llvm-project/lld/ELF/InputFiles.cpp

1248 lines
46 KiB
C++
Raw Normal View History

//===- InputFiles.cpp -----------------------------------------------------===//
//
// The LLVM Linker
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
#include "InputFiles.h"
2016-02-11 23:24:48 +08:00
#include "InputSection.h"
#include "LinkerScript.h"
ELF: New symbol table design. This patch implements a new design for the symbol table that stores SymbolBodies within a memory region of the Symbol object. Symbols are mutated by constructing SymbolBodies in place over existing SymbolBodies, rather than by mutating pointers. As mentioned in the initial proposal [1], this memory layout helps reduce the cache miss rate by improving memory locality. Performance numbers: old(s) new(s) Without debug info: chrome 7.178 6.432 (-11.5%) LLVMgold.so 0.505 0.502 (-0.5%) clang 0.954 0.827 (-15.4%) llvm-as 0.052 0.045 (-15.5%) With debug info: scylla 5.695 5.613 (-1.5%) clang 14.396 14.143 (-1.8%) Performance counter results show that the fewer required indirections is indeed the cause of the improved performance. For example, when linking chrome, stalled cycles decreases from 14,556,444,002 to 12,959,238,310, and instructions per cycle increases from 0.78 to 0.83. We are also executing many fewer instructions (15,516,401,933 down to 15,002,434,310), probably because we spend less time allocating SymbolBodies. The new mechanism by which symbols are added to the symbol table is by calling add* functions on the SymbolTable. In this patch, I handle local symbols by storing them inside "unparented" SymbolBodies. This is suboptimal, but if we do want to try to avoid allocating these SymbolBodies, we can probably do that separately. I also removed a few members from the SymbolBody class that were only being used to pass information from the input file to the symbol table. This patch implements the new design for the ELF linker only. I intend to prepare a similar patch for the COFF linker. [1] http://lists.llvm.org/pipermail/llvm-dev/2016-April/098832.html Differential Revision: http://reviews.llvm.org/D19752 llvm-svn: 268178
2016-05-01 12:55:03 +08:00
#include "SymbolTable.h"
#include "Symbols.h"
#include "SyntheticSections.h"
#include "lld/Common/ErrorHandler.h"
#include "lld/Common/Memory.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/CodeGen/Analysis.h"
#include "llvm/DebugInfo/DWARF/DWARFContext.h"
#include "llvm/IR/LLVMContext.h"
2016-03-02 23:43:50 +08:00
#include "llvm/IR/Module.h"
#include "llvm/LTO/LTO.h"
#include "llvm/MC/StringTableBuilder.h"
#include "llvm/Object/ELFObjectFile.h"
#include "llvm/Support/ARMAttributeParser.h"
#include "llvm/Support/ARMBuildAttributes.h"
#include "llvm/Support/Path.h"
#include "llvm/Support/TarWriter.h"
#include "llvm/Support/raw_ostream.h"
using namespace llvm;
using namespace llvm::ELF;
using namespace llvm::object;
using namespace llvm::sys;
using namespace llvm::sys::fs;
using namespace lld;
2016-02-28 08:25:54 +08:00
using namespace lld::elf;
std::vector<BinaryFile *> elf::BinaryFiles;
std::vector<BitcodeFile *> elf::BitcodeFiles;
std::vector<InputFile *> elf::ObjectFiles;
std::vector<InputFile *> elf::SharedFiles;
TarWriter *elf::Tar;
InputFile::InputFile(Kind K, MemoryBufferRef M) : MB(M), FileKind(K) {}
Optional<MemoryBufferRef> elf::readFile(StringRef Path) {
// The --chroot option changes our virtual root directory.
// This is useful when you are dealing with files created by --reproduce.
if (!Config->Chroot.empty() && Path.startswith("/"))
Path = Saver.save(Config->Chroot + Path);
log(Path);
auto MBOrErr = MemoryBuffer::getFile(Path);
if (auto EC = MBOrErr.getError()) {
error("cannot open " + Path + ": " + EC.message());
return None;
}
std::unique_ptr<MemoryBuffer> &MB = *MBOrErr;
MemoryBufferRef MBRef = MB->getMemBufferRef();
make<std::unique_ptr<MemoryBuffer>>(std::move(MB)); // take MB ownership
if (Tar)
Tar->append(relativeToRoot(Path), MBRef.getBuffer());
return MBRef;
}
// Concatenates arguments to construct a string representing an error location.
static std::string createFileLineMsg(StringRef Path, unsigned Line) {
std::string Filename = path::filename(Path);
std::string Lineno = ":" + std::to_string(Line);
if (Filename == Path)
return Filename + Lineno;
return Filename + Lineno + " (" + Path.str() + Lineno + ")";
}
template <class ELFT>
static std::string getSrcMsgAux(ObjFile<ELFT> &File, const Symbol &Sym,
InputSectionBase &Sec, uint64_t Offset) {
// In DWARF, functions and variables are stored to different places.
// First, lookup a function for a given offset.
if (Optional<DILineInfo> Info = File.getDILineInfo(&Sec, Offset))
return createFileLineMsg(Info->FileName, Info->Line);
// If it failed, lookup again as a variable.
if (Optional<std::pair<std::string, unsigned>> FileLine =
File.getVariableLoc(Sym.getName()))
return createFileLineMsg(FileLine->first, FileLine->second);
// File.SourceFile contains STT_FILE symbol, and that is a last resort.
return File.SourceFile;
}
std::string InputFile::getSrcMsg(const Symbol &Sym, InputSectionBase &Sec,
uint64_t Offset) {
if (kind() != ObjKind)
return "";
switch (Config->EKind) {
default:
llvm_unreachable("Invalid kind");
case ELF32LEKind:
return getSrcMsgAux(cast<ObjFile<ELF32LE>>(*this), Sym, Sec, Offset);
case ELF32BEKind:
return getSrcMsgAux(cast<ObjFile<ELF32BE>>(*this), Sym, Sec, Offset);
case ELF64LEKind:
return getSrcMsgAux(cast<ObjFile<ELF64LE>>(*this), Sym, Sec, Offset);
case ELF64BEKind:
return getSrcMsgAux(cast<ObjFile<ELF64BE>>(*this), Sym, Sec, Offset);
}
}
template <class ELFT> void ObjFile<ELFT>::initializeDwarf() {
2018-03-23 08:42:47 +08:00
Dwarf = llvm::make_unique<DWARFContext>(make_unique<LLDDwarfObj<ELFT>>(this));
const DWARFObject &Obj = Dwarf->getDWARFObj();
DwarfLine.reset(new DWARFDebugLine);
DWARFDataExtractor LineData(Obj, Obj.getLineSection(), Config->IsLE,
Config->Wordsize);
for (std::unique_ptr<DWARFCompileUnit> &CU : Dwarf->compile_units()) {
const DWARFDebugLine::LineTable *LT = Dwarf->getLineTableForUnit(CU.get());
if (!LT)
continue;
LineTables.push_back(LT);
// Loop over variable records and insert them to VariableLoc.
for (const auto &Entry : CU->dies()) {
DWARFDie Die(CU.get(), &Entry);
// Skip all tags that are not variables.
if (Die.getTag() != dwarf::DW_TAG_variable)
continue;
// Skip if a local variable because we don't need them for generating
// error messages. In general, only non-local symbols can fail to be
// linked.
if (!dwarf::toUnsigned(Die.find(dwarf::DW_AT_external), 0))
continue;
// Get the source filename index for the variable.
unsigned File = dwarf::toUnsigned(Die.find(dwarf::DW_AT_decl_file), 0);
if (!LT->hasFileAtIndex(File))
continue;
// Get the line number on which the variable is declared.
unsigned Line = dwarf::toUnsigned(Die.find(dwarf::DW_AT_decl_line), 0);
// Get the name of the variable and add the collected information to
// VariableLoc. Usually Name is non-empty, but it can be empty if the
// input object file lacks some debug info.
