//===- InputSection.cpp ---------------------------------------------------===// // // The LLVM Linker // // This file is distributed under the University of Illinois Open Source // License. See LICENSE.TXT for details. // //===----------------------------------------------------------------------===// #include "InputSection.h" #include "Config.h" #include "EhFrame.h" #include "Error.h" #include "InputFiles.h" #include "LinkerScript.h" #include "OutputSections.h" #include "Target.h" #include "Thunks.h" #include "llvm/Support/Compression.h" #include "llvm/Support/Endian.h" using namespace llvm; using namespace llvm::ELF; using namespace llvm::object; using namespace llvm::support::endian; using namespace lld; using namespace lld::elf; template bool elf::isDiscarded(InputSectionBase *S) { return !S || S == &InputSection::Discarded || !S->Live || Script::X->isDiscarded(S); } template InputSectionBase::InputSectionBase(elf::ObjectFile *File, const Elf_Shdr *Header, Kind SectionKind) : Header(Header), File(File), SectionKind(SectionKind), Repl(this), Compressed(Header->sh_flags & SHF_COMPRESSED) { // The garbage collector sets sections' Live bits. // If GC is disabled, all sections are considered live by default. Live = !Config->GcSections; // The ELF spec states that a value of 0 means the section has // no alignment constraits. Alignment = std::max(Header->sh_addralign, 1); } template size_t InputSectionBase::getSize() const { if (auto *D = dyn_cast>(this)) if (D->getThunksSize() > 0) return D->getThunkOff() + D->getThunksSize(); return Header->sh_size; } template StringRef InputSectionBase::getSectionName() const { return check(File->getObj().getSectionName(this->Header)); } template ArrayRef InputSectionBase::getSectionData() const { if (Compressed) return ArrayRef((const uint8_t *)Uncompressed.data(), Uncompressed.size()); return check(this->File->getObj().getSectionContents(this->Header)); } template typename ELFT::uint InputSectionBase::getOffset(uintX_t Offset) const { switch (SectionKind) { case Regular: return cast>(this)->OutSecOff + Offset; case EHFrame: return cast>(this)->getOffset(Offset); case Merge: return cast>(this)->getOffset(Offset); case MipsReginfo: case MipsOptions: // MIPS .reginfo and .MIPS.options sections are consumed by the linker, // and the linker produces a single output section. It is possible that // input files contain section symbol points to the corresponding input // section. Redirect it to the produced output section. if (Offset != 0) fatal("Unsupported reference to the middle of '" + getSectionName() + "' section"); return this->OutSec->getVA(); } llvm_unreachable("invalid section kind"); } template void InputSectionBase::uncompress() { if (!zlib::isAvailable()) fatal("build lld with zlib to enable compressed sections support"); // A compressed section consists of a header of Elf_Chdr type // followed by compressed data. ArrayRef Data = check(this->File->getObj().getSectionContents(this->Header)); if (Data.size() < sizeof(Elf_Chdr)) fatal("corrupt compressed section"); auto *Hdr = reinterpret_cast(Data.data()); Data = Data.slice(sizeof(Elf_Chdr)); if (Hdr->ch_type != ELFCOMPRESS_ZLIB) fatal("unsupported compression type"); StringRef Buf((const char *)Data.data(), Data.size()); if (zlib::uncompress(Buf, Uncompressed, Hdr->ch_size) != zlib::StatusOK) fatal("error uncompressing section"); } template typename ELFT::uint InputSectionBase::getOffset(const DefinedRegular &Sym) const { return getOffset(Sym.Value); } template InputSection::InputSection(elf::ObjectFile *F, const Elf_Shdr *Header) : InputSectionBase(F, Header, Base::Regular) {} template bool InputSection::classof(const InputSectionBase *S) { return S->SectionKind == Base::Regular; } template InputSectionBase *InputSection::getRelocatedSection() { assert(this->Header->sh_type == SHT_RELA || this->Header->sh_type == SHT_REL); ArrayRef *> Sections = this->File->getSections(); return Sections[this->Header->sh_info]; } template void InputSection::addThunk(const Thunk *T) { Thunks.push_back(T); } template uint64_t InputSection::getThunkOff() const { return this->Header->sh_size; } template uint64_t InputSection::getThunksSize() const { uint64_t Total = 0; for (const Thunk *T : Thunks) Total += T->size(); return Total; } // This is used for -r. We can't use memcpy to copy relocations because we need // to update symbol table offset and section index for each relocation. So we // copy relocations one by one. template template void InputSection::copyRelocations(uint8_t *Buf, ArrayRef Rels) { InputSectionBase *RelocatedSection = getRelocatedSection(); for (const RelTy &Rel : Rels) { uint32_t Type = Rel.getType(Config->Mips64EL); SymbolBody &Body = this->File->getRelocTargetSym(Rel); RelTy *P = reinterpret_cast(Buf); Buf += sizeof(RelTy); P->r_offset = RelocatedSection->getOffset(Rel.r_offset); P->setSymbolAndType(Body.DynsymIndex, Type, Config->Mips64EL); } } // Page(Expr) is the page address of the expression Expr, defined // as (Expr & ~0xFFF). (This applies even if the machine page size // supported by the platform has a different value.) static uint64_t getAArch64Page(uint64_t Expr) { return Expr & (~static_cast(0xFFF)); } template static typename ELFT::uint getSymVA(uint32_t Type, typename ELFT::uint A, typename ELFT::uint P, const SymbolBody &Body, RelExpr Expr) { typedef typename ELFT::uint uintX_t; switch (Expr) { case R_HINT: llvm_unreachable("cannot relocate hint relocs"); case R_TLSLD: return Out::Got->getTlsIndexOff() + A - Out::Got->getNumEntries() * sizeof(uintX_t); case R_TLSLD_PC: return Out::Got->getTlsIndexVA() + A - P; case R_THUNK_ABS: return Body.getThunkVA() + A; case R_THUNK_PC: case R_THUNK_PLT_PC: return Body.getThunkVA() + A - P; case R_PPC_TOC: return getPPC64TocBase() + A; case R_TLSGD: return Out::Got->getGlobalDynOffset(Body) + A - Out::Got->getNumEntries() * sizeof(uintX_t); case R_TLSGD_PC: return Out::Got->getGlobalDynAddr(Body) + A - P; case R_TLSDESC: return Out::Got->getGlobalDynAddr(Body) + A; case R_TLSDESC_PAGE: return getAArch64Page(Out::Got->getGlobalDynAddr(Body) + A) - getAArch64Page(P); case R_PLT: return Body.getPltVA() + A; case R_PLT_PC: case R_PPC_PLT_OPD: return Body.getPltVA() + A - P; case R_SIZE: return Body.getSize() + A; case R_GOTREL: return Body.getVA(A) - Out::Got->getVA(); case R_RELAX_TLS_GD_TO_IE_END: case R_GOT_FROM_END: return Body.getGotOffset() + A - Out::Got->getNumEntries() * sizeof(uintX_t); case R_RELAX_TLS_GD_TO_IE_ABS: case R_GOT: return Body.getGotVA() + A; case R_RELAX_TLS_GD_TO_IE_PAGE_PC: case R_GOT_PAGE_PC: return getAArch64Page(Body.getGotVA() + A) - getAArch64Page(P); case R_RELAX_TLS_GD_TO_IE: case R_GOT_PC: return Body.getGotVA() + A - P; case R_GOTONLY_PC: return Out::Got->getVA() + A - P; case R_RELAX_TLS_LD_TO_LE: case R_RELAX_TLS_IE_TO_LE: case R_RELAX_TLS_GD_TO_LE: case R_TLS: if (Target->TcbSize) return Body.getVA(A) + alignTo(Target->TcbSize, Out::TlsPhdr->p_align); return