//===- Writer.cpp ---------------------------------------------------------===// // // The LLVM Linker // // This file is distributed under the University of Illinois Open Source // License. See LICENSE.TXT for details. // //===----------------------------------------------------------------------===// #include "Writer.h" #include "Config.h" #include "LinkerScript.h" #include "Memory.h" #include "OutputSections.h" #include "Relocations.h" #include "Strings.h" #include "SymbolTable.h" #include "SyntheticSections.h" #include "Target.h" #include "llvm/ADT/StringMap.h" #include "llvm/ADT/StringSwitch.h" #include "llvm/Support/FileOutputBuffer.h" #include "llvm/Support/FileSystem.h" #include "llvm/Support/raw_ostream.h" #include #include using namespace llvm; using namespace llvm::ELF; using namespace llvm::object; using namespace llvm::support; using namespace llvm::support::endian; using namespace lld; using namespace lld::elf; namespace { // The writer writes a SymbolTable result to a file. template class Writer { public: typedef typename ELFT::uint uintX_t; typedef typename ELFT::Shdr Elf_Shdr; typedef typename ELFT::Ehdr Elf_Ehdr; typedef typename ELFT::Phdr Elf_Phdr; typedef typename ELFT::Sym Elf_Sym; typedef typename ELFT::SymRange Elf_Sym_Range; typedef typename ELFT::Rela Elf_Rela; void run(); private: void createSyntheticSections(); void copyLocalSymbols(); void addReservedSymbols(); void addInputSec(InputSectionBase *S); void createSections(); void forEachRelSec(std::function &)> Fn); void sortSections(); void finalizeSections(); void addPredefinedSections(); std::vector createPhdrs(); void removeEmptyPTLoad(); void addPtArmExid(std::vector &Phdrs); void assignAddresses(); void assignFileOffsets(); void assignFileOffsetsBinary(); void setPhdrs(); void fixHeaders(); void fixSectionAlignments(); void fixAbsoluteSymbols(); void openFile(); void writeHeader(); void writeSections(); void writeSectionsBinary(); void writeBuildId(); std::unique_ptr Buffer; std::vector OutputSections; OutputSectionFactory Factory; void addRelIpltSymbols(); void addStartEndSymbols(); void addStartStopSymbols(OutputSectionBase *Sec); uintX_t getEntryAddr(); OutputSectionBase *findSection(StringRef Name); std::vector Phdrs; uintX_t FileSize; uintX_t SectionHeaderOff; bool AllocateHeader = true; }; } // anonymous namespace StringRef elf::getOutputSectionName(StringRef Name) { if (Config->Relocatable) return Name; for (StringRef V : {".text.", ".rodata.", ".data.rel.ro.", ".data.", ".bss.", ".init_array.", ".fini_array.", ".ctors.", ".dtors.", ".tbss.", ".gcc_except_table.", ".tdata.", ".ARM.exidx."}) { StringRef Prefix = V.drop_back(); if (Name.startswith(V) || Name == Prefix) return Prefix; } // CommonSection is identified as "COMMON" in linker scripts. // By default, it should go to .bss section. if (Name == "COMMON") return ".bss"; // ".zdebug_" is a prefix for ZLIB-compressed sections. // Because we decompressed input sections, we want to remove 'z'. if (Name.startswith(".zdebug_")) return Saver.save(Twine(".") + Name.substr(2)); return Name; } template void elf::reportDiscarded(InputSectionBase *IS) { if (!Config->PrintGcSections) return; errs() << "removing unused section from '" << IS->Name << "' in file '" << IS->getFile()->getName() << "'\n"; } template static bool needsInterpSection() { return !Symtab::X->getSharedFiles().empty() && !Config->DynamicLinker.empty() && !Script::X->ignoreInterpSection(); } template void elf::writeResult() { Writer().run(); } template void Writer::removeEmptyPTLoad() { auto I = std::remove_if(Phdrs.begin(), Phdrs.end(), [&](const PhdrEntry &P) { if (P.p_type != PT_LOAD) return false; if (!P.First) return true; uintX_t Size = P.Last->Addr + P.Last->Size - P.First->Addr; return Size == 0; }); Phdrs.erase(I, Phdrs.end()); } // The main function of the writer. template void Writer::run() { // Create linker-synthesized sections such as .got or .plt. // Such sections are of type input section. createSyntheticSections(); // We need to create some reserved symbols such as _end. Create them. if (!Config->Relocatable) addReservedSymbols(); // Some architectures use small displacements for jump instructions. // It is linker's responsibility to create thunks containing long // jump instructions if jump targets are too far. Create thunks. if (Target->NeedsThunks) forEachRelSec(createThunks); // Create output sections. Script::X->OutputSections = &OutputSections; if (ScriptConfig->HasSections) { // If linker script contains SECTIONS commands, let it create sections. Script::X->processCommands(Factory); // Linker scripts may have left some input sections unassigned. // Assign such sections using the default rule. Script::X->addOrphanSections(Factory); } else { // If linker script does not contain SECTIONS commands, create // output sections by default rules. We still need to give the // linker script a chance to run, because it might contain // non-SECTIONS commands such as ASSERT. createSections(); Script::X->processCommands(Factory); } if (Config->Discard != DiscardPolicy::All) copyLocalSymbols(); // Now that we have a complete set of output sections. This function // completes section contents. For example, we need to add strings // to the string table, and add entries to .got and .plt. // finalizeSections does that. finalizeSections(); if (ErrorCount) return; if (Config->Relocatable) { assignFileOffsets(); } else { if (ScriptConfig->HasSections) { Script::X->assignAddresses(Phdrs); } else { fixSectionAlignments(); assignAddresses(); } // Remove empty PT_LOAD to avoid causing the dynamic linker to try to mmap a // 0 sized region. This has to be done late since only after assignAddresses // we know the size of the sections. removeEmptyPTLoad(); if (!Config->OFormatBinary) assignFileOffsets(); else assignFileOffsetsBinary(); setPhdrs(); fixAbsoluteSymbols(); } // Write the result down to a file. openFile(); if (ErrorCount) return; if (!Config->OFormatBinary) { writeHeader(); writeSections(); } else { writeSectionsBinary(); } // Backfill .note.gnu.build-id section content. This is done at last // because the content is usually a hash value of the entire output file. writeBuildId(); if (ErrorCount) return; if (auto EC = Buffer->commit()) error(EC, "failed to write to the output file"); // Flush the output streams and exit immediately. A full shutdown // is a good test that we are keeping track of all allocated memory, // but actually freeing it is a waste of time in a regular linker run. if (Config->ExitEarly) exitLld(0); } // Initialize Out members. template void Writer::createSyntheticSections() { // Initialize all pointers with NULL. This is needed because // you can call lld::elf::main more than once as a library. memset(&Out::First, 0, sizeof(Out)); // Create