//===- 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 "OutputSections.h" #include "Relocations.h" #include "Strings.h" #include "SymbolTable.h" #include "Target.h" #include "llvm/ADT/StringMap.h" #include "llvm/ADT/StringSwitch.h" #include "llvm/Support/FileOutputBuffer.h" #include "llvm/Support/StringSaver.h" #include "llvm/Support/raw_ostream.h" using namespace llvm; using namespace llvm::ELF; using namespace llvm::object; 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; Writer(SymbolTable &S) : Symtab(S) {} void run(); private: // This describes a program header entry. // Each contains type, access flags and range of output sections that will be // placed in it. struct Phdr { Phdr(unsigned Type, unsigned Flags) { H.p_type = Type; H.p_flags = Flags; } Elf_Phdr H = {}; OutputSectionBase *First = nullptr; OutputSectionBase *Last = nullptr; }; void copyLocalSymbols(); void addReservedSymbols(); void createSections(); void addPredefinedSections(); bool needsGot(); void createPhdrs(); void assignAddresses(); void assignFileOffsets(); void setPhdrs(); void fixHeaders(); void fixSectionAlignments(); void fixAbsoluteSymbols(); void openFile(); void writeHeader(); void writeSections(); void writeBuildId(); bool needsInterpSection() const { return !Symtab.getSharedFiles().empty() && !Config->DynamicLinker.empty(); } bool isOutputDynamic() const { return !Symtab.getSharedFiles().empty() || Config->Pic; } void addCommonSymbols(std::vector &Syms); std::unique_ptr Buffer; BumpPtrAllocator Alloc; std::vector *> OutputSections; std::vector>> OwningSections; void addRelIpltSymbols(); void addStartEndSymbols(); void addStartStopSymbols(OutputSectionBase *Sec); SymbolTable &Symtab; std::vector Phdrs; uintX_t FileSize; uintX_t SectionHeaderOff; }; } // anonymous namespace template StringRef elf::getOutputSectionName(InputSectionBase *S) { StringRef Dest = Script::X->getOutputSection(S); if (!Dest.empty()) return Dest; StringRef Name = S->getSectionName(); for (StringRef V : {".text.", ".rodata.", ".data.rel.ro.", ".data.", ".bss.", ".init_array.", ".fini_array.", ".ctors.", ".dtors.", ".tbss.", ".gcc_except_table.", ".tdata."}) if (Name.startswith(V)) return V.drop_back(); return Name; } template void elf::reportDiscarded(InputSectionBase *IS, const std::unique_ptr> &File) { if (!Config->PrintGcSections || !IS || IS->Live) return; errs() << "removing unused section from '" << IS->getSectionName() << "' in file '" << File->getName() << "'\n"; } template void elf::writeResult(SymbolTable *Symtab) { typedef typename ELFT::uint uintX_t; typedef typename ELFT::Ehdr Elf_Ehdr; // Create singleton output sections. OutputSection Bss(".bss", SHT_NOBITS, SHF_ALLOC | SHF_WRITE); DynamicSection Dynamic; EhOutputSection EhFrame; GotSection Got; InterpSection Interp; PltSection Plt; RelocationSection RelaDyn(Config->Rela ? ".rela.dyn" : ".rel.dyn", Config->ZCombreloc); StringTableSection DynStrTab(".dynstr", true); StringTableSection ShStrTab(".shstrtab", false); SymbolTableSection DynSymTab(DynStrTab); VersionTableSection VerSym; VersionNeedSection VerNeed; OutputSectionBase ElfHeader("", 0, SHF_ALLOC); ElfHeader.setSize(sizeof(Elf_Ehdr)); OutputSectionBase ProgramHeaders("", 0, SHF_ALLOC); ProgramHeaders.updateAlignment(sizeof(uintX_t)); // Instantiate optional output sections if they are needed. std::unique_ptr> BuildId; std::unique_ptr> EhFrameHdr; std::unique_ptr> GnuHashTab; std::unique_ptr> GotPlt; std::unique_ptr> HashTab; std::unique_ptr> RelaPlt; std::unique_ptr> StrTab; std::unique_ptr> SymTabSec; std::unique_ptr> MipsRldMap; std::unique_ptr> VerDef; if (Config->BuildId == BuildIdKind::Fnv1) BuildId.reset(new BuildIdFnv1); else if (Config->BuildId == BuildIdKind::Md5) BuildId.reset(new BuildIdMd5); else if (Config->BuildId == BuildIdKind::Sha1) BuildId.reset(new BuildIdSha1); else if (Config->BuildId == BuildIdKind::Hexstring) BuildId.reset(new BuildIdHexstring); if (Config->EhFrameHdr) EhFrameHdr.reset(new EhFrameHeader); if (Config->GnuHash) GnuHashTab.reset(new GnuHashTableSection); if (Config->SysvHash) HashTab.reset(new HashTableSection); StringRef S = Config->Rela ? ".rela.plt" : ".rel.plt"; GotPlt.reset(new GotPltSection); RelaPlt.reset(new RelocationSection(S, false /*Sort*/)); if (!Config->StripAll) { StrTab.reset(new StringTableSection(".strtab", false)); SymTabSec.reset(new SymbolTableSection(*StrTab)); } if (Config->EMachine == EM_MIPS && !Config->Shared) { // This is a MIPS specific section to hold a space within the data segment // of executable file which is pointed to by the DT_MIPS_RLD_MAP entry. // See "Dynamic section" in Chapter 5 in the following document: // ftp://www.linux-mips.org/pub/linux/mips/doc/ABI/mipsabi.pdf MipsRldMap.reset(new OutputSection(".rld_map", SHT_PROGBITS, SHF_ALLOC | SHF_WRITE)); MipsRldMap->setSize(sizeof(uintX_t)); MipsRldMap->updateAlignment(sizeof(uintX_t)); } if (!Config->VersionDefinitions.empty()) VerDef.reset(new VersionDefinitionSection()); Out::Bss = &Bss; Out::BuildId = BuildId.get(); Out::DynStrTab = &DynStrTab; Out::DynSymTab = &DynSymTab; Out::Dynamic = &Dynamic; Out::EhFrame = &EhFrame; Out::EhFrameHdr = EhFrameHdr.get(); Out::GnuHashTab = GnuHashTab.get(); Out::Got = &Got; Out::GotPlt = GotPlt.get(); Out::HashTab = HashTab.get(); Out::Interp = &Interp; Out::Plt = &Plt; Out::RelaDyn = &RelaDyn; Out::RelaPlt = RelaPlt.get(); Out::ShStrTab = &ShStrTab; Out::StrTab = StrTab.get(); Out::SymTab = SymTabSec.get(); Out::VerDef = VerDef.get(); Out::VerSym = &VerSym; Out::VerNeed = &VerNeed; Out::MipsRldMap = MipsRldMap.get(); Out::Opd = nullptr; Out::OpdBuf = nullptr; Out::TlsPhdr = nullptr; Out::ElfHeader = &ElfHeader; Out::ProgramHeaders = &ProgramHeaders; Writer(*Symtab).run(); } // The main function of the writer. template void Writer::run() { if (!Config->DiscardAll) copyLocalSymbols(); addReservedSymbols(); createSections(); if (HasError) return; if (Config->Relocatable) { assignFileOffsets(); } else { createPhdrs(); fixHeaders(); if (ScriptConfig->DoLayout) { Script::X->assignAddresses(OutputSections); } else { fixSectionAlignments(); assignAddresses(); } assignFileOffsets(); setPhdrs(); fixAbsoluteSymbols(); } openFile(); if (HasError) return; writeHeader(); writeSections(); writeBuildId(); if (HasError) return; if (auto EC = Buffer->commit()) error(EC, "failed to write to the output file"); } template static void reportUndefined(SymbolTable &Symtab, SymbolBody *Sym) { if (Config->UnresolvedSymbols == UnresolvedPolicy::Ignore) return; if (Config->Shared && Sym->symbol()->Visibility == STV_DEFAULT && Config->UnresolvedSymbols != UnresolvedPolicy::NoUndef) return; std::string Msg = "undefined symbol: " + Sym->getName().str(); if (Sym->File) Msg += " in " + getFilename(Sym->File); if (Config->UnresolvedSymbols == UnresolvedPolicy::Warn) warning(Msg); else error(Msg); } 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->DiscardNone) 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->DiscardLocals) return false; return !