Body.getVA(A) - Out::TlsPhdr->p_memsz; case R_RELAX_TLS_GD_TO_LE_NEG: case R_NEG_TLS: return Out::TlsPhdr->p_memsz - Body.getVA(A); case R_ABS: case R_RELAX_GOT_PC_NOPIC: return Body.getVA(A); case R_GOT_OFF: return Body.getGotOffset() + A; case R_MIPS_GOT_LOCAL_PAGE: // If relocation against MIPS local symbol requires GOT entry, this entry // should be initialized by 'page address'. This address is high 16-bits // of sum the symbol's value and the addend. return Out::Got->getMipsLocalPageOffset(Body.getVA(A)); case R_MIPS_GOT_OFF: // In case of MIPS if a GOT relocation has non-zero addend this addend // should be applied to the GOT entry content not to the GOT entry offset. // That is why we use separate expression type. return Out::Got->getMipsGotOffset(Body, A); case R_MIPS_TLSGD: return Out::Got->getGlobalDynOffset(Body) + Out::Got->getMipsTlsOffset() - MipsGPOffset; case R_MIPS_TLSLD: return Out::Got->getTlsIndexOff() + Out::Got->getMipsTlsOffset() - MipsGPOffset; case R_PPC_OPD: { uint64_t SymVA = Body.getVA(A); // If we have an undefined weak symbol, we might get here with a symbol // address of zero. That could overflow, but the code must be unreachable, // so don't bother doing anything at all. if (!SymVA) return 0; if (Out::Opd) { // If this is a local call, and we currently have the address of a // function-descriptor, get the underlying code address instead. uint64_t OpdStart = Out::Opd->getVA(); uint64_t OpdEnd = OpdStart + Out::Opd->getSize(); bool InOpd = OpdStart <= SymVA && SymVA < OpdEnd; if (InOpd) SymVA = read64be(&Out::OpdBuf[SymVA - OpdStart]); } return SymVA - P; } case R_PC: case R_RELAX_GOT_PC: return Body.getVA(A) - P; case R_PLT_PAGE_PC: case R_PAGE_PC: return getAArch64Page(Body.getVA(A)) - getAArch64Page(P); } llvm_unreachable("Invalid expression"); } // This function applies relocations to sections without SHF_ALLOC bit. // Such sections are never mapped to memory at runtime. Debug sections are // an example. Relocations in non-alloc sections are much easier to // handle than in allocated sections because it will never need complex // treatement such as GOT or PLT (because at runtime no one refers them). // So, we handle relocations for non-alloc sections directly in this // function as a performance optimization. template template void InputSection::relocateNonAlloc(uint8_t *Buf, ArrayRef Rels) { const unsigned Bits = sizeof(uintX_t) * 8; for (const RelTy &Rel : Rels) { uint32_t Type = Rel.getType(Config->Mips64EL); uintX_t Offset = this->getOffset(Rel.r_offset); uint8_t *BufLoc = Buf + Offset; uintX_t Addend = getAddend(Rel); if (!RelTy::IsRela) Addend += Target->getImplicitAddend(BufLoc, Type); SymbolBody &Sym = this->File->getRelocTargetSym(Rel); if (Target->getRelExpr(Type, Sym) != R_ABS) { error(this->getSectionName() + " has non-ABS reloc"); return; } uintX_t AddrLoc = this->OutSec->getVA() + Offset; uint64_t SymVA = SignExtend64(getSymVA(Type, Addend, AddrLoc, Sym, R_ABS)); Target->relocateOne(BufLoc, Type, SymVA); } } template void InputSectionBase::relocate(uint8_t *Buf, uint8_t *BufEnd) { // scanReloc function in Writer.cpp constructs Relocations // vector only for SHF_ALLOC'ed sections. For other sections, // we handle relocations directly here. auto *IS = dyn_cast>(this); if (IS && !