singleton output sections. Out::Bss = make>(".bss", SHT_NOBITS, SHF_ALLOC | SHF_WRITE); Out::BssRelRo = make>(".bss.rel.ro", SHT_NOBITS, SHF_ALLOC | SHF_WRITE); In::DynStrTab = make>(".dynstr", true); In::Dynamic = make>(); Out::EhFrame = make>(); In::RelaDyn = make>( Config->Rela ? ".rela.dyn" : ".rel.dyn", Config->ZCombreloc); In::ShStrTab = make>(".shstrtab", false); Out::ElfHeader = make("", 0, SHF_ALLOC); Out::ElfHeader->Size = sizeof(Elf_Ehdr); Out::ProgramHeaders = make("", 0, SHF_ALLOC); Out::ProgramHeaders->updateAlignment(sizeof(uintX_t)); if (needsInterpSection()) { In::Interp = createInterpSection(); Symtab::X->Sections.push_back(In::Interp); } else { In::Interp = nullptr; } if (!Config->Relocatable) Symtab::X->Sections.push_back(createCommentSection()); if (Config->Strip != StripPolicy::All) { In::StrTab = make>(".strtab", false); In::SymTab = make>(*In::StrTab); } if (Config->BuildId != BuildIdKind::None) { In::BuildId = make>(); Symtab::X->Sections.push_back(In::BuildId); } InputSection *Common = createCommonSection(); if (!Common->Data.empty()) { In::Common = Common; Symtab::X->Sections.push_back(Common); } // Add MIPS-specific sections. bool HasDynSymTab = !Symtab::X->getSharedFiles().empty() || Config->Pic; if (Config->EMachine == EM_MIPS) { if (!Config->Shared && HasDynSymTab) { In::MipsRldMap = make>(); Symtab::X->Sections.push_back(In::MipsRldMap); } if (auto *Sec = MipsAbiFlagsSection::create()) Symtab::X->Sections.push_back(Sec); if (auto *Sec = MipsOptionsSection::create()) Symtab::X->Sections.push_back(Sec); if (auto *Sec = MipsReginfoSection::create()) Symtab::X->Sections.push_back(Sec); } if (HasDynSymTab) { In::DynSymTab = make>(*In::DynStrTab); Symtab::X->Sections.push_back(In::DynSymTab); In::VerSym = make>(); Symtab::X->Sections.push_back(In::VerSym); if (!Config->VersionDefinitions.empty()) { In::VerDef = make>(); Symtab::X->Sections.push_back(In::VerDef); } In::VerNeed = make>(); Symtab::X->Sections.push_back(In::VerNeed); if (Config->GnuHash) { In::GnuHashTab = make>(); Symtab::X->Sections.push_back(In::GnuHashTab); } if (Config->SysvHash) { In::HashTab = make>(); Symtab::X->Sections.push_back(In::HashTab); } Symtab::X->Sections.push_back(In::Dynamic); Symtab::X->Sections.push_back(In::DynStrTab); Symtab::X->Sections.push_back(In::RelaDyn); } // Add .got. MIPS' .got is so different from the other archs, // it has its own class. if (Config->EMachine == EM_MIPS) { In::MipsGot = make>(); Symtab::X->Sections.push_back(In::MipsGot); } else { In::Got = make>(); Symtab::X->Sections.push_back(In::Got); } In::GotPlt = make>(); Symtab::X->Sections.push_back(In::GotPlt); In::IgotPlt = make>(); Symtab::X->Sections.push_back(In::IgotPlt); if (Config->GdbIndex) { In::GdbIndex = make>(); Symtab::X->Sections.push_back(In::GdbIndex); } // We always need to add rel[a].plt to output if it has entries. // Even for static linking it can contain R_[*]_IRELATIVE relocations. In::RelaPlt = make>( Config->Rela ? ".rela.plt" : ".rel.plt", false /*Sort*/); Symtab::X->Sections.push_back(In::RelaPlt); // The RelaIplt immediately follows .rel.plt (.rel.dyn for ARM) to ensure // that the IRelative relocations are processed last by the dynamic loader In::RelaIplt = make>( (Config->EMachine == EM_ARM) ? ".rel.dyn" : In::RelaPlt->Name, false /*Sort*/); Symtab::X->Sections.push_back(In::RelaIplt); In::Plt = make>(); Symtab::X->Sections.push_back(In::Plt); In::Iplt = make>(); Symtab::X->Sections.push_back(In::Iplt); if (Config->EhFrameHdr) { In::EhFrameHdr = make>(); Symtab::X->Sections.push_back(In::EhFrameHdr); } } template static bool shouldKeepInSymtab(InputSectionBase *Sec, StringRef SymName, const SymbolBody &B) { if (B.isFile()) return false; // We keep sections in symtab for relocatable output. if (B.isSection()) return Config->Relocatable; // If sym references a section in a discarded group, don't keep it. if (Sec == &InputSection::Discarded) return false; if (Config->Discard == DiscardPolicy::None) return true; // In ELF assembly .L symbols are normally discarded by the assembler. // If the assembler fails to do so, the linker discards them if // * --discard-locals is used. // * The symbol is in a SHF_MERGE section, which is normally the reason for // the assembler keeping the .L symbol. if (!SymName.startswith(".L") && !SymName.empty()) return true; if (Config->Discard == DiscardPolicy::Locals) return false; return !Sec || !(Sec->Flags & SHF_MERGE); } template static bool includeInSymtab(const SymbolBody &B) { if (!B.isLocal() && !B.symbol()->IsUsedInRegularObj) return false; // If --retain-symbols-file is given, we'll keep only symbols listed in that // file. if (Config->Discard == DiscardPolicy::RetainFile && !Config->RetainSymbolsFile.count(B.getName())) return false; if (auto *D = dyn_cast>(&B)) { // Always include absolute symbols. if (!D->Section) return true; // Exclude symbols pointing to garbage-collected sections. if (!D->Section->Live) return false; if (auto *S = dyn_cast>(D->Section)) if (!S->getSectionPiece(D->Value)->Live) return false; } return true; } // Local symbols are not in the linker's symbol table. This function scans // each object file's symbol table to copy local symbols to the output. template void Writer::copyLocalSymbols() { if (!In::SymTab) return; for (elf::ObjectFile *F : Symtab::X->getObjectFiles()) { for (SymbolBody *B : F->getLocalSymbols()) { if (!B->IsLocal) fatal(toString(F) + ": broken object: getLocalSymbols returns a non-local symbol"); auto *DR = dyn_cast>(B); // No reason to keep local undefined symbol in symtab. if (!DR) continue; if (!includeInSymtab(*B)) continue; InputSectionBase *Sec = DR->Section; if (!shouldKeepInSymtab(Sec, B->getName(), *B)) continue; In::SymTab->addLocal(B); } } } // PPC64 has a number of special SHT_PROGBITS+SHF_ALLOC+SHF_WRITE sections that // we would like to make sure appear is a specific order to maximize their // coverage by a single signed 16-bit offset from the TOC base pointer. // Conversely, the special .tocbss section should be first among all SHT_NOBITS // sections. This will put it next to the loaded special PPC64 sections (and, // thus, within reach of the TOC base pointer). static int getPPC64SectionRank(StringRef SectionName) { return StringSwitch(SectionName) .Case(".tocbss", 0) .Case(".branch_lt", 2) .Case(".toc", 3) .Case(".toc1", 4) .Case(".opd", 5) .Default(1); } // All sections with SHF_MIPS_GPREL flag should be grouped together // because data in these sections is addressable with a gp relative address. static int getMipsSectionRank(const OutputSectionBase *S) { if ((S->Flags & SHF_MIPS_GPREL) == 0) return 0; if (S->getName() == ".got") return 1; return 2; } template bool elf::isRelroSection(const OutputSectionBase *Sec) { if (!Config->ZRelro) return false; uint64_t Flags = Sec->Flags; if (!