diff options
Diffstat (limited to 'lib/MC')
44 files changed, 4624 insertions, 2278 deletions
diff --git a/lib/MC/CMakeLists.txt b/lib/MC/CMakeLists.txt index 5856e07..6aed059 100644 --- a/lib/MC/CMakeLists.txt +++ b/lib/MC/CMakeLists.txt @@ -8,6 +8,8 @@ add_llvm_library(LLVMMC MCCodeEmitter.cpp MCContext.cpp MCDisassembler.cpp + MCELF.cpp + MCELFObjectTargetWriter.cpp MCELFStreamer.cpp MCExpr.cpp MCInst.cpp @@ -16,10 +18,11 @@ add_llvm_library(LLVMMC MCDwarf.cpp MCLoggingStreamer.cpp MCMachOStreamer.cpp + MCMachObjectTargetWriter.cpp MCNullStreamer.cpp MCObjectStreamer.cpp - MCObjectFormat.cpp MCObjectWriter.cpp + MCPureStreamer.cpp MCSection.cpp MCSectionCOFF.cpp MCSectionELF.cpp @@ -32,3 +35,6 @@ add_llvm_library(LLVMMC WinCOFFObjectWriter.cpp TargetAsmBackend.cpp ) + +add_subdirectory(MCParser) +add_subdirectory(MCDisassembler) diff --git a/lib/MC/ELFObjectWriter.cpp b/lib/MC/ELFObjectWriter.cpp index b15c2d4..3492cbc 100644 --- a/lib/MC/ELFObjectWriter.cpp +++ b/lib/MC/ELFObjectWriter.cpp @@ -11,73 +11,35 @@ // //===----------------------------------------------------------------------===// -#include "llvm/ADT/SmallPtrSet.h" +#include "ELFObjectWriter.h" #include "llvm/ADT/STLExtras.h" #include "llvm/ADT/StringMap.h" #include "llvm/ADT/Twine.h" -#include "llvm/MC/MCAssembler.h" #include "llvm/MC/MCAsmLayout.h" #include "llvm/MC/MCContext.h" -#include "llvm/MC/MCELFSymbolFlags.h" #include "llvm/MC/MCExpr.h" -#include "llvm/MC/MCObjectWriter.h" #include "llvm/MC/MCSectionELF.h" -#include "llvm/MC/MCSymbol.h" #include "llvm/MC/MCValue.h" #include "llvm/Support/Debug.h" #include "llvm/Support/ErrorHandling.h" #include "llvm/Support/ELF.h" #include "llvm/Target/TargetAsmBackend.h" +#include "llvm/ADT/StringSwitch.h" #include "../Target/X86/X86FixupKinds.h" +#include "../Target/ARM/ARMFixupKinds.h" #include <vector> using namespace llvm; -static unsigned GetType(const MCSymbolData &SD) { - uint32_t Type = (SD.getFlags() & (0xf << ELF_STT_Shift)) >> ELF_STT_Shift; - assert(Type == ELF::STT_NOTYPE || Type == ELF::STT_OBJECT || - Type == ELF::STT_FUNC || Type == ELF::STT_SECTION || - Type == ELF::STT_FILE || Type == ELF::STT_COMMON || - Type == ELF::STT_TLS); - return Type; -} - -static unsigned GetBinding(const MCSymbolData &SD) { - uint32_t Binding = (SD.getFlags() & (0xf << ELF_STB_Shift)) >> ELF_STB_Shift; - assert(Binding == ELF::STB_LOCAL || Binding == ELF::STB_GLOBAL || - Binding == ELF::STB_WEAK); - return Binding; -} - -static void SetBinding(MCSymbolData &SD, unsigned Binding) { - assert(Binding == ELF::STB_LOCAL || Binding == ELF::STB_GLOBAL || - Binding == ELF::STB_WEAK); - uint32_t OtherFlags = SD.getFlags() & ~(0xf << ELF_STB_Shift); - SD.setFlags(OtherFlags | (Binding << ELF_STB_Shift)); -} +bool ELFObjectWriter::isFixupKindPCRel(const MCAssembler &Asm, unsigned Kind) { + const MCFixupKindInfo &FKI = + Asm.getBackend().getFixupKindInfo((MCFixupKind) Kind); -static unsigned GetVisibility(MCSymbolData &SD) { - unsigned Visibility = - (SD.getFlags() & (0xf << ELF_STV_Shift)) >> ELF_STV_Shift; - assert(Visibility == ELF::STV_DEFAULT || Visibility == ELF::STV_INTERNAL || - Visibility == ELF::STV_HIDDEN || Visibility == ELF::STV_PROTECTED); - return Visibility; + return FKI.Flags & MCFixupKindInfo::FKF_IsPCRel; } -static bool isFixupKindX86PCRel(unsigned Kind) { - switch (Kind) { - default: - return false; - case X86::reloc_pcrel_1byte: - case X86::reloc_pcrel_4byte: - case X86::reloc_riprel_4byte: - case X86::reloc_riprel_4byte_movq_load: - return true; - } -} - -static bool RelocNeedsGOT(MCSymbolRefExpr::VariantKind Variant) { +bool ELFObjectWriter::RelocNeedsGOT(MCSymbolRefExpr::VariantKind Variant) { switch (Variant) { default: return false; @@ -97,294 +59,6 @@ static bool RelocNeedsGOT(MCSymbolRefExpr::VariantKind Variant) { } } -namespace { - class ELFObjectWriter : public MCObjectWriter { - protected: - /*static bool isFixupKindX86RIPRel(unsigned Kind) { - return Kind == X86::reloc_riprel_4byte || - Kind == X86::reloc_riprel_4byte_movq_load; - }*/ - - - /// ELFSymbolData - Helper struct for containing some precomputed information - /// on symbols. - struct ELFSymbolData { - MCSymbolData *SymbolData; - uint64_t StringIndex; - uint32_t SectionIndex; - - // Support lexicographic sorting. - bool operator<(const ELFSymbolData &RHS) const { - if (GetType(*SymbolData) == ELF::STT_FILE) - return true; - if (GetType(*RHS.SymbolData) == ELF::STT_FILE) - return false; - return SymbolData->getSymbol().getName() < - RHS.SymbolData->getSymbol().getName(); - } - }; - - /// @name Relocation Data - /// @{ - - struct ELFRelocationEntry { - // Make these big enough for both 32-bit and 64-bit - uint64_t r_offset; - int Index; - unsigned Type; - const MCSymbol *Symbol; - uint64_t r_addend; - - // Support lexicographic sorting. - bool operator<(const ELFRelocationEntry &RE) const { - return RE.r_offset < r_offset; - } - }; - - SmallPtrSet<const MCSymbol *, 16> UsedInReloc; - SmallPtrSet<const MCSymbol *, 16> WeakrefUsedInReloc; - DenseMap<const MCSymbol *, const MCSymbol *> Renames; - - llvm::DenseMap<const MCSectionData*, - std::vector<ELFRelocationEntry> > Relocations; - DenseMap<const MCSection*, uint64_t> SectionStringTableIndex; - - /// @} - /// @name Symbol Table Data - /// @{ - - SmallString<256> StringTable; - std::vector<ELFSymbolData> LocalSymbolData; - std::vector<ELFSymbolData> ExternalSymbolData; - std::vector<ELFSymbolData> UndefinedSymbolData; - - /// @} - - bool NeedsGOT; - - bool NeedsSymtabShndx; - - unsigned Is64Bit : 1; - - bool HasRelocationAddend; - - Triple::OSType OSType; - - uint16_t EMachine; - - // This holds the symbol table index of the last local symbol. - unsigned LastLocalSymbolIndex; - // This holds the .strtab section index. - unsigned StringTableIndex; - // This holds the .symtab section index. - unsigned SymbolTableIndex; - - unsigned ShstrtabIndex; - - - const MCSymbol *SymbolToReloc(const MCAssembler &Asm, - const MCValue &Target, - const MCFragment &F) const; - - public: - ELFObjectWriter(raw_ostream &_OS, bool _Is64Bit, bool IsLittleEndian, - uint16_t _EMachine, bool _HasRelAddend, - Triple::OSType _OSType) - : MCObjectWriter(_OS, IsLittleEndian), - NeedsGOT(false), NeedsSymtabShndx(false), - Is64Bit(_Is64Bit), HasRelocationAddend(_HasRelAddend), - OSType(_OSType), EMachine(_EMachine) { - } - - virtual ~ELFObjectWriter(); - - void WriteWord(uint64_t W) { - if (Is64Bit) - Write64(W); - else - Write32(W); - } - - void StringLE16(char *buf, uint16_t Value) { - buf[0] = char(Value >> 0); - buf[1] = char(Value >> 8); - } - - void StringLE32(char *buf, uint32_t Value) { - StringLE16(buf, uint16_t(Value >> 0)); - StringLE16(buf + 2, uint16_t(Value >> 16)); - } - - void StringLE64(char *buf, uint64_t Value) { - StringLE32(buf, uint32_t(Value >> 0)); - StringLE32(buf + 4, uint32_t(Value >> 32)); - } - - void StringBE16(char *buf ,uint16_t Value) { - buf[0] = char(Value >> 8); - buf[1] = char(Value >> 0); - } - - void StringBE32(char *buf, uint32_t Value) { - StringBE16(buf, uint16_t(Value >> 16)); - StringBE16(buf + 2, uint16_t(Value >> 0)); - } - - void StringBE64(char *buf, uint64_t Value) { - StringBE32(buf, uint32_t(Value >> 32)); - StringBE32(buf + 4, uint32_t(Value >> 0)); - } - - void String8(MCDataFragment &F, uint8_t Value) { - char buf[1]; - buf[0] = Value; - F.getContents() += StringRef(buf, 1); - } - - void String16(MCDataFragment &F, uint16_t Value) { - char buf[2]; - if (isLittleEndian()) - StringLE16(buf, Value); - else - StringBE16(buf, Value); - F.getContents() += StringRef(buf, 2); - } - - void String32(MCDataFragment &F, uint32_t Value) { - char buf[4]; - if (isLittleEndian()) - StringLE32(buf, Value); - else - StringBE32(buf, Value); - F.getContents() += StringRef(buf, 4); - } - - void String64(MCDataFragment &F, uint64_t Value) { - char buf[8]; - if (isLittleEndian()) - StringLE64(buf, Value); - else - StringBE64(buf, Value); - F.getContents() += StringRef(buf, 8); - } - - virtual void WriteHeader(uint64_t SectionDataSize, unsigned NumberOfSections); - - virtual void WriteSymbolEntry(MCDataFragment *SymtabF, MCDataFragment *ShndxF, - uint64_t name, uint8_t info, - uint64_t value, uint64_t size, - uint8_t other, uint32_t shndx, - bool Reserved); - - virtual void WriteSymbol(MCDataFragment *SymtabF, MCDataFragment *ShndxF, - ELFSymbolData &MSD, - const MCAsmLayout &Layout); - - typedef DenseMap<const MCSectionELF*, uint32_t> SectionIndexMapTy; - virtual void WriteSymbolTable(MCDataFragment *SymtabF, MCDataFragment *ShndxF, - const MCAssembler &Asm, - const MCAsmLayout &Layout, - const SectionIndexMapTy &SectionIndexMap); - - virtual void RecordRelocation(const MCAssembler &Asm, const MCAsmLayout &Layout, - const MCFragment *Fragment, const MCFixup &Fixup, - MCValue Target, uint64_t &FixedValue) { - assert(0 && "RecordRelocation is not specific enough"); - } - - virtual uint64_t getSymbolIndexInSymbolTable(const MCAssembler &Asm, - const MCSymbol *S); - - // Map from a group section to the signature symbol - typedef DenseMap<const MCSectionELF*, const MCSymbol*> GroupMapTy; - // Map from a signature symbol to the group section - typedef DenseMap<const MCSymbol*, const MCSectionELF*> RevGroupMapTy; - - /// ComputeSymbolTable - Compute the symbol table data - /// - /// \param StringTable [out] - The string table data. - /// \param StringIndexMap [out] - Map from symbol names to offsets in the - /// string table. - virtual void ComputeSymbolTable(MCAssembler &Asm, - const SectionIndexMapTy &SectionIndexMap, - RevGroupMapTy RevGroupMap); - - virtual void ComputeIndexMap(MCAssembler &Asm, - SectionIndexMapTy &SectionIndexMap); - - virtual void WriteRelocation(MCAssembler &Asm, MCAsmLayout &Layout, - const MCSectionData &SD); - - virtual void WriteRelocations(MCAssembler &Asm, MCAsmLayout &Layout) { - for (MCAssembler::const_iterator it = Asm.begin(), - ie = Asm.end(); it != ie; ++it) { - WriteRelocation(Asm, Layout, *it); - } - } - - virtual void CreateMetadataSections(MCAssembler &Asm, MCAsmLayout &Layout, - const SectionIndexMapTy &SectionIndexMap); - - virtual void CreateGroupSections(MCAssembler &Asm, MCAsmLayout &Layout, - GroupMapTy &GroupMap, RevGroupMapTy &RevGroupMap); - - virtual void ExecutePostLayoutBinding(MCAssembler &Asm); - - virtual void WriteSecHdrEntry(uint32_t Name, uint32_t Type, uint64_t Flags, - uint64_t Address, uint64_t Offset, - uint64_t Size, uint32_t Link, uint32_t Info, - uint64_t Alignment, uint64_t EntrySize); - - virtual void WriteRelocationsFragment(const MCAssembler &Asm, MCDataFragment *F, - const MCSectionData *SD); - - virtual bool IsFixupFullyResolved(const MCAssembler &Asm, - const MCValue Target, - bool IsPCRel, - const MCFragment *DF) const; - - virtual void WriteObject(MCAssembler &Asm, const MCAsmLayout &Layout); - virtual void WriteSection(MCAssembler &Asm, - const SectionIndexMapTy &SectionIndexMap, - uint32_t GroupSymbolIndex, - uint64_t Offset, uint64_t Size, uint64_t Alignment, - const MCSectionELF &Section); - }; - - //===- X86ELFObjectWriter -------------------------------------------===// - - class X86ELFObjectWriter : public ELFObjectWriter { - public: - X86ELFObjectWriter(raw_ostream &_OS, bool _Is64Bit, bool IsLittleEndian, - uint16_t _EMachine, bool _HasRelAddend, - Triple::OSType _OSType); - - virtual ~X86ELFObjectWriter(); - virtual void RecordRelocation(const MCAssembler &Asm, - const MCAsmLayout &Layout, - const MCFragment *Fragment, - const MCFixup &Fixup, MCValue Target, - uint64_t &FixedValue); - }; - - - //===- ARMELFObjectWriter -------------------------------------------===// - - class ARMELFObjectWriter : public ELFObjectWriter { - public: - ARMELFObjectWriter(raw_ostream &_OS, bool _Is64Bit, bool IsLittleEndian, - uint16_t _EMachine, bool _HasRelAddend, - Triple::OSType _OSType); - - virtual ~ARMELFObjectWriter(); - virtual void RecordRelocation(const MCAssembler &Asm, - const MCAsmLayout &Layout, - const MCFragment *Fragment, - const MCFixup &Fixup, MCValue Target, - uint64_t &FixedValue); - }; -} - ELFObjectWriter::~ELFObjectWriter() {} @@ -404,14 +78,14 @@ void ELFObjectWriter::WriteHeader(uint64_t SectionDataSize, Write8('L'); // e_ident[EI_MAG2] Write8('F'); // e_ident[EI_MAG3] - Write8(Is64Bit ? ELF::ELFCLASS64 : ELF::ELFCLASS32); // e_ident[EI_CLASS] + Write8(is64Bit() ? ELF::ELFCLASS64 : ELF::ELFCLASS32); // e_ident[EI_CLASS] // e_ident[EI_DATA] Write8(isLittleEndian() ? ELF::ELFDATA2LSB : ELF::ELFDATA2MSB); Write8(ELF::EV_CURRENT); // e_ident[EI_VERSION] // e_ident[EI_OSABI] - switch (OSType) { + switch (TargetObjectWriter->getOSType()) { case Triple::FreeBSD: Write8(ELF::ELFOSABI_FREEBSD); break; case Triple::Linux: Write8(ELF::ELFOSABI_LINUX); break; default: Write8(ELF::ELFOSABI_NONE); break; @@ -422,25 +96,25 @@ void ELFObjectWriter::WriteHeader(uint64_t SectionDataSize, Write16(ELF::ET_REL); // e_type - Write16(EMachine); // e_machine = target + Write16(TargetObjectWriter->getEMachine()); // e_machine = target Write32(ELF::EV_CURRENT); // e_version WriteWord(0); // e_entry, no entry point in .o file WriteWord(0); // e_phoff, no program header for .o - WriteWord(SectionDataSize + (Is64Bit ? sizeof(ELF::Elf64_Ehdr) : + WriteWord(SectionDataSize + (is64Bit() ? sizeof(ELF::Elf64_Ehdr) : sizeof(ELF::Elf32_Ehdr))); // e_shoff = sec hdr table off in bytes - // FIXME: Make this configurable. - Write32(0); // e_flags = whatever the target wants + // e_flags = whatever the target wants + WriteEFlags(); // e_ehsize = ELF header size - Write16(Is64Bit ? sizeof(ELF::Elf64_Ehdr) : sizeof(ELF::Elf32_Ehdr)); + Write16(is64Bit() ? sizeof(ELF::Elf64_Ehdr) : sizeof(ELF::Elf32_Ehdr)); Write16(0); // e_phentsize = prog header entry size Write16(0); // e_phnum = # prog header entries = 0 // e_shentsize = Section header entry size - Write16(Is64Bit ? sizeof(ELF::Elf64_Shdr) : sizeof(ELF::Elf32_Shdr)); + Write16(is64Bit() ? sizeof(ELF::Elf64_Shdr) : sizeof(ELF::Elf32_Shdr)); // e_shnum = # of section header ents if (NumberOfSections >= ELF::SHN_LORESERVE) @@ -472,7 +146,7 @@ void ELFObjectWriter::WriteSymbolEntry(MCDataFragment *SymtabF, uint16_t Index = (shndx >= ELF::SHN_LORESERVE && !Reserved) ? uint16_t(ELF::SHN_XINDEX) : shndx; - if (Is64Bit) { + if (is64Bit()) { String32(*SymtabF, name); // st_name String8(*SymtabF, info); // st_info String8(*SymtabF, other); // st_other @@ -489,22 +163,32 @@ void ELFObjectWriter::WriteSymbolEntry(MCDataFragment *SymtabF, } } -static uint64_t SymbolValue(MCSymbolData &Data, const MCAsmLayout &Layout) { +uint64_t ELFObjectWriter::SymbolValue(MCSymbolData &Data, + const MCAsmLayout &Layout) { if (Data.isCommon() && Data.isExternal()) return Data.getCommonAlignment(); const MCSymbol &Symbol = Data.getSymbol(); + + if (Symbol.isAbsolute() && Symbol.isVariable()) { + if (const MCExpr *Value = Symbol.getVariableValue()) { + int64_t IntValue; + if (Value->EvaluateAsAbsolute(IntValue, Layout)) + return (uint64_t)IntValue; + } + } + if (!Symbol.isInSection()) return 0; - if (MCFragment *FF = Data.getFragment()) - return Layout.getSymbolAddress(&Data) - - Layout.getSectionAddress(FF->getParent()); + if (Data.getFragment()) + return Layout.getSymbolOffset(&Data); return 0; } -void ELFObjectWriter::ExecutePostLayoutBinding(MCAssembler &Asm) { +void ELFObjectWriter::ExecutePostLayoutBinding(MCAssembler &Asm, + const MCAsmLayout &Layout) { // The presence of symbol versions causes undefined symbols and // versions declared with @@@ to be renamed. @@ -526,7 +210,7 @@ void ELFObjectWriter::ExecutePostLayoutBinding(MCAssembler &Asm) { // Aliases defined with .symvar copy the binding from the symbol they alias. // This is the first place we are able to copy this information. it->setExternal(SD.isExternal()); - SetBinding(*it, GetBinding(SD)); + MCELF::SetBinding(*it, MCELF::GetBinding(SD)); StringRef Rest = AliasName.substr(Pos); if (!Symbol.isUndefined() && !Rest.startswith("@@@")) @@ -552,35 +236,24 @@ void ELFObjectWriter::WriteSymbol(MCDataFragment *SymtabF, bool IsReserved = Data.isCommon() || Data.getSymbol().isAbsolute() || Data.getSymbol().isVariable(); - uint8_t Binding = GetBinding(OrigData); - uint8_t Visibility = GetVisibility(OrigData); - uint8_t Type = GetType(Data); + uint8_t Binding = MCELF::GetBinding(OrigData); + uint8_t Visibility = MCELF::GetVisibility(OrigData); + uint8_t Type = MCELF::GetType(Data); uint8_t Info = (Binding << ELF_STB_Shift) | (Type << ELF_STT_Shift); uint8_t Other = Visibility; uint64_t Value = SymbolValue(Data, Layout); uint64_t Size = 0; - const MCExpr *ESize; assert(!(Data.isCommon() && !Data.isExternal())); - ESize = Data.getSize(); - if (Data.getSize()) { - MCValue Res; - if (ESize->getKind() == MCExpr::Binary) { - const MCBinaryExpr *BE = static_cast<const MCBinaryExpr *>(ESize); - - if (BE->EvaluateAsRelocatable(Res, &Layout)) { - assert(!Res.getSymA() || !Res.getSymA()->getSymbol().isDefined()); - assert(!Res.getSymB() || !Res.getSymB()->getSymbol().isDefined()); - Size = Res.getConstant(); - } - } else if (ESize->getKind() == MCExpr::Constant) { - Size = static_cast<const MCConstantExpr *>(ESize)->getValue(); - } else { - assert(0 && "Unsupported size expression"); - } + const MCExpr *ESize = Data.getSize(); + if (ESize) { + int64_t Res; + if (!ESize->EvaluateAsAbsolute(Res, Layout)) + report_fatal_error("Size expression must be absolute."); + Size = Res; } // Write out the symbol table entry @@ -631,7 +304,7 @@ void ELFObjectWriter::WriteSymbolTable(MCDataFragment *SymtabF, (Data.getFlags() & ELF_STB_Weak)) && "External symbol requires STB_GLOBAL or STB_WEAK flag"); WriteSymbol(SymtabF, ShndxF, MSD, Layout); - if (GetBinding(Data) == ELF::STB_LOCAL) + if (MCELF::GetBinding(Data) == ELF::STB_LOCAL) LastLocalSymbolIndex++; } @@ -639,7 +312,7 @@ void ELFObjectWriter::WriteSymbolTable(MCDataFragment *SymtabF, ELFSymbolData &MSD = UndefinedSymbolData[i]; MCSymbolData &Data = *MSD.SymbolData; WriteSymbol(SymtabF, ShndxF, MSD, Layout); - if (GetBinding(Data) == ELF::STB_LOCAL) + if (MCELF::GetBinding(Data) == ELF::STB_LOCAL) LastLocalSymbolIndex++; } } @@ -666,9 +339,16 @@ const MCSymbol *ELFObjectWriter::SymbolToReloc(const MCAssembler &Asm, const MCSectionELF &Section = static_cast<const MCSectionELF&>(ASymbol.getSection()); + const SectionKind secKind = Section.getKind(); - if (Section.getKind().isBSS()) - return NULL; + if (secKind.isBSS()) + return ExplicitRelSym(Asm, Target, F, true); + + if (secKind.isThreadLocal()) { + if (Renamed) + return Renamed; + return &Symbol; + } MCSymbolRefExpr::VariantKind Kind = Target.getSymA()->getKind(); const MCSectionELF &Sec2 = @@ -683,7 +363,7 @@ const MCSymbol *ELFObjectWriter::SymbolToReloc(const MCAssembler &Asm, return &Symbol; } - if (Section.getFlags() & MCSectionELF::SHF_MERGE) { + if (Section.getFlags() & ELF::SHF_MERGE) { if (Target.getConstant() == 0) return NULL; if (Renamed) @@ -691,7 +371,72 @@ const MCSymbol *ELFObjectWriter::SymbolToReloc(const MCAssembler &Asm, return &Symbol; } - return NULL; + return ExplicitRelSym(Asm, Target, F, false); +} + + +void ELFObjectWriter::RecordRelocation(const MCAssembler &Asm, + const MCAsmLayout &Layout, + const MCFragment *Fragment, + const MCFixup &Fixup, + MCValue Target, + uint64_t &FixedValue) { + int64_t Addend = 0; + int Index = 0; + int64_t Value = Target.getConstant(); + const MCSymbol *RelocSymbol = NULL; + + bool IsPCRel = isFixupKindPCRel(Asm, Fixup.getKind()); + if (!Target.isAbsolute()) { + const MCSymbol &Symbol = Target.getSymA()->getSymbol(); + const MCSymbol &ASymbol = Symbol.AliasedSymbol(); + RelocSymbol = SymbolToReloc(Asm, Target, *Fragment); + + if (const MCSymbolRefExpr *RefB = Target.getSymB()) { + const MCSymbol &SymbolB = RefB->getSymbol(); + MCSymbolData &SDB = Asm.getSymbolData(SymbolB); + IsPCRel = true; + + // Offset of the symbol in the section + int64_t a = Layout.getSymbolOffset(&SDB); + + // Ofeset of the relocation in the section + int64_t b = Layout.getFragmentOffset(Fragment) + Fixup.getOffset(); + Value += b - a; + } + + if (!RelocSymbol) { + MCSymbolData &SD = Asm.getSymbolData(ASymbol); + MCFragment *F = SD.getFragment(); + + Index = F->getParent()->getOrdinal() + 1; + + // Offset of the symbol in the section + Value += Layout.getSymbolOffset(&SD); + } else { + if (Asm.getSymbolData(Symbol).getFlags() & ELF_Other_Weakref) + WeakrefUsedInReloc.insert(RelocSymbol); + else + UsedInReloc.insert(RelocSymbol); + Index = -1; + } + Addend = Value; + // Compensate for the addend on i386. + if (is64Bit()) + Value = 0; + } + + FixedValue = Value; + unsigned Type = GetRelocType(Target, Fixup, IsPCRel, + (RelocSymbol != 0), Addend); + + uint64_t RelocOffset = Layout.getFragmentOffset(Fragment) + + Fixup.getOffset(); + + if (!hasRelocationAddend()) + Addend = 0; + ELFRelocationEntry ERE(RelocOffset, Index, Type, RelocSymbol, Addend); + Relocations[Fragment->getParent()].push_back(ERE); } @@ -702,8 +447,9 @@ ELFObjectWriter::getSymbolIndexInSymbolTable(const MCAssembler &Asm, return SD.getIndex(); } -static bool isInSymtab(const MCAssembler &Asm, const MCSymbolData &Data, - bool Used, bool Renamed) { +bool ELFObjectWriter::isInSymtab(const MCAssembler &Asm, + const MCSymbolData &Data, + bool Used, bool Renamed) { if (Data.getFlags() & ELF_Other_Weakref) return false; @@ -719,7 +465,11 @@ static bool isInSymtab(const MCAssembler &Asm, const MCSymbolData &Data, return true; const MCSymbol &A = Symbol.AliasedSymbol(); - if (!A.isVariable() && A.isUndefined() && !Data.isCommon()) + if (Symbol.isVariable() && !A.isVariable() && A.isUndefined()) + return false; + + bool IsGlobal = MCELF::GetBinding(Data) == ELF::STB_GLOBAL; + if (!Symbol.isVariable() && Symbol.isUndefined() && !IsGlobal) return false; if (!Asm.isSymbolLinkerVisible(Symbol) && !Symbol.isUndefined()) @@ -731,8 +481,8 @@ static bool isInSymtab(const MCAssembler &Asm, const MCSymbolData &Data, return true; } -static bool isLocal(const MCSymbolData &Data, bool isSignature, - bool isUsedInReloc) { +bool ELFObjectWriter::isLocal(const MCSymbolData &Data, bool isSignature, + bool isUsedInReloc) { if (Data.isExternal()) return false; @@ -780,7 +530,7 @@ void ELFObjectWriter::ComputeSymbolTable(MCAssembler &Asm, MCSymbol *Sym = Asm.getContext().GetOrCreateSymbol(Name); MCSymbolData &Data = Asm.getOrCreateSymbolData(*Sym); Data.setExternal(true); - SetBinding(Data, ELF::STB_GLOBAL); + MCELF::SetBinding(Data, ELF::STB_GLOBAL); } // Build section lookup table. @@ -811,14 +561,14 @@ void ELFObjectWriter::ComputeSymbolTable(MCAssembler &Asm, // Undefined symbols are global, but this is the first place we // are able to set it. bool Local = isLocal(*it, isSignature, Used); - if (!Local && GetBinding(*it) == ELF::STB_LOCAL) { + if (!Local && MCELF::GetBinding(*it) == ELF::STB_LOCAL) { MCSymbolData &SD = Asm.getSymbolData(RefSymbol); - SetBinding(*it, ELF::STB_GLOBAL); - SetBinding(SD, ELF::STB_GLOBAL); + MCELF::SetBinding(*it, ELF::STB_GLOBAL); + MCELF::SetBinding(SD, ELF::STB_GLOBAL); } if (RefSymbol.isUndefined() && !Used && WeakrefUsed) - SetBinding(*it, ELF::STB_WEAK); + MCELF::SetBinding(*it, ELF::STB_WEAK); if (it->isCommon()) { assert(!Local); @@ -895,28 +645,26 @@ void ELFObjectWriter::WriteRelocation(MCAssembler &Asm, MCAsmLayout &Layout, static_cast<const MCSectionELF&>(SD.getSection()); const StringRef SectionName = Section.getSectionName(); - std::string RelaSectionName = HasRelocationAddend ? ".rela" : ".rel"; + std::string RelaSectionName = hasRelocationAddend() ? ".rela" : ".rel"; RelaSectionName += SectionName; unsigned EntrySize; - if (HasRelocationAddend) - EntrySize = Is64Bit ? sizeof(ELF::Elf64_Rela) : sizeof(ELF::Elf32_Rela); + if (hasRelocationAddend()) + EntrySize = is64Bit() ? sizeof(ELF::Elf64_Rela) : sizeof(ELF::Elf32_Rela); else - EntrySize = Is64Bit ? sizeof(ELF::Elf64_Rel) : sizeof(ELF::Elf32_Rel); + EntrySize = is64Bit() ? sizeof(ELF::Elf64_Rel) : sizeof(ELF::Elf32_Rel); - RelaSection = Ctx.getELFSection(RelaSectionName, HasRelocationAddend ? + RelaSection = Ctx.getELFSection(RelaSectionName, hasRelocationAddend() ? ELF::SHT_RELA : ELF::SHT_REL, 0, SectionKind::getReadOnly(), EntrySize, ""); MCSectionData &RelaSD = Asm.getOrCreateSectionData(*RelaSection); - RelaSD.setAlignment(Is64Bit ? 8 : 4); + RelaSD.setAlignment(is64Bit() ? 