StringRef Name = dwarf::toString(Die.find(dwarf::DW_AT_name), "");
if (!Name.empty())
VariableLoc.insert({Name, {LT, File, Line}});
}
}
}
// Returns the pair of file name and line number describing location of data
// object (variable, array, etc) definition.
template <class ELFT>
Optional<std::pair<std::string, unsigned>>
ObjFile<ELFT>::getVariableLoc(StringRef Name) {
llvm::call_once(InitDwarfLine, [this]() { initializeDwarf(); });
// Return if we have no debug information about data object.
auto It = VariableLoc.find(Name);
if (It == VariableLoc.end())
return None;
// Take file name string from line table.
std::string FileName;
if (!It->second.LT->getFileNameByIndex(
It->second.File, nullptr,
DILineInfoSpecifier::FileLineInfoKind::AbsoluteFilePath, FileName))
return None;
return std::make_pair(FileName, It->second.Line);
}
// Returns source line information for a given offset
// using DWARF debug info.
template <class ELFT>
Optional<DILineInfo> ObjFile<ELFT>::getDILineInfo(InputSectionBase *S,
uint64_t Offset) {
llvm::call_once(InitDwarfLine, [this]() { initializeDwarf(); });
// Use fake address calcuated by adding section file offset and offset in
// section. See comments for ObjectInfo class.
DILineInfo Info;
for (const llvm::DWARFDebugLine::LineTable *LT : LineTables)
if (LT->getFileLineInfoForAddress(
S->getOffsetInFile() + Offset, nullptr,
DILineInfoSpecifier::FileLineInfoKind::AbsoluteFilePath, Info))
return Info;
return None;
}
// Returns source line information for a given offset using DWARF debug info.
template <class ELFT>
std::string ObjFile<ELFT>::getLineInfo(InputSectionBase *S, uint64_t Offset) {
if (Optional<DILineInfo> Info = getDILineInfo(S, Offset))
return Info->FileName + ":" + std::to_string(Info->Line);
return "";
}
// Returns "<internal>", "foo.a(bar.o)" or "baz.o".
std::string lld::toString(const InputFile *F) {
if (!F)
return "<internal>";
if (F->ToStringCache.empty()) {
if (F->ArchiveName.empty())
F->ToStringCache = F->getName();
else
F->ToStringCache = (F->ArchiveName + "(" + F->getName() + ")").str();
}
return F->ToStringCache;
}
template <class ELFT>
ELFFileBase<ELFT>::ELFFileBase(Kind K, MemoryBufferRef MB) : InputFile(K, MB) {
if (ELFT::TargetEndianness == support::little)
EKind = ELFT::Is64Bits ? ELF64LEKind : ELF32LEKind;
else
EKind = ELFT::Is64Bits ? ELF64BEKind : ELF32BEKind;
EMachine = getObj().getHeader()->e_machine;
OSABI = getObj().getHeader()->e_ident[llvm::ELF::EI_OSABI];
}
template <class ELFT>
typename ELFT::SymRange ELFFileBase<ELFT>::getGlobalELFSyms() {
return makeArrayRef(ELFSyms.begin() + FirstNonLocal, ELFSyms.end());
}
template <class ELFT>
uint32_t ELFFileBase<ELFT>::getSectionIndex(const Elf_Sym &Sym) const {
return CHECK(getObj().getSectionIndex(&Sym, ELFSyms, SymtabSHNDX), this);
}
template <class ELFT>
void ELFFileBase<ELFT>::initSymtab(ArrayRef<Elf_Shdr> Sections,
const Elf_Shdr *Symtab) {
FirstNonLocal = Symtab->sh_info;
ELFSyms = CHECK(getObj().symbols(Symtab), this);
if (FirstNonLocal == 0 || FirstNonLocal > ELFSyms.size())
fatal(toString(this) + ": invalid sh_info in symbol table");
StringTable =
CHECK(getObj().getStringTableForSymtab(*Symtab, Sections), this);
}
template <class ELFT>
ObjFile<ELFT>::ObjFile(MemoryBufferRef M, StringRef ArchiveName)
: ELFFileBase<ELFT>(Base::ObjKind, M) {
this->ArchiveName = ArchiveName;
}
template <class ELFT> ArrayRef<Symbol *> ObjFile<ELFT>::getLocalSymbols() {
if (this->Symbols.empty())
return {};
return makeArrayRef(this->Symbols).slice(1, this->FirstNonLocal - 1);
}
template <class ELFT>
void ObjFile<ELFT>::parse(DenseSet<CachedHashStringRef> &ComdatGroups) {
// Read section and symbol tables.
initializeSections(ComdatGroups);
initializeSymbols();
}
// Sections with SHT_GROUP and comdat bits define comdat section groups.
// They are identified and deduplicated by group name. This function
// returns a group name.
template <class ELFT>
StringRef ObjFile<ELFT>::getShtGroupSignature(ArrayRef<Elf_Shdr> Sections,
const Elf_Shdr &Sec) {
// Group signatures are stored as symbol names in object files.
// sh_info contains a symbol index, so we fetch a symbol and read its name.
if (this->ELFSyms.empty())
this->initSymtab(
Sections, CHECK(object::getSection<ELFT>(Sections, Sec.sh_link), this));
const Elf_Sym *Sym =
CHECK(object::getSymbol<ELFT>(this->ELFSyms, Sec.sh_info), this);
StringRef Signature = CHECK(Sym->getName(this->StringTable), this);
// As a special case, if a symbol is a section symbol and has no name,
// we use a section name as a signature.
//
// Such SHT_GROUP sections are invalid from the perspective of the ELF
// standard, but GNU gold 1.14 (the newest version as of July 2017) or
// older produce such sections as outputs for the -r option, so we need
// a bug-compatibility.
if (Signature.empty() && Sym->getType() == STT_SECTION)
return getSectionName(Sec);
return Signature;
}
template <class ELFT>
ArrayRef<typename ObjFile<ELFT>::Elf_Word>
ObjFile<ELFT>::getShtGroupEntries(const Elf_Shdr &Sec) {
const ELFFile<ELFT> &Obj = this->getObj();
ArrayRef<Elf_Word> Entries =
CHECK(Obj.template getSectionContentsAsArray<Elf_Word>(&Sec), this);
if (Entries.empty() || Entries[0] != GRP_COMDAT)
fatal(toString(this) + ": unsupported SHT_GROUP format");
return Entries.slice(1);
}
template <class ELFT> bool ObjFile<ELFT>::shouldMerge(const Elf_Shdr &Sec) {
// We don't merge sections if -O0 (default is -O1). This makes sometimes
// the linker significantly faster, although the output will be bigger.
if (Config->Optimize == 0)
return false;
// A mergeable section with size 0 is useless because they don't have
// any data to merge. A mergeable string section with size 0 can be
// argued as invalid because it doesn't end with a null character.
