tools/llvm-objcopy/COFF/Reader.cpp

   1 //===- Reader.cpp ---------------------------------------------------------===//
   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 "Reader.h"
  10 #include "Object.h"
  11 #include "llvm/ADT/ArrayRef.h"
  12 #include "llvm/ADT/StringRef.h"
  13 #include "llvm/BinaryFormat/COFF.h"
  14 #include "llvm/Object/COFF.h"
  15 #include "llvm/Support/ErrorHandling.h"
  16 #include <cstddef>
  17 #include <cstdint>
  18
  19 namespace llvm {
  20 namespace objcopy {
  21 namespace coff {
  22
  23 using namespace object;
  24 using namespace COFF;
  25
  26 Error COFFReader::readExecutableHeaders(Object &Obj) const {
  27   const dos_header *DH = COFFObj.getDOSHeader();
  28   Obj.Is64 = COFFObj.is64();
  29   if (!DH)
  30     return Error::success();
  31
  32   Obj.IsPE = true;
  33   Obj.DosHeader = *DH;
  34   if (DH->AddressOfNewExeHeader > sizeof(*DH))
  35     Obj.DosStub = ArrayRef<uint8_t>(reinterpret_cast<const uint8_t *>(&DH[1]),
  36                                     DH->AddressOfNewExeHeader - sizeof(*DH));
  37
  38   if (COFFObj.is64()) {
  39     Obj.PeHeader = *COFFObj.getPE32PlusHeader();
  40   } else {
  41     const pe32_header *PE32 = COFFObj.getPE32Header();
  42     copyPeHeader(Obj.PeHeader, *PE32);
  43     // The pe32plus_header (stored in Object) lacks the BaseOfData field.
  44     Obj.BaseOfData = PE32->BaseOfData;
  45   }
  46
  47   for (size_t I = 0; I < Obj.PeHeader.NumberOfRvaAndSize; I++) {
  48     const data_directory *Dir;
  49     if (auto EC = COFFObj.getDataDirectory(I, Dir))
  50       return errorCodeToError(EC);
  51     Obj.DataDirectories.emplace_back(*Dir);
  52   }
  53   return Error::success();
  54 }
  55
  56 Error COFFReader::readSections(Object &Obj) const {
  57   std::vector<Section> Sections;
  58   // Section indexing starts from 1.
  59   for (size_t I = 1, E = COFFObj.getNumberOfSections(); I <= E; I++) {
  60     const coff_section *Sec;
  61     if (auto EC = COFFObj.getSection(I, Sec))
  62       return errorCodeToError(EC);
  63     Sections.push_back(Section());
  64     Section &S = Sections.back();
  65     S.Header = *Sec;
  66     ArrayRef<uint8_t> Contents;
  67     if (Error E = COFFObj.getSectionContents(Sec, Contents))
  68       return E;
  69     S.setContentsRef(Contents);
  70     ArrayRef<coff_relocation> Relocs = COFFObj.getRelocations(Sec);
  71     for (const coff_relocation &R : Relocs)
  72       S.Relocs.push_back(R);
  73     if (Expected<StringRef> NameOrErr = COFFObj.getSectionName(Sec))
  74       S.Name = *NameOrErr;
  75     else
  76       return NameOrErr.takeError();
  77     if (Sec->hasExtendedRelocations())
  78       return createStringError(object_error::parse_failed,
  79                                "extended relocations not supported yet");
  80   }
  81   Obj.addSections(Sections);
  82   return Error::success();
  83 }
  84
  85 Error COFFReader::readSymbols(Object &Obj, bool IsBigObj) const {
  86   std::vector<Symbol> Symbols;
  87   Symbols.reserve(COFFObj.getRawNumberOfSymbols());
  88   ArrayRef<Section> Sections = Obj.getSections();
  89   for (uint32_t I = 0, E = COFFObj.getRawNumberOfSymbols(); I < E;) {
  90     Expected<COFFSymbolRef> SymOrErr = COFFObj.getSymbol(I);
  91     if (!SymOrErr)
  92       return SymOrErr.takeError();
  93     COFFSymbolRef SymRef = *SymOrErr;
  94
  95     Symbols.push_back(Symbol());
  96     Symbol &Sym = Symbols.back();
  97     // Copy symbols from the original form into an intermediate coff_symbol32.
  98     if (IsBigObj)
  99       copySymbol(Sym.Sym,
 100                  *reinterpret_cast<const coff_symbol32 *>(SymRef.getRawPtr()));
 101     else
 102       copySymbol(Sym.Sym,
 103                  *reinterpret_cast<const coff_symbol16 *>(SymRef.getRawPtr()));
 104     if (auto EC = COFFObj.getSymbolName(SymRef, Sym.Name))
 105       return errorCodeToError(EC);
 106
 107     ArrayRef<uint8_t> AuxData = COFFObj.getSymbolAuxData(SymRef);
 108     size_t SymSize = IsBigObj ? sizeof(coff_symbol32) : sizeof(coff_symbol16);
 109     assert(AuxData.size() == SymSize * SymRef.getNumberOfAuxSymbols());
 110     // The auxillary symbols are structs of sizeof(coff_symbol16) each.
 111     // In the big object format (where symbols are coff_symbol32), each
 112     // auxillary symbol is padded with 2 bytes at the end. Copy each
 113     // auxillary symbol to the Sym.AuxData vector. For file symbols,
 114     // the whole range of aux symbols are interpreted as one null padded
 115     // string instead.