(Flags & SHF_ALLOC) || !(Flags & SHF_WRITE)) return false; if (Flags & SHF_TLS) return true; uint32_t Type = Sec->Type; if (Type == SHT_INIT_ARRAY || Type == SHT_FINI_ARRAY || Type == SHT_PREINIT_ARRAY) return true; if (Sec == In::GotPlt->OutSec) return Config->ZNow; if (Sec == In::Dynamic->OutSec) return true; if (In::Got && Sec == In::Got->OutSec) return true; if (Sec == Out::BssRelRo) return true; StringRef S = Sec->getName(); return S == ".data.rel.ro" || S == ".ctors" || S == ".dtors" || S == ".jcr" || S == ".eh_frame" || S == ".openbsd.randomdata"; } template static bool compareSectionsNonScript(const OutputSectionBase *A, const OutputSectionBase *B) { // Put .interp first because some loaders want to see that section // on the first page of the executable file when loaded into memory. bool AIsInterp = A->getName() == ".interp"; bool BIsInterp = B->getName() == ".interp"; if (AIsInterp != BIsInterp) return AIsInterp; // Allocatable sections go first to reduce the total PT_LOAD size and // so debug info doesn't change addresses in actual code. bool AIsAlloc = A->Flags & SHF_ALLOC; bool BIsAlloc = B->Flags & SHF_ALLOC; if (AIsAlloc != BIsAlloc) return AIsAlloc; // We don't have any special requirements for the relative order of two non // allocatable sections. if (!AIsAlloc) return false; // We want to put section specified by -T option first, so we // can start assigning VA starting from them later. auto AAddrSetI = Config->SectionStartMap.find(A->getName()); auto BAddrSetI = Config->SectionStartMap.find(B->getName()); bool AHasAddrSet = AAddrSetI != Config->SectionStartMap.end(); bool BHasAddrSet = BAddrSetI != Config->SectionStartMap.end(); if (AHasAddrSet != BHasAddrSet) return AHasAddrSet; if (AHasAddrSet) return AAddrSetI->second < BAddrSetI->second; // We want the read only sections first so that they go in the PT_LOAD // covering the program headers at the start of the file. bool AIsWritable = A->Flags & SHF_WRITE; bool BIsWritable = B->Flags & SHF_WRITE; if (AIsWritable != BIsWritable) return BIsWritable; if (!Config->SingleRoRx) { // For a corresponding reason, put non exec sections first (the program // header PT_LOAD is not executable). // We only do that if we are not using linker scripts, since with linker // scripts ro and rx sections are in the same PT_LOAD, so their relative // order is not important. The same applies for -no-rosegment. bool AIsExec = A->Flags & SHF_EXECINSTR; bool BIsExec = B->Flags & SHF_EXECINSTR; if (AIsExec != BIsExec) return BIsExec; } // If we got here we know that both A and B are in the same PT_LOAD. bool AIsTls = A->Flags & SHF_TLS; bool BIsTls = B->Flags & SHF_TLS; bool AIsNoBits = A->Type == SHT_NOBITS; bool BIsNoBits = B->Type == SHT_NOBITS; // The first requirement we have is to put (non-TLS) nobits sections last. The // reason is that the only thing the dynamic linker will see about them is a // p_memsz that is larger than p_filesz. Seeing that it zeros the end of the // PT_LOAD, so that has to correspond to the nobits sections. bool AIsNonTlsNoBits = AIsNoBits && !AIsTls; bool BIsNonTlsNoBits = BIsNoBits && !BIsTls; if (AIsNonTlsNoBits != BIsNonTlsNoBits) return BIsNonTlsNoBits; // We place nobits RelRo sections before plain r/w ones, and non-nobits RelRo // sections after r/w ones, so that the RelRo sections are contiguous. bool AIsRelRo = isRelroSection(A); bool BIsRelRo = isRelroSection(B); if (AIsRelRo != BIsRelRo) return AIsNonTlsNoBits ? AIsRelRo : BIsRelRo; // The TLS initialization block needs to be a single contiguous block in a R/W // PT_LOAD, so stick TLS sections directly before the other RelRo R/W // sections. The TLS NOBITS sections are placed here as they don't take up // virtual address space in the PT_LOAD. if (AIsTls != BIsTls) return AIsTls; // Within the TLS initialization block, the non-nobits sections need to appear // first. if (AIsNoBits != BIsNoBits) return BIsNoBits; // Some architectures have additional ordering restrictions for sections // within the same PT_LOAD. if (Config->EMachine == EM_PPC64) return getPPC64SectionRank(A->getName()) < getPPC64SectionRank(B->getName()); if (Config->EMachine == EM_MIPS) return getMipsSectionRank(A) < getMipsSectionRank(B); return false; } // Output section ordering is determined by this function. template static bool compareSections(const OutputSectionBase *A, const OutputSectionBase *B) { // For now, put sections mentioned in a linker script first. int AIndex = Script::X->getSectionIndex(A->getName()); int BIndex = Script::X->getSectionIndex(B->getName()); bool AInScript = AIndex != INT_MAX; bool BInScript = BIndex != INT_MAX; if (AInScript != BInScript) return AInScript; // If both are in the script, use that order. if (AInScript) return AIndex < BIndex; return compareSectionsNonScript(A, B); } // Program header entry PhdrEntry::PhdrEntry(unsigned Type, unsigned Flags) { p_type = Type; p_flags = Flags; } void PhdrEntry::add(OutputSectionBase *Sec) { Last = Sec; if (!First) First = Sec; p_align = std::max(p_align, Sec->Addralign); if (p_type == PT_LOAD) Sec->FirstInPtLoad = First; } template static void addOptionalSynthetic(StringRef Name, OutputSectionBase *Sec, typename ELFT::uint Val, uint8_t StOther = STV_HIDDEN) { if (SymbolBody *S = Symtab::X->find(Name)) if (!S->isInCurrentDSO()) Symtab::X->addSynthetic(Name, Sec, Val, StOther); } template static Symbol *addRegular(StringRef Name, InputSectionBase *Sec, typename ELFT::uint Value) { // The linker generated symbols are added as STB_WEAK to allow user defined // ones to override them. return Symtab::X->addRegular(Name, STV_HIDDEN, STT_NOTYPE, Value, /*Size=*/0, STB_WEAK, Sec, /*File=*/nullptr); } template static Symbol *addOptionalRegular(StringRef Name, InputSectionBase *IS, typename ELFT::uint Value) { SymbolBody *S = Symtab::X->find(Name); if (!S) return nullptr; if (S->isInCurrentDSO()) return S->symbol(); return addRegular(Name, IS, Value); } // The beginning and the ending of .rel[a].plt section are marked // with __rel[a]_iplt_{start,end} symbols if it is a statically linked // executable. The runtime needs these symbols in order to resolve // all IRELATIVE relocs on startup. For dynamic executables, we don't // need these symbols, since IRELATIVE relocs are resolved through GOT // and PLT. For details, see http://www.airs.com/blog/archives/403. template void Writer::addRelIpltSymbols() { if (In::DynSymTab) return; StringRef S = Config->Rela ? "__rela_iplt_start" : "__rel_iplt_start"; addOptionalRegular(S, In::RelaIplt, 0); S = Config->Rela ? "__rela_iplt_end" : "__rel_iplt_end"; addOptionalRegular(S, In::RelaIplt, -1); } // The linker is expected to define some symbols depending on // the linking result. This function defines such symbols. template