(IS->Header->sh_flags & SHF_ALLOC)) { for (const Elf_Shdr *RelSec : IS->RelocSections) { if (RelSec->sh_type == SHT_RELA) IS->relocateNonAlloc(Buf, IS->File->getObj().relas(RelSec)); else IS->relocateNonAlloc(Buf, IS->File->getObj().rels(RelSec)); } return; } const unsigned Bits = sizeof(uintX_t) * 8; for (const Relocation &Rel : Relocations) { uintX_t Offset = Rel.InputSec->getOffset(Rel.Offset); uint8_t *BufLoc = Buf + Offset; uint32_t Type = Rel.Type; uintX_t A = Rel.Addend; uintX_t AddrLoc = OutSec->getVA() + Offset; RelExpr Expr = Rel.Expr; uint64_t SymVA = SignExtend64(getSymVA(Type, A, AddrLoc, *Rel.Sym, Expr)); switch (Expr) { case R_RELAX_GOT_PC: case R_RELAX_GOT_PC_NOPIC: Target->relaxGot(BufLoc, SymVA); break; case R_RELAX_TLS_IE_TO_LE: Target->relaxTlsIeToLe(BufLoc, Type, SymVA); break; case R_RELAX_TLS_LD_TO_LE: Target->relaxTlsLdToLe(BufLoc, Type, SymVA); break; case R_RELAX_TLS_GD_TO_LE: case R_RELAX_TLS_GD_TO_LE_NEG: Target->relaxTlsGdToLe(BufLoc, Type, SymVA); break; case R_RELAX_TLS_GD_TO_IE: case R_RELAX_TLS_GD_TO_IE_ABS: case R_RELAX_TLS_GD_TO_IE_PAGE_PC: case R_RELAX_TLS_GD_TO_IE_END: Target->relaxTlsGdToIe(BufLoc, Type, SymVA); break; case R_PPC_PLT_OPD: // Patch a nop (0x60000000) to a ld. if (BufLoc + 8 <= BufEnd && read32be(BufLoc + 4) == 0x60000000) write32be(BufLoc + 4, 0xe8410028); // ld %r2, 40(%r1) // fallthrough default: Target->relocateOne(BufLoc, Type, SymVA); break; } } } template void InputSection::writeTo(uint8_t *Buf) { if (this->Header->sh_type == SHT_NOBITS) return; ELFFile &EObj = this->File->getObj(); // If -r is given, then an InputSection may be a relocation section. if (this->Header->sh_type == SHT_RELA) { copyRelocations(Buf + OutSecOff, EObj.relas(this->Header)); return; } if (this->Header->sh_type == SHT_REL) { copyRelocations(Buf + OutSecOff, EObj.rels(this->Header)); return; } // Copy section contents from source object file to output file. ArrayRef Data = this->getSectionData(); memcpy(Buf + OutSecOff, Data.data(), Data.size()); // Iterate over all relocation sections that apply to this section. uint8_t *BufEnd = Buf + OutSecOff + Data.size(); this->relocate(Buf, BufEnd); // The section might have a data/code generated by the linker and need // to be written after the section. Usually these are thunks - small piece // of code used to jump between "incompatible" functions like PIC and non-PIC // or if the jump target too far and its address does not fit to the short // jump istruction. if (!Thunks.empty()) { Buf += OutSecOff + getThunkOff(); for (const Thunk *T : Thunks) { T->writeTo(Buf); Buf += T->size(); } } } template void InputSection::replace(InputSection *Other) { this->Alignment = std::max(this->Alignment, Other->Alignment); Other->Repl = this->Repl; Other->Live = false; } template SplitInputSection::SplitInputSection( elf::ObjectFile *File, const Elf_Shdr *Header, typename InputSectionBase::Kind SectionKind) : InputSectionBase(File, Header, SectionKind) {} template EhInputSection::EhInputSection(elf::ObjectFile *F, const Elf_Shdr *Header) : SplitInputSection(F, Header, InputSectionBase::EHFrame) { // Mark .eh_frame sections as live by default because there are // usually no relocations that point to .eh_frames. Otherwise, // the garbage collector would drop all .eh_frame sections. this->Live = true; } template bool EhInputSection::classof(const InputSectionBase *S) { return S->SectionKind == InputSectionBase::EHFrame; } // .eh_frame is a sequence of CIE or FDE records. // This function splits an input section into records and returns them. template void EhInputSection::split() { ArrayRef Data = this->getSectionData(); for (size_t Off = 0, End = Data.size(); Off != End;) { size_t Size = readEhRecordSize(Data.slice(Off)); this->Pieces.emplace_back(Off, Data.slice(Off, Size)); // The empty record is the end marker. if (Size == 4) break; Off += Size; } } template typename ELFT::uint EhInputSection::getOffset(uintX_t Offset) const { // The file crtbeginT.o has relocations pointing to the start of an empty // .eh_frame that is known to be the first in the link. It does that to // identify the start of the output .eh_frame. Handle this special case. if (this->getSectionHdr()->sh_size == 0) return Offset; const SectionPiece *Piece = this->getSectionPiece(Offset); if (Piece->OutputOff == size_t(-1)) return -1; // Not in the output uintX_t Addend = Offset - Piece->InputOff; return Piece->OutputOff + Addend; } static size_t findNull(ArrayRef A, size_t EntSize) { // Optimize the common case. StringRef S((const char *)A.data(), A.size()); if (EntSize == 1) return S.find(0); for (unsigned I = 0, N = S.size(); I != N; I += EntSize) { const char *B = S.begin() + I; if (std::all_of(B, B + EntSize, [](char C) { return C == 0; })) return I; } return StringRef::npos; } // Split SHF_STRINGS section. Such section is a sequence of // null-terminated strings. static std::vector splitStrings(ArrayRef Data, size_t EntSize) { std::vector V; size_t Off = 0; while (!Data.empty()) { size_t End = findNull(Data, EntSize); if (End == StringRef::npos) fatal("string is not null terminated"); size_t Size = End + EntSize; V.emplace_back(Off, Data.slice(0, Size)); Data = Data.slice(Size); Off += Size; } return V; } // Split non-SHF_STRINGS section. Such section is a sequence of // fixed size records. static std::vector splitNonStrings(ArrayRef Data, size_t EntSize) { std::vector V; size_t Size = Data.size(); assert((Size % EntSize) == 0); for (unsigned I = 0, N = Size; I != N; I += EntSize) V.emplace_back(I, Data.slice(I, EntSize)); return V; } template MergeInputSection::MergeInputSection(elf::ObjectFile *F, const Elf_Shdr *Header) : SplitInputSection(F, Header, InputSectionBase::Merge) {} template void MergeInputSection::splitIntoPieces() { ArrayRef Data = this->getSectionData(); uintX_t EntSize = this->Header->sh_entsize; if (this->Header->sh_flags & SHF_STRINGS) this->Pieces = splitStrings(Data, EntSize); else this->Pieces = splitNonStrings(Data, EntSize); if (Config->GcSections) for (uintX_t Off : LiveOffsets) this->getSectionPiece(Off)->Live = true; } template bool MergeInputSection::classof(const InputSectionBase *S) { return S->SectionKind == InputSectionBase::Merge; } // Do binary search to get a section piece at a given input offset. template SectionPiece *SplitInputSection::getSectionPiece(uintX_t Offset) { auto *This = static_cast *>(this); return const_cast(This->getSectionPiece(Offset)); } template const SectionPiece * SplitInputSection::getSectionPiece(uintX_t Offset) const { ArrayRef D = this->getSectionData(); StringRef Data((const char *)D.data(), D.size()); uintX_t Size = Data.size(); if (Offset >= Size) fatal("entry is past the end of the section"); // Find the element this offset points to. auto I = std::upper_bound( Pieces.begin(), Pieces.end(), Offset, [](const uintX_t &A, const SectionPiece &B) { return A < B.InputOff; }); --I; return &*I; } // Returns the offset in an output section for a given input offset. // Because contents of a mergeable section is not contiguous in output, // it is not just an addition to a base output offset. template