(Sec->getSectionHdr()->sh_flags & SHF_MERGE); } template static bool includeInSymtab(const SymbolBody &B) { if (!B.isLocal() && !B.symbol()->IsUsedInRegularObj) 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 (!Out::SymTab) return; for (const std::unique_ptr> &F : Symtab.getObjectFiles()) { const char *StrTab = F->getStringTable().data(); for (SymbolBody *B : F->getLocalSymbols()) { auto *DR = dyn_cast>(B); // No reason to keep local undefined symbol in symtab. if (!DR) continue; if (!includeInSymtab(*B)) continue; StringRef SymName(StrTab + B->getNameOffset()); InputSectionBase *Sec = DR->Section; if (!shouldKeepInSymtab(Sec, SymName, *B)) continue; ++Out::SymTab->NumLocals; if (Config->Relocatable) B->DynsymIndex = Out::SymTab->NumLocals; F->KeptLocalSyms.push_back( std::make_pair(DR, Out::SymTab->StrTabSec.addString(SymName))); } } } // 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); } template static bool isRelroSection(OutputSectionBase *Sec) { if (!Config->ZRelro) return false; typename ELFT::uint Flags = Sec->getFlags(); if (!(Flags & SHF_ALLOC) || !(Flags & SHF_WRITE)) return false; if (Flags & SHF_TLS) return true; uint32_t Type = Sec->getType(); if (Type == SHT_INIT_ARRAY || Type == SHT_FINI_ARRAY || Type == SHT_PREINIT_ARRAY) return true; if (Sec == Out::GotPlt) return Config->ZNow; if (Sec == Out::Dynamic || Sec == Out::Got) return true; StringRef S = Sec->getName(); return S == ".data.rel.ro" || S == ".ctors" || S == ".dtors" || S == ".jcr" || S == ".eh_frame"; } // Output section ordering is determined by this function. template static bool compareSections(OutputSectionBase *A, OutputSectionBase *B) { typedef typename ELFT::uint uintX_t; int Comp = Script::X->compareSections(A->getName(), B->getName()); if (Comp != 0) return Comp < 0; uintX_t AFlags = A->getFlags(); uintX_t BFlags = B->getFlags(); // Allocatable sections go first to reduce the total PT_LOAD size and // so debug info doesn't change addresses in actual code. bool AIsAlloc = AFlags & SHF_ALLOC; bool BIsAlloc = BFlags & 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 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 = AFlags & SHF_WRITE; bool BIsWritable = BFlags & SHF_WRITE; if (AIsWritable != BIsWritable) return BIsWritable; // For a corresponding reason, put non exec sections first (the program // header PT_LOAD is not executable). bool AIsExec = AFlags & SHF_EXECINSTR; bool BIsExec = BFlags & SHF_EXECINSTR; if (AIsExec != BIsExec) return BIsExec; // If we got here we know that both A and B are in the same PT_LOAD. // The TLS initialization block needs to be a single contiguous block in a R/W // PT_LOAD, so stick TLS sections directly before R/W sections. The TLS NOBITS // sections are placed here as they don't take up virtual address space in the // PT_LOAD. bool AIsTls = AFlags & SHF_TLS; bool BIsTls = BFlags & SHF_TLS; if (AIsTls != BIsTls) return AIsTls; // The next requirement we have is to put 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 AIsNoBits = A->getType() == SHT_NOBITS; bool BIsNoBits = B->getType() == SHT_NOBITS; if (AIsNoBits != BIsNoBits) return BIsNoBits; // We place RelRo section before plain r/w ones. bool AIsRelRo = isRelroSection(A); bool BIsRelRo = isRelroSection(B); if (AIsRelRo != BIsRelRo) return AIsRelRo; // 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()); return false; } // Until this function is called, common symbols do not belong to any section. // This function adds them to end of BSS section. template void Writer::addCommonSymbols(std::vector &Syms) { if (Syms.empty()) return; // Sort the common symbols by alignment as an heuristic to pack them better. std::stable_sort(Syms.begin(), Syms.end(), [](const DefinedCommon *A, const DefinedCommon *B) { return A->Alignment > B->Alignment; }); uintX_t Off = Out::Bss->getSize(); for (DefinedCommon *C : Syms) { Off = alignTo(Off, C->Alignment); Out::Bss->updateAlignment(C->Alignment); C->OffsetInBss = Off; Off += C->Size; } Out::Bss->setSize(Off); } template static Symbol *addOptionalSynthetic(SymbolTable &Table, StringRef Name, OutputSectionBase *Sec, typename ELFT::uint Val) { SymbolBody *S = Table.find(Name); if (!S) return nullptr; if (!S->isUndefined() && !S->isShared()) return S->symbol(); return Table.addSynthetic(Name, Sec, Val); } // 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 (isOutputDynamic() || !Out::RelaPlt) return; StringRef S = Config->Rela ? "__rela_iplt_start" : "__rel_iplt_start"; addOptionalSynthetic(Symtab, S, Out::RelaPlt, 0); S = Config->Rela ? "__rela_iplt_end" : "__rel_iplt_end"; addOptionalSynthetic(Symtab, S, Out::RelaPlt, DefinedSynthetic::SectionEnd); } // 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 is relative to GOT. // See "Global Data Symbols" in Chapter 6 in the following document: // ftp://www.linux-mips.org/pub/linux/mips/doc/ABI/mipsabi.pdf Symtab.addSynthetic("_gp", Out::Got, MipsGPOffset); // On MIPS O32 ABI, _gp_disp is a magic symbol designates offset between // start of function and 'gp' pointer into GOT. Symbol *Sym = addOptionalSynthetic(Symtab, "_gp_disp", Out::Got, MipsGPOffset); if (Sym) ElfSym::MipsGpDisp = Sym->body(); // 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 addOptionalSynthetic(Symtab, "__gnu_local_gp", Out::Got, MipsGPOffset); } // 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. if (!Config->Relocatable) Symtab.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. if (!isOutputDynamic()) Symtab.addIgnored("__tls_get_addr"); auto Define = [this](StringRef S, DefinedRegular *&Sym1, DefinedRegular *&Sym2) { Sym1 = Symtab.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.find(S)) if (B->isUndefined()) Sym2 = Symtab.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(); } // Create output section objects and add them to OutputSections. template void Writer::createSections() { // Create output sections for input object file sections. std::vector *> RegularSections; OutputSectionFactory Factory; for (const std::unique_ptr> &F : Symtab.getObjectFiles()) { for (InputSectionBase *C : F->getSections()) { if (isDiscarded(C)) { reportDiscarded(C, F); continue; } OutputSectionBase *Sec; bool IsNew; std::tie(Sec, IsNew) = Factory.create(C, getOutputSectionName(C)); if (IsNew) { OwningSections.emplace_back(Sec); OutputSections.push_back(Sec); RegularSections.push_back(Sec); } Sec->addSection(C); } } // If we have a .opd section (used under PPC64 for function descriptors), // store a pointer to it here so that we can use it later when processing // relocations. Out::Opd = Factory.lookup(".opd", SHT_PROGBITS, SHF_WRITE | SHF_ALLOC); Out::Dynamic->PreInitArraySec = Factory.lookup( ".preinit_array", SHT_PREINIT_ARRAY, SHF_WRITE | SHF_ALLOC); Out::Dynamic->InitArraySec = Factory.lookup(".init_array", SHT_INIT_ARRAY, SHF_WRITE | SHF_ALLOC); Out::Dynamic->FiniArraySec = Factory.lookup(".fini_array", SHT_FINI_ARRAY, SHF_WRITE | SHF_ALLOC); // Sort section contents for __attribute__((init_priority(N)). sortInitFini(Out::Dynamic->InitArraySec); sortInitFini(Out::Dynamic->FiniArraySec); sortCtorsDtors(Factory.lookup(".ctors", SHT_PROGBITS, SHF_WRITE | SHF_ALLOC)); sortCtorsDtors(Factory.lookup(".dtors", SHT_PROGBITS, SHF_WRITE | SHF_ALLOC)); // 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 : RegularSections) 