void Writer::addReservedSymbols() { if (Config->EMachine == EM_MIPS) { // Define _gp for MIPS. st_value of _gp symbol will be updated by Writer // so that it points to an absolute address which by default is relative // to GOT. Default offset is 0x7ff0. // See "Global Data Symbols" in Chapter 6 in the following document: // ftp://www.linux-mips.org/pub/linux/mips/doc/ABI/mipsabi.pdf ElfSym::MipsGp = Symtab::X->addAbsolute("_gp", STV_HIDDEN, STB_LOCAL); // On MIPS O32 ABI, _gp_disp is a magic symbol designates offset between // start of function and 'gp' pointer into GOT. To simplify relocation // calculation we assign _gp value to it and calculate corresponding // relocations as relative to this value. if (Symtab::X->find("_gp_disp")) ElfSym::MipsGpDisp = Symtab::X->addAbsolute("_gp_disp", STV_HIDDEN, STB_LOCAL); // The __gnu_local_gp is a magic symbol equal to the current value of 'gp' // pointer. This symbol is used in the code generated by .cpload pseudo-op // in case of using -mno-shared option. // https://sourceware.org/ml/binutils/2004-12/msg00094.html if (Symtab::X->find("__gnu_local_gp")) ElfSym::MipsLocalGp = Symtab::X->addAbsolute("__gnu_local_gp", STV_HIDDEN, STB_LOCAL); } // In the assembly for 32 bit x86 the _GLOBAL_OFFSET_TABLE_ symbol // is magical and is used to produce a R_386_GOTPC relocation. // The R_386_GOTPC relocation value doesn't actually depend on the // symbol value, so it could use an index of STN_UNDEF which, according // to the spec, means the symbol value is 0. // Unfortunately both gas and MC keep the _GLOBAL_OFFSET_TABLE_ symbol in // the object file. // The situation is even stranger on x86_64 where the assembly doesn't // need the magical symbol, but gas still puts _GLOBAL_OFFSET_TABLE_ as // an undefined symbol in the .o files. // Given that the symbol is effectively unused, we just create a dummy // hidden one to avoid the undefined symbol error. Symtab::X->addIgnored("_GLOBAL_OFFSET_TABLE_"); // __tls_get_addr is defined by the dynamic linker for dynamic ELFs. For // static linking the linker is required to optimize away any references to // __tls_get_addr, so it's not defined anywhere. Create a hidden definition // to avoid the undefined symbol error. As usual special cases are ARM and // MIPS - the libc for these targets defines __tls_get_addr itself because // there are no TLS optimizations for these targets. if (!In::DynSymTab && (Config->EMachine != EM_MIPS && Config->EMachine != EM_ARM)) Symtab::X->addIgnored("__tls_get_addr"); // If linker script do layout we do not need to create any standart symbols. if (ScriptConfig->HasSections) return; ElfSym::EhdrStart = Symtab::X->addIgnored("__ehdr_start"); auto Define = [this](StringRef S, DefinedRegular *&Sym1, DefinedRegular *&Sym2) { Sym1 = Symtab::X->addIgnored(S, STV_DEFAULT); // The name without the underscore is not a reserved name, // so it is defined only when there is a reference against it. assert(S.startswith("_")); S = S.substr(1); if (SymbolBody *B = Symtab::X->find(S)) if (B->isUndefined()) Sym2 = Symtab::X->addAbsolute(S, STV_DEFAULT); }; Define("_end", ElfSym::End, ElfSym::End2); Define("_etext", ElfSym::Etext, ElfSym::Etext2); Define("_edata", ElfSym::Edata, ElfSym::Edata2); } // Sort input sections by section name suffixes for // __attribute__((init_priority(N))). template static void sortInitFini(OutputSectionBase *S) { if (S) reinterpret_cast *>(S)->sortInitFini(); } // Sort input sections by the special rule for .ctors and .dtors. template static void sortCtorsDtors(OutputSectionBase *S) { if (S) reinterpret_cast *>(S)->sortCtorsDtors(); } // Sort input sections using the list provided by --symbol-ordering-file. template static void sortBySymbolsOrder(ArrayRef OutputSections) { if (Config->SymbolOrderingFile.empty()) return; // Build a map from symbols to their priorities. Symbols that didn't // appear in the symbol ordering file have the lowest priority 0. // All explicitly mentioned symbols have negative (higher) priorities. DenseMap SymbolOrder; int Priority = -Config->SymbolOrderingFile.size(); for (StringRef S : Config->SymbolOrderingFile) SymbolOrder.insert({S, Priority++}); // Build a map from sections to their priorities. DenseMap *, int> SectionOrder; for (elf::ObjectFile *File : Symtab::X->getObjectFiles()) { for (SymbolBody *Body : File->getSymbols()) { auto *D = dyn_cast>(Body); if (!D || !D->Section) continue; int &Priority = SectionOrder[D->Section]; Priority = std::min(Priority, SymbolOrder.lookup(D->getName())); } } // Sort sections by priority. for (OutputSectionBase *Base : OutputSections) if (auto *Sec = dyn_cast>(Base)) Sec->sort([&](InputSection *S) { return SectionOrder.lookup(S); }); } template void Writer::forEachRelSec( std::function &)> Fn) { for (InputSectionBase *IS : Symtab::X->Sections) { if (!IS->Live) continue; // Scan all relocations. Each relocation goes through a series // of tests to determine if it needs special treatment, such as // creating GOT, PLT, copy relocations, etc. // Note that relocations for non-alloc sections are directly // processed by InputSection::relocateNonAlloc. if (!(IS->Flags & SHF_ALLOC)) continue; if (isa>(IS) || isa>(IS)) Fn(*IS); } } template void Writer::addInputSec(InputSectionBase *IS) { if (!IS) return; if (!IS->Live) { reportDiscarded(IS); return; } OutputSectionBase *Sec; bool IsNew; StringRef OutsecName = getOutputSectionName(IS->Name); std::tie(Sec, IsNew) = Factory.create(IS, OutsecName); if (IsNew) OutputSections.push_back(Sec); Sec->addSection(IS); } template void Writer::createSections() { for (InputSectionBase *IS : Symtab::X->Sections) addInputSec(IS); sortBySymbolsOrder(OutputSections); sortInitFini(findSection(".init_array")); sortInitFini(findSection(".fini_array")); sortCtorsDtors(findSection(".ctors")); sortCtorsDtors(findSection(".dtors")); for (OutputSectionBase *Sec : OutputSections) Sec->assignOffsets(); } template static bool canSharePtLoad(const OutputSectionBase &S1, const OutputSectionBase &S2) { if (!(S1.Flags & SHF_ALLOC) || !