// We'll avoid a mess by handling them as if they were non-mergeable.
if (Sec.sh_size == 0)
return false;
// Check for sh_entsize. The ELF spec is not clear about the zero
// sh_entsize. It says that "the member [sh_entsize] contains 0 if
// the section does not hold a table of fixed-size entries". We know
// that Rust 1.13 produces a string mergeable section with a zero
// sh_entsize. Here we just accept it rather than being picky about it.
uint64_t EntSize = Sec.sh_entsize;
if (EntSize == 0)
return false;
if (Sec.sh_size % EntSize)
fatal(toString(this) +
": SHF_MERGE section size must be a multiple of sh_entsize");
uint64_t Flags = Sec.sh_flags;
if (!(Flags & SHF_MERGE))
return false;
if (Flags & SHF_WRITE)
fatal(toString(this) + ": writable SHF_MERGE section is not supported");
return true;
}
template <class ELFT>
void ObjFile<ELFT>::initializeSections(
DenseSet<CachedHashStringRef> &ComdatGroups) {
2017-05-26 10:17:30 +08:00
const ELFFile<ELFT> &Obj = this->getObj();
ArrayRef<Elf_Shdr> ObjSections = CHECK(this->getObj().sections(), this);
uint64_t Size = ObjSections.size();
this->Sections.resize(Size);
this->SectionStringTable =
CHECK(Obj.getSectionStringTable(ObjSections), this);
2017-05-26 10:17:30 +08:00
for (size_t I = 0, E = ObjSections.size(); I < E; I++) {
if (this->Sections[I] == &InputSection::Discarded)
continue;
2017-05-26 10:17:30 +08:00
const Elf_Shdr &Sec = ObjSections[I];
2016-10-05 00:47:49 +08:00
// SHF_EXCLUDE'ed sections are discarded by the linker. However,
// if -r is given, we'll let the final link discard such sections.
// This is compatible with GNU.
if ((Sec.sh_flags & SHF_EXCLUDE) && !Config->Relocatable) {
this->Sections[I] = &InputSection::Discarded;
continue;
}
switch (Sec.sh_type) {
case SHT_GROUP: {
2017-06-09 10:42:20 +08:00
// De-duplicate section groups by their signatures.
StringRef Signature = getShtGroupSignature(ObjSections, Sec);
bool IsNew = ComdatGroups.insert(CachedHashStringRef(Signature)).second;
this->Sections[I] = &InputSection::Discarded;
// If it is a new section group, we want to keep group members.
// Group leader sections, which contain indices of group members, are
// discarded because they are useless beyond this point. The only
// exception is the -r option because in order to produce re-linkable
// object files, we want to pass through basically everything.
if (IsNew) {
if (Config->Relocatable)
this->Sections[I] = createInputSection(Sec);
continue;
2017-06-09 10:42:20 +08:00
}
2017-06-09 10:42:20 +08:00
// Otherwise, discard group members.
for (uint32_t SecIndex : getShtGroupEntries(Sec)) {
if (SecIndex >= Size)
fatal(toString(this) +
": invalid section index in group: " + Twine(SecIndex));
this->Sections[SecIndex] = &InputSection::Discarded;
}
break;
}
case SHT_SYMTAB:
this->initSymtab(ObjSections, &Sec);
break;
case SHT_SYMTAB_SHNDX:
this->SymtabSHNDX = CHECK(Obj.getSHNDXTable(Sec, ObjSections), this);
break;
case SHT_STRTAB:
case SHT_NULL:
break;
default:
this->Sections[I] = createInputSection(Sec);
}
// .ARM.exidx sections have a reverse dependency on the InputSection they
// have a SHF_LINK_ORDER dependency, this is identified by the sh_link.
if (Sec.sh_flags & SHF_LINK_ORDER) {
if (Sec.sh_link >= this->Sections.size())
2017-12-21 03:59:47 +08:00
fatal(toString(this) +
": invalid sh_link index: " + Twine(Sec.sh_link));
InputSectionBase *LinkSec = this->Sections[Sec.sh_link];
InputSection *IS = cast<InputSection>(this->Sections[I]);
LinkSec->DependentSections.push_back(IS);
if (!isa<InputSection>(LinkSec))
error("a section " + IS->Name +
" with SHF_LINK_ORDER should not refer a non-regular "
"section: " +
toString(LinkSec));
}
}
}
// The ARM support in lld makes some use of instructions that are not available
// on all ARM architectures. Namely:
// - Use of BLX instruction for interworking between ARM and Thumb state.
// - Use of the extended Thumb branch encoding in relocation.
// - Use of the MOVT/MOVW instructions in Thumb Thunks.
// The ARM Attributes section contains information about the architecture chosen
// at compile time. We follow the convention that if at least one input object
// is compiled with an architecture that supports these features then lld is
// permitted to use them.
static void updateSupportedARMFeatures(const ARMAttributeParser &Attributes) {
if (!Attributes.hasAttribute(ARMBuildAttrs::CPU_arch))
return;
auto Arch = Attributes.getAttributeValue(ARMBuildAttrs::CPU_arch);
switch (Arch) {
case ARMBuildAttrs::Pre_v4:
case ARMBuildAttrs::v4:
case ARMBuildAttrs::v4T:
// Architectures prior to v5 do not support BLX instruction
break;
case ARMBuildAttrs::v5T:
case ARMBuildAttrs::v5TE:
case ARMBuildAttrs::v5TEJ:
case ARMBuildAttrs::v6:
case ARMBuildAttrs::v6KZ:
case ARMBuildAttrs::v6K:
Config->ARMHasBlx = true;
// Architectures used in pre-Cortex processors do not support
// The J1 = 1 J2 = 1 Thumb branch range extension, with the exception
// of Architecture v6T2 (arm1156t2-s and arm1156t2f-s) that do.
break;
default:
// All other Architectures have BLX and extended branch encoding
Config->ARMHasBlx = true;
Config->ARMJ1J2BranchEncoding = true;
if (Arch != ARMBuildAttrs::v6_M && Arch != ARMBuildAttrs::v6S_M)
// All Architectures used in Cortex processors with the exception
// of v6-M and v6S-M have the MOVT and MOVW instructions.
Config->ARMHasMovtMovw = true;
break;
}
}
template <class ELFT>
InputSectionBase *ObjFile<ELFT>::getRelocTarget(const Elf_Shdr &Sec) {
uint32_t Idx = Sec.sh_info;
if (Idx >= this->Sections.size())
fatal(toString(this) + ": invalid relocated section index: " + Twine(Idx));
InputSectionBase *Target = this->Sections[Idx];
// Strictly speaking, a relocation section must be included in the
// group of the section it relocates. However, LLVM 3.3 and earlier
// would fail to do so, so we gracefully handle that case.
if (Target == &InputSection::Discarded)
return nullptr;
if (!Target)
fatal(toString(this) + ": unsupported relocation reference");
return Target;
}
// Create a regular InputSection class that has the same contents
// as a given section.
static InputSection *toRegularSection(MergeInputSection *Sec) {
return make<InputSection>(Sec->File, Sec->Flags, Sec->Type, Sec->Alignment,
Sec->Data, Sec->Name);
}
template <class ELFT>
InputSectionBase *ObjFile<ELFT>::createInputSection(const Elf_Shdr &Sec) {
StringRef Name = getSectionName(Sec);
switch (Sec.sh_type) {
case SHT_ARM_ATTRIBUTES: {
if (Config->EMachine != EM_ARM)
break;
ARMAttributeParser Attributes;
ArrayRef<uint8_t> Contents = check(this->getObj().getSectionContents(&Sec));
2017-12-21 03:59:47 +08:00
Attributes.Parse(Contents, /*isLittle*/ Config->EKind == ELF32LEKind);
updateSupportedARMFeatures(Attributes);
// FIXME: Retain the first attribute section we see. The eglibc ARM
// dynamic loaders require the presence of an attribute section for dlopen
// to work. In a full implementation we would merge all attribute sections.
if (InX::ARMAttributes == nullptr) {
InX::ARMAttributes = make<InputSection>(*this, Sec, Name);
return InX::ARMAttributes;
}
return &InputSection::Discarded;
}
case SHT_RELA:
case SHT_REL: {
2018-03-27 10:52:58 +08:00
// Find a relocation target section and associate this section with that.
// Target may have been discarded if it is in a different section group
// and the group is discarded, even though it's a violation of the
// spec. We handle that situation gracefully by discarding dangling
// relocation sections.
InputSectionBase *Target = getRelocTarget(Sec);
if (!Target)
return nullptr;
// This section contains relocation information.