8 : 4); MCDataFragment *F = new MCDataFragment(&RelaSD); WriteRelocationsFragment(Asm, F, &SD); - - Asm.AddSectionToTheEnd(*this, RelaSD, Layout); } } @@ -948,18 +696,20 @@ void ELFObjectWriter::WriteRelocationsFragment(const MCAssembler &Asm, for (unsigned i = 0, e = Relocs.size(); i != e; ++i) { ELFRelocationEntry entry = Relocs[e - i - 1]; - if (entry.Index < 0) + if (!entry.Index) + ; + else if (entry.Index < 0) entry.Index = getSymbolIndexInSymbolTable(Asm, entry.Symbol); else - entry.Index += LocalSymbolData.size() + 1; - if (Is64Bit) { + entry.Index += LocalSymbolData.size(); + if (is64Bit()) { String64(*F, entry.r_offset); struct ELF::Elf64_Rela ERE64; ERE64.setSymbolAndType(entry.Index, entry.Type); String64(*F, ERE64.r_info); - if (HasRelocationAddend) + if (hasRelocationAddend()) String64(*F, entry.r_addend); } else { String32(*F, entry.r_offset); @@ -968,7 +718,7 @@ void ELFObjectWriter::WriteRelocationsFragment(const MCAssembler &Asm, ERE32.setSymbolAndType(entry.Index, entry.Type); String32(*F, ERE32.r_info); - if (HasRelocationAddend) + if (hasRelocationAddend()) String32(*F, entry.r_addend); } } @@ -980,7 +730,7 @@ void ELFObjectWriter::CreateMetadataSections(MCAssembler &Asm, MCContext &Ctx = Asm.getContext(); MCDataFragment *F; - unsigned EntrySize = Is64Bit ? ELF::SYMENTRY_SIZE64 : ELF::SYMENTRY_SIZE32; + unsigned EntrySize = is64Bit() ? ELF::SYMENTRY_SIZE64 : ELF::SYMENTRY_SIZE32; // We construct .shstrtab, .symtab and .strtab in this order to match gnu as. const MCSectionELF *ShstrtabSection = @@ -995,7 +745,7 @@ void ELFObjectWriter::CreateMetadataSections(MCAssembler &Asm, SectionKind::getReadOnly(), EntrySize, ""); MCSectionData &SymtabSD = Asm.getOrCreateSectionData(*SymtabSection); - SymtabSD.setAlignment(Is64Bit ? 8 : 4); + SymtabSD.setAlignment(is64Bit() ? 8 : 4); SymbolTableIndex = Asm.size(); MCSectionData *SymtabShndxSD = NULL; @@ -1022,14 +772,11 @@ void ELFObjectWriter::CreateMetadataSections(MCAssembler &Asm, MCDataFragment *ShndxF = NULL; if (NeedsSymtabShndx) { ShndxF = new MCDataFragment(SymtabShndxSD); - Asm.AddSectionToTheEnd(*this, *SymtabShndxSD, Layout); } WriteSymbolTable(F, ShndxF, Asm, Layout, SectionIndexMap); - Asm.AddSectionToTheEnd(*this, SymtabSD, Layout); F = new MCDataFragment(&StrtabSD); F->getContents().append(StringTable.begin(), StringTable.end()); - Asm.AddSectionToTheEnd(*this, StrtabSD, Layout); F = new MCDataFragment(&ShstrtabSD); @@ -1061,61 +808,34 @@ void ELFObjectWriter::CreateMetadataSections(MCAssembler &Asm, F->getContents() += Name; F->getContents() += '\x00'; } - - Asm.AddSectionToTheEnd(*this, ShstrtabSD, Layout); } -bool ELFObjectWriter::IsFixupFullyResolved(const MCAssembler &Asm, - const MCValue Target, - bool IsPCRel, - const MCFragment *DF) const { - // If this is a PCrel relocation, find the section this fixup value is - // relative to. - const MCSection *BaseSection = 0; - if (IsPCRel) { - BaseSection = &DF->getParent()->getSection(); - assert(BaseSection); - } - - const MCSection *SectionA = 0; - const MCSymbol *SymbolA = 0; - if (const MCSymbolRefExpr *A = Target.getSymA()) { - SymbolA = &A->getSymbol().AliasedSymbol(); - SectionA = &SymbolA->getSection(); - } - - const MCSection *SectionB = 0; - if (const MCSymbolRefExpr *B = Target.getSymB()) { - SectionB = &B->getSymbol().AliasedSymbol().getSection(); +void ELFObjectWriter::CreateIndexedSections(MCAssembler &Asm, + MCAsmLayout &Layout, + GroupMapTy &GroupMap, + RevGroupMapTy &RevGroupMap) { + // Create the .note.GNU-stack section if needed. + MCContext &Ctx = Asm.getContext(); + if (Asm.getNoExecStack()) { + const MCSectionELF *GnuStackSection = + Ctx.getELFSection(".note.GNU-stack", ELF::SHT_PROGBITS, 0, + SectionKind::getReadOnly()); + Asm.getOrCreateSectionData(*GnuStackSection); } - if (!BaseSection) - return SectionA == SectionB; - - const MCSymbolData &DataA = Asm.getSymbolData(*SymbolA); - if (DataA.isExternal()) - return false; - - return !SectionB && BaseSection == SectionA; -} - -void ELFObjectWriter::CreateGroupSections(MCAssembler &Asm, - MCAsmLayout &Layout, - GroupMapTy &GroupMap, - RevGroupMapTy &RevGroupMap) { // Build the groups for (MCAssembler::const_iterator it = Asm.begin(), ie = Asm.end(); it != ie; ++it) { const MCSectionELF &Section = static_cast<const MCSectionELF&>(it->getSection()); - if (!(Section.getFlags() & MCSectionELF::SHF_GROUP)) + if (!(Section.getFlags() & ELF::SHF_GROUP)) continue; const MCSymbol *SignatureSymbol = Section.getGroup(); Asm.getOrCreateSymbolData(*SignatureSymbol); const MCSectionELF *&Group = RevGroupMap[SignatureSymbol]; if (!Group) { - Group = Asm.getContext().CreateELFGroupSection(); + Group = Ctx.CreateELFGroupSection(); MCSectionData &Data = Asm.getOrCreateSectionData(*Group); Data.setAlignment(4); MCDataFragment *F = new MCDataFragment(&Data); @@ -1131,7 +851,7 @@ void ELFObjectWriter::CreateGroupSections(MCAssembler &Asm, it != ie; ++it, ++Index) { const MCSectionELF &Section = static_cast<const MCSectionELF&>(it->getSection()); - if (!(Section.getFlags() & MCSectionELF::SHF_GROUP)) + if (!(Section.getFlags() & ELF::SHF_GROUP)) continue; const MCSectionELF *Group = RevGroupMap[Section.getGroup()]; MCSectionData &Data = Asm.getOrCreateSectionData(*Group); @@ -1139,13 +859,6 @@ void ELFObjectWriter::CreateGroupSections(MCAssembler &Asm, MCDataFragment *F = new MCDataFragment(&Data); String32(*F, NumGroups + Index); } - - for (RevGroupMapTy::const_iterator i = RevGroupMap.begin(), - e = RevGroupMap.end(); i != e; ++i) { - const MCSectionELF *Group = i->second; - MCSectionData &Data = Asm.getOrCreateSectionData(*Group); - Asm.AddSectionToTheEnd(*this, Data, Layout); - } } void ELFObjectWriter::WriteSection(MCAssembler &Asm, @@ -1197,8 +910,13 @@ void ELFObjectWriter::WriteSection(MCAssembler &Asm, case ELF::SHT_PROGBITS: case ELF::SHT_STRTAB: case ELF::SHT_NOBITS: + case ELF::SHT_NOTE: case ELF::SHT_NULL: case ELF::SHT_ARM_ATTRIBUTES: + case ELF::SHT_INIT_ARRAY: + case ELF::SHT_FINI_ARRAY: + case ELF::SHT_PREINIT_ARRAY: + case ELF::SHT_X86_64_UNWIND: // Nothing to do. break; @@ -1218,12 +936,52 @@ void ELFObjectWriter::WriteSection(MCAssembler &Asm, Alignment, Section.getEntrySize()); } +bool ELFObjectWriter::IsELFMetaDataSection(const MCSectionData &SD) { + return SD.getOrdinal() == ~UINT32_C(0) && + !SD.getSection().isVirtualSection(); +} + +uint64_t ELFObjectWriter::DataSectionSize(const MCSectionData &SD) { + uint64_t Ret = 0; + for (MCSectionData::const_iterator i = SD.begin(), e = SD.end(); i != e; + ++i) { + const MCFragment &F = *i; + assert(F.getKind() == MCFragment::FT_Data); + Ret += cast<MCDataFragment>(F).getContents().size(); + } + return Ret; +} + +uint64_t ELFObjectWriter::GetSectionFileSize(const MCAsmLayout &Layout, + const MCSectionData &SD) { + if (IsELFMetaDataSection(SD)) + return DataSectionSize(SD); + return Layout.getSectionFileSize(&SD); +} + +uint64_t ELFObjectWriter::GetSectionAddressSize(const MCAsmLayout &Layout, + const MCSectionData &SD) { + if (IsELFMetaDataSection(SD)) + return DataSectionSize(SD); + return Layout.getSectionAddressSize(&SD); +} + +void ELFObjectWriter::WriteDataSectionData(ELFObjectWriter *W, + const MCSectionData &SD) { + for (MCSectionData::const_iterator i = SD.begin(), e = SD.end(); i != e; + ++i) { + const MCFragment &F = *i; + assert(F.getKind() == MCFragment::FT_Data); + W->WriteBytes(cast<MCDataFragment>(F).getContents().str()); + } +} + void ELFObjectWriter::WriteObject(MCAssembler &Asm, const MCAsmLayout &Layout) { GroupMapTy GroupMap; RevGroupMapTy RevGroupMap; - CreateGroupSections(Asm, const_cast<MCAsmLayout&>(Layout), GroupMap, - RevGroupMap); + CreateIndexedSections(Asm, const_cast<MCAsmLayout&>(Layout), GroupMap, + RevGroupMap); SectionIndexMapTy SectionIndexMap; @@ -1241,8 +999,9 @@ void ELFObjectWriter::WriteObject(MCAssembler &Asm, // Add 1 for the null section. unsigned NumSections = Asm.size() + 1; - uint64_t NaturalAlignment = Is64Bit ? 8 : 4; - uint64_t HeaderSize = Is64Bit ? sizeof(ELF::Elf64_Ehdr) : sizeof(ELF::Elf32_Ehdr); + uint64_t NaturalAlignment = is64Bit() ? 8 : 4; + uint64_t HeaderSize = is64Bit() ? sizeof(ELF::Elf64_Ehdr) : + sizeof(ELF::Elf32_Ehdr); uint64_t FileOff = HeaderSize; std::vector<const MCSectionELF*> Sections; @@ -1261,9 +1020,7 @@ void ELFObjectWriter::WriteObject(MCAssembler &Asm, FileOff = RoundUpToAlignment(FileOff, SD.getAlignment()); // Get the size of the section in the output file (including padding). - uint64_t Size = Layout.getSectionFileSize(&SD); - - FileOff += Size; + FileOff += GetSectionFileSize(Layout, SD); } FileOff = RoundUpToAlignment(FileOff, NaturalAlignment); @@ -1287,9 +1044,12 @@ void ELFObjectWriter::WriteObject(MCAssembler &Asm, // Remember the offset into the file for this section. SectionOffsetMap[&Section] = FileOff; - FileOff += Layout.getSectionFileSize(&SD); + FileOff += GetSectionFileSize(Layout, SD); - Asm.WriteSectionData(&SD, Layout, this); + if (IsELFMetaDataSection(SD)) + WriteDataSectionData(this, SD); + else + Asm.WriteSectionData(&SD, Layout); } uint64_t Padding = OffsetToAlignment(FileOff, NaturalAlignment); @@ -1315,26 +1075,37 @@ void ELFObjectWriter::WriteObject(MCAssembler &Asm, else GroupSymbolIndex = getSymbolIndexInSymbolTable(Asm, GroupMap[&Section]); + uint64_t Size = GetSectionAddressSize(Layout, SD); + WriteSection(Asm, SectionIndexMap, GroupSymbolIndex, - SectionOffsetMap[&Section], Layout.getSectionSize(&SD), + SectionOffsetMap[&Section], Size, SD.getAlignment(), Section); } } -MCObjectWriter *llvm::createELFObjectWriter(raw_ostream &OS, - bool Is64Bit, - Triple::OSType OSType, - uint16_t EMachine, - bool IsLittleEndian, - bool HasRelocationAddend) { - switch (EMachine) { +bool +ELFObjectWriter::IsSymbolRefDifferenceFullyResolvedImpl(const MCAssembler &Asm, + const MCSymbolData &DataA, + const MCFragment &FB, + bool InSet, + bool IsPCRel) const { + if (DataA.getFlags() & ELF_STB_Weak) + return false; + return MCObjectWriter::IsSymbolRefDifferenceFullyResolvedImpl( + Asm, DataA, FB,InSet, IsPCRel); +} + +MCObjectWriter *llvm::createELFObjectWriter(MCELFObjectTargetWriter *MOTW, + raw_ostream &OS, + bool IsLittleEndian) { + switch (MOTW->getEMachine()) { case ELF::EM_386: case ELF::EM_X86_64: - return new X86ELFObjectWriter(OS, Is64Bit, IsLittleEndian, EMachine, - HasRelocationAddend, OSType); break; + return new X86ELFObjectWriter(MOTW, OS, IsLittleEndian); break; case ELF::EM_ARM: - return new ARMELFObjectWriter(OS, Is64Bit, IsLittleEndian, EMachine, - HasRelocationAddend, OSType); break; + return new ARMELFObjectWriter(MOTW, OS, IsLittleEndian); break; + case ELF::EM_MBLAZE: + return new MBlazeELFObjectWriter(MOTW, OS, IsLittleEndian); break; default: llvm_unreachable("Unsupported architecture"); break; } } @@ -1343,121 +1114,276 @@ MCObjectWriter *llvm::createELFObjectWriter(raw_ostream &OS, /// START OF SUBCLASSES for ELFObjectWriter //===- ARMELFObjectWriter -------------------------------------------===// -ARMELFObjectWriter::ARMELFObjectWriter(raw_ostream &_OS, bool _Is64Bit, - bool _IsLittleEndian, - uint16_t _EMachine, bool _HasRelocationAddend, - Triple::OSType _OSType) - : ELFObjectWriter(_OS, _Is64Bit, _IsLittleEndian, _EMachine, - _HasRelocationAddend, _OSType) +ARMELFObjectWriter::ARMELFObjectWriter(MCELFObjectTargetWriter *MOTW, + raw_ostream &_OS, + bool IsLittleEndian) + : ELFObjectWriter(MOTW, _OS, IsLittleEndian) {} ARMELFObjectWriter::~ARMELFObjectWriter() {} -void ARMELFObjectWriter::RecordRelocation(const MCAssembler &Asm, - const MCAsmLayout &Layout, - const MCFragment *Fragment, - const MCFixup &Fixup, - MCValue Target, - uint64_t &FixedValue) { - assert(0 && "ARMELFObjectWriter::RecordRelocation() unimplemented"); +// FIXME: get the real EABI Version from the Triple. +void ARMELFObjectWriter::WriteEFlags() { + Write32(ELF::EF_ARM_EABIMASK & DefaultEABIVersion); } +// In ARM, _MergedGlobals and other most symbols get emitted directly. +// I.e. not as an offset to a section symbol. +// This code is a first-cut approximation of what ARM/gcc does. +const MCSymbol *ARMELFObjectWriter::ExplicitRelSym(const MCAssembler &Asm, + const MCValue &Target, + const MCFragment &F, + bool IsBSS) const { + const MCSymbol &Symbol = Target.getSymA()->getSymbol(); + bool EmitThisSym = false; -//===- X86ELFObjectWriter -------------------------------------------===// + if (IsBSS) { + EmitThisSym = StringSwitch<bool>(Symbol.getName()) + .Case("_MergedGlobals", true) + .Default(false); + } else { + EmitThisSym = StringSwitch<bool>(Symbol.getName()) + .Case("_MergedGlobals", true) + .StartsWith(".L.str", true) + .Default(false); + } + if (EmitThisSym) + return &Symbol; + if (! Symbol.isTemporary()) + return &Symbol; + return NULL; +} +unsigned