
  15
  16 namespace llvm {
  17 namespace objcopy {
  18 namespace macho {
  19
  20 void MachOReader::readHeader(Object &O) const {
  21   O.Header.Magic = MachOObj.getHeader().magic;
  22   O.Header.CPUType = MachOObj.getHeader().cputype;
  23   O.Header.CPUSubType = MachOObj.getHeader().cpusubtype;
  24   O.Header.FileType = MachOObj.getHeader().filetype;
  25   O.Header.NCmds = MachOObj.getHeader().ncmds;
  26   O.Header.SizeOfCmds = MachOObj.getHeader().sizeofcmds;
  27   O.Header.Flags = MachOObj.getHeader().flags;
  28 }
  29
  30 template <typename SectionType>
  31 Section constructSectionCommon(SectionType Sec) {
  32   Section S;
  33   S.Sectname =
  34       StringRef(Sec.sectname, strnlen(Sec.sectname, sizeof(Sec.sectname)))
  35           .str();
  36   S.Segname =
  37       StringRef(Sec.segname, strnlen(Sec.segname, sizeof(Sec.sectname))).str();
  38   S.Addr = Sec.addr;
  39   S.Size = Sec.size;
  40   S.Offset = Sec.offset;
  41   S.Align = Sec.align;
  42   S.RelOff = Sec.reloff;
  43   S.NReloc = Sec.nreloc;
  44   S.Flags = Sec.flags;
  45   S.Reserved1 = Sec.reserved1;
  46   S.Reserved2 = Sec.reserved2;
  47   S.Reserved3 = 0;
  48   return S;
  49 }
  50
  51 template <typename SectionType> Section constructSection(SectionType Sec);
  52
  53 template <> Section constructSection(MachO::section Sec) {
  54   return constructSectionCommon(Sec);
  55 }
  56
  57 template <> Section constructSection(MachO::section_64 Sec) {
  58   Section S = constructSectionCommon(Sec);
  59   S.Reserved3 = Sec.reserved3;
  60   return S;
  61 }
  62
  63 // TODO: get rid of reportError and make MachOReader return Expected<> instead.
  64 template <typename SectionType, typename SegmentType>
  65 std::vector<Section>
  66 extractSections(const object::MachOObjectFile::LoadCommandInfo &LoadCmd,
  67                 const object::MachOObjectFile &MachOObj,
  68                 size_t &NextSectionIndex) {
  69   auto End = LoadCmd.Ptr + LoadCmd.C.cmdsize;
  70   const SectionType *Curr =
  71       reinterpret_cast<const SectionType *>(LoadCmd.Ptr + sizeof(SegmentType));
  72   std::vector<Section> Sections;
  73   for (; reinterpret_cast<const void *>(Curr) < End; Curr++) {
  74     if (MachOObj.isLittleEndian() != sys::IsLittleEndianHost) {
  75       SectionType Sec;
  76       memcpy((void *)&Sec, Curr, sizeof(SectionType));
  77       MachO::swapStruct(Sec);
  78       Sections.push_back(constructSection(Sec));
  79     } else {
  80       Sections.push_back(constructSection(*Curr));
  81     }
  82
  83     Section &S = Sections.back();
  84
  85     Expected<object::SectionRef> SecRef =
  86         MachOObj.getSection(NextSectionIndex++);
  87     if (!SecRef)
  88       reportError(MachOObj.getFileName(), SecRef.takeError());
  89
  90     if (Expected<ArrayRef<uint8_t>> E =
  91             MachOObj.getSectionContents(SecRef->getRawDataRefImpl()))
  92       S.Content =
  93           StringRef(reinterpret_cast<const char *>(E->data()), E->size());
  94     else
  95       reportError(MachOObj.getFileName(), E.takeError());
  96
  97     S.Relocations.reserve(S.NReloc);
  98     for (auto RI = MachOObj.section_rel_begin(SecRef->getRawDataRefImpl()),
  99               RE = MachOObj.section_rel_end(SecRef->getRawDataRefImpl());
 100          RI != RE; ++RI) {
 101       RelocationInfo R;
 102       R.Symbol = nullptr; // We'll fill this field later.
 103       R.Info = MachOObj.getRelocation(RI->getRawDataRefImpl());
 104       R.Scattered = MachOObj.isRelocationScattered(R.Info);
 105       S.Relocations.push_back(R);
 106     }
 107
 108     assert(S.NReloc == S.Relocations.size() &&
 109            "Incorrect number of relocations");
 110   }
 111   return Sections;
 112 }
 113
 114 void MachOReader::readLoadCommands(Object &O) const {
 115   // For MachO sections indices start from 1.