// If -r is given, we do not interpret or apply relocation
// but just copy relocation sections to output.
if (Config->Relocatable)
return make<InputSection>(*this, Sec, Name);
if (Target->FirstRelocation)
fatal(toString(this) +
": multiple relocation sections to one section are not supported");
2018-03-27 10:52:58 +08:00
// ELF spec allows mergeable sections with relocations, but they are
// rare, and it is in practice hard to merge such sections by contents,
// because applying relocations at end of linking changes section
// contents. So, we simply handle such sections as non-mergeable ones.
// Degrading like this is acceptable because section merging is optional.
if (auto *MS = dyn_cast<MergeInputSection>(Target)) {
Target = toRegularSection(MS);
this->Sections[Sec.sh_info] = Target;
}
if (Sec.sh_type == SHT_RELA) {
ArrayRef<Elf_Rela> Rels = CHECK(this->getObj().relas(&Sec), this);
Target->FirstRelocation = Rels.begin();
Target->NumRelocations = Rels.size();
Target->AreRelocsRela = true;
} else {
ArrayRef<Elf_Rel> Rels = CHECK(this->getObj().rels(&Sec), this);
Target->FirstRelocation = Rels.begin();
Target->NumRelocations = Rels.size();
Target->AreRelocsRela = false;
}
assert(isUInt<31>(Target->NumRelocations));
// Relocation sections processed by the linker are usually removed
// from the output, so returning `nullptr` for the normal case.
// However, if -emit-relocs is given, we need to leave them in the output.
// (Some post link analysis tools need this information.)
if (Config->EmitRelocs) {
InputSection *RelocSec = make<InputSection>(*this, Sec, Name);
// We will not emit relocation section if target was discarded.
Target->DependentSections.push_back(RelocSec);
return RelocSec;
}
return nullptr;
}
}
// The GNU linker uses .note.GNU-stack section as a marker indicating
// that the code in the object file does not expect that the stack is
// executable (in terms of NX bit). If all input files have the marker,
// the GNU linker adds a PT_GNU_STACK segment to tells the loader to
// make the stack non-executable. Most object files have this section as
// of 2017.
//
// But making the stack non-executable is a norm today for security
// reasons. Failure to do so may result in a serious security issue.
// Therefore, we make LLD always add PT_GNU_STACK unless it is
// explicitly told to do otherwise (by -z execstack). Because the stack
// executable-ness is controlled solely by command line options,
// .note.GNU-stack sections are simply ignored.
if (Name == ".note.GNU-stack")
return &InputSection::Discarded;
// Split stacks is a feature to support a discontiguous stack. At least
// as of 2017, it seems that the feature is not being used widely.
// Only GNU gold supports that. We don't. For the details about that,
// see https://gcc.gnu.org/wiki/SplitStacks
if (Name == ".note.GNU-split-stack") {
error(toString(this) +
": object file compiled with -fsplit-stack is not supported");
return &InputSection::Discarded;
}
// The linkonce feature is a sort of proto-comdat. Some glibc i386 object
// files contain definitions of symbol "__x86.get_pc_thunk.bx" in linkonce
// sections. Drop those sections to avoid duplicate symbol errors.
// FIXME: This is glibc PR20543, we should remove this hack once that has been
// fixed for a while.
if (Name.startswith(".gnu.linkonce."))
return &InputSection::Discarded;
// If we are creating a new .build-id section, strip existing .build-id
// sections so that the output won't have more than one .build-id.
// This is not usually a problem because input object files normally don't
// have .build-id sections, but you can create such files by
// "ld.{bfd,gold,lld} -r --build-id", and we want to guard against it.
if (Name == ".note.gnu.build-id" && Config->BuildId != BuildIdKind::None)
return &InputSection::Discarded;
2016-07-15 12:57:44 +08:00
// The linker merges EH (exception handling) frames and creates a
// .eh_frame_hdr section for runtime. So we handle them with a special
// class. For relocatable outputs, they are just passed through.
if (Name == ".eh_frame" && !Config->Relocatable)
return make<EhInputSection>(*this, Sec, Name);
2016-07-15 12:57:44 +08:00
if (shouldMerge(Sec))
return make<MergeInputSection>(*this, Sec, Name);
return make<InputSection>(*this, Sec, Name);
}
template <class ELFT>
StringRef ObjFile<ELFT>::getSectionName(const Elf_Shdr &Sec) {
return CHECK(this->getObj().getSectionName(&Sec, SectionStringTable), this);
}
template <class ELFT> void ObjFile<ELFT>::initializeSymbols() {
this->Symbols.reserve(this->ELFSyms.size());
for (const Elf_Sym &Sym : this->ELFSyms)
this->Symbols.push_back(createSymbol(&Sym));
}
template <class ELFT> Symbol *ObjFile<ELFT>::createSymbol(const Elf_Sym *Sym) {
int Binding = Sym->getBinding();
uint32_t SecIdx = this->getSectionIndex(*Sym);
if (SecIdx >= this->Sections.size())
fatal(toString(this) + ": invalid section index: " + Twine(SecIdx));
InputSectionBase *Sec = this->Sections[SecIdx];
uint8_t StOther = Sym->st_other;
uint8_t Type = Sym->getType();
uint64_t Value = Sym->st_value;
uint64_t Size = Sym->st_size;
if (Binding == STB_LOCAL) {
if (Sym->getType() == STT_FILE)
SourceFile = CHECK(Sym->getName(this->StringTable), this);
if (this->StringTable.size() <= Sym->st_name)
fatal(toString(this) + ": invalid symbol name offset");
StringRefZ Name = this->StringTable.data() + Sym->st_name;
if (Sym->st_shndx == SHN_UNDEF)
return make<Undefined>(this, Name, Binding, StOther, Type);
return make<Defined>(this, Name, Binding, StOther, Type, Value, Size, Sec);
}
StringRef Name = CHECK(Sym->getName(this->StringTable), this);
switch (Sym->st_shndx) {
case SHN_UNDEF:
return Symtab->addUndefined<ELFT>(Name, Binding, StOther, Type,
/*CanOmitFromDynSym=*/false, this);
case SHN_COMMON:
if (Value == 0 || Value >= UINT32_MAX)
fatal(toString(this) + ": common symbol '" + Name +
"' has invalid alignment: " + Twine(Value));
return Symtab->addCommon(Name, Size, Value, Binding, StOther, Type, *this);
}
switch (Binding) {
default:
fatal(toString(this) + ": unexpected binding: " + Twine(Binding));
case STB_GLOBAL:
case STB_WEAK:
case STB_GNU_UNIQUE:
if (Sec == &InputSection::Discarded)
return Symtab->addUndefined<ELFT>(Name, Binding, StOther, Type,
/*CanOmitFromDynSym=*/false, this);
return Symtab->addRegular(Name, StOther, Type, Value, Size, Binding, Sec,
this);
}
}
ArchiveFile::ArchiveFile(std::unique_ptr<Archive> &&File)
: InputFile(ArchiveKind, File->getMemoryBufferRef()),
File(std::move(File)) {}
template <class ELFT> void ArchiveFile::parse() {
for (const Archive::Symbol &Sym : File->symbols())
Symtab->addLazyArchive<ELFT>(Sym.getName(), *this, Sym);
}
// Returns a buffer pointing to a member file containing a given symbol.