ARMELFObjectWriter::GetRelocType(const MCValue &Target, + const MCFixup &Fixup, + bool IsPCRel, + bool IsRelocWithSymbol, + int64_t Addend) { + MCSymbolRefExpr::VariantKind Modifier = Target.isAbsolute() ? + MCSymbolRefExpr::VK_None : Target.getSymA()->getKind(); -X86ELFObjectWriter::X86ELFObjectWriter(raw_ostream &_OS, bool _Is64Bit, - bool _IsLittleEndian, - uint16_t _EMachine, bool _HasRelocationAddend, - Triple::OSType _OSType) - : ELFObjectWriter(_OS, _Is64Bit, _IsLittleEndian, _EMachine, - _HasRelocationAddend, _OSType) -{} + unsigned Type = 0; + if (IsPCRel) { + switch ((unsigned)Fixup.getKind()) { + default: assert(0 && "Unimplemented"); + case FK_Data_4: + switch (Modifier) { + default: llvm_unreachable("Unsupported Modifier"); + case MCSymbolRefExpr::VK_None: + Type = ELF::R_ARM_BASE_PREL; + break; + case MCSymbolRefExpr::VK_ARM_TLSGD: + assert(0 && "unimplemented"); + break; + case MCSymbolRefExpr::VK_ARM_GOTTPOFF: + Type = ELF::R_ARM_TLS_IE32; + break; + } + break; + case ARM::fixup_arm_uncondbranch: + switch (Modifier) { + case MCSymbolRefExpr::VK_ARM_PLT: + Type = ELF::R_ARM_PLT32; + break; + default: + Type = ELF::R_ARM_CALL; + break; + } + break; + case ARM::fixup_arm_condbranch: + Type = ELF::R_ARM_JUMP24; + break; + case ARM::fixup_arm_movt_hi16: + case ARM::fixup_arm_movt_hi16_pcrel: + Type = ELF::R_ARM_MOVT_PREL; + break; + case ARM::fixup_arm_movw_lo16: + case ARM::fixup_arm_movw_lo16_pcrel: + Type = ELF::R_ARM_MOVW_PREL_NC; + break; + case ARM::fixup_t2_movt_hi16: + case ARM::fixup_t2_movt_hi16_pcrel: + Type = ELF::R_ARM_THM_MOVT_PREL; + break; + case ARM::fixup_t2_movw_lo16: + case ARM::fixup_t2_movw_lo16_pcrel: + Type = ELF::R_ARM_THM_MOVW_PREL_NC; + break; + } + } else { + switch ((unsigned)Fixup.getKind()) { + default: llvm_unreachable("invalid fixup kind!"); + case FK_Data_4: + switch (Modifier) { + default: llvm_unreachable("Unsupported Modifier"); break; + case MCSymbolRefExpr::VK_ARM_GOT: + Type = ELF::R_ARM_GOT_BREL; + break; + case MCSymbolRefExpr::VK_ARM_TLSGD: + Type = ELF::R_ARM_TLS_GD32; + break; + case MCSymbolRefExpr::VK_ARM_TPOFF: + Type = ELF::R_ARM_TLS_LE32; + break; + case MCSymbolRefExpr::VK_ARM_GOTTPOFF: + Type = ELF::R_ARM_TLS_IE32; + break; + case MCSymbolRefExpr::VK_None: + Type = ELF::R_ARM_ABS32; + break; + case MCSymbolRefExpr::VK_ARM_GOTOFF: + Type = ELF::R_ARM_GOTOFF32; + break; + } + break; + case ARM::fixup_arm_ldst_pcrel_12: + case ARM::fixup_arm_pcrel_10: + case ARM::fixup_arm_adr_pcrel_12: + case ARM::fixup_arm_thumb_bl: + case ARM::fixup_arm_thumb_cb: + case ARM::fixup_arm_thumb_cp: + case ARM::fixup_arm_thumb_br: + assert(0 && "Unimplemented"); + break; + case ARM::fixup_arm_uncondbranch: + Type = ELF::R_ARM_CALL; + break; + case ARM::fixup_arm_condbranch: + Type = ELF::R_ARM_JUMP24; + break; + case ARM::fixup_arm_movt_hi16: + Type = ELF::R_ARM_MOVT_ABS; + break; + case ARM::fixup_arm_movw_lo16: + Type = ELF::R_ARM_MOVW_ABS_NC; + break; + case ARM::fixup_t2_movt_hi16: + Type = ELF::R_ARM_THM_MOVT_ABS; + break; + case ARM::fixup_t2_movw_lo16: + Type = ELF::R_ARM_THM_MOVW_ABS_NC; + break; + } + } -X86ELFObjectWriter::~X86ELFObjectWriter() -{} + if (RelocNeedsGOT(Modifier)) + NeedsGOT = true; -void X86ELFObjectWriter::RecordRelocation(const MCAssembler &Asm, - const MCAsmLayout &Layout, - const MCFragment *Fragment, - const MCFixup &Fixup, - MCValue Target, - uint64_t &FixedValue) { - int64_t Addend = 0; - int Index = 0; - int64_t Value = Target.getConstant(); - const MCSymbol &Symbol = Target.getSymA()->getSymbol(); - const MCSymbol &ASymbol = Symbol.AliasedSymbol(); - const MCSymbol *RelocSymbol = SymbolToReloc(Asm, Target, *Fragment); + return Type; +} - bool IsPCRel = isFixupKindX86PCRel(Fixup.getKind()); - if (!Target.isAbsolute()) { - if (const MCSymbolRefExpr *RefB = Target.getSymB()) { - const MCSymbol &SymbolB = RefB->getSymbol(); - MCSymbolData &SDB = Asm.getSymbolData(SymbolB); - IsPCRel = true; - MCSectionData *Sec = Fragment->getParent(); +//===- MBlazeELFObjectWriter -------------------------------------------===// - // Offset of the symbol in the section - int64_t a = Layout.getSymbolAddress(&SDB) - Layout.getSectionAddress(Sec); +MBlazeELFObjectWriter::MBlazeELFObjectWriter(MCELFObjectTargetWriter *MOTW, + raw_ostream &_OS, + bool IsLittleEndian) + : ELFObjectWriter(MOTW, _OS, IsLittleEndian) { +} - // Ofeset of the relocation in the section - int64_t b = Layout.getFragmentOffset(Fragment) + Fixup.getOffset(); - Value += b - a; +MBlazeELFObjectWriter::~MBlazeELFObjectWriter() { +} + +unsigned MBlazeELFObjectWriter::GetRelocType(const MCValue &Target, + const MCFixup &Fixup, + bool IsPCRel, + bool IsRelocWithSymbol, + int64_t Addend) { + // determine the type of the relocation + unsigned Type; + if (IsPCRel) { + switch ((unsigned)Fixup.getKind()) { + default: + llvm_unreachable("Unimplemented"); + case FK_PCRel_4: + Type = ELF::R_MICROBLAZE_64_PCREL; + break; + case FK_PCRel_2: + Type = ELF::R_MICROBLAZE_32_PCREL; + break; } + } else { + switch ((unsigned)Fixup.getKind()) { + default: llvm_unreachable("invalid fixup kind!"); + case FK_Data_4: + Type = ((IsRelocWithSymbol || Addend !=0) + ? ELF::R_MICROBLAZE_32 + : ELF::R_MICROBLAZE_64); + break; + case FK_Data_2: + Type = ELF::R_MICROBLAZE_32; + break; + } + } + return Type; +} - if (!RelocSymbol) { - MCSymbolData &SD = Asm.getSymbolData(ASymbol); - MCFragment *F = SD.getFragment(); +//===- X86ELFObjectWriter -------------------------------------------===// - Index = F->getParent()->getOrdinal(); - MCSectionData *FSD = F->getParent(); - // Offset of the symbol in the section - Value += Layout.getSymbolAddress(&SD) - Layout.getSectionAddress(FSD); - } else { - if (Asm.getSymbolData(Symbol).getFlags() & ELF_Other_Weakref) - WeakrefUsedInReloc.insert(RelocSymbol); - else - UsedInReloc.insert(RelocSymbol); - Index = -1; - } - Addend = Value; - // Compensate for the addend on i386. - if (Is64Bit) - Value = 0; - } +X86ELFObjectWriter::X86ELFObjectWriter(MCELFObjectTargetWriter *MOTW, + raw_ostream &_OS, + bool IsLittleEndian) + : ELFObjectWriter(MOTW, _OS, IsLittleEndian) +{} - FixedValue = Value; +X86ELFObjectWriter::~X86ELFObjectWriter() +{} +unsigned X86ELFObjectWriter::GetRelocType(const MCValue &Target, + const MCFixup &Fixup, + bool IsPCRel, + bool IsRelocWithSymbol, + int64_t Addend) { // determine the type of the relocation - MCSymbolRefExpr::VariantKind Modifier = Target.getSymA()->getKind(); + MCSymbolRefExpr::VariantKind Modifier = Target.isAbsolute() ? + MCSymbolRefExpr::VK_None : Target.getSymA()->getKind(); unsigned Type; - if (Is64Bit) { + if (is64Bit()) { if (IsPCRel) { - switch (Modifier) { - default: - llvm_unreachable("Unimplemented"); - case MCSymbolRefExpr::VK_None: - Type = ELF::R_X86_64_PC32; - break; - case MCSymbolRefExpr::VK_PLT: - Type = ELF::R_X86_64_PLT32; + switch ((unsigned)Fixup.getKind()) { + default: llvm_unreachable("invalid fixup kind!"); + case FK_PCRel_8: + assert(Modifier == MCSymbolRefExpr::VK_None); + Type = ELF::R_X86_64_PC64; break; - case MCSymbolRefExpr::VK_GOTPCREL: - Type = ELF::R_X86_64_GOTPCREL; + case X86::reloc_signed_4byte: + case X86::reloc_riprel_4byte_movq_load: + case FK_Data_4: // FIXME? + case X86::reloc_riprel_4byte: + case FK_PCRel_4: + switch (Modifier) { + default: + llvm_unreachable("Unimplemented"); + case MCSymbolRefExpr::VK_None: + Type = ELF::R_X86_64_PC32; + break; + case MCSymbolRefExpr::VK_PLT: + Type = ELF::R_X86_64_PLT32; + break; + case MCSymbolRefExpr::VK_GOTPCREL: + Type = ELF::R_X86_64_GOTPCREL; + break; + case MCSymbolRefExpr::VK_GOTTPOFF: + Type = ELF::R_X86_64_GOTTPOFF; break; - case MCSymbolRefExpr::VK_GOTTPOFF: - Type = ELF::R_X86_64_GOTTPOFF; + case MCSymbolRefExpr::VK_TLSGD: + Type = ELF::R_X86_64_TLSGD; + break; + case MCSymbolRefExpr::VK_TLSLD: + Type = ELF::R_X86_64_TLSLD; + break; + } break; - case MCSymbolRefExpr::VK_TLSGD: - Type = ELF::R_X86_64_TLSGD; + case FK_PCRel_2: + assert(Modifier == MCSymbolRefExpr::VK_None); + Type = ELF::R_X86_64_PC16; break; - case MCSymbolRefExpr::VK_TLSLD: - Type = ELF::R_X86_64_TLSLD; + case FK_PCRel_1: + assert(Modifier == MCSymbolRefExpr::VK_None); + Type = ELF::R_X86_64_PC8; break; } } else { @@ -1465,7 +1391,6 @@ void X86ELFObjectWriter::RecordRelocation(const MCAssembler &Asm, default: llvm_unreachable("invalid fixup kind!"); case FK_Data_8: Type = ELF::R_X86_64_64; break; case X86::reloc_signed_4byte: - case X86::reloc_pcrel_4byte: assert(isInt<32>(Target.getConstant())); switch (Modifier) { default: @@ -1491,7 +1416,7 @@ void X86ELFObjectWriter::RecordRelocation(const MCAssembler &Asm, Type = ELF::R_X86_64_32; break; case FK_Data_2: Type = ELF::R_X86_64_16; break; - case X86::reloc_pcrel_1byte: + case FK_PCRel_1: case FK_Data_1: Type = ELF::R_X86_64_8; break; } } @@ -1518,7 +1443,7 @@ void X86ELFObjectWriter::RecordRelocation(const MCAssembler &Asm, // FIXME: Should we avoid selecting reloc_signed_4byte in 32 bit mode // instead? case X86::reloc_signed_4byte: - case X86::reloc_pcrel_4byte: + case FK_PCRel_4: case FK_Data_4: switch (Modifier) { default: @@ -1556,7 +1481,7 @@ void X86ELFObjectWriter::RecordRelocation(const MCAssembler &Asm, } break; case FK_Data_2: Type = ELF::R_386_16; break; - case X86::reloc_pcrel_1byte: + case FK_PCRel_1: case FK_Data_1: Type = ELF::R_386_8; break; } } @@ -1565,18 +1490,5 @@ void X86ELFObjectWriter::RecordRelocation(const MCAssembler &Asm, if (RelocNeedsGOT(Modifier)) NeedsGOT = true; - ELFRelocationEntry ERE; - - ERE.Index = Index; - ERE.Type = Type; - ERE.Symbol = RelocSymbol; - - ERE.r_offset = Layout.getFragmentOffset(Fragment) + Fixup.getOffset(); - - if (HasRelocationAddend) - ERE.r_addend = Addend; - else - ERE.r_addend = 0; // Silence compiler warning. - - Relocations[Fragment->getParent()].push_back(ERE); + return Type; } diff --git a/lib/MC/ELFObjectWriter.h b/lib/MC/ELFObjectWriter.h new file mode 100644 index 0000000..9457623 --- /dev/null +++ b/lib/MC/ELFObjectWriter.h @@ -0,0 +1,391 @@ +//===- lib/MC/ELFObjectWriter.h - ELF File Writer -------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file implements ELF object file writer information. +// +//===----------------------------------------------------------------------===// + +#ifndef LLVM_MC_ELFOBJECTWRITER_H +#define LLVM_MC_ELFOBJECTWRITER_H + +#include "MCELF.h" +#include "llvm/ADT/OwningPtr.h" +#include "llvm/ADT/SmallPtrSet.h" +#include "llvm/ADT/SmallString.h" +#include "llvm/ADT/STLExtras.h" +#include "llvm/MC/MCAssembler.h" +#include "llvm/MC/MCELFObjectWriter.h" +#include "llvm/MC/MCELFSymbolFlags.h" +#include "llvm/MC/MCObjectWriter.h" +#include "llvm/MC/MCExpr.h" +#include "llvm/MC/MCSymbol.h" + +#include <vector> + +namespace llvm { + +class MCSection; +class MCDataFragment; +class MCSectionELF; + +class ELFObjectWriter : public MCObjectWriter { + protected: + + static bool isFixupKindPCRel(const MCAssembler &Asm, unsigned Kind); + static bool RelocNeedsGOT(MCSymbolRefExpr::VariantKind Variant); + static uint64_t SymbolValue(MCSymbolData &Data, const MCAsmLayout &Layout); + static bool isInSymtab(const MCAssembler &Asm, const MCSymbolData &Data, + bool Used, bool Renamed); + static bool isLocal(const MCSymbolData &Data, bool isSignature, + bool isUsedInReloc); + static bool IsELFMetaDataSection(const MCSectionData &SD); + static uint64_t DataSectionSize(const MCSectionData &SD); + static uint64_t GetSectionFileSize(const MCAsmLayout &Layout, + const MCSectionData &SD); + static uint64_t GetSectionAddressSize(const MCAsmLayout &Layout, + const MCSectionData &SD); + static void WriteDataSectionData(ELFObjectWriter *W, + const MCSectionData &SD); + + /*static bool isFixupKindX86RIPRel(unsigned Kind) { + return Kind == X86::reloc_riprel_4byte || + Kind == X86::reloc_riprel_4byte_movq_load; + }*/ + + /// ELFSymbolData - Helper struct for containing some precomputed + /// information on symbols. + struct ELFSymbolData { + MCSymbolData *SymbolData; + uint64_t StringIndex; + uint32_t SectionIndex; + + // Support lexicographic sorting. + bool operator<(const ELFSymbolData &RHS) const { + if (MCELF::GetType(*SymbolData) == ELF::STT_FILE) + return true; + if (MCELF::GetType(*RHS.SymbolData) == ELF::STT_FILE) + return false; + return SymbolData->getSymbol().getName() < + RHS.SymbolData->getSymbol().getName(); + } + }; + + /// @name