(S2.Flags & SHF_ALLOC)) return false; bool S1IsWrite = S1.Flags & SHF_WRITE; bool S2IsWrite = S2.Flags & SHF_WRITE; if (S1IsWrite != S2IsWrite) return false; if (!S1IsWrite) return true; // RO and RX share a PT_LOAD with linker scripts. return (S1.Flags & SHF_EXECINSTR) == (S2.Flags & SHF_EXECINSTR); } template void Writer::sortSections() { // Don't sort if using -r. It is not necessary and we want to preserve the // relative order for SHF_LINK_ORDER sections. if (Config->Relocatable) return; if (!ScriptConfig->HasSections) { std::stable_sort(OutputSections.begin(), OutputSections.end(), compareSectionsNonScript); return; } Script::X->adjustSectionsBeforeSorting(); // The order of the sections in the script is arbitrary and may not agree with // compareSectionsNonScript. This means that we cannot easily define a // strict weak ordering. To see why, consider a comparison of a section in the // script and one not in the script. We have a two simple options: // * Make them equivalent (a is not less than b, and b is not less than a). // The problem is then that equivalence has to be transitive and we can // have sections a, b and c with only b in a script and a less than c // which breaks this property. // * Use compareSectionsNonScript. Given that the script order doesn't have // to match, we can end up with sections a, b, c, d where b and c are in the // script and c is compareSectionsNonScript less than b. In which case d // can be equivalent to c, a to b and d < a. As a concrete example: // .a (rx) # not in script // .b (rx) # in script // .c (ro) # in script // .d (ro) # not in script // // The way we define an order then is: // * First put script sections at the start and sort the script and // non-script sections independently. // * Move each non-script section to its preferred position. We try // to put each section in the last position where it it can share // a PT_LOAD. std::stable_sort(OutputSections.begin(), OutputSections.end(), compareSections); auto I = OutputSections.begin(); auto E = OutputSections.end(); auto NonScriptI = std::find_if(OutputSections.begin(), E, [](OutputSectionBase *S) { return Script::X->getSectionIndex(S->getName()) == INT_MAX; }); while (NonScriptI != E) { auto BestPos = std::max_element( I, NonScriptI, [&](OutputSectionBase *&A, OutputSectionBase *&B) { bool ACanSharePtLoad = canSharePtLoad(**NonScriptI, *A); bool BCanSharePtLoad = canSharePtLoad(**NonScriptI, *B); if (ACanSharePtLoad != BCanSharePtLoad) return BCanSharePtLoad; bool ACmp = compareSectionsNonScript(*NonScriptI, A); bool BCmp = compareSectionsNonScript(*NonScriptI, B); if (ACmp != BCmp) return BCmp; // FIXME: missing test size_t PosA = &A - &OutputSections[0]; size_t PosB = &B - &OutputSections[0]; return ACmp ? PosA > PosB : PosA < PosB; }); // max_element only returns NonScriptI if the range is empty. If the range // is not empty we should consider moving the the element forward one // position. if (BestPos != NonScriptI && !compareSectionsNonScript(*NonScriptI, *BestPos)) ++BestPos; std::rotate(BestPos, NonScriptI, NonScriptI + 1); ++NonScriptI; } Script::X->adjustSectionsAfterSorting(); } template static void finalizeSynthetic(const std::vector *> &Sections) { for (SyntheticSection *SS : Sections) if (SS && SS->OutSec && !SS->empty()) { SS->finalize(); SS->OutSec->Size = 0; SS->OutSec->assignOffsets(); } } // We need to add input synthetic sections early in createSyntheticSections() // to make them visible from linkescript side. But not all sections are always // required to be in output. For example we don't need dynamic section content // sometimes. This function filters out such unused sections from output. template static void removeUnusedSyntheticSections(std::vector &V) { // Input synthetic sections are placed after all regular ones. We iterate over // them all and exit at first non-synthetic. for (InputSectionBase *S : llvm::reverse(Symtab::X->Sections)) { SyntheticSection *SS = dyn_cast>(S); if (!SS) return; if (!SS->empty() || !SS->OutSec) continue; OutputSection *OutSec = cast>(SS->OutSec); OutSec->Sections.erase( std::find(OutSec->Sections.begin(), OutSec->Sections.end(), SS)); // If there is no other sections in output section, remove it from output. if (OutSec->Sections.empty()) V.erase(std::find(V.begin(), V.end(), OutSec)); } } // Create output section objects and add them to OutputSections. template void Writer::finalizeSections() { Out::DebugInfo = findSection(".debug_info"); Out::PreinitArray = findSection(".preinit_array"); Out::InitArray = findSection(".init_array"); Out::FiniArray = findSection(".fini_array"); // The linker needs to define SECNAME_start, SECNAME_end and SECNAME_stop // symbols for sections, so that the runtime can get the start and end // addresses of each section by section name. Add such symbols. if (!Config->Relocatable) { addStartEndSymbols(); for (OutputSectionBase *Sec : OutputSections) addStartStopSymbols(Sec); } // Add _DYNAMIC symbol. Unlike GNU gold, our _DYNAMIC symbol has no type. // It should be okay as no one seems to care about the type. // Even the author of gold doesn't remember why gold behaves that way. // https://sourceware.org/ml/binutils/2002-03/msg00360.html if (In::DynSymTab) addRegular("_DYNAMIC", In::Dynamic, 0); // Define __rel[a]_iplt_{start,end} symbols if needed. addRelIpltSymbols(); if (!Out::EhFrame->empty()) { OutputSections.push_back(Out::EhFrame); Out::EhFrame->finalize(); } // Scan relocations. This must be done after every symbol is declared so that // we can correctly decide if a dynamic relocation is needed. forEachRelSec(scanRelocations); // Now that we have defined all possible symbols including linker- // synthesized ones. Visit all symbols to give the finishing touches. for (Symbol *S : Symtab::X->getSymbols()) { SymbolBody *Body = S->body(); if (!includeInSymtab(*Body)) continue; if (In::SymTab) In::SymTab->addGlobal(Body); if (In::DynSymTab && S->includeInDynsym()) { In::DynSymTab->addGlobal(Body); if (auto *SS = dyn_cast>(Body)) if (SS->file()->isNeeded()) In::VerNeed->addSymbol(SS); } } // Do not proceed if there was an undefined symbol. if (ErrorCount) return; // So far we have added sections from input object files. // This function adds linker-created Out::* sections. addPredefinedSections(); removeUnusedSyntheticSections(OutputSections); sortSections(); unsigned I = 1; for (OutputSectionBase *Sec : OutputSections) { Sec->SectionIndex = I++; Sec->ShName = In::ShStrTab->addString(Sec->getName()); } // Binary and relocatable output does not have PHDRS. // The headers have to be created before finalize as that can influence the // image base and the dynamic section on mips includes the image base. if (!Config->Relocatable && !Config->OFormatBinary) { Phdrs = Script::X->hasPhdrsCommands() ? Script::X->createPhdrs() : createPhdrs(); addPtArmExid(Phdrs); fixHeaders(); } // Fill other section headers. The dynamic table is finalized // at the end because some tags like RELSZ depend on result // of finalizing other sections. for (OutputSectionBase *Sec : OutputSections) Sec->finalize(); // Dynamic section must be the last one in this list and dynamic // symbol table section (DynSymTab) must be the first one. finalizeSynthetic( {In::DynSymTab, In::GnuHashTab, In::HashTab, In::SymTab, In::ShStrTab, In::StrTab, In::VerDef, In::DynStrTab, In::GdbIndex, In::Got, In::MipsGot, In::IgotPlt, In::GotPlt, In::RelaDyn, In::RelaIplt, In::RelaPlt, In::Plt, In::Iplt, In::Plt, In::EhFrameHdr, In::VerSym, In::VerNeed, In::Dynamic}); } template