std::pair<MemoryBufferRef, uint64_t>
ArchiveFile::getMember(const Archive::Symbol *Sym) {
Archive::Child C =
CHECK(Sym->getMember(), toString(this) +
": could not get the member for symbol " +
Sym->getName());
if (!Seen.insert(C.getChildOffset()).second)
return {MemoryBufferRef(), 0};
MemoryBufferRef Ret =
CHECK(C.getMemoryBufferRef(),
toString(this) +
": could not get the buffer for the member defining symbol " +
Sym->getName());
if (C.getParent()->isThin() && Tar)
Tar->append(relativeToRoot(CHECK(C.getFullName(), this)), Ret.getBuffer());
if (C.getParent()->isThin())
return {Ret, 0};
return {Ret, C.getChildOffset()};
}
template <class ELFT>
SharedFile<ELFT>::SharedFile(MemoryBufferRef M, StringRef DefaultSoName)
: ELFFileBase<ELFT>(Base::SharedKind, M), SoName(DefaultSoName),
IsNeeded(!Config->AsNeeded) {}
// Partially parse the shared object file so that we can call
// getSoName on this object.
template <class ELFT> void SharedFile<ELFT>::parseSoName() {
const Elf_Shdr *DynamicSec = nullptr;
const ELFFile<ELFT> Obj = this->getObj();
ArrayRef<Elf_Shdr> Sections = CHECK(Obj.sections(), this);
// Search for .dynsym, .dynamic, .symtab, .gnu.version and .gnu.version_d.
for (const Elf_Shdr &Sec : Sections) {
switch (Sec.sh_type) {
default:
continue;
case SHT_DYNSYM:
this->initSymtab(Sections, &Sec);
break;
case SHT_DYNAMIC:
DynamicSec = &Sec;
break;
case SHT_SYMTAB_SHNDX:
this->SymtabSHNDX = CHECK(Obj.getSHNDXTable(Sec, Sections), this);
break;
case SHT_GNU_versym:
this->VersymSec = &Sec;
break;
case SHT_GNU_verdef:
this->VerdefSec = &Sec;
break;
}
}
if (this->VersymSec && this->ELFSyms.empty())
error("SHT_GNU_versym should be associated with symbol table");
// Search for a DT_SONAME tag to initialize this->SoName.
2015-10-12 23:49:02 +08:00
if (!DynamicSec)
return;
ArrayRef<Elf_Dyn> Arr =
CHECK(Obj.template getSectionContentsAsArray<Elf_Dyn>(DynamicSec), this);
for (const Elf_Dyn &Dyn : Arr) {
2015-10-12 23:49:02 +08:00
if (Dyn.d_tag == DT_SONAME) {
uint64_t Val = Dyn.getVal();
2015-10-12 23:49:02 +08:00
if (Val >= this->StringTable.size())
fatal(toString(this) + ": invalid DT_SONAME entry");
SoName = this->StringTable.data() + Val;
2015-10-12 23:49:02 +08:00
return;
}
}
}
// Parses ".gnu.version" section which is a parallel array for the symbol table.
// If a given file doesn't have ".gnu.version" section, returns VER_NDX_GLOBAL.
template <class ELFT> std::vector<uint32_t> SharedFile<ELFT>::parseVersyms() {
size_t Size = this->ELFSyms.size() - this->FirstNonLocal;
if (!VersymSec)
return std::vector<uint32_t>(Size, VER_NDX_GLOBAL);
const char *Base = this->MB.getBuffer().data();
const Elf_Versym *Versym =
reinterpret_cast<const Elf_Versym *>(Base + VersymSec->sh_offset) +
this->FirstNonLocal;
std::vector<uint32_t> Ret(Size);
for (size_t I = 0; I < Size; ++I)
Ret[I] = Versym[I].vs_index;
return Ret;
}
// Parse the version definitions in the object file if present. Returns a vector
// whose nth element contains a pointer to the Elf_Verdef for version identifier
// n. Version identifiers that are not definitions map to nullptr.
template <class ELFT>
std::vector<const typename ELFT::Verdef *> SharedFile<ELFT>::parseVerdefs() {
if (!VerdefSec)
return {};
// We cannot determine the largest verdef identifier without inspecting
// every Elf_Verdef, but both bfd and gold assign verdef identifiers
// sequentially starting from 1, so we predict that the largest identifier
// will be VerdefCount.
unsigned VerdefCount = VerdefSec->sh_info;
std::vector<const Elf_Verdef *> Verdefs(VerdefCount + 1);
// Build the Verdefs array by following the chain of Elf_Verdef objects
// from the start of the .gnu.version_d section.
const char *Base = this->MB.getBuffer().data();
const char *Verdef = Base + VerdefSec->sh_offset;
for (unsigned I = 0; I != VerdefCount; ++I) {
auto *CurVerdef = reinterpret_cast<const Elf_Verdef *>(Verdef);
Verdef += CurVerdef->vd_next;
unsigned VerdefIndex = CurVerdef->vd_ndx;
if (Verdefs.size() <= VerdefIndex)
Verdefs.resize(VerdefIndex + 1);
Verdefs[VerdefIndex] = CurVerdef;
}
return Verdefs;
}
// We do not usually care about alignments of data in shared object
// files because the loader takes care of it. However, if we promote a
// DSO symbol to point to .bss due to copy relocation, we need to keep
// the original alignment requirements. We infer it in this function.
template <class ELFT>
uint32_t SharedFile<ELFT>::getAlignment(ArrayRef<Elf_Shdr> Sections,
const Elf_Sym &Sym) {
uint64_t Ret = 1;
if (Sym.st_value)
Ret = 1ULL << countTrailingZeros((uint64_t)Sym.st_value);
if (0 < Sym.st_shndx && Sym.st_shndx < Sections.size())
Ret = std::min<uint64_t>(Ret, Sections[Sym.st_shndx].sh_addralign);
if (Ret > UINT32_MAX)
error(toString(this) + ": alignment too large: " +
CHECK(Sym.getName(this->StringTable), this));
return Ret;
}
// Fully parse the shared object file. This must be called after parseSoName().
//
// This function parses symbol versions. If a DSO has version information,
// the file has a ".gnu.version_d" section which contains symbol version
// definitions. Each symbol is associated to one version through a table in
// ".gnu.version" section. That table is a parallel array for the symbol
// table, and each table entry contains an index in ".gnu.version_d".
//
// The special index 0 is reserved for VERF_NDX_LOCAL and 1 is for
// VER_NDX_GLOBAL. There's no table entry for these special versions in
// ".gnu.version_d".
//
// The file format for symbol versioning is perhaps a bit more complicated
// than necessary, but you can easily understand the code if you wrap your
// head around the data structure described above.
template <class ELFT> void SharedFile<ELFT>::parseRest() {
Verdefs = parseVerdefs(); // parse .gnu.version_d
std::vector<uint32_t> Versyms = parseVersyms(); // parse .gnu.version
ArrayRef<Elf_Shdr> Sections = CHECK(this->getObj().sections(), this);
// Add symbols to the symbol table.