Relocation Data + /// @{ + + struct ELFRelocationEntry { + // Make these big enough for both 32-bit and 64-bit + uint64_t r_offset; + int Index; + unsigned Type; + const MCSymbol *Symbol; + uint64_t r_addend; + + ELFRelocationEntry() + : r_offset(0), Index(0), Type(0), Symbol(0), r_addend(0) {} + + ELFRelocationEntry(uint64_t RelocOffset, int Idx, + unsigned RelType, const MCSymbol *Sym, + uint64_t Addend) + : r_offset(RelocOffset), Index(Idx), Type(RelType), + Symbol(Sym), r_addend(Addend) {} + + // Support lexicographic sorting. + bool operator<(const ELFRelocationEntry &RE) const { + return RE.r_offset < r_offset; + } + }; + + /// The target specific ELF writer instance. + llvm::OwningPtr<MCELFObjectTargetWriter> TargetObjectWriter; + + SmallPtrSet<const MCSymbol *, 16> UsedInReloc; + SmallPtrSet<const MCSymbol *, 16> WeakrefUsedInReloc; + DenseMap<const MCSymbol *, const MCSymbol *> Renames; + + llvm::DenseMap<const MCSectionData*, + std::vector<ELFRelocationEntry> > Relocations; + DenseMap<const MCSection*, uint64_t> SectionStringTableIndex; + + /// @} + /// @name Symbol Table Data + /// @{ + + SmallString<256> StringTable; + std::vector<ELFSymbolData> LocalSymbolData; + std::vector<ELFSymbolData> ExternalSymbolData; + std::vector<ELFSymbolData> UndefinedSymbolData; + + /// @} + + bool NeedsGOT; + + bool NeedsSymtabShndx; + + // This holds the symbol table index of the last local symbol. + unsigned LastLocalSymbolIndex; + // This holds the .strtab section index. + unsigned StringTableIndex; + // This holds the .symtab section index. + unsigned SymbolTableIndex; + + unsigned ShstrtabIndex; + + + const MCSymbol *SymbolToReloc(const MCAssembler &Asm, + const MCValue &Target, + const MCFragment &F) const; + + // For arch-specific emission of explicit reloc symbol + virtual const MCSymbol *ExplicitRelSym(const MCAssembler &Asm, + const MCValue &Target, + const MCFragment &F, + bool IsBSS) const { + return NULL; + } + + bool is64Bit() const { return TargetObjectWriter->is64Bit(); } + bool hasRelocationAddend() const { + return TargetObjectWriter->hasRelocationAddend(); + } + + public: + ELFObjectWriter(MCELFObjectTargetWriter *MOTW, + raw_ostream &_OS, bool IsLittleEndian) + : MCObjectWriter(_OS, IsLittleEndian), + TargetObjectWriter(MOTW), + NeedsGOT(false), NeedsSymtabShndx(false){ + } + + virtual ~ELFObjectWriter(); + + void WriteWord(uint64_t W) { + if (is64Bit()) + Write64(W); + else + Write32(W); + } + + void StringLE16(char *buf, uint16_t Value) { + buf[0] = char(Value >> 0); + buf[1] = char(Value >> 8); + } + + void StringLE32(char *buf, uint32_t Value) { + StringLE16(buf, uint16_t(Value >> 0)); + StringLE16(buf + 2, uint16_t(Value >> 16)); + } + + void StringLE64(char *buf, uint64_t Value) { + StringLE32(buf, uint32_t(Value >> 0)); + StringLE32(buf + 4, uint32_t(Value >> 32)); + } + + void StringBE16(char *buf ,uint16_t Value) { + buf[0] = char(Value >> 8); + buf[1] = char(Value >> 0); + } + + void StringBE32(char *buf, uint32_t Value) { + StringBE16(buf, uint16_t(Value >> 16)); + StringBE16(buf + 2, uint16_t(Value >> 0)); + } + + void StringBE64(char *buf, uint64_t Value) { + StringBE32(buf, uint32_t(Value >> 32)); + StringBE32(buf + 4, uint32_t(Value >> 0)); + } + + void String8(MCDataFragment &F, uint8_t Value) { + char buf[1]; + buf[0] = Value; + F.getContents() += StringRef(buf, 1); + } + + void String16(MCDataFragment &F, uint16_t Value) { + char buf[2]; + if (isLittleEndian()) + StringLE16(buf, Value); + else + StringBE16(buf, Value); + F.getContents() += StringRef(buf, 2); + } + + void String32(MCDataFragment &F, uint32_t Value) { + char buf[4]; + if (isLittleEndian()) + StringLE32(buf, Value); + else + StringBE32(buf, Value); + F.getContents() += StringRef(buf, 4); + } + + void String64(MCDataFragment &F, uint64_t Value) { + char buf[8]; + if (isLittleEndian()) + StringLE64(buf, Value); + else + StringBE64(buf, Value); + F.getContents() += StringRef(buf, 8); + } + + virtual void WriteHeader(uint64_t SectionDataSize, unsigned NumberOfSections); + + /// Default e_flags = 0 + virtual void WriteEFlags() { Write32(0); } + + virtual void WriteSymbolEntry(MCDataFragment *SymtabF, MCDataFragment *ShndxF, + uint64_t name, uint8_t info, + uint64_t value, uint64_t size, + uint8_t other, uint32_t shndx, + bool Reserved); + + virtual void WriteSymbol(MCDataFragment *SymtabF, MCDataFragment *ShndxF, + ELFSymbolData &MSD, + const MCAsmLayout &Layout); + + typedef DenseMap<const MCSectionELF*, uint32_t> SectionIndexMapTy; + virtual void WriteSymbolTable(MCDataFragment *SymtabF, MCDataFragment *ShndxF, + const MCAssembler &Asm, + const MCAsmLayout &Layout, + const SectionIndexMapTy &SectionIndexMap); + + virtual void RecordRelocation(const MCAssembler &Asm, const MCAsmLayout &Layout, + const MCFragment *Fragment, const MCFixup &Fixup, + MCValue Target, uint64_t &FixedValue); + + virtual uint64_t getSymbolIndexInSymbolTable(const MCAssembler &Asm, + const MCSymbol *S); + + // Map from a group section to the signature symbol + typedef DenseMap<const MCSectionELF*, const MCSymbol*> GroupMapTy; + // Map from a signature symbol to the group section + typedef DenseMap<const MCSymbol*, const MCSectionELF*> RevGroupMapTy; + + /// ComputeSymbolTable - Compute the symbol table data + /// + /// \param StringTable [out] - The string table data. + /// \param StringIndexMap [out] - Map from symbol names to offsets in the + /// string table. + virtual void ComputeSymbolTable(MCAssembler &Asm, + const SectionIndexMapTy &SectionIndexMap, + RevGroupMapTy RevGroupMap); + + virtual void ComputeIndexMap(MCAssembler &Asm, + SectionIndexMapTy &SectionIndexMap); + + virtual void WriteRelocation(MCAssembler &Asm, MCAsmLayout &Layout, + const MCSectionData &SD); + + virtual void WriteRelocations(MCAssembler &Asm, MCAsmLayout &Layout) { + for (MCAssembler::const_iterator it = Asm.begin(), + ie = Asm.end(); it != ie; ++it) { + WriteRelocation(Asm, Layout, *it); + } + } + + virtual void CreateMetadataSections(MCAssembler &Asm, MCAsmLayout &Layout, + const SectionIndexMapTy &SectionIndexMap); + + // Create the sections that show up in the symbol table. Currently + // those are the .note.GNU-stack section and the group sections. + virtual void CreateIndexedSections(MCAssembler &Asm, MCAsmLayout &Layout, + GroupMapTy &GroupMap, + RevGroupMapTy &RevGroupMap); + + virtual void ExecutePostLayoutBinding(MCAssembler &Asm, + const MCAsmLayout &Layout); + + virtual void WriteSecHdrEntry(uint32_t Name, uint32_t Type, uint64_t Flags, + uint64_t Address, uint64_t Offset, + uint64_t Size, uint32_t Link, uint32_t Info, + uint64_t Alignment, uint64_t EntrySize); + + virtual void WriteRelocationsFragment(const MCAssembler &Asm, + MCDataFragment *F, + const MCSectionData *SD); + + virtual bool + IsSymbolRefDifferenceFullyResolvedImpl(const MCAssembler &Asm, + const MCSymbolData &DataA, + const MCFragment &FB, + bool InSet, + bool IsPCRel) const; + + virtual void WriteObject(MCAssembler &Asm, const MCAsmLayout &Layout); + virtual void WriteSection(MCAssembler &Asm, + const SectionIndexMapTy &SectionIndexMap, + uint32_t GroupSymbolIndex, + uint64_t Offset, uint64_t Size, uint64_t Alignment, + const MCSectionELF &Section); + + protected: + virtual unsigned GetRelocType(const MCValue &Target, const MCFixup &Fixup, + bool IsPCRel, bool IsRelocWithSymbol, + int64_t Addend) = 0; + }; + + //===- X86ELFObjectWriter -------------------------------------------===// + + class X86ELFObjectWriter : public ELFObjectWriter { + public: + X86ELFObjectWriter(MCELFObjectTargetWriter *MOTW, + raw_ostream &_OS, + bool IsLittleEndian); + + virtual ~X86ELFObjectWriter(); + protected: + virtual unsigned GetRelocType(const MCValue &Target, const MCFixup &Fixup, + bool IsPCRel, bool IsRelocWithSymbol, + int64_t Addend); + }; + + + //===- ARMELFObjectWriter -------------------------------------------===// + + class ARMELFObjectWriter : public ELFObjectWriter { + public: + // FIXME: MCAssembler can't yet return the Subtarget, + enum { DefaultEABIVersion = 0x05000000U }; + + ARMELFObjectWriter(MCELFObjectTargetWriter *MOTW, + raw_ostream &_OS, + bool IsLittleEndian); + + virtual ~ARMELFObjectWriter(); + + virtual void WriteEFlags(); + protected: + virtual const MCSymbol *ExplicitRelSym(const MCAssembler &Asm, + const MCValue &Target, + const MCFragment &F, + bool IsBSS) const; + + virtual unsigned GetRelocType(const MCValue &Target, const MCFixup &Fixup, + bool IsPCRel, bool IsRelocWithSymbol, + int64_t Addend); + }; + + //===- MBlazeELFObjectWriter -------------------------------------------===// + + class MBlazeELFObjectWriter : public ELFObjectWriter { + public: + MBlazeELFObjectWriter(MCELFObjectTargetWriter *MOTW, + raw_ostream &_OS, + bool IsLittleEndian); + + virtual ~MBlazeELFObjectWriter(); + protected: + virtual unsigned GetRelocType(const MCValue &Target, const MCFixup &Fixup, + bool IsPCRel, bool IsRelocWithSymbol, + int64_t Addend); + }; +} + +#endif diff --git a/lib/MC/MCAsmInfo.cpp b/lib/MC/MCAsmInfo.cpp index e8902e7..8199fb2 100644 --- a/lib/MC/MCAsmInfo.cpp +++ b/lib/MC/MCAsmInfo.cpp @@ -13,7 +13,7 @@ //===----------------------------------------------------------------------===// #include "llvm/MC/MCAsmInfo.h" -#include "llvm/System/DataTypes.h" +#include "llvm/Support/DataTypes.h" #include <cctype> #include <cstring> using namespace llvm; @@ -23,6 +23,7 @@ MCAsmInfo::MCAsmInfo() { HasMachoZeroFillDirective = false; HasMachoTBSSDirective = false; HasStaticCtorDtorReferenceInStaticMode = false; + LinkerRequiresNonEmptyDwarfLines = false; MaxInstLength = 4; PCSymbol = "$"; SeparatorChar = ';'; @@ -53,18 +54,20 @@ MCAsmInfo::MCAsmInfo() { GPRel32Directive = 0; GlobalDirective = "\t.globl\t"; HasSetDirective = true; + HasAggressiveSymbolFolding = true; HasLCOMMDirective = false; COMMDirectiveAlignmentIsInBytes = true; HasDotTypeDotSizeDirective = true; HasSingleParameterDotFile = true; HasNoDeadStrip = false; + HasSymbolResolver = false; WeakRefDirective = 0; WeakDefDirective = 0; LinkOnceDirective = 0; HiddenVisibilityAttr = MCSA_Hidden; + HiddenDeclarationVisibilityAttr = MCSA_Hidden; ProtectedVisibilityAttr = MCSA_Protected; HasLEB128 = false; - HasDotLocAndDotFile = false; SupportsDebugInformation = false; ExceptionsType = ExceptionHandling::None; DwarfRequiresFrameSection = true; diff --git a/lib/MC/MCAsmInfoDarwin.cpp b/lib/MC/MCAsmInfoDarwin.cpp index e0e261a..526ad0d 100644 --- a/lib/MC/MCAsmInfoDarwin.cpp +++ b/lib/MC/MCAsmInfoDarwin.cpp @@ -37,13 +37,21 @@ MCAsmInfoDarwin::MCAsmInfoDarwin() { HasMachoZeroFillDirective = true; // Uses .zerofill HasMachoTBSSDirective = true; // Uses .tbss HasStaticCtorDtorReferenceInStaticMode = true; - + + // FIXME: Darwin 10 and newer don't need this. + LinkerRequiresNonEmptyDwarfLines = true; + + // FIXME: Change this once MC is the system assembler. + HasAggressiveSymbolFolding = false; + HiddenVisibilityAttr = MCSA_PrivateExtern; + HiddenDeclarationVisibilityAttr = MCSA_Invalid; // Doesn't support protected visibility. ProtectedVisibilityAttr = MCSA_Global; HasDotTypeDotSizeDirective = false; HasNoDeadStrip = true; + HasSymbolResolver = true; DwarfUsesAbsoluteLabelForStmtList = false; DwarfUsesLabelOffsetForRanges = false; diff --git a/lib/MC/MCAsmStreamer.cpp b/lib/MC/MCAsmStreamer.cpp index 692f178..c7ecf53 100644 --- a/lib/MC/MCAsmStreamer.cpp +++ b/lib/MC/MCAsmStreamer.cpp @@ -12,6 +12,7 @@ #include "llvm/MC/MCCodeEmitter.h" #include "llvm/MC/MCContext.h" #include "llvm/MC/MCExpr.h" +#include "llvm/MC/MCFixupKindInfo.h" #include "llvm/MC/MCInst.h" #include "llvm/MC/MCInstPrinter.h" #include "llvm/MC/MCSectionMachO.h" @@ -23,7 +24,10 @@ #include "llvm/Support/MathExtras.h" #include "llvm/Support/Format.h" #include "llvm/Support/FormattedStream.h" -#include <ctype.h> +#include "llvm/Target/TargetAsmBackend.h" +#include "llvm/Target/TargetAsmInfo.h" +#include "llvm/Target/TargetLoweringObjectFile.h" +#include <cctype> using namespace llvm; namespace { @@ -33,29 +37,33 @@ class MCAsmStreamer : public MCStreamer { const MCAsmInfo &MAI; OwningPtr<MCInstPrinter> InstPrinter; OwningPtr<MCCodeEmitter> Emitter; + OwningPtr<TargetAsmBackend> AsmBackend; SmallString<128> CommentToEmit; raw_svector_ostream CommentStream; - unsigned IsLittleEndian : 1; unsigned IsVerboseAsm : 1; unsigned ShowInst : 1; + unsigned