typename ELFT::uint MergeInputSection::getOffset(uintX_t Offset) const { auto It = OffsetMap.find(Offset); if (It != OffsetMap.end()) return It->second; // If Offset is not at beginning of a section piece, it is not in the map. // In that case we need to search from the original section piece vector. const SectionPiece &Piece = *this->getSectionPiece(Offset); assert(Piece.Live); uintX_t Addend = Offset - Piece.InputOff; return Piece.OutputOff + Addend; } // Create a map from input offsets to output offsets for all section pieces. // It is called after finalize(). template void MergeInputSection::finalizePieces() { OffsetMap.grow(this->Pieces.size()); for (SectionPiece &Piece : this->Pieces) { if (!Piece.Live) continue; if (Piece.OutputOff == size_t(-1)) { // Offsets of tail-merged strings are computed lazily. auto *OutSec = static_cast *>(this->OutSec); ArrayRef D = Piece.data(); StringRef S((const char *)D.data(), D.size()); Piece.OutputOff = OutSec->getOffset(S); } OffsetMap[Piece.InputOff] = Piece.OutputOff; } } template MipsReginfoInputSection::MipsReginfoInputSection(elf::ObjectFile *F, const Elf_Shdr *Hdr) : InputSectionBase(F, Hdr, InputSectionBase::MipsReginfo) { // Initialize this->Reginfo. ArrayRef D = this->getSectionData(); if (D.size() != sizeof(Elf_Mips_RegInfo)) { error("invalid size of .reginfo section"); return; } Reginfo = reinterpret_cast *>(D.data()); } template bool MipsReginfoInputSection::classof(const InputSectionBase *S) { return S->SectionKind == InputSectionBase::MipsReginfo; } template MipsOptionsInputSection::MipsOptionsInputSection(elf::ObjectFile *F, const Elf_Shdr *Hdr) : InputSectionBase(F, Hdr, InputSectionBase::MipsOptions) { // Find ODK_REGINFO option in the section's content. ArrayRef D = this->getSectionData(); while (!D.empty()) { if (D.size() < sizeof(Elf_Mips_Options)) { error("invalid size of .MIPS.options section"); break; } auto *O = reinterpret_cast *>(D.data()); if (O->kind == ODK_REGINFO) { Reginfo = &O->getRegInfo(); break; } D = D.slice(O->size); } } template bool MipsOptionsInputSection::classof(const InputSectionBase *S) { return S->SectionKind == InputSectionBase::MipsOptions; } template bool elf::isDiscarded(InputSectionBase *); template bool elf::isDiscarded(InputSectionBase *); template bool elf::isDiscarded(InputSectionBase *); template bool elf::isDiscarded(InputSectionBase *); template class elf::InputSectionBase; template class elf::InputSectionBase; template class elf::InputSectionBase; template class elf::InputSectionBase; template class elf::InputSection; template class elf::InputSection; template class elf::InputSection; template class elf::InputSection; template class elf::SplitInputSection; template class elf::SplitInputSection; template class elf::SplitInputSection; template class elf::SplitInputSection; template class elf::EhInputSection; template class elf::EhInputSection; template class elf::EhInputSection; template class elf::EhInputSection; template class elf::MergeInputSection; template class elf::MergeInputSection; template class elf::MergeInputSection; template class elf::MergeInputSection; template class elf::MipsReginfoInputSection; template class elf::MipsReginfoInputSection; template class elf::MipsReginfoInputSection; template class elf::MipsReginfoInputSection; template class elf::MipsOptionsInputSection; template class elf::MipsOptionsInputSection; template class elf::MipsOptionsInputSection; template class elf::MipsOptionsInputSection;