ArrayRef<Elf_Sym> Syms = this->getGlobalELFSyms();
for (size_t I = 0; I < Syms.size(); ++I) {
const Elf_Sym &Sym = Syms[I];
StringRef Name = CHECK(Sym.getName(this->StringTable), this);
if (Sym.isUndefined()) {
Symbol *S = Symtab->addUndefined<ELFT>(Name, Sym.getBinding(),
Sym.st_other, Sym.getType(),
/*CanOmitFromDynSym=*/false, this);
S->ExportDynamic = true;
continue;
}
// ELF spec requires that all local symbols precede weak or global
// symbols in each symbol table, and the index of first non-local symbol
// is stored to sh_info. If a local symbol appears after some non-local
// symbol, that's a violation of the spec.
if (Sym.getBinding() == STB_LOCAL) {
2017-12-15 08:07:15 +08:00
warn("found local symbol '" + Name +
"' in global part of symbol table in file " + toString(this));
continue;
}
// MIPS BFD linker puts _gp_disp symbol into DSO files and incorrectly
// assigns VER_NDX_LOCAL to this section global symbol. Here is a
// workaround for this bug.
uint32_t Idx = Versyms[I] & ~VERSYM_HIDDEN;
if (Config->EMachine == EM_MIPS && Idx == VER_NDX_LOCAL &&
Name == "_gp_disp")
continue;
uint64_t Alignment = getAlignment(Sections, Sym);
if (!(Versyms[I] & VERSYM_HIDDEN))
Symtab->addShared(Name, *this, Sym, Alignment, Idx);
// Also add the symbol with the versioned name to handle undefined symbols
// with explicit versions.
if (Idx == VER_NDX_GLOBAL)
continue;
if (Idx >= Verdefs.size() || Idx == VER_NDX_LOCAL) {
error("corrupt input file: version definition index " + Twine(Idx) +
" for symbol " + Name + " is out of bounds\n>>> defined in " +
toString(this));
continue;
}
StringRef VerName =
this->StringTable.data() + Verdefs[Idx]->getAux()->vda_name;
Name = Saver.save(Name + "@" + VerName);
Symtab->addShared(Name, *this, Sym, Alignment, Idx);
}
}
static ELFKind getBitcodeELFKind(const Triple &T) {
if (T.isLittleEndian())
return T.isArch64Bit() ? ELF64LEKind : ELF32LEKind;
return T.isArch64Bit() ? ELF64BEKind : ELF32BEKind;
}
static uint8_t getBitcodeMachineKind(StringRef Path, const Triple &T) {
switch (T.getArch()) {
2016-07-07 10:46:30 +08:00
case Triple::aarch64:
return EM_AARCH64;
case Triple::arm:
case Triple::thumb:
2016-07-07 10:46:30 +08:00
return EM_ARM;
case Triple::avr:
return EM_AVR;
2016-07-07 10:46:30 +08:00
case Triple::mips:
case Triple::mipsel:
case Triple::mips64:
case Triple::mips64el:
return EM_MIPS;
case Triple::ppc:
return EM_PPC;
case Triple::ppc64:
return EM_PPC64;
case Triple::x86:
return T.isOSIAMCU() ? EM_IAMCU : EM_386;
2016-07-07 10:46:30 +08:00
case Triple::x86_64:
return EM_X86_64;
default:
fatal(Path + ": could not infer e_machine from bitcode target triple " +
T.str());
}
}
BitcodeFile::BitcodeFile(MemoryBufferRef MB, StringRef ArchiveName,
uint64_t OffsetInArchive)
: InputFile(BitcodeKind, MB) {
this->ArchiveName = ArchiveName;
// Here we pass a new MemoryBufferRef which is identified by ArchiveName
// (the fully resolved path of the archive) + member name + offset of the
// member in the archive.
// ThinLTO uses the MemoryBufferRef identifier to access its internal
// data structures and if two archives define two members with the same name,
// this causes a collision which result in only one of the objects being
// taken into consideration at LTO time (which very likely causes undefined
// symbols later in the link stage).
MemoryBufferRef MBRef(MB.getBuffer(),
Saver.save(ArchiveName + MB.getBufferIdentifier() +
utostr(OffsetInArchive)));
Obj = CHECK(lto::InputFile::create(MBRef), this);
Triple T(Obj->getTargetTriple());
EKind = getBitcodeELFKind(T);
EMachine = getBitcodeMachineKind(MB.getBufferIdentifier(), T);
}
static uint8_t mapVisibility(GlobalValue::VisibilityTypes GvVisibility) {
switch (GvVisibility) {
case GlobalValue::DefaultVisibility:
return STV_DEFAULT;
case GlobalValue::HiddenVisibility:
return STV_HIDDEN;
case GlobalValue::ProtectedVisibility:
return STV_PROTECTED;
}
llvm_unreachable("unknown visibility");
}
ELF: New symbol table design. This patch implements a new design for the symbol table that stores SymbolBodies within a memory region of the Symbol object. Symbols are mutated by constructing SymbolBodies in place over existing SymbolBodies, rather than by mutating pointers. As mentioned in the initial proposal [1], this memory layout helps reduce the cache miss rate by improving memory locality. Performance numbers: old(s) new(s) Without debug info: chrome 7.178 6.432 (-11.5%) LLVMgold.so 0.505 0.502 (-0.5%) clang 0.954 0.827 (-15.4%) llvm-as 0.052 0.045 (-15.5%) With debug info: scylla 5.695 5.613 (-1.5%) clang 14.396 14.143 (-1.8%) Performance counter results show that the fewer required indirections is indeed the cause of the improved performance. For example, when linking chrome, stalled cycles decreases from 14,556,444,002 to 12,959,238,310, and instructions per cycle increases from 0.78 to 0.83. We are also executing many fewer instructions (15,516,401,933 down to 15,002,434,310), probably because we spend less time allocating SymbolBodies. The new mechanism by which symbols are added to the symbol table is by calling add* functions on the SymbolTable. In this patch, I handle local symbols by storing them inside "unparented" SymbolBodies. This is suboptimal, but if we do want to try to avoid allocating these SymbolBodies, we can probably do that separately. I also removed a few members from the SymbolBody class that were only being used to pass information from the input file to the symbol table. This patch implements the new design for the ELF linker only. I intend to prepare a similar patch for the COFF linker. [1] http://lists.llvm.org/pipermail/llvm-dev/2016-April/098832.html Differential Revision: http://reviews.llvm.org/D19752 llvm-svn: 268178
2016-05-01 12:55:03 +08:00
template <class ELFT>
static Symbol *createBitcodeSymbol(const std::vector<bool> &KeptComdats,
const lto::InputFile::Symbol &ObjSym,
BitcodeFile &F) {
StringRef NameRef = Saver.save(ObjSym.getName());
uint32_t Binding = ObjSym.isWeak() ? STB_WEAK : STB_GLOBAL;
uint8_t Type = ObjSym.isTLS() ? STT_TLS : STT_NOTYPE;
uint8_t Visibility = mapVisibility(ObjSym.getVisibility());
bool CanOmitFromDynSym = ObjSym.canBeOmittedFromSymbolTable();
int C = ObjSym.getComdatIndex();
if (C != -1 && !KeptComdats[C])
return Symtab->addUndefined<ELFT>(NameRef, Binding, Visibility, Type,
CanOmitFromDynSym, &F);
if (ObjSym.isUndefined())
return Symtab->addUndefined<ELFT>(NameRef, Binding, Visibility, Type,
CanOmitFromDynSym, &F);
if (ObjSym.isCommon())
return Symtab->addCommon(NameRef, ObjSym.getCommonSize(),
ObjSym.getCommonAlignment(), Binding, Visibility,
STT_OBJECT, F);
return Symtab->addBitcode(NameRef, Binding, Visibility, Type,
CanOmitFromDynSym, F);
}