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 (isOutputDynamic()) Symtab.addSynthetic("_DYNAMIC", Out::Dynamic, 0); // Define __rel[a]_iplt_{start,end} symbols if needed. addRelIpltSymbols(); // Add scripted symbols with zero values now. // Real values will be assigned later Script::X->addScriptedSymbols(); 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. for (const std::unique_ptr> &F : Symtab.getObjectFiles()) { for (InputSectionBase *C : F->getSections()) { if (isDiscarded(C)) continue; if (auto *S = dyn_cast>(C)) { scanRelocations(*S); continue; } if (auto *S = dyn_cast>(C)) if (S->RelocSection) scanRelocations(*S, *S->RelocSection); } } for (OutputSectionBase *Sec : OutputSections) Sec->assignOffsets(); // Now that we have defined all possible symbols including linker- // synthesized ones. Visit all symbols to give the finishing touches. std::vector CommonSymbols; for (Symbol *S : Symtab.getSymbols()) { SymbolBody *Body = S->body(); // We only report undefined symbols in regular objects. This means that we // will accept an undefined reference in bitcode if it can be optimized out. if (S->IsUsedInRegularObj && Body->isUndefined() && !S->isWeak()) reportUndefined(Symtab, Body); if (auto *C = dyn_cast(Body)) CommonSymbols.push_back(C); if (!includeInSymtab(*Body)) continue; if (Out::SymTab) Out::SymTab->addSymbol(Body); if (isOutputDynamic() && S->includeInDynsym()) { Out::DynSymTab->addSymbol(Body); if (auto *SS = dyn_cast>(Body)) if (SS->file()->isNeeded()) Out::VerNeed->addSymbol(SS); } } // Do not proceed if there was an undefined symbol. if (HasError) return; addCommonSymbols(CommonSymbols); // So far we have added sections from input object files. // This function adds linker-created Out::* sections. addPredefinedSections(); std::stable_sort(OutputSections.begin(), OutputSections.end(), compareSections); unsigned I = 1; for (OutputSectionBase *Sec : OutputSections) { Sec->SectionIndex = I++; Sec->setSHName(Out::ShStrTab->addString(Sec->getName())); } // Finalizers fix each section's size. // .dynsym is finalized early since that may fill up .gnu.hash. if (isOutputDynamic()) Out::DynSymTab->finalize(); // Fill other section headers. The dynamic table is finalized // at the end because some tags like RELSZ depend on result // of finalizing other sections. The dynamic string table is // finalized once the .dynamic finalizer has added a few last // strings. See DynamicSection::finalize() for (OutputSectionBase *Sec : OutputSections) if (Sec != Out::DynStrTab && Sec != Out::Dynamic) Sec->finalize(); if (isOutputDynamic()) Out::Dynamic->finalize(); // Now that all output offsets are fixed. Finalize mergeable sections // to fix their maps from input offsets to output offsets. for (OutputSectionBase *Sec : OutputSections) Sec->finalizePieces(); } template bool Writer::needsGot() { if (!Out::Got->empty()) return true; // We add the .got section to the result for dynamic MIPS target because // its address and properties are mentioned in the .dynamic section. if (Config->EMachine == EM_MIPS) return true; // If we have a relocation that is relative to GOT (such as GOTOFFREL), // we need to emit a GOT even if it's empty. return Out::Got->HasGotOffRel; } // This function add Out::* sections to OutputSections. template void Writer::addPredefinedSections() { auto Add = [&](OutputSectionBase *C) { if (C) OutputSections.push_back(C); }; // A core file does not usually contain unmodified segments except // the first page of the executable. Add the build ID section to beginning of // the file so that the section is included in the first