UseLoc : 1; + + bool needsSet(const MCExpr *Value); public: MCAsmStreamer(MCContext &Context, formatted_raw_ostream &os, - bool isLittleEndian, bool isVerboseAsm, MCInstPrinter *printer, - MCCodeEmitter *emitter, bool showInst) + bool isVerboseAsm, + bool useLoc, + MCInstPrinter *printer, MCCodeEmitter *emitter, + TargetAsmBackend *asmbackend, + bool showInst) : MCStreamer(Context), OS(os), MAI(Context.getAsmInfo()), - InstPrinter(printer), Emitter(emitter), CommentStream(CommentToEmit), - IsLittleEndian(isLittleEndian), IsVerboseAsm(isVerboseAsm), - ShowInst(showInst) { + InstPrinter(printer), Emitter(emitter), AsmBackend(asmbackend), + CommentStream(CommentToEmit), IsVerboseAsm(isVerboseAsm), + ShowInst(showInst), UseLoc(useLoc) { if (InstPrinter && IsVerboseAsm) InstPrinter->setCommentStream(CommentStream); } ~MCAsmStreamer() {} - bool isLittleEndian() const { return IsLittleEndian; } - inline void EmitEOL() { // If we don't have any comments, just emit a \n. if (!IsVerboseAsm) { @@ -99,7 +107,7 @@ public: /// @name MCStreamer Interface /// @{ - virtual void SwitchSection(const MCSection *Section); + virtual void ChangeSection(const MCSection *Section); virtual void InitSections() { // FIXME, this is MachO specific, but the testsuite @@ -116,6 +124,9 @@ public: virtual void EmitAssignment(MCSymbol *Symbol, const MCExpr *Value); virtual void EmitWeakReference(MCSymbol *Alias, const MCSymbol *Symbol); + virtual void EmitDwarfAdvanceLineAddr(int64_t LineDelta, + const MCSymbol *LastLabel, + const MCSymbol *Label); virtual void EmitSymbolAttribute(MCSymbol *Symbol, MCSymbolAttr Attribute); @@ -142,9 +153,10 @@ public: virtual void EmitBytes(StringRef Data, unsigned AddrSpace); - virtual void EmitValue(const MCExpr *Value, unsigned Size,unsigned AddrSpace); - - virtual void EmitIntValue(uint64_t Value, unsigned Size, unsigned AddrSpace); + virtual void EmitValueImpl(const MCExpr *Value, unsigned Size, + bool isPCRel, unsigned AddrSpace); + virtual void EmitIntValue(uint64_t Value, unsigned Size, + unsigned AddrSpace = 0); virtual void EmitULEB128Value(const MCExpr *Value, unsigned AddrSpace = 0); @@ -167,7 +179,28 @@ public: unsigned char Value = 0); virtual void EmitFileDirective(StringRef Filename); - virtual void EmitDwarfFileDirective(unsigned FileNo, StringRef Filename); + virtual bool EmitDwarfFileDirective(unsigned FileNo, StringRef Filename); + virtual void EmitDwarfLocDirective(unsigned FileNo, unsigned Line, + unsigned Column, unsigned Flags, + unsigned Isa, unsigned Discriminator); + + virtual bool EmitCFIStartProc(); + virtual bool EmitCFIEndProc(); + virtual bool EmitCFIDefCfaOffset(int64_t Offset); + virtual bool EmitCFIDefCfaRegister(int64_t Register); + virtual bool EmitCFIOffset(int64_t Register, int64_t Offset); + virtual bool EmitCFIPersonality(const MCSymbol *Sym, unsigned Encoding); + virtual bool EmitCFILsda(const MCSymbol *Sym, unsigned Encoding); + + virtual void EmitFnStart(); + virtual void EmitFnEnd(); + virtual void EmitCantUnwind(); + virtual void EmitPersonality(const MCSymbol *Personality); + virtual void EmitHandlerData(); + virtual void EmitSetFP(unsigned FpReg, unsigned SpReg, int64_t Offset = 0); + virtual void EmitPad(int64_t Offset); + virtual void EmitRegSave(const SmallVectorImpl<unsigned> &RegList, bool); + virtual void EmitInstruction(const MCInst &Inst); @@ -231,23 +264,19 @@ static inline int64_t truncateToSize(int64_t Value, unsigned Bytes) { return Value & ((uint64_t) (int64_t) -1 >> (64 - Bytes * 8)); } -void MCAsmStreamer::SwitchSection(const MCSection *Section) { +void MCAsmStreamer::ChangeSection(const MCSection *Section) { assert(Section && "Cannot switch to a null section!"); - if (Section != CurSection) { - PrevSection = CurSection; - CurSection = Section; - Section->PrintSwitchToSection(MAI, OS); - } + Section->PrintSwitchToSection(MAI, OS); } void MCAsmStreamer::EmitLabel(MCSymbol *Symbol) { assert(Symbol->isUndefined() && "Cannot define a symbol twice!"); assert(!Symbol->isVariable() && "Cannot emit a variable symbol!"); - assert(CurSection && "Cannot emit before setting section!"); + assert(getCurrentSection() && "Cannot emit before setting section!"); OS << *Symbol << MAI.getLabelSuffix(); EmitEOL(); - Symbol->setSection(*CurSection); + Symbol->setSection(*getCurrentSection()); } void MCAsmStreamer::EmitAssemblerFlag(MCAssemblerFlag Flag) { @@ -283,6 +312,13 @@ void MCAsmStreamer::EmitWeakReference(MCSymbol *Alias, const MCSymbol *Symbol) { EmitEOL(); } +void MCAsmStreamer::EmitDwarfAdvanceLineAddr(int64_t LineDelta, + const MCSymbol *LastLabel, + const MCSymbol *Label) { + EmitDwarfSetLineAddr(LineDelta, Label, + getContext().getTargetAsmInfo().getPointerSize()); +} + void MCAsmStreamer::EmitSymbolAttribute(MCSymbol *Symbol, MCSymbolAttr Attribute) { switch (Attribute) { @@ -318,6 +354,7 @@ void MCAsmStreamer::EmitSymbolAttribute(MCSymbol *Symbol, case MCSA_LazyReference: OS << "\t.lazy_reference\t"; break; case MCSA_Local: OS << "\t.local\t"; break; case MCSA_NoDeadStrip: OS << "\t.no_dead_strip\t"; break; + case MCSA_SymbolResolver: OS << "\t.symbol_resolver\t"; break; case MCSA_PrivateExtern: OS << "\t.private_extern\t"; break; case MCSA_Protected: OS << "\t.protected\t"; break; case MCSA_Reference: OS << "\t.reference\t"; break; @@ -455,7 +492,7 @@ static void PrintQuotedString(StringRef Data, raw_ostream &OS) { void MCAsmStreamer::EmitBytes(StringRef Data, unsigned AddrSpace) { - assert(CurSection && "Cannot emit contents before setting section!"); + assert(getCurrentSection() && "Cannot emit contents before setting section!"); if (Data.empty()) return; if (Data.size() == 1) { @@ -479,11 +516,15 @@ void MCAsmStreamer::EmitBytes(StringRef Data, unsigned AddrSpace) { EmitEOL(); } -/// EmitIntValue - Special case of EmitValue that avoids the client having -/// to pass in a MCExpr for constant integers. void MCAsmStreamer::EmitIntValue(uint64_t Value, unsigned Size, unsigned AddrSpace) { - assert(CurSection && "Cannot emit contents before setting section!"); + EmitValue(MCConstantExpr::Create(Value, getContext()), Size, AddrSpace); +} + +void MCAsmStreamer::EmitValueImpl(const MCExpr *Value, unsigned Size, + bool isPCRel, unsigned AddrSpace) { + assert(getCurrentSection() && "Cannot emit contents before setting section!"); + assert(!isPCRel && "Cannot emit pc relative relocations!"); const char *Directive = 0; switch (Size) { default: break; @@ -494,45 +535,41 @@ void MCAsmStreamer::EmitIntValue(uint64_t Value, unsigned Size, Directive = MAI.getData64bitsDirective(AddrSpace); // If the target doesn't support 64-bit data, emit as two 32-bit halves. if (Directive) break; - if (isLittleEndian()) { - EmitIntValue((uint32_t)(Value >> 0 ), 4, AddrSpace); - EmitIntValue((uint32_t)(Value >> 32), 4, AddrSpace); + int64_t IntValue; + if (!Value->EvaluateAsAbsolute(IntValue)) + report_fatal_error("Don't know how to emit this value."); + if (getContext().getTargetAsmInfo().isLittleEndian()) { + EmitIntValue((uint32_t)(IntValue >> 0 ), 4, AddrSpace); + EmitIntValue((uint32_t)(IntValue >> 32), 4, AddrSpace); } else { - EmitIntValue((uint32_t)(Value >> 32), 4, AddrSpace); - EmitIntValue((uint32_t)(Value >> 0 ), 4, AddrSpace); + EmitIntValue((uint32_t)(IntValue >> 32), 4, AddrSpace); + EmitIntValue((uint32_t)(IntValue >> 0 ), 4, AddrSpace); } return; } assert(Directive && "Invalid size for machine code value!"); - OS << Directive << truncateToSize(Value, Size); - EmitEOL(); -} - -void MCAsmStreamer::EmitValue(const MCExpr *Value, unsigned Size, - unsigned AddrSpace) { - assert(CurSection && "Cannot emit contents before setting section!"); - const char *Directive = 0; - switch (Size) { - default: break; - case 1: Directive = MAI.getData8bitsDirective(AddrSpace); break; - case 2: Directive = MAI.getData16bitsDirective(AddrSpace); break; - case 4: Directive = MAI.getData32bitsDirective(AddrSpace); break; - case 8: Directive = MAI.getData64bitsDirective(AddrSpace); break; - } - - assert(Directive && "Invalid size for machine code value!"); OS << Directive << *Value; EmitEOL(); } void MCAsmStreamer::EmitULEB128Value(const MCExpr *Value, unsigned AddrSpace) { + int64_t IntValue; + if (Value->EvaluateAsAbsolute(IntValue)) { + EmitULEB128IntValue(IntValue, AddrSpace); + return; + } assert(MAI.hasLEB128() && "Cannot print a .uleb"); OS << ".uleb128 " << *Value; EmitEOL(); } void MCAsmStreamer::EmitSLEB128Value(const MCExpr *Value, unsigned AddrSpace) { + int64_t IntValue; + if (Value->EvaluateAsAbsolute(IntValue)) { + EmitSLEB128IntValue(IntValue, AddrSpace); + return; + } assert(MAI.hasLEB128() && "Cannot print a .sleb"); OS << ".sleb128 " << *Value; EmitEOL(); @@ -634,10 +671,118 @@ void MCAsmStreamer::EmitFileDirective(StringRef Filename) { EmitEOL(); } -void MCAsmStreamer::EmitDwarfFileDirective(unsigned FileNo, StringRef Filename){ - OS << "\t.file\t" << FileNo << ' '; - PrintQuotedString(Filename, OS); +bool MCAsmStreamer::EmitDwarfFileDirective(unsigned FileNo, StringRef Filename){ + if (UseLoc) { + OS << "\t.file\t" << FileNo << ' '; + PrintQuotedString(Filename, OS); + EmitEOL(); + } + return this->MCStreamer::EmitDwarfFileDirective(FileNo, Filename); +} + +void MCAsmStreamer::EmitDwarfLocDirective(unsigned FileNo, unsigned Line, + unsigned Column, unsigned Flags, + unsigned Isa, + unsigned Discriminator) { + this->MCStreamer::EmitDwarfLocDirective(FileNo, Line, Column, Flags, + Isa, Discriminator); + if (!UseLoc) + return; + + OS << "\t.loc\t" << FileNo << " " << Line << " " << Column; + if (Flags & DWARF2_FLAG_BASIC_BLOCK) + OS << " basic_block"; + if (Flags & DWARF2_FLAG_PROLOGUE_END) + OS << " prologue_end"; + if (Flags & DWARF2_FLAG_EPILOGUE_BEGIN) + OS << " epilogue_begin"; + + unsigned OldFlags = getContext().getCurrentDwarfLoc().getFlags(); + if ((Flags & DWARF2_FLAG_IS_STMT) != (OldFlags & DWARF2_FLAG_IS_STMT)) { + OS << " is_stmt "; + + if (Flags & DWARF2_FLAG_IS_STMT) + OS << "1"; + else + OS << "0"; + } + + if (Isa) + OS << "isa " << Isa; + if (Discriminator) + OS << "discriminator " << Discriminator; + EmitEOL(); +} + +bool MCAsmStreamer::EmitCFIStartProc() { + if (this->MCStreamer::EmitCFIStartProc()) + return true; + + OS << "\t.cfi_startproc"; + EmitEOL(); + + return false; +} + +bool MCAsmStreamer::EmitCFIEndProc() { + if (this->MCStreamer::EmitCFIEndProc()) + return true; + + OS << "\t.cfi_endproc"; EmitEOL(); + + return false; +} + +bool MCAsmStreamer::EmitCFIDefCfaOffset(int64_t Offset) { + if (this->MCStreamer::EmitCFIDefCfaOffset(Offset)) + return true; + + OS << "\t.cfi_def_cfa_offset " << Offset; + EmitEOL(); + + return false; +} + +bool MCAsmStreamer::EmitCFIDefCfaRegister(int64_t Register) { + if (this->MCStreamer::EmitCFIDefCfaRegister(Register)) + return true; + + OS << "\t.cfi_def_cfa_register " << Register; + EmitEOL(); + + return false; +} + +bool MCAsmStreamer::EmitCFIOffset(int64_t Register, int64_t Offset) { + if (this->MCStreamer::EmitCFIOffset(Register, Offset)) + return true; + + OS << "\t.cfi_offset " << Register << ", " << Offset; + EmitEOL(); + + return false; +} + +bool MCAsmStreamer::EmitCFIPersonality(const MCSymbol *Sym, + unsigned Encoding) { + if (this->MCStreamer::EmitCFIPersonality(Sym, Encoding)) + return true; + + OS << "\t.cfi_personality " << Encoding << ", " << *Sym; + EmitEOL(); + + return false; +} + +bool MCAsmStreamer::EmitCFILsda(const MCSymbol *Sym, unsigned Encoding) { + if (this->MCStreamer::EmitCFILsda(Sym, Encoding)) + return true; + + OS << "\t.cfi_lsda " << Encoding << ", " << *Sym; + EmitEOL(); + + return false; } void MCAsmStreamer::AddEncodingComment(const MCInst &Inst) { @@ -658,7 +803,7 @@ void MCAsmStreamer::AddEncodingComment(const MCInst &Inst) { for (unsigned i = 0, e = Fixups.size(); i != e; ++i) { MCFixup &F = Fixups[i]; - const MCFixupKindInfo &Info = Emitter->getFixupKindInfo(F.getKind()); + const MCFixupKindInfo &Info = AsmBackend->getFixupKindInfo(F.getKind()); for (unsigned j = 0; j != Info.TargetSize; ++j) { unsigned Index = F.getOffset() * 8 + Info.TargetOffset + j; assert(Index < Code.size() * 8 && "Invalid offset in fixup!"); @@ -666,6 +811,8 @@ void MCAsmStreamer::AddEncodingComment(const MCInst &Inst) { } } + // FIXME: Node the fixup comments for Thumb2 are completely bogus since the + // high order halfword of a 32-bit Thumb2 instruction is emitted first. OS << "encoding: ["; for (unsigned i = 0, e = Code.size(); i != e; ++i) { if (i) @@ -685,8 +832,12 @@ void MCAsmStreamer::AddEncodingComment(const MCInst &Inst) { if (MapEntry == 0) { OS << format("0x%02x", uint8_t(Code[i])); } else { - assert(Code[i] == 0 && "Encoder wrote into fixed up bit!"); - OS << char('A' + MapEntry - 1); + if (Code[i]) { + // FIXME: Some of the 8 bits require fix up. + OS << format("0x%02x", uint8_t(Code[i])) << '\'' + << char('A' + MapEntry - 1) << '\''; + } else + OS << char('A' + MapEntry - 1); } } else { // Otherwise, write out in binary. @@ -695,7 +846,7 @@ void MCAsmStreamer::AddEncodingComment(const MCInst &Inst) { unsigned Bit = (Code[i] >> j) & 1; unsigned FixupBit; - if (IsLittleEndian) + if (getContext().getTargetAsmInfo().isLittleEndian()) FixupBit = i * 8 + j; else FixupBit = i * 8 + (7-j); @@ -712,14 +863,72 @@ void MCAsmStreamer::AddEncodingComment(const MCInst &Inst) { for (unsigned i = 0, e = Fixups.size(); i != e; ++i) { MCFixup &F = Fixups[i]; - const MCFixupKindInfo &Info = Emitter->getFixupKindInfo(F.getKind()); + const MCFixupKindInfo &Info = AsmBackend->getFixupKindInfo(F.getKind()); OS << " fixup " << char('A' + i) << " - " << "offset: " << F.getOffset() << ", value: " << *F.getValue() << ", kind: " << Info.Name << "\n"; } } +void MCAsmStreamer::EmitFnStart() { + OS << "\t.fnstart"; + EmitEOL(); +} + +void MCAsmStreamer::EmitFnEnd() { + OS << "\t.fnend"; + EmitEOL(); +} + +void MCAsmStreamer::EmitCantUnwind() { + OS << "\t.cantunwind"; + EmitEOL(); +} + +void MCAsmStreamer::EmitHandlerData() { + OS << "\t.handlerdata"; + EmitEOL(); +} + +void MCAsmStreamer::EmitPersonality(const MCSymbol *Personality) { + OS << "\t.personality " << Personality->getName(); + EmitEOL(); +} + +void MCAsmStreamer::EmitSetFP(unsigned FpReg, unsigned SpReg, int64_t Offset) { + OS << "\t.setfp\t" << InstPrinter->getRegName(FpReg) + << ", " << InstPrinter->getRegName(SpReg); + if (Offset) + OS << ", #" << Offset; + EmitEOL(); +} + +void MCAsmStreamer::EmitPad(int64_t Offset) { + OS << "\t.pad\t#" << Offset; + EmitEOL(); +} + +void MCAsmStreamer::EmitRegSave(const SmallVectorImpl<unsigned> &RegList, + bool isVector) { + assert(RegList.size() && "RegList should not be empty"); + if (isVector) + OS << "\t.vsave\t{"; + else + OS << "\t.save\t{"; + + OS << InstPrinter->getRegName(RegList[0]); + + for (unsigned i = 1, e = RegList.size(); i != e; ++i) + OS << ", " << InstPrinter->getRegName(RegList[i]); + + OS << "}"; + EmitEOL(); +} + void MCAsmStreamer::EmitInstruction(const MCInst &Inst) { - assert(CurSection && "Cannot emit contents before setting section!"); + assert(getCurrentSection() && "Cannot emit contents before setting section!"); + + if (!UseLoc) + MCLineEntry::Make(this, getCurrentSection()); // Show the encoding in a comment if we have a code emitter. if (Emitter) @@ -750,13 +959,16 @@ void MCAsmStreamer::EmitRawText(StringRef String) { } void MCAsmStreamer::Finish() { + // Dump out the dwarf file & directory tables and line tables. + if (getContext().hasDwarfFiles() && !UseLoc) + MCDwarfFileTable::Emit(this); } MCStreamer *llvm::createAsmStreamer(MCContext &Context, formatted_raw_ostream &OS, - bool isLittleEndian, - bool isVerboseAsm, MCInstPrinter *IP, - MCCodeEmitter *CE, bool ShowInst) { - return new MCAsmStreamer(Context, OS, isLittleEndian, isVerboseAsm, - IP, CE, ShowInst); + bool isVerboseAsm, bool useLoc, + MCInstPrinter *IP, MCCodeEmitter *CE, + TargetAsmBackend *TAB, bool ShowInst) { + return new MCAsmStreamer(Context, OS, isVerboseAsm, useLoc, + IP, CE, TAB, ShowInst); } diff --git a/lib/MC/MCAssembler.cpp b/lib/MC/MCAssembler.cpp index c80dc3c..9992646 100644 --- a/lib/MC/MCAssembler.cpp +++ b/lib/MC/MCAssembler.cpp @@ -11,6 +11,7 @@ #include "llvm/MC/MCAssembler.h" #include "llvm/MC/MCAsmLayout.h" #include "llvm/MC/MCCodeEmitter.h" +#include "llvm/MC/MCContext.h" #include "llvm/MC/MCExpr.h" #include "llvm/MC/MCObjectWriter.h" #include "llvm/MC/MCSection.h" @@ -38,7 +39,6 @@ STATISTIC(FragmentLayouts, "Number of fragment layouts"); STATISTIC(ObjectBytes, "Number of emitted object file bytes"); STATISTIC(RelaxationSteps, "Number of assembler layout and relaxation steps"); STATISTIC(RelaxedInstructions, "Number of relaxed instructions"); -STATISTIC(SectionLayouts, "Number of section layouts"); } } @@ -50,131 +50,78 @@ STATISTIC(SectionLayouts, "Number of section layouts"); /* *** */ MCAsmLayout::MCAsmLayout(MCAssembler &Asm) - : Assembler(Asm), LastValidFragment(0) + : Assembler(Asm), LastValidFragment() { // Compute the section layout order. Virtual sections must go last. for (MCAssembler::iterator it = Asm.begin(), ie = Asm.end(); it != ie; ++it) - if (!Asm.getBackend().isVirtualSection(it->getSection())) + if (!it->getSection().isVirtualSection()) SectionOrder.push_back(&*it); for (MCAssembler::iterator it = Asm.begin(), ie = Asm.end(); it != ie; ++it) - if (Asm.getBackend().isVirtualSection(it->getSection())) + if (it->getSection().isVirtualSection()) SectionOrder.push_back(&*it); } -bool MCAsmLayout::isSectionUpToDate(const MCSectionData *SD) const { - // The first section is always up-to-date. - unsigned Index = SD->getLayoutOrder(); - if (!Index) - return true; - - // Otherwise, sections are always implicitly computed when the preceeding - // fragment is layed out. - const MCSectionData *Prev = getSectionOrder()[Index - 1]; - return isFragmentUpToDate(&(Prev->getFragmentList().back())); -} - bool MCAsmLayout::isFragmentUpToDate(const MCFragment *F) const { - return (LastValidFragment && - F->getLayoutOrder() <= LastValidFragment->getLayoutOrder()); + const MCSectionData &SD = *F->getParent(); + const MCFragment *LastValid = LastValidFragment.lookup(&SD); + if (!LastValid) + return false; + assert(LastValid->getParent() == F->getParent()); + return F->getLayoutOrder() <= LastValid->getLayoutOrder(); } -void MCAsmLayout::UpdateForSlide(MCFragment *F, int SlideAmount) { +void MCAsmLayout::Invalidate(MCFragment *F) { // If this fragment wasn't already up-to-date, we don't need to do anything. if (!isFragmentUpToDate(F)) return; - // Otherwise, reset the last valid fragment to the predecessor of the - // invalidated fragment. - LastValidFragment = F->getPrevNode(); - if (!LastValidFragment) { - unsigned Index = F->getParent()->getLayoutOrder(); - if (Index != 0) { - MCSectionData *Prev = getSectionOrder()[Index - 1]; - LastValidFragment = &(Prev->getFragmentList().back()); - } - } + // Otherwise, reset the last valid fragment to this fragment. + const MCSectionData &SD = *F->getParent(); + LastValidFragment[&SD] = F; } void MCAsmLayout::EnsureValid(const MCFragment *F) const { + MCSectionData &SD = *F->getParent(); + + MCFragment *Cur = LastValidFragment[&SD]; + if (!Cur) + Cur = &*SD.begin(); + else + Cur = Cur->getNextNode(); + // Advance the layout position until the fragment is up-to-date. while (!isFragmentUpToDate(F)) { - // Advance to the next fragment. - MCFragment *Cur = LastValidFragment; - if (Cur) - Cur = Cur->getNextNode(); - if (!Cur) { - unsigned NextIndex = 0; - if (LastValidFragment) - NextIndex = LastValidFragment->getParent()->getLayoutOrder() + 1; - Cur = SectionOrder[NextIndex]->begin(); - } - const_cast<MCAsmLayout*>(this)->LayoutFragment(Cur); + Cur = Cur->getNextNode(); } } -void MCAsmLayout::FragmentReplaced(MCFragment *Src, MCFragment *Dst) { - if (LastValidFragment == Src) - LastValidFragment = Dst; - - Dst->Offset = Src->Offset; - Dst->EffectiveSize = Src->EffectiveSize; -} - -uint64_t MCAsmLayout::getFragmentAddress(const MCFragment *F) const { - assert(F->getParent() && "Missing section()!"); - return getSectionAddress(F->getParent()) + getFragmentOffset(F); -} - -uint64_t MCAsmLayout::getFragmentEffectiveSize(const MCFragment *F) const { - EnsureValid(F); - assert(F->EffectiveSize != ~UINT64_C(0) && "Address not set!"); - return F->EffectiveSize; -} - uint64_t MCAsmLayout::getFragmentOffset(const MCFragment *F) const { EnsureValid(F); assert(F->Offset != ~UINT64_C(0) && "Address not set!"); return F->Offset; } -uint64_t MCAsmLayout::getSymbolAddress(const MCSymbolData *SD) const { - assert(SD->getFragment() && "Invalid getAddress() on undefined symbol!"); - return getFragmentAddress(SD->getFragment()) + SD->getOffset(); -} - -uint64_t MCAsmLayout::getSectionAddress(const MCSectionData *SD) const { - EnsureValid(SD->begin()); - assert(SD->Address != ~UINT64_C(0) && "Address not set!"); - return SD->Address; +uint64_t MCAsmLayout::getSymbolOffset(const MCSymbolData *SD) const { + assert(SD->getFragment() && "Invalid getOffset() on undefined symbol!"); + return getFragmentOffset(SD->getFragment()) + SD->getOffset(); } uint64_t MCAsmLayout::getSectionAddressSize(const MCSectionData *SD) const { // The size is the last fragment's end offset. const MCFragment &F = SD->getFragmentList().back(); - return getFragmentOffset(&F) + getFragmentEffectiveSize(&F); + return getFragmentOffset(&F) + getAssembler().ComputeFragmentSize(*this, F); } uint64_t MCAsmLayout::getSectionFileSize(const MCSectionData *SD) const { // Virtual sections have no file size. - if (getAssembler().getBackend().isVirtualSection(SD->getSection())) + if (SD->getSection().isVirtualSection()) return 0; // Otherwise, the file size is the same as the address space size. return getSectionAddressSize(SD); } -uint64_t MCAsmLayout::getSectionSize(const MCSectionData *SD) const { - // The logical size is the address space size minus any tail padding. - uint64_t Size = getSectionAddressSize(SD); - const MCAlignFragment *AF = - dyn_cast<MCAlignFragment>(&(SD->getFragmentList().back())); - if (AF && AF->hasOnlyAlignAddress()) - Size -= getFragmentEffectiveSize(AF); - - return Size; -} - /* *** */ MCFragment::MCFragment() : Kind(FragmentType(~0)) { @@ -184,8 +131,7 @@ MCFragment::~MCFragment() { } MCFragment::MCFragment(FragmentType _Kind, MCSectionData *_Parent) - : Kind(_Kind), Parent(_Parent), Atom(0), Offset(~UINT64_C(0)), - EffectiveSize(~UINT64_C(0)) + : Kind(_Kind), Parent(_Parent), Atom(0), Offset(~UINT64_C(0)) { if (Parent) Parent->getFragmentList().push_back(this); @@ -197,8 +143,8 @@ MCSectionData::MCSectionData() : Section(0) {} MCSectionData::MCSectionData(const MCSection &_Section, MCAssembler *A) : Section(&_Section), + Ordinal(~UINT32_C(0)), Alignment(1), - Address(~UINT64_C(0)), HasInstructions(false) { if (A) @@ -222,12 +168,11 @@ MCSymbolData::MCSymbolData(const MCSymbol &_Symbol, MCFragment *_Fragment, /* *** */ -MCAssembler::MCAssembler(MCContext &_Context, TargetAsmBackend &_Backend, - MCCodeEmitter &_Emitter, bool _PadSectionToAlignment, - raw_ostream &_OS) - : Context(_Context), Backend(_Backend), Emitter(_Emitter), - OS(_OS), RelaxAll(false), SubsectionsViaSymbols(false), - PadSectionToAlignment(_PadSectionToAlignment) +MCAssembler::MCAssembler(MCContext &Context_, TargetAsmBackend &Backend_, + MCCodeEmitter &Emitter_, MCObjectWriter &Writer_, + raw_ostream &OS_) + : Context(Context_), Backend(Backend_), Emitter(Emitter_), Writer(Writer_), + OS(OS_), RelaxAll(false), NoExecStack(false), SubsectionsViaSymbols(false) { } @@ -266,52 +211,83 @@ const MCSymbolData *MCAssembler::getAtom(const MCSymbolData *SD) const { return SD->getFragment()->getAtom(); } -bool MCAssembler::EvaluateFixup(const MCObjectWriter &Writer, - const MCAsmLayout &Layout, +bool MCAssembler::EvaluateFixup(const MCAsmLayout &Layout, const MCFixup &Fixup, const MCFragment *DF, MCValue &Target, uint64_t &Value) const { ++stats::EvaluateFixup; - if (!Fixup.getValue()->EvaluateAsRelocatable(Target, &Layout)) + if (!Fixup.getValue()->EvaluateAsRelocatable(Target, Layout)) report_fatal_error("expected relocatable expression"); - // FIXME: How do non-scattered symbols work in ELF? I presume the linker - // doesn't support small relocations, but then under what criteria does the - // assembler allow symbol differences? + bool IsPCRel = Backend.getFixupKindInfo( + Fixup.getKind()).Flags & MCFixupKindInfo::FKF_IsPCRel; + + bool IsResolved; + if (IsPCRel) { + if (Target.getSymB()) { + IsResolved = false; + } else if (!Target.getSymA()) { + IsResolved = false; + } else { + const MCSymbolRefExpr *A = Target.getSymA(); + const MCSymbol &SA = A->getSymbol(); + if (A->getKind() != MCSymbolRefExpr::VK_None || + SA.AliasedSymbol().isUndefined()) { + IsResolved = false; + } else { + const