void Writer::addPredefinedSections() { if (Out::Bss->Size > 0) OutputSections.push_back(Out::Bss); if (Out::BssRelRo->Size > 0) OutputSections.push_back(Out::BssRelRo); auto OS = dyn_cast_or_null>(findSection(".ARM.exidx")); if (OS && !OS->Sections.empty() && !Config->Relocatable) OS->addSection(make>()); addInputSec(In::SymTab); addInputSec(In::ShStrTab); addInputSec(In::StrTab); } // The linker is expected to define SECNAME_start and SECNAME_end // symbols for a few sections. This function defines them. template void Writer::addStartEndSymbols() { auto Define = [&](StringRef Start, StringRef End, OutputSectionBase *OS) { // These symbols resolve to the image base if the section does not exist. // A special value -1 indicates end of the section. addOptionalSynthetic(Start, OS, 0); addOptionalSynthetic(End, OS, OS ? -1 : 0); }; Define("__preinit_array_start", "__preinit_array_end", Out::PreinitArray); Define("__init_array_start", "__init_array_end", Out::InitArray); Define("__fini_array_start", "__fini_array_end", Out::FiniArray); if (OutputSectionBase *Sec = findSection(".ARM.exidx")) Define("__exidx_start", "__exidx_end", Sec); } // If a section name is valid as a C identifier (which is rare because of // the leading '.'), linkers are expected to define __start_ and // __stop_ symbols. They are at beginning and end of the section, // respectively. This is not requested by the ELF standard, but GNU ld and // gold provide the feature, and used by many programs. template void Writer::addStartStopSymbols(OutputSectionBase *Sec) { StringRef S = Sec->getName(); if (!isValidCIdentifier(S)) return; addOptionalSynthetic(Saver.save("__start_" + S), Sec, 0, STV_DEFAULT); addOptionalSynthetic(Saver.save("__stop_" + S), Sec, -1, STV_DEFAULT); } template OutputSectionBase *Writer::findSection(StringRef Name) { for (OutputSectionBase *Sec : OutputSections) if (Sec->getName() == Name) return Sec; return nullptr; } template static bool needsPtLoad(OutputSectionBase *Sec) { if (!(Sec->Flags & SHF_ALLOC)) return false; // Don't allocate VA space for TLS NOBITS sections. The PT_TLS PHDR is // responsible for allocating space for them, not the PT_LOAD that // contains the TLS initialization image. if (Sec->Flags & SHF_TLS && Sec->Type == SHT_NOBITS) return false; return true; } // Linker scripts are responsible for aligning addresses. Unfortunately, most // linker scripts are designed for creating two PT_LOADs only, one RX and one // RW. This means that there is no alignment in the RO to RX transition and we // cannot create a PT_LOAD there. template static typename ELFT::uint computeFlags(typename ELFT::uint F) { if (Config->OMagic) return PF_R | PF_W | PF_X; if (Config->SingleRoRx && !(F & PF_W)) return F | PF_X; return F; } // Decide which program headers to create and which sections to include in each // one. template std::vector Writer::createPhdrs() { std::vector Ret; auto AddHdr = [&](unsigned Type, unsigned Flags) -> PhdrEntry * { Ret.emplace_back(Type, Flags); return &Ret.back(); }; // The first phdr entry is PT_PHDR which describes the program header itself. PhdrEntry &Hdr = *AddHdr(PT_PHDR, PF_R); Hdr.add(Out::ProgramHeaders); // PT_INTERP must be the second entry if exists. if (OutputSectionBase *Sec = findSection(".interp")) { PhdrEntry &Hdr = *AddHdr(PT_INTERP, Sec->getPhdrFlags()); Hdr.add(Sec); } // Add the first PT_LOAD segment for regular output sections. uintX_t Flags = computeFlags(PF_R); PhdrEntry *Load = AddHdr(PT_LOAD, Flags); PhdrEntry TlsHdr(PT_TLS, PF_R); PhdrEntry RelRo(PT_GNU_RELRO, PF_R); PhdrEntry Note(PT_NOTE, PF_R); for (OutputSectionBase *Sec : OutputSections) { if (!(Sec->Flags & SHF_ALLOC)) break; // If we meet TLS section then we create TLS header // and put all TLS sections inside for further use when // assign addresses. if (Sec->Flags & SHF_TLS) TlsHdr.add(Sec); if (!needsPtLoad(Sec)) continue; // Segments are contiguous memory regions that has the same attributes // (e.g. executable or writable). There is one phdr for each segment. // Therefore, we need to create a new phdr when the next section has // different flags or is loaded at a discontiguous address using AT linker // script command. uintX_t NewFlags = computeFlags(Sec->getPhdrFlags()); if (Script::X->hasLMA(Sec->getName()) || Flags != NewFlags) { Load = AddHdr(PT_LOAD, NewFlags); Flags = NewFlags; } Load->add(Sec); if (isRelroSection(Sec)) RelRo.add(Sec); if (Sec->Type == SHT_NOTE) Note.add(Sec); } // Add the TLS segment unless it's empty. if (TlsHdr.First) Ret.push_back(std::move(TlsHdr)); // Add an entry for .dynamic. if (In::DynSymTab) { PhdrEntry &H = *AddHdr(PT_DYNAMIC, In::Dynamic->OutSec->getPhdrFlags()); H.add(In::Dynamic->OutSec); } // PT_GNU_RELRO includes all sections that should be marked as // read-only by dynamic linker after proccessing relocations. if (RelRo.First) Ret.push_back(std::move(RelRo)); // PT_GNU_EH_FRAME is a special section pointing on .eh_frame_hdr. if (!Out::EhFrame->empty() && In::EhFrameHdr) { PhdrEntry &Hdr = *AddHdr(PT_GNU_EH_FRAME, In::EhFrameHdr->OutSec->getPhdrFlags()); Hdr.add(In::EhFrameHdr->OutSec); } // PT_OPENBSD_RANDOMIZE specifies the location and size of a part of the // memory image of the program that must be filled with random data before any // code in the object is executed. if (OutputSectionBase *Sec = findSection(".openbsd.randomdata")) { PhdrEntry &Hdr = *AddHdr(PT_OPENBSD_RANDOMIZE, Sec->getPhdrFlags()); Hdr.add(Sec); } // PT_GNU_STACK is a special section to tell the loader to make the // pages for the stack non-executable. if (!Config->ZExecstack) { PhdrEntry &Hdr = *AddHdr(PT_GNU_STACK, PF_R | PF_W); if (Config->ZStackSize != uint64_t(-1)) Hdr.p_memsz = Config->ZStackSize; } // PT_OPENBSD_WXNEEDED is a OpenBSD-specific header to mark the executable // is expected to perform W^X violations, such as calling mprotect(2) or // mmap(2) with PROT_WRITE | PROT_EXEC, which is prohibited by default on // OpenBSD. if (Config->ZWxneeded) AddHdr(PT_OPENBSD_WXNEEDED, PF_X); if (Note.First) Ret.push_back(std::move(Note)); return Ret; } template void Writer::addPtArmExid(std::vector &Phdrs) { if (Config->EMachine != EM_ARM) return; auto I = std::find_if( OutputSections.begin(), OutputSections.end(), [](OutputSectionBase *Sec) { return Sec->Type == SHT_ARM_EXIDX; }); if (I == OutputSections.end()) return; // PT_ARM_EXIDX is the ARM EHABI equivalent of PT_GNU_EH_FRAME PhdrEntry ARMExidx(PT_ARM_EXIDX, PF_R); ARMExidx.add(*I); Phdrs.push_back(ARMExidx); } // The first section of each PT_LOAD, the first section in PT_GNU_RELRO and the // first section after PT_GNU_RELRO have to be page aligned so that the dynamic // linker can set the permissions. template void Writer::fixSectionAlignments() { for (const PhdrEntry &P : Phdrs) if (P.p_type == PT_LOAD && P.First) P.First->PageAlign = true; for (const PhdrEntry &P : Phdrs) { if (P.p_type != PT_GNU_RELRO) continue; if (P.First) P.First->PageAlign = true; // Find the first section after PT_GNU_RELRO. If it is in a PT_LOAD we // have to align it to a page. auto End = OutputSections.end(); auto I = std::find(OutputSections.begin(), End, P.Last); if (I == End || (I + 1) == End) continue; OutputSectionBase *Sec = *(I + 1); if (needsPtLoad(Sec)) Sec->PageAlign = true; } } template void elf::allocateHeaders(MutableArrayRef Phdrs, ArrayRef OutputSections) { auto FirstPTLoad = std::find_if(Phdrs.begin(), Phdrs.end(), [](const PhdrEntry &E) { return E.p_type == PT_LOAD; }); if (FirstPTLoad == Phdrs.end()) return; if (FirstPTLoad->First) for (OutputSectionBase *Sec : OutputSections) if (Sec->FirstInPtLoad == FirstPTLoad->First) Sec->FirstInPtLoad = Out::ElfHeader; FirstPTLoad->First = Out::ElfHeader; if (!FirstPTLoad->Last) FirstPTLoad->Last = Out::ProgramHeaders; } // We should set file offsets and VAs for elf header and program headers // sections. These are special, we do not include them into output sections // list, but have them to simplify the code. template void Writer::fixHeaders() { Out::ProgramHeaders->Size = sizeof(Elf_Phdr) * Phdrs.size(); // If the script has SECTIONS, assignAddresses will compute the values. if (ScriptConfig->HasSections) return; uintX_t HeaderSize = getHeaderSize(); // When -T
option is specified, lower the base to make room for those // sections. if (!Config->SectionStartMap.empty()) { uint64_t Min = -1; for (const auto &P : Config->SectionStartMap) Min = std::min(Min, P.second); if (HeaderSize < Min) Min -= HeaderSize; else AllocateHeader = false; if (Min < Config->ImageBase) Config->ImageBase = alignDown(Min, Config->MaxPageSize); } if (AllocateHeader) allocateHeaders(Phdrs, OutputSections); uintX_t BaseVA = Config->ImageBase; Out::ElfHeader->Addr = BaseVA; Out::ProgramHeaders->Addr = BaseVA + Out::ElfHeader->Size; } // Assign VAs (addresses at run-time) to output sections. template void Writer::assignAddresses() { uintX_t VA = Config->ImageBase; if (AllocateHeader) VA += getHeaderSize(); uintX_t ThreadBssOffset = 0; for (OutputSectionBase *Sec : OutputSections) { uintX_t Alignment = Sec->Addralign; if (Sec->PageAlign) Alignment = std::max(Alignment, Config->MaxPageSize); auto I = Config->SectionStartMap.find(Sec->getName()); if (I != Config->SectionStartMap.end()) VA = I->second; // We only assign VAs to allocated sections. if (needsPtLoad(Sec)) { VA = alignTo(VA, Alignment); Sec->Addr = VA; VA += Sec->Size; } else if (Sec->Flags & SHF_TLS && Sec->Type == SHT_NOBITS) { uintX_t TVA = VA + ThreadBssOffset; TVA = alignTo(TVA, Alignment); Sec->Addr = TVA; ThreadBssOffset = TVA - VA + Sec->Size; } } } // Adjusts the file alignment for a given output section and returns // its new file offset. The file offset must be the same with its // virtual address (modulo the page size) so that the loader can load // executables without any address adjustment. template static uintX_t getFileAlignment(uintX_t Off, OutputSectionBase *Sec) { OutputSectionBase *First = Sec->FirstInPtLoad; // If the section is not in a PT_LOAD, we just have to align it. if (!First) return alignTo(Off, Sec->Addralign); // The first section in a PT_LOAD has to have congruent offset and address // module the page size. if (Sec == First) return alignTo(Off, Config->MaxPageSize, Sec->Addr); // If two sections share the same PT_LOAD the file offset is calculated // using this formula: Off2 = Off1 + (VA2 - VA1). return First->Offset + Sec->Addr - First->Addr; } template void setOffset(OutputSectionBase *Sec, uintX_t &Off) { if (Sec->Type == SHT_NOBITS) { Sec->Offset = Off; return; } Off = getFileAlignment(Off, Sec); Sec->Offset = Off; Off += Sec->Size; } template void Writer::assignFileOffsetsBinary() { uintX_t Off = 0; for (OutputSectionBase *Sec : OutputSections) if (Sec->Flags & SHF_ALLOC) setOffset(Sec, Off); FileSize = alignTo(Off, sizeof(uintX_t)); } // Assign file offsets to output sections. template void Writer::assignFileOffsets() { uintX_t Off = 0; setOffset(Out::ElfHeader, Off); setOffset(Out::ProgramHeaders, Off); for (OutputSectionBase *Sec : OutputSections) setOffset(Sec, Off); SectionHeaderOff = alignTo(Off, sizeof(uintX_t)); FileSize = SectionHeaderOff + (OutputSections.size() + 1) * sizeof(Elf_Shdr); } // Finalize the program headers. We call this function after we assign // file offsets and VAs to all sections. template void Writer::setPhdrs() { for (PhdrEntry &P : Phdrs) { OutputSectionBase *First = P.First; OutputSectionBase *Last = P.Last; if (First) { P.p_filesz = Last->Offset - First->Offset; if (Last->Type != SHT_NOBITS) P.p_filesz += Last->Size; P.p_memsz = Last->Addr + Last->Size - First->Addr; P.p_offset = First->Offset; P.p_vaddr = First->Addr; if (!P.HasLMA) P.p_paddr = First->getLMA(); } if (P.p_type == PT_LOAD) P.p_align = Config->MaxPageSize; else if (P.p_type == PT_GNU_RELRO) P.p_align = 1; // The TLS pointer goes after PT_TLS. At least glibc will align it, // so round up the size to make sure the offsets are correct. if (P.p_type == PT_TLS) { Out::TlsPhdr = &P; if (P.p_memsz) P.p_memsz = alignTo(P.p_memsz, P.p_align); } } } // The entry point address is chosen in the following ways. // // 1. the '-e' entry command-line option; // 2. the ENTRY(symbol) command in a linker control script; // 3. the value of the symbol start, if present; // 4. the address of the first byte of the .text section, if present; // 5. the address 0. template typename ELFT::uint Writer::getEntryAddr() { // Case 1, 2 or 3. As a special case, if the symbol is actually // a number, we'll use that number as an address. if (SymbolBody *B = Symtab::X->find(Config->Entry)) return B->getVA(); uint64_t Addr; if (!Config->Entry.getAsInteger(0, Addr)) return Addr; // Case 4 if (OutputSectionBase *Sec = findSection(".text")) { if (Config->WarnMissingEntry) warn("cannot find entry symbol " + Config->Entry + "; defaulting to 0x" + utohexstr(Sec->Addr)); return Sec->Addr; } // Case 5 if (Config->WarnMissingEntry) warn("cannot find entry symbol " + Config->Entry + "; not setting start address"); return 0; } template static uint8_t getELFEncoding() { if (ELFT::TargetEndianness == llvm::support::little) return ELFDATA2LSB; return ELFDATA2MSB; } static uint16_t getELFType() { if (Config->Pic) return ET_DYN; if (Config->Relocatable) return ET_REL; return ET_EXEC; } // This function is called after we have assigned address and size // to each section. This function fixes some predefined absolute // symbol values that depend on section address and size. template void Writer::fixAbsoluteSymbols() { // __ehdr_start is the location of program headers. if (ElfSym::EhdrStart) ElfSym::EhdrStart->Value = Out::ProgramHeaders->Addr; auto Set = [](DefinedRegular *S1, DefinedRegular *S2, uintX_t V) { if (S1) S1->Value = V; if (S2) S2->Value = V; }; // _etext is the first location after the last read-only loadable segment. // _edata is the first location after the last read-write loadable segment. // _end is the first location after the uninitialized data region. for (PhdrEntry &P : Phdrs) { if (P.p_type != PT_LOAD) continue; Set(ElfSym::End, ElfSym::End2, P.p_vaddr + P.p_memsz); uintX_t Val = P.p_vaddr + P.p_filesz; if (P.p_flags & PF_W) Set(ElfSym::Edata, ElfSym::Edata2, Val); else Set(ElfSym::Etext, ElfSym::Etext2, Val); } // Setup MIPS _gp_disp/__gnu_local_gp symbols which should // be equal to the _gp symbol's value. if (Config->EMachine == EM_MIPS) { if (!ElfSym::MipsGp->Value) { // Find GP-relative section with the lowest address // and use this address to calculate default _gp value. uintX_t Gp = -1; for (const OutputSectionBase * OS : OutputSections) if ((OS->Flags & SHF_MIPS_GPREL) && OS->Addr < Gp) Gp = OS->Addr; if (Gp != (uintX_t)-1) ElfSym::MipsGp->Value = Gp + 0x7ff0; } if (ElfSym::MipsGpDisp) ElfSym::MipsGpDisp->Value = ElfSym::MipsGp->Value; if (ElfSym::MipsLocalGp) ElfSym::MipsLocalGp->Value = ElfSym::MipsGp->Value; } } template void Writer::writeHeader() { uint8_t *Buf = Buffer->getBufferStart(); memcpy(Buf, "\177ELF", 4); // Write the ELF header. auto *EHdr = reinterpret_cast(Buf); EHdr->e_ident[EI_CLASS] = ELFT::Is64Bits ? ELFCLASS64 : ELFCLASS32; EHdr->e_ident[EI_DATA] = getELFEncoding(); EHdr->e_ident[EI_VERSION] = EV_CURRENT; EHdr->e_ident[EI_OSABI] = Config->OSABI; EHdr->e_type = getELFType(); EHdr->e_machine = Config->EMachine; EHdr->e_version = EV_CURRENT; EHdr->e_entry = getEntryAddr(); EHdr->e_shoff = SectionHeaderOff; EHdr->e_ehsize = sizeof(Elf_Ehdr); EHdr->e_phnum = Phdrs.size(); EHdr->e_shentsize = sizeof(Elf_Shdr); EHdr->e_shnum = OutputSections.size() + 1; EHdr->e_shstrndx = In::ShStrTab->OutSec->SectionIndex; if (Config->EMachine == EM_ARM) // We don't currently use any features incompatible with EF_ARM_EABI_VER5, // but we don't have any firm guarantees of conformance. Linux AArch64 // kernels (as of 2016) require an EABI version to be set. EHdr->e_flags = EF_ARM_EABI_VER5; else if (Config->EMachine == EM_MIPS) EHdr->e_flags = getMipsEFlags(); if (!Config->Relocatable) { EHdr->e_phoff = sizeof(Elf_Ehdr); EHdr->e_phentsize = sizeof(Elf_Phdr); } // Write the program header table. auto *HBuf = reinterpret_cast(Buf + EHdr->e_phoff); for (PhdrEntry &P : Phdrs) { HBuf->p_type = P.p_type; HBuf->p_flags = P.p_flags; HBuf->p_offset = P.p_offset; HBuf->p_vaddr = P.p_vaddr; HBuf->p_paddr = P.p_paddr; HBuf->p_filesz = P.p_filesz; HBuf->p_memsz = P.p_memsz; HBuf->p_align = P.p_align; ++HBuf; } // Write the section header table. Note that the first table entry is null. auto *SHdrs = reinterpret_cast(Buf + EHdr->e_shoff); for (OutputSectionBase *Sec : OutputSections) Sec->writeHeaderTo(++SHdrs); } // Removes a given file asynchronously. This is a performance hack, // so remove this when operating systems are improved. // // On Linux (and probably on other Unix-like systems), unlink(2) is a // noticeably slow system call. As of 2016, unlink takes 250 // milliseconds to remove a 1 GB file on ext4 filesystem on my machine. // // To create a new result file, we first remove existing file. So, if // you repeatedly link a 1 GB program in a regular compile-link-debug // cycle, every cycle wastes 250 milliseconds only to remove a file. // Since LLD can link a 1 GB binary in about 5 seconds, that waste // actually counts. // // This function spawns a background thread to call unlink. // The calling thread returns almost immediately. static void unlinkAsync(StringRef Path) { if (!Config->Threads || !sys::fs::exists(Config->OutputFile)) return; // First, rename Path to avoid race condition. We cannot remove // Path from a different thread because we are now going to create // Path as a new file. If we do that in a different thread, the new // thread can remove the new file. SmallString<128> TempPath; if (sys::fs::createUniqueFile(Path + "tmp%%%%%%%%", TempPath)) return; if (sys::fs::rename(Path, TempPath)) { sys::fs::remove(TempPath); return; } // Remove TempPath in background. std::thread([=] { ::remove(TempPath.str().str().c_str()); }).detach(); } // Open a result file. template void Writer::openFile() { unlinkAsync(Config->OutputFile); ErrorOr> BufferOrErr = FileOutputBuffer::create(Config->OutputFile, FileSize, FileOutputBuffer::F_executable); if (auto EC = BufferOrErr.getError()) error(EC, "failed to open " + Config->OutputFile); else Buffer = std::move(*BufferOrErr); } template void Writer::writeSectionsBinary() { uint8_t *Buf = Buffer->getBufferStart(); for (OutputSectionBase *Sec : OutputSections) if (Sec->Flags & SHF_ALLOC) Sec->writeTo(Buf + Sec->Offset); } // Write section contents to a mmap'ed file. template void Writer::writeSections() { uint8_t *Buf = Buffer->getBufferStart(); // PPC64 needs to process relocations in the .opd section // before processing relocations in code-containing sections. Out::Opd = findSection(".opd"); if (Out::Opd) { Out::OpdBuf = Buf + Out::Opd->Offset; Out::Opd->writeTo(Buf + Out::Opd->Offset); } OutputSectionBase *EhFrameHdr = In::EhFrameHdr ? In::EhFrameHdr->OutSec : nullptr; for (OutputSectionBase *Sec : OutputSections) if (Sec != Out::Opd && Sec != EhFrameHdr) Sec->writeTo(Buf + Sec->Offset); // The .eh_frame_hdr depends on .eh_frame section contents, therefore // it should be written after .eh_frame is written. if (!Out::EhFrame->empty() && EhFrameHdr) EhFrameHdr->writeTo(Buf + EhFrameHdr->Offset); } template void Writer::writeBuildId() { if (!In::BuildId || !In::BuildId->OutSec) return; // Compute a hash of all sections of the output file. uint8_t *Start = Buffer->getBufferStart(); uint8_t *End = Start + FileSize; In::BuildId->writeBuildId({Start, End}); } template void elf::writeResult(); template void elf::writeResult(); template void elf::writeResult(); template void elf::writeResult(); template void elf::allocateHeaders(MutableArrayRef, ArrayRef); template void elf::allocateHeaders(MutableArrayRef, ArrayRef); template void elf::allocateHeaders(MutableArrayRef, ArrayRef); template void elf::allocateHeaders(MutableArrayRef, ArrayRef); template bool elf::isRelroSection(const OutputSectionBase *); template bool elf::isRelroSection(const OutputSectionBase *); template bool elf::isRelroSection(const OutputSectionBase *); template bool elf::isRelroSection(const OutputSectionBase *); template void elf::reportDiscarded(InputSectionBase *); template void elf::reportDiscarded(InputSectionBase *); template void elf::reportDiscarded(InputSectionBase *); template void elf::reportDiscarded(InputSectionBase *);