ELF: New symbol table design. This patch implements a new design for the symbol table that stores SymbolBodies within a memory region of the Symbol object. Symbols are mutated by constructing SymbolBodies in place over existing SymbolBodies, rather than by mutating pointers. As mentioned in the initial proposal [1], this memory layout helps reduce the cache miss rate by improving memory locality. Performance numbers: old(s) new(s) Without debug info: chrome 7.178 6.432 (-11.5%) LLVMgold.so 0.505 0.502 (-0.5%) clang 0.954 0.827 (-15.4%) llvm-as 0.052 0.045 (-15.5%) With debug info: scylla 5.695 5.613 (-1.5%) clang 14.396 14.143 (-1.8%) Performance counter results show that the fewer required indirections is indeed the cause of the improved performance. For example, when linking chrome, stalled cycles decreases from 14,556,444,002 to 12,959,238,310, and instructions per cycle increases from 0.78 to 0.83. We are also executing many fewer instructions (15,516,401,933 down to 15,002,434,310), probably because we spend less time allocating SymbolBodies. The new mechanism by which symbols are added to the symbol table is by calling add* functions on the SymbolTable. In this patch, I handle local symbols by storing them inside "unparented" SymbolBodies. This is suboptimal, but if we do want to try to avoid allocating these SymbolBodies, we can probably do that separately. I also removed a few members from the SymbolBody class that were only being used to pass information from the input file to the symbol table. This patch implements the new design for the ELF linker only. I intend to prepare a similar patch for the COFF linker. [1] http://lists.llvm.org/pipermail/llvm-dev/2016-April/098832.html Differential Revision: http://reviews.llvm.org/D19752 llvm-svn: 268178
2016-05-01 12:55:03 +08:00
template <class ELFT>
void BitcodeFile::parse(DenseSet<CachedHashStringRef> &ComdatGroups) {
std::vector<bool> KeptComdats;
for (StringRef S : Obj->getComdatTable())
KeptComdats.push_back(ComdatGroups.insert(CachedHashStringRef(S)).second);
for (const lto::InputFile::Symbol &ObjSym : Obj->symbols())
Symbols.push_back(createBitcodeSymbol<ELFT>(KeptComdats, ObjSym, *this));
}
static ELFKind getELFKind(MemoryBufferRef MB) {
unsigned char Size;
unsigned char Endian;
std::tie(Size, Endian) = getElfArchType(MB.getBuffer());
if (Endian != ELFDATA2LSB && Endian != ELFDATA2MSB)
fatal(MB.getBufferIdentifier() + ": invalid data encoding");
if (Size != ELFCLASS32 && Size != ELFCLASS64)
fatal(MB.getBufferIdentifier() + ": invalid file class");
size_t BufSize = MB.getBuffer().size();
if ((Size == ELFCLASS32 && BufSize < sizeof(Elf32_Ehdr)) ||
(Size == ELFCLASS64 && BufSize < sizeof(Elf64_Ehdr)))
fatal(MB.getBufferIdentifier() + ": file is too short");
if (Size == ELFCLASS32)
return (Endian == ELFDATA2LSB) ? ELF32LEKind : ELF32BEKind;
return (Endian == ELFDATA2LSB) ? ELF64LEKind : ELF64BEKind;
}
void BinaryFile::parse() {
ArrayRef<uint8_t> Data = toArrayRef(MB.getBuffer());
auto *Section = make<InputSection>(this, SHF_ALLOC | SHF_WRITE, SHT_PROGBITS,
8, Data, ".data");
Sections.push_back(Section);
// For each input file foo that is embedded to a result as a binary
// blob, we define _binary_foo_{start,end,size} symbols, so that
// user programs can access blobs by name. Non-alphanumeric
// characters in a filename are replaced with underscore.
std::string S = "_binary_" + MB.getBufferIdentifier().str();
for (size_t I = 0; I < S.size(); ++I)
if (!isAlnum(S[I]))
S[I] = '_';
Symtab->addRegular(Saver.save(S + "_start"), STV_DEFAULT, STT_OBJECT, 0, 0,
STB_GLOBAL, Section, nullptr);
Symtab->addRegular(Saver.save(S + "_end"), STV_DEFAULT, STT_OBJECT,
Data.size(), 0, STB_GLOBAL, Section, nullptr);
Symtab->addRegular(Saver.save(S + "_size"), STV_DEFAULT, STT_OBJECT,
Data.size(), 0, STB_GLOBAL, nullptr, nullptr);
}
static bool isBitcode(MemoryBufferRef MB) {
using namespace sys::fs;
return identify_magic(MB.getBuffer()) == file_magic::bitcode;
}
InputFile *elf::createObjectFile(MemoryBufferRef MB, StringRef ArchiveName,
uint64_t OffsetInArchive) {
if (isBitcode(MB))
return make<BitcodeFile>(MB, ArchiveName, OffsetInArchive);
switch (getELFKind(MB)) {
case ELF32LEKind:
return make<ObjFile<ELF32LE>>(MB, ArchiveName);
case ELF32BEKind:
return make<ObjFile<ELF32BE>>(MB, ArchiveName);
case ELF64LEKind:
return make<ObjFile<ELF64LE>>(MB, ArchiveName);
case ELF64BEKind:
return make<ObjFile<ELF64BE>>(MB, ArchiveName);
default:
llvm_unreachable("getELFKind");
}
}
InputFile *elf::createSharedFile(MemoryBufferRef MB, StringRef DefaultSoName) {
switch (getELFKind(MB)) {
case ELF32LEKind:
return make<SharedFile<ELF32LE>>(MB, DefaultSoName);
case ELF32BEKind:
return make<SharedFile<ELF32BE>>(MB, DefaultSoName);
case ELF64LEKind:
return make<SharedFile<ELF64LE>>(MB, DefaultSoName);
case ELF64BEKind:
return make<SharedFile<ELF64BE>>(MB, DefaultSoName);
default:
llvm_unreachable("getELFKind");
}
}
MemoryBufferRef LazyObjFile::getBuffer() {
if (Seen)
return MemoryBufferRef();
Seen = true;
return MB;
}
InputFile *LazyObjFile::fetch() {
MemoryBufferRef MBRef = getBuffer();
if (MBRef.getBuffer().empty())
return nullptr;
return createObjectFile(MBRef, ArchiveName, OffsetInArchive);
}
template <class ELFT> void LazyObjFile::parse() {
// A lazy object file wraps either a bitcode file or an ELF file.
if (isBitcode(this->MB)) {
std::unique_ptr<lto::InputFile> Obj =
CHECK(lto::InputFile::create(this->MB), this);
for (const lto::InputFile::Symbol &Sym : Obj->symbols())
if (!Sym.isUndefined())
Symtab->addLazyObject<ELFT>(Saver.save(Sym.getName()), *this);
return;
}
switch (getELFKind(this->MB)) {
case ELF32LEKind:
addElfSymbols<ELF32LE>();
return;
case ELF32BEKind:
addElfSymbols<ELF32BE>();
return;
case ELF64LEKind:
addElfSymbols<ELF64LE>();
return;
case ELF64BEKind:
addElfSymbols<ELF64BE>();
return;
default:
llvm_unreachable("getELFKind");
}
}
template <class ELFT> void LazyObjFile::addElfSymbols() {
ELFFile<ELFT> Obj = check(ELFFile<ELFT>::create(MB.getBuffer()));
ArrayRef<typename ELFT::Shdr> Sections = CHECK(Obj.sections(), this);
for (const typename ELFT::Shdr &Sec : Sections) {
if (Sec.sh_type != SHT_SYMTAB)
continue;
typename ELFT::SymRange Syms = CHECK(Obj.symbols(&Sec), this);
uint32_t FirstNonLocal = Sec.sh_info;
StringRef StringTable =
CHECK(Obj.getStringTableForSymtab(Sec, Sections), this);
for (const typename ELFT::Sym &Sym : Syms.slice(FirstNonLocal))
if (Sym.st_shndx != SHN_UNDEF)
Symtab->addLazyObject<ELFT>(CHECK(Sym.getName(StringTable), this),
*this);
return;
}
}
// This is for --just-symbols.
//
// This option allows you to link your output against other existing
// program, so that if you load both your program and the other program
// into memory, your output can use program's symbols.