 116     if (SymRef.isFileRecord())
 117       Sym.AuxFile = StringRef(reinterpret_cast<const char *>(AuxData.data()),
 118                               AuxData.size())
 119                         .rtrim('\0');
 120     else
 121       for (size_t I = 0; I < SymRef.getNumberOfAuxSymbols(); I++)
 122         Sym.AuxData.push_back(AuxData.slice(I * SymSize, sizeof(AuxSymbol)));
 123
 124     // Find the unique id of the section
 125     if (SymRef.getSectionNumber() <=
 126         0) // Special symbol (undefined/absolute/debug)
 127       Sym.TargetSectionId = SymRef.getSectionNumber();
 128     else if (static_cast<uint32_t>(SymRef.getSectionNumber() - 1) <
 129              Sections.size())
 130       Sym.TargetSectionId = Sections[SymRef.getSectionNumber() - 1].UniqueId;
 131     else
 132       return createStringError(object_error::parse_failed,
 133                                "section number out of range");
 134     // For section definitions, check if it is comdat associative, and if
 135     // it is, find the target section unique id.
 136     const coff_aux_section_definition *SD = SymRef.getSectionDefinition();
 137     const coff_aux_weak_external *WE = SymRef.getWeakExternal();
 138     if (SD && SD->Selection == IMAGE_COMDAT_SELECT_ASSOCIATIVE) {
 139       int32_t Index = SD->getNumber(IsBigObj);
 140       if (Index <= 0 || static_cast<uint32_t>(Index - 1) >= Sections.size())
 141         return createStringError(object_error::parse_failed,
 142                                  "unexpected associative section index");
 143       Sym.AssociativeComdatTargetSectionId = Sections[Index - 1].UniqueId;
 144     } else if (WE) {
 145       // This is a raw symbol index for now, but store it in the Symbol
 146       // until we've added them to the Object, which assigns the final
 147       // unique ids.
 148       Sym.WeakTargetSymbolId = WE->TagIndex;
 149     }
 150     I += 1 + SymRef.getNumberOfAuxSymbols();
 151   }
 152   Obj.addSymbols(Symbols);
 153   return Error::success();
 154 }
 155
 156 Error COFFReader::setSymbolTargets(Object &Obj) const {
 157   std::vector<const Symbol *> RawSymbolTable;
 158   for (const Symbol &Sym : Obj.getSymbols()) {
 159     RawSymbolTable.push_back(&Sym);
 160     for (size_t I = 0; I < Sym.Sym.NumberOfAuxSymbols; I++)
 161       RawSymbolTable.push_back(nullptr);
 162   }
 163   for (Symbol &Sym : Obj.getMutableSymbols()) {
 164     // Convert WeakTargetSymbolId from the original raw symbol index to
 165     // a proper unique id.
 166     if (Sym.WeakTargetSymbolId) {
 167       if (*Sym.WeakTargetSymbolId >= RawSymbolTable.size())
 168         return createStringError(object_error::parse_failed,
 169                                  "weak external reference out of range");
 170       const Symbol *Target = RawSymbolTable[*Sym.WeakTargetSymbolId];
 171       if (Target == nullptr)
 172         return createStringError(object_error::parse_failed,
 173                                  "invalid SymbolTableIndex");
 174       Sym.WeakTargetSymbolId = Target->UniqueId;
 175     }
 176   }
 177   for (Section &Sec : Obj.getMutableSections()) {
 178     for (Relocation &R : Sec.Relocs) {
 179       if (R.Reloc.SymbolTableIndex >= RawSymbolTable.size())
 180         return createStringError(object_error::parse_failed,
 181                                  "SymbolTableIndex out of range");
 182       const Symbol *Sym = RawSymbolTable[R.Reloc.SymbolTableIndex];
 183       if (Sym == nullptr)
 184         return createStringError(object_error::parse_failed,
 185                                  "invalid SymbolTableIndex");
 186       R.Target = Sym->UniqueId;
 187       R.TargetName = Sym->Name;
 188     }
 189   }
 190   return Error::success();
 191 }
 192
 193 Expected<std::unique_ptr<Object>> COFFReader::create() const {
 194   auto Obj = std::make_unique<Object>();
 195
 196   bool IsBigObj = false;
 197   if (const coff_file_header *CFH = COFFObj.getCOFFHeader()) {
 198     Obj->CoffFileHeader = *CFH;
 199   } else {
 200     const coff_bigobj_file_header *CBFH = COFFObj.getCOFFBigObjHeader();
 201     if (!CBFH)
 202       return createStringError(object_error::parse_failed,
 203                                "no COFF file header returned");
 204     // Only copying the few fields from the bigobj header that we need
 205     // and won't recreate in the end.
 206     Obj->CoffFileHeader.Machine = CBFH->Machine;
 207     Obj->CoffFileHeader.TimeDateStamp = CBFH->TimeDateStamp;
 208     IsBigObj = true;
 209   }
 210
 211   if (Error E = readExecutableHeaders(*Obj))
 212     return std::move(E);
 213   if (Error E = readSections(*Obj))
 214     return std::move(E);
 215   if (Error E = readSymbols(*Obj, IsBigObj))
 216     return std::move(E);
 217   if (Error E = setSymbolTargets(*Obj))
 218     return std::move(E);
 219
 220   return std::move(Obj);
 221 }
 222
 223 } // end namespace coff
 224 } // end namespace objcopy
 225 } // end namespace llvm