page. if (Out::BuildId) OutputSections.insert(OutputSections.begin(), Out::BuildId); // Add .interp at first because some loaders want to see that section // on the first page of the executable file when loaded into memory. if (needsInterpSection()) OutputSections.insert(OutputSections.begin(), Out::Interp); // This order is not the same as the final output order // because we sort the sections using their attributes below. Add(Out::SymTab); Add(Out::ShStrTab); Add(Out::StrTab); if (isOutputDynamic()) { Add(Out::DynSymTab); bool HasVerNeed = Out::VerNeed->getNeedNum() != 0; if (Out::VerDef || HasVerNeed) Add(Out::VerSym); Add(Out::VerDef); if (HasVerNeed) Add(Out::VerNeed); Add(Out::GnuHashTab); Add(Out::HashTab); Add(Out::Dynamic); Add(Out::DynStrTab); if (Out::RelaDyn->hasRelocs()) Add(Out::RelaDyn); Add(Out::MipsRldMap); } // We always need to add rel[a].plt to output if it has entries. // Even during static linking it can contain R_[*]_IRELATIVE relocations. if (Out::RelaPlt && Out::RelaPlt->hasRelocs()) { Add(Out::RelaPlt); Out::RelaPlt->Static = !isOutputDynamic(); } if (needsGot()) Add(Out::Got); if (Out::GotPlt && !Out::GotPlt->empty()) Add(Out::GotPlt); if (!Out::Plt->empty()) Add(Out::Plt); if (!Out::EhFrame->empty()) Add(Out::EhFrameHdr); if (Out::Bss->getSize() > 0) Add(Out::Bss); } // 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) { if (OS) { this->Symtab.addSynthetic(Start, OS, 0); this->Symtab.addSynthetic(End, OS, DefinedSynthetic::SectionEnd); } else { addOptionalSynthetic(this->Symtab, Start, (OutputSectionBase *)nullptr, 0); addOptionalSynthetic(this->Symtab, End, (OutputSectionBase *)nullptr, 0); } }; Define("__preinit_array_start", "__preinit_array_end", Out::Dynamic->PreInitArraySec); Define("__init_array_start", "__init_array_end", Out::Dynamic->InitArraySec); Define("__fini_array_start", "__fini_array_end", Out::Dynamic->FiniArraySec); } // 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; StringSaver Saver(Alloc); StringRef Start = Saver.save("__start_" + S); StringRef Stop = Saver.save("__stop_" + S); if (SymbolBody *B = Symtab.find(Start)) if (B->isUndefined()) Symtab.addSynthetic(Start, Sec, 0); if (SymbolBody *B = Symtab.find(Stop)) if (B->isUndefined()) Symtab.addSynthetic(Stop, Sec, DefinedSynthetic::SectionEnd); } template static bool needsPtLoad(OutputSectionBase *Sec) { if (!(Sec->getFlags() & 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->getFlags() & SHF_TLS && Sec->getType() == SHT_NOBITS) return false; return true; } static uint32_t toPhdrFlags(uint64_t Flags) { uint32_t Ret = PF_R; if (Flags & SHF_WRITE) Ret |= PF_W; if (Flags & SHF_EXECINSTR) Ret |= PF_X; return Ret; } // Decide which program headers to create and which sections to include in each // one. template void Writer::createPhdrs() { auto AddHdr = [this](unsigned Type, unsigned Flags) { return &*Phdrs.emplace(Phdrs.end(), Type, Flags); }; auto AddSec = [](Phdr &Hdr, OutputSectionBase *Sec) { Hdr.Last = Sec; if (!Hdr.First) Hdr.First = Sec; Hdr.H.p_align = std::max(Hdr.H.p_align, Sec->getAlignment()); }; // The first phdr entry is PT_PHDR which describes the program header itself. Phdr &Hdr = *AddHdr(PT_PHDR, PF_R); AddSec(Hdr, Out::ProgramHeaders); // PT_INTERP must be the second entry if exists. if (needsInterpSection()) { Phdr &Hdr = *AddHdr(PT_INTERP, toPhdrFlags(Out::Interp->getFlags())); AddSec(Hdr, Out::Interp); } // Add the first PT_LOAD segment for regular output sections. uintX_t Flags = PF_R; Phdr *Load = AddHdr(PT_LOAD, Flags); AddSec(*Load, Out::ElfHeader); AddSec(*Load, Out::ProgramHeaders); Phdr TlsHdr(PT_TLS, PF_R); Phdr RelRo(PT_GNU_RELRO, PF_R); Phdr Note(PT_NOTE, PF_R); for (OutputSectionBase *Sec : OutputSections) { if (!