//
// What we are doing here is to read defined symbols from a given ELF
// file and add them as absolute symbols.
template <class ELFT> void elf::readJustSymbolsFile(MemoryBufferRef MB) {
typedef typename ELFT::Shdr Elf_Shdr;
typedef typename ELFT::Sym Elf_Sym;
typedef typename ELFT::SymRange Elf_Sym_Range;
StringRef ObjName = MB.getBufferIdentifier();
ELFFile<ELFT> Obj = check(ELFFile<ELFT>::create(MB.getBuffer()));
ArrayRef<Elf_Shdr> Sections = CHECK(Obj.sections(), ObjName);
for (const Elf_Shdr &Sec : Sections) {
if (Sec.sh_type != SHT_SYMTAB)
continue;
Elf_Sym_Range Syms = CHECK(Obj.symbols(&Sec), ObjName);
uint32_t FirstNonLocal = Sec.sh_info;
StringRef StringTable =
CHECK(Obj.getStringTableForSymtab(Sec, Sections), ObjName);
for (const Elf_Sym &Sym : Syms.slice(FirstNonLocal))
if (Sym.st_shndx != SHN_UNDEF)
Symtab->addRegular(CHECK(Sym.getName(StringTable), ObjName),
Sym.st_other, Sym.getType(), Sym.st_value,
Sym.st_size, Sym.getBinding(), nullptr, nullptr);
return;
}
}
ELF: New symbol table design. This patch implements a new design for the symbol table that stores SymbolBodies within a memory region of the Symbol object. Symbols are mutated by constructing SymbolBodies in place over existing SymbolBodies, rather than by mutating pointers. As mentioned in the initial proposal [1], this memory layout helps reduce the cache miss rate by improving memory locality. Performance numbers: old(s) new(s) Without debug info: chrome 7.178 6.432 (-11.5%) LLVMgold.so 0.505 0.502 (-0.5%) clang 0.954 0.827 (-15.4%) llvm-as 0.052 0.045 (-15.5%) With debug info: scylla 5.695 5.613 (-1.5%) clang 14.396 14.143 (-1.8%) Performance counter results show that the fewer required indirections is indeed the cause of the improved performance. For example, when linking chrome, stalled cycles decreases from 14,556,444,002 to 12,959,238,310, and instructions per cycle increases from 0.78 to 0.83. We are also executing many fewer instructions (15,516,401,933 down to 15,002,434,310), probably because we spend less time allocating SymbolBodies. The new mechanism by which symbols are added to the symbol table is by calling add* functions on the SymbolTable. In this patch, I handle local symbols by storing them inside "unparented" SymbolBodies. This is suboptimal, but if we do want to try to avoid allocating these SymbolBodies, we can probably do that separately. I also removed a few members from the SymbolBody class that were only being used to pass information from the input file to the symbol table. This patch implements the new design for the ELF linker only. I intend to prepare a similar patch for the COFF linker. [1] http://lists.llvm.org/pipermail/llvm-dev/2016-April/098832.html Differential Revision: http://reviews.llvm.org/D19752 llvm-svn: 268178
2016-05-01 12:55:03 +08:00
template void ArchiveFile::parse<ELF32LE>();
template void ArchiveFile::parse<ELF32BE>();
template void ArchiveFile::parse<ELF64LE>();
template void ArchiveFile::parse<ELF64BE>();
template void BitcodeFile::parse<ELF32LE>(DenseSet<CachedHashStringRef> &);
template void BitcodeFile::parse<ELF32BE>(DenseSet<CachedHashStringRef> &);
template void BitcodeFile::parse<ELF64LE>(DenseSet<CachedHashStringRef> &);
template void BitcodeFile::parse<ELF64BE>(DenseSet<CachedHashStringRef> &);
ELF: New symbol table design. This patch implements a new design for the symbol table that stores SymbolBodies within a memory region of the Symbol object. Symbols are mutated by constructing SymbolBodies in place over existing SymbolBodies, rather than by mutating pointers. As mentioned in the initial proposal [1], this memory layout helps reduce the cache miss rate by improving memory locality. Performance numbers: old(s) new(s) Without debug info: chrome 7.178 6.432 (-11.5%) LLVMgold.so 0.505 0.502 (-0.5%) clang 0.954 0.827 (-15.4%) llvm-as 0.052 0.045 (-15.5%) With debug info: scylla 5.695 5.613 (-1.5%) clang 14.396 14.143 (-1.8%) Performance counter results show that the fewer required indirections is indeed the cause of the improved performance. For example, when linking chrome, stalled cycles decreases from 14,556,444,002 to 12,959,238,310, and instructions per cycle increases from 0.78 to 0.83. We are also executing many fewer instructions (15,516,401,933 down to 15,002,434,310), probably because we spend less time allocating SymbolBodies. The new mechanism by which symbols are added to the symbol table is by calling add* functions on the SymbolTable. In this patch, I handle local symbols by storing them inside "unparented" SymbolBodies. This is suboptimal, but if we do want to try to avoid allocating these SymbolBodies, we can probably do that separately. I also removed a few members from the SymbolBody class that were only being used to pass information from the input file to the symbol table. This patch implements the new design for the ELF linker only. I intend to prepare a similar patch for the COFF linker. [1] http://lists.llvm.org/pipermail/llvm-dev/2016-April/098832.html Differential Revision: http://reviews.llvm.org/D19752 llvm-svn: 268178
2016-05-01 12:55:03 +08:00
template void LazyObjFile::parse<ELF32LE>();
template void LazyObjFile::parse<ELF32BE>();
template void LazyObjFile::parse<ELF64LE>();
template void LazyObjFile::parse<ELF64BE>();
ELF: New symbol table design. This patch implements a new design for the symbol table that stores SymbolBodies within a memory region of the Symbol object. Symbols are mutated by constructing SymbolBodies in place over existing SymbolBodies, rather than by mutating pointers. As mentioned in the initial proposal [1], this memory layout helps reduce the cache miss rate by improving memory locality. Performance numbers: old(s) new(s) Without debug info: chrome 7.178 6.432 (-11.5%) LLVMgold.so 0.505 0.502 (-0.5%) clang 0.954 0.827 (-15.4%) llvm-as 0.052 0.045 (-15.5%) With debug info: scylla 5.695 5.613 (-1.5%) clang 14.396 14.143 (-1.8%) Performance counter results show that the fewer required indirections is indeed the cause of the improved performance. For example, when linking chrome, stalled cycles decreases from 14,556,444,002 to 12,959,238,310, and instructions per cycle increases from 0.78 to 0.83. We are also executing many fewer instructions (15,516,401,933 down to 15,002,434,310), probably because we spend less time allocating SymbolBodies. The new mechanism by which symbols are added to the symbol table is by calling add* functions on the SymbolTable. In this patch, I handle local symbols by storing them inside "unparented" SymbolBodies. This is suboptimal, but if we do want to try to avoid allocating these SymbolBodies, we can probably do that separately. I also removed a few members from the SymbolBody class that were only being used to pass information from the input file to the symbol table. This patch implements the new design for the ELF linker only. I intend to prepare a similar patch for the COFF linker. [1] http://lists.llvm.org/pipermail/llvm-dev/2016-April/098832.html Differential Revision: http://reviews.llvm.org/D19752 llvm-svn: 268178
2016-05-01 12:55:03 +08:00
2016-02-28 08:25:54 +08:00
template class elf::ELFFileBase<ELF32LE>;
template class elf::ELFFileBase<ELF32BE>;
template class elf::ELFFileBase<ELF64LE>;
template class elf::ELFFileBase<ELF64BE>;
template class elf::ObjFile<ELF32LE>;
template class elf::ObjFile<ELF32BE>;
template class elf::ObjFile<ELF64LE>;
template class elf::ObjFile<ELF64BE>;
2016-02-28 08:25:54 +08:00
template class elf::SharedFile<ELF32LE>;
template class elf::SharedFile<ELF32BE>;
template class elf::SharedFile<ELF64LE>;
template class elf::SharedFile<ELF64BE>;
template void elf::readJustSymbolsFile<ELF32LE>(MemoryBufferRef);
template void elf::readJustSymbolsFile<ELF32BE>(MemoryBufferRef);
template void elf::readJustSymbolsFile<ELF64LE>(MemoryBufferRef);
template void elf::readJustSymbolsFile<ELF64BE>(MemoryBufferRef);