 116   size_t NextSectionIndex = 1;
 117   for (auto LoadCmd : MachOObj.load_commands()) {
 118     LoadCommand LC;
 119     switch (LoadCmd.C.cmd) {
 120     case MachO::LC_SEGMENT:
 121       LC.Sections = extractSections<MachO::section, MachO::segment_command>(
 122           LoadCmd, MachOObj, NextSectionIndex);
 123       break;
 124     case MachO::LC_SEGMENT_64:
 125       LC.Sections =
 126           extractSections<MachO::section_64, MachO::segment_command_64>(
 127               LoadCmd, MachOObj, NextSectionIndex);
 128       break;
 129     case MachO::LC_SYMTAB:
 130       O.SymTabCommandIndex = O.LoadCommands.size();
 131       break;
 132     case MachO::LC_DYLD_INFO:
 133     case MachO::LC_DYLD_INFO_ONLY:
 134       O.DyLdInfoCommandIndex = O.LoadCommands.size();
 135       break;
 136     }
 137 #define HANDLE_LOAD_COMMAND(LCName, LCValue, LCStruct)                         \
 138   case MachO::LCName:                                                          \
 139     memcpy((void *)&(LC.MachOLoadCommand.LCStruct##_data), LoadCmd.Ptr,        \
 140            sizeof(MachO::LCStruct));                                           \
 141     if (MachOObj.isLittleEndian() != sys::IsLittleEndianHost)                  \
 142       MachO::swapStruct(LC.MachOLoadCommand.LCStruct##_data);                  \
 143     LC.Payload = ArrayRef<uint8_t>(                                            \
 144         reinterpret_cast<uint8_t *>(const_cast<char *>(LoadCmd.Ptr)) +         \
 145             sizeof(MachO::LCStruct),                                           \
 146         LoadCmd.C.cmdsize - sizeof(MachO::LCStruct));                          \
 147     break;
 148
 149     switch (LoadCmd.C.cmd) {
 150     default:
 151       memcpy((void *)&(LC.MachOLoadCommand.load_command_data), LoadCmd.Ptr,
 152              sizeof(MachO::load_command));
 153       if (MachOObj.isLittleEndian() != sys::IsLittleEndianHost)
 154         MachO::swapStruct(LC.MachOLoadCommand.load_command_data);
 155       LC.Payload = ArrayRef<uint8_t>(
 156           reinterpret_cast<uint8_t *>(const_cast<char *>(LoadCmd.Ptr)) +
 157               sizeof(MachO::load_command),
 158           LoadCmd.C.cmdsize - sizeof(MachO::load_command));
 159       break;
 160 #include "llvm/BinaryFormat/MachO.def"
 161     }
 162     O.LoadCommands.push_back(std::move(LC));
 163   }
 164 }
 165
 166 template <typename nlist_t>
 167 SymbolEntry constructSymbolEntry(StringRef StrTable, const nlist_t &nlist) {
 168   assert(nlist.n_strx < StrTable.size() &&
 169          "n_strx exceeds the size of the string table");
 170   SymbolEntry SE;
 171   SE.Name = StringRef(StrTable.data() + nlist.n_strx).str();
 172   SE.n_type = nlist.n_type;
 173   SE.n_sect = nlist.n_sect;
 174   SE.n_desc = nlist.n_desc;
 175   SE.n_value = nlist.n_value;
 176   return SE;
 177 }
 178
 179 void MachOReader::readSymbolTable(Object &O) const {
 180   StringRef StrTable = MachOObj.getStringTableData();
 181   for (auto Symbol : MachOObj.symbols()) {
 182     SymbolEntry SE =
 183         (MachOObj.is64Bit()
 184              ? constructSymbolEntry(
 185                    StrTable,
 186                    MachOObj.getSymbol64TableEntry(Symbol.getRawDataRefImpl()))
 187              : constructSymbolEntry(
 188                    StrTable,
 189                    MachOObj.getSymbolTableEntry(Symbol.getRawDataRefImpl())));
 190
 191     O.SymTable.Symbols.push_back(llvm::make_unique<SymbolEntry>(SE));
 192   }
 193 }
 194
 195 void MachOReader::setSymbolInRelocationInfo(Object &O) const {
 196   for (auto &LC : O.LoadCommands)
 197     for (auto &Sec : LC.Sections)
 198       for (auto &Reloc : Sec.Relocations)
 199         if (!Reloc.Scattered) {
 200           auto *Info = reinterpret_cast<MachO::relocation_info *>(&Reloc.Info);
 201           Reloc.Symbol = O.SymTable.getSymbolByIndex(Info->r_symbolnum);
 202         }
 203 }
 204
 205 void MachOReader::readRebaseInfo(Object &O) const {
 206   O.Rebases.Opcodes = MachOObj.getDyldInfoRebaseOpcodes();
 207 }
 208
 209 void MachOReader::readBindInfo(Object &O) const {
 210   O.Binds.Opcodes = MachOObj.getDyldInfoBindOpcodes();
 211 }
 212
 213 void MachOReader::readWeakBindInfo(Object &O) const {
 214   O.WeakBinds.Opcodes = MachOObj.getDyldInfoWeakBindOpcodes();
 215 }
 216
 217 void MachOReader::readLazyBindInfo(Object &O) const {
 218   O.LazyBinds.Opcodes = MachOObj.getDyldInfoLazyBindOpcodes();
 219 }
 220
 221 void MachOReader::readExportInfo(Object &O) const {
 222   O.Exports.Trie = MachOObj.getDyldInfoExportsTrie();
 223 }
 224
 225 std::unique_ptr<Object> MachOReader::create() const {
 226   auto Obj = llvm::make_unique<Object>();
 227   readHeader(*Obj);
 228   readLoadCommands(*Obj);
 229   readSymbolTable(*Obj);
 230   setSymbolInRelocationInfo(*Obj);
 231   readRebaseInfo(*Obj);
 232   readBindInfo(*Obj);
 233   readWeakBindInfo(*Obj);
 234   readLazyBindInfo(*Obj);
 235   readExportInfo(*Obj);
 236   return Obj;
 237 }
 238
 239 } // end namespace macho
 240 } // end namespace objcopy
 241 } // end namespace llvm