(Sec->getFlags() & SHF_ALLOC)) break; // If we meet TLS section then we create TLS header // and put all TLS sections inside for futher use when // assign addresses. if (Sec->getFlags() & SHF_TLS) AddSec(TlsHdr, Sec); if (!needsPtLoad(Sec)) continue; // If flags changed then we want new load segment. uintX_t NewFlags = toPhdrFlags(Sec->getFlags()); if (Flags != NewFlags) { Load = AddHdr(PT_LOAD, NewFlags); Flags = NewFlags; } AddSec(*Load, Sec); if (isRelroSection(Sec)) AddSec(RelRo, Sec); if (Sec->getType() == SHT_NOTE) AddSec(Note, Sec); } // Add the TLS segment unless it's empty. if (TlsHdr.First) Phdrs.push_back(std::move(TlsHdr)); // Add an entry for .dynamic. if (isOutputDynamic()) { Phdr &H = *AddHdr(PT_DYNAMIC, toPhdrFlags(Out::Dynamic->getFlags())); AddSec(H, Out::Dynamic); } // PT_GNU_RELRO includes all sections that should be marked as // read-only by dynamic linker after proccessing relocations. if (RelRo.First) Phdrs.push_back(std::move(RelRo)); // PT_GNU_EH_FRAME is a special section pointing on .eh_frame_hdr. if (!Out::EhFrame->empty() && Out::EhFrameHdr) { Phdr &Hdr = *AddHdr(PT_GNU_EH_FRAME, toPhdrFlags(Out::EhFrameHdr->getFlags())); AddSec(Hdr, Out::EhFrameHdr); } // PT_GNU_STACK is a special section to tell the loader to make the // pages for the stack non-executable. if (!Config->ZExecStack) AddHdr(PT_GNU_STACK, PF_R | PF_W); if (Note.First) Phdrs.push_back(std::move(Note)); Out::ProgramHeaders->setSize(sizeof(Elf_Phdr) * Phdrs.size()); } // The first section of each PT_LOAD 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 Phdr &P : Phdrs) if (P.H.p_type == PT_LOAD) P.First->PageAlign = true; for (const Phdr &P : Phdrs) { if (P.H.p_type != PT_GNU_RELRO) continue; // 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; } } // 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() { uintX_t BaseVA = ScriptConfig->DoLayout ? 0 : Config->ImageBase; Out::ElfHeader->setVA(BaseVA); uintX_t Off = Out::ElfHeader->getSize(); Out::ProgramHeaders->setVA(Off + BaseVA); } // Assign VAs (addresses at run-time) to output sections. template void Writer::assignAddresses() { uintX_t VA = Config->ImageBase + Out::ElfHeader->getSize() + Out::ProgramHeaders->getSize(); uintX_t ThreadBssOffset = 0; for (OutputSectionBase *Sec : OutputSections) { uintX_t Alignment = Sec->getAlignment(); if (Sec->PageAlign) Alignment = std::max(Alignment, Target->PageSize); // We only assign VAs to allocated sections. if (needsPtLoad(Sec)) { VA = alignTo(VA, Alignment); Sec->setVA(VA); VA += Sec->getSize(); } else if (Sec->getFlags() & SHF_TLS && Sec->getType() == SHT_NOBITS) { uintX_t TVA = VA + ThreadBssOffset; TVA = alignTo(TVA, Alignment); Sec->setVA(TVA); ThreadBssOffset = TVA - VA + Sec->getSize(); } } } // 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) { uintX_t Alignment = Sec->getAlignment(); if (Sec->PageAlign) Alignment = std::max(Alignment, Target->PageSize); Off = alignTo(Off, Alignment); // Relocatable output does not have program headers // and does not need any other offset adjusting. if (Config->Relocatable || !(Sec->getFlags() & SHF_ALLOC)) return Off; return alignTo(Off, Target->PageSize, Sec->getVA()); } // Assign file offsets to output sections. template void Writer::assignFileOffsets() { uintX_t Off = 0; auto Set = [&](OutputSectionBase *Sec) { if (Sec->getType() == SHT_NOBITS) { Sec->setFileOffset(Off); return; } Off = getFileAlignment(Off, Sec); Sec->setFileOffset(Off); Off += Sec->getSize(); }; Set(Out::ElfHeader); Set(Out::ProgramHeaders); for (OutputSectionBase *Sec : OutputSections) Set(Sec); 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 (Phdr &P : Phdrs) { Elf_Phdr &H = P.H; OutputSectionBase *First = P.First; OutputSectionBase *Last = P.Last; if (First) { H.p_filesz = Last->getFileOff() - First->getFileOff(); if (Last->getType() != SHT_NOBITS) H.p_filesz += Last->getSize(); H.p_memsz = Last->getVA() + Last->getSize() - First->getVA(); H.p_offset = First->getFileOff(); H.p_vaddr = First->getVA(); } if (H.p_type == PT_LOAD) H.p_align = Target->PageSize; else if (H.p_type == PT_GNU_RELRO) H.p_align = 1; H.p_paddr = H.p_vaddr; // 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 (H.p_type == PT_TLS) { Out::TlsPhdr = &H; H.p_memsz = alignTo(H.p_memsz, H.p_align); } } } static uint32_t getMipsEFlags(bool Is64Bits) { // FIXME: In fact ELF flags depends on ELF flags of input object files // and selected emulation. For now just use hard coded values. if (Is64Bits) return EF_MIPS_CPIC | EF_MIPS_PIC | EF_MIPS_ARCH_64R2; uint32_t V = EF_MIPS_CPIC | EF_MIPS_ABI_O32 | EF_MIPS_ARCH_32R2; if (Config->Shared) V |= EF_MIPS_PIC; return V; } template static typename ELFT::uint getEntryAddr() { if (Symbol *S = Config->EntrySym) return S->body()->getVA(); if (Config->EntryAddr != uint64_t(-1)) return Config->EntryAddr; 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() { 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 (Phdr &P : Phdrs) { Elf_Phdr &H = P.H; if (H.p_type != PT_LOAD) continue; Set(ElfSym::End, ElfSym::End2, H.p_vaddr + H.p_memsz); uintX_t Val = H.p_vaddr + H.p_filesz; if (H.p_flags & PF_W) Set(ElfSym::Edata, ElfSym::Edata2, Val); else Set(ElfSym::Etext, ElfSym::Etext2, Val); } } template void Writer::writeHeader() { uint8_t *Buf = Buffer->getBufferStart(); memcpy(Buf, "\177ELF", 4); auto &FirstObj = cast>(*Config->FirstElf); // 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] = FirstObj.getOSABI(); EHdr->e_type = getELFType(); EHdr->e_machine = FirstObj.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 = Out::ShStrTab->SectionIndex; if (Config->EMachine == EM_MIPS) EHdr->e_flags = getMipsEFlags(ELFT::Is64Bits); 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 (Phdr &P : Phdrs) *HBuf++ = P.H; // 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); } template void Writer::openFile() { 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); } // 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. if (OutputSectionBase *Sec = Out::Opd) { Out::OpdBuf = Buf + Sec->getFileOff(); Sec->writeTo(Buf + Sec->getFileOff()); } for (OutputSectionBase *Sec : OutputSections) if (Sec != Out::Opd) Sec->writeTo(Buf + Sec->getFileOff()); } template void Writer::writeBuildId() { if (!Out::BuildId) return; // Compute a hash of all sections except .debug_* sections. // We skip debug sections because they tend to be very large // and their contents are very likely to be the same as long as // other sections are the same. uint8_t *Start = Buffer->getBufferStart(); uint8_t *Last = Start; std::vector> Regions; for (OutputSectionBase *Sec : OutputSections) { uint8_t *End = Start + Sec->getFileOff(); if (!Sec->getName().startswith(".debug_")) Regions.push_back({Last, End}); Last = End; } Regions.push_back({Last, Start + FileSize}); Out::BuildId->writeBuildId(Regions); } template void elf::writeResult(SymbolTable *Symtab); template void elf::writeResult(SymbolTable *Symtab); template void elf::writeResult(SymbolTable *Symtab); template void elf::writeResult(SymbolTable *Symtab);