MCSymbolData &DataA = getSymbolData(SA); + IsResolved = + getWriter().IsSymbolRefDifferenceFullyResolvedImpl(*this, DataA, + *DF, false, true); + } + } + } else { + IsResolved = Target.isAbsolute(); + } Value = Target.getConstant(); - bool IsPCRel = Emitter.getFixupKindInfo( - Fixup.getKind()).Flags & MCFixupKindInfo::FKF_IsPCRel; - bool IsResolved = true; + bool IsThumb = false; if (const MCSymbolRefExpr *A = Target.getSymA()) { const MCSymbol &Sym = A->getSymbol().AliasedSymbol(); if (Sym.isDefined()) - Value += Layout.getSymbolAddress(&getSymbolData(Sym)); - else - IsResolved = false; + Value += Layout.getSymbolOffset(&getSymbolData(Sym)); + if (isThumbFunc(&Sym)) + IsThumb = true; } if (const MCSymbolRefExpr *B = Target.getSymB()) { const MCSymbol &Sym = B->getSymbol().AliasedSymbol(); if (Sym.isDefined()) - Value -= Layout.getSymbolAddress(&getSymbolData(Sym)); - else - IsResolved = false; + Value -= Layout.getSymbolOffset(&getSymbolData(Sym)); } - if (IsResolved) - IsResolved = Writer.IsFixupFullyResolved(*this, Target, IsPCRel, DF); - if (IsPCRel) - Value -= Layout.getFragmentAddress(DF) + Fixup.getOffset(); + bool ShouldAlignPC = Backend.getFixupKindInfo(Fixup.getKind()).Flags & + MCFixupKindInfo::FKF_IsAlignedDownTo32Bits; + assert((ShouldAlignPC ? IsPCRel : true) && + "FKF_IsAlignedDownTo32Bits is only allowed on PC-relative fixups!"); + + if (IsPCRel) { + uint32_t Offset = Layout.getFragmentOffset(DF) + Fixup.getOffset(); + + // A number of ARM fixups in Thumb mode require that the effective PC + // address be determined as the 32-bit aligned version of the actual offset. + if (ShouldAlignPC) Offset &= ~0x3; + Value -= Offset; + } + + // ARM fixups based from a thumb function address need to have the low + // bit set. The actual value is always at least 16-bit aligned, so the + // low bit is normally clear and available for use as an ISA flag for + // interworking. + if (IsThumb) + Value |= 1; return IsResolved; } -uint64_t MCAssembler::ComputeFragmentSize(MCAsmLayout &Layout, - const MCFragment &F, - uint64_t SectionAddress, - uint64_t FragmentOffset) const { +uint64_t MCAssembler::ComputeFragmentSize(const MCAsmLayout &Layout, + const MCFragment &F) const { switch (F.getKind()) { case MCFragment::FT_Data: return cast<MCDataFragment>(F).getContents().size(); @@ -321,52 +297,47 @@ uint64_t MCAssembler::ComputeFragmentSize(MCAsmLayout &Layout, return cast<MCInstFragment>(F).getInstSize(); case MCFragment::FT_LEB: - return cast<MCLEBFragment>(F).getSize(); + return cast<MCLEBFragment>(F).getContents().size(); case MCFragment::FT_Align: { const MCAlignFragment &AF = cast<MCAlignFragment>(F); - - assert((!AF.hasOnlyAlignAddress() || !AF.getNextNode()) && - "Invalid OnlyAlignAddress bit, not the last fragment!"); - - uint64_t Size = OffsetToAlignment(SectionAddress + FragmentOffset, - AF.getAlignment()); - - // Honor MaxBytesToEmit. + unsigned Offset = Layout.getFragmentOffset(&AF); + unsigned Size = OffsetToAlignment(Offset, AF.getAlignment()); if (Size > AF.getMaxBytesToEmit()) return 0; - return Size; } - case MCFragment::FT_Org: - return cast<MCOrgFragment>(F).getSize(); + case MCFragment::FT_Org: { + MCOrgFragment &OF = cast<MCOrgFragment>(F); + int64_t TargetLocation; + if (!OF.getOffset().EvaluateAsAbsolute(TargetLocation, Layout)) + report_fatal_error("expected assembly-time absolute expression"); + + // FIXME: We need a way to communicate this error. + uint64_t FragmentOffset = Layout.getFragmentOffset(&OF); + int64_t Size = TargetLocation - FragmentOffset; + if (Size < 0 || Size >= 0x40000000) + report_fatal_error("invalid .org offset '" + Twine(TargetLocation) + + "' (at offset '" + Twine(FragmentOffset) + "')"); + return Size; + } case MCFragment::FT_Dwarf: - return cast<MCDwarfLineAddrFragment>(F).getSize(); + return cast<MCDwarfLineAddrFragment>(F).getContents().size(); + case MCFragment::FT_DwarfFrame: + return cast<MCDwarfCallFrameFragment>(F).getContents().size(); } assert(0 && "invalid fragment kind"); return 0; } -void MCAsmLayout::LayoutFile() { - // Initialize the first section and set the valid fragment layout point. All - // actual layout computations are done lazily. - LastValidFragment = 0; - if (!getSectionOrder().empty()) - getSectionOrder().front()->Address = 0; -} - void MCAsmLayout::LayoutFragment(MCFragment *F) { MCFragment *Prev = F->getPrevNode(); // We should never try to recompute something which is up-to-date. assert(!isFragmentUpToDate(F) && "Attempt to recompute up-to-date fragment!"); - // We should never try to compute the fragment layout if the section isn't - // up-to-date. - assert(isSectionUpToDate(F->getParent()) && - "Attempt to compute fragment before it's section!"); // We should never try to compute the fragment layout if it's predecessor // isn't up-to-date. assert((!Prev || isFragmentUpToDate(Prev)) && @@ -374,55 +345,26 @@ void MCAsmLayout::LayoutFragment(MCFragment *F) { ++stats::FragmentLayouts; - // Compute the fragment start address. - uint64_t StartAddress = F->getParent()->Address; - uint64_t Address = StartAddress; - if (Prev) - Address += Prev->Offset + Prev->EffectiveSize; - // Compute fragment offset and size. - F->Offset = Address - StartAddress; - F->EffectiveSize = getAssembler().ComputeFragmentSize(*this, *F, StartAddress, - F->Offset); - LastValidFragment = F; - - // If this is the last fragment in a section, update the next section address. - if (!F->getNextNode()) { - unsigned NextIndex = F->getParent()->getLayoutOrder() + 1; - if (NextIndex != getSectionOrder().size()) - LayoutSection(getSectionOrder()[NextIndex]); - } -} - -void MCAsmLayout::LayoutSection(MCSectionData *SD) { - unsigned SectionOrderIndex = SD->getLayoutOrder(); - - ++stats::SectionLayouts; - - // Compute the section start address. - uint64_t StartAddress = 0; - if (SectionOrderIndex) { - MCSectionData *Prev = getSectionOrder()[SectionOrderIndex - 1]; - StartAddress = getSectionAddress(Prev) + getSectionAddressSize(Prev); - } - - // Honor the section alignment requirements. - StartAddress = RoundUpToAlignment(StartAddress, SD->getAlignment()); + uint64_t Offset = 0; + if (Prev) + Offset += Prev->Offset + getAssembler().ComputeFragmentSize(*this, *Prev); - // Set the section address. - SD->Address = StartAddress; + F->Offset = Offset; + LastValidFragment[F->getParent()] = F; } /// WriteFragmentData - Write the \arg F data to the output file. static void WriteFragmentData(const MCAssembler &Asm, const MCAsmLayout &Layout, - const MCFragment &F, MCObjectWriter *OW) { + const MCFragment &F) { + MCObjectWriter *OW = &Asm.getWriter(); uint64_t Start = OW->getStream().tell(); (void) Start; ++stats::EmittedFragments; // FIXME: Embed in fragments instead? - uint64_t FragmentSize = Layout.getFragmentEffectiveSize(&F); + uint64_t FragmentSize = Asm.ComputeFragmentSize(Layout, F); switch (F.getKind()) { case MCFragment::FT_Align: { MCAlignFragment &AF = cast<MCAlignFragment>(F); @@ -489,24 +431,15 @@ static void WriteFragmentData(const MCAssembler &Asm, const MCAsmLayout &Layout, break; } - case MCFragment::FT_Inst: - llvm_unreachable("unexpected inst fragment after lowering"); + case MCFragment::FT_Inst: { + MCInstFragment &IF = cast<MCInstFragment>(F); + OW->WriteBytes(StringRef(IF.getCode().begin(), IF.getCode().size())); break; + } case MCFragment::FT_LEB: { MCLEBFragment &LF = cast<MCLEBFragment>(F); - - // FIXME: It is probably better if we don't call EvaluateAsAbsolute in - // here. - int64_t Value; - LF.getValue().EvaluateAsAbsolute(Value, &Layout); - SmallString<32> Tmp; - raw_svector_ostream OSE(Tmp); - if (LF.isSigned()) - MCObjectWriter::EncodeSLEB128(Value, OSE); - else - MCObjectWriter::EncodeULEB128(Value, OSE); - OW->WriteBytes(OSE.str()); + OW->WriteBytes(LF.getContents().str()); break; } @@ -521,15 +454,12 @@ static void WriteFragmentData(const MCAssembler &Asm, const MCAsmLayout &Layout, case MCFragment::FT_Dwarf: { const MCDwarfLineAddrFragment &OF = cast<MCDwarfLineAddrFragment>(F); - - // The AddrDelta is really unsigned and it can only increase. - int64_t AddrDelta; - OF.getAddrDelta().EvaluateAsAbsolute(AddrDelta, &Layout); - - int64_t LineDelta; - LineDelta = OF.getLineDelta(); - - MCDwarfLineAddr::Write(OW, LineDelta, (uint64_t)AddrDelta); + OW->WriteBytes(OF.getContents().str()); + break; + } + case MCFragment::FT_DwarfFrame: { + const MCDwarfCallFrameFragment &CF = cast<MCDwarfCallFrameFragment>(F); + OW->WriteBytes(CF.getContents().str()); break; } } @@ -538,10 +468,9 @@ static void WriteFragmentData(const MCAssembler &Asm, const MCAsmLayout &Layout, } void MCAssembler::WriteSectionData(const MCSectionData *SD, - const MCAsmLayout &Layout, - MCObjectWriter *OW) const { + const MCAsmLayout &Layout) const { // Ignore virtual sections. - if (getBackend().isVirtualSection(SD->getSection())) { + if (SD->getSection().isVirtualSection()) { assert(Layout.getSectionFileSize(SD) == 0 && "Invalid size for section!"); // Check that contents are only things legal inside a virtual section. @@ -579,37 +508,34 @@ void MCAssembler::WriteSectionData(const MCSectionData *SD, return; } - uint64_t Start = OW->getStream().tell(); + uint64_t Start = getWriter().getStream().tell(); (void) Start; for (MCSectionData::const_iterator it = SD->begin(), ie = SD->end(); it != ie; ++it) - WriteFragmentData(*this, Layout, *it, OW); + WriteFragmentData(*this, Layout, *it); - assert(OW->getStream().tell() - Start == Layout.getSectionFileSize(SD)); + assert(getWriter().getStream().tell() - Start == + Layout.getSectionAddressSize(SD)); } -void MCAssembler::AddSectionToTheEnd(const MCObjectWriter &Writer, - MCSectionData &SD, MCAsmLayout &Layout) { - // Create dummy fragments and assign section ordinals. - unsigned SectionIndex = size(); - SD.setOrdinal(SectionIndex); - // Assign layout order indices to sections and fragments. - const MCFragment &Last = *Layout.getSectionOrder().back()->rbegin(); - unsigned FragmentIndex = Last.getLayoutOrder() + 1; +uint64_t MCAssembler::HandleFixup(const MCAsmLayout &Layout, + MCFragment &F, + const MCFixup &Fixup) { + // Evaluate the fixup. + MCValue Target; + uint64_t FixedValue; + if (!EvaluateFixup(Layout, Fixup, &F, Target, FixedValue)) { + // The fixup was unresolved, we need a relocation. Inform the object + // writer of the relocation, and give it an opportunity to adjust the + // fixup value if need be. + getWriter().RecordRelocation(*this, Layout, &F, Fixup, Target, FixedValue); + } + return FixedValue; + } - SD.setLayoutOrder(Layout.getSectionOrder().size()); - for (MCSectionData::iterator it2 = SD.begin(), - ie2 = SD.end(); it2 != ie2; ++it2) { - it2->setLayoutOrder(FragmentIndex++); - } - Layout.getSectionOrder().push_back(&SD); - - Layout.LayoutSection(&SD); -} - -void MCAssembler::Finish(MCObjectWriter *Writer) { +void MCAssembler::Finish() { DEBUG_WITH_TYPE("mc-dump", { llvm::errs() << "assembler backend - pre-layout\n--\n"; dump(); }); @@ -617,30 +543,6 @@ void MCAssembler::Finish(MCObjectWriter *Writer) { // Create the layout object. MCAsmLayout Layout(*this); - // Insert additional align fragments for concrete sections to explicitly pad - // the previous section to match their alignment requirements. This is for - // 'gas' compatibility, it shouldn't strictly be necessary. - if (PadSectionToAlignment) { - for (unsigned i = 1, e = Layout.getSectionOrder().size(); i < e; ++i) { - MCSectionData *SD = Layout.getSectionOrder()[i]; - - // Ignore sections without alignment requirements. - unsigned Align = SD->getAlignment(); - if (Align <= 1) - continue; - - // Ignore virtual sections, they don't cause file size modifications. - if (getBackend().isVirtualSection(SD->getSection())) - continue; - - // Otherwise, create a new align fragment at the end of the previous - // section. - MCAlignFragment *AF = new MCAlignFragment(Align, 0, 1, Align, - Layout.getSectionOrder()[i - 1]); - AF->setOnlyAlignAddress(true); - } - } - // Create dummy fragments and assign section ordinals. unsigned SectionIndex = 0; for (MCAssembler::iterator it = begin(), ie = end(); it != ie; ++it) { @@ -653,27 +555,18 @@ void MCAssembler::Finish(MCObjectWriter *Writer) { } // Assign layout order indices to sections and fragments. - unsigned FragmentIndex = 0; for (unsigned i = 0, e = Layout.getSectionOrder().size(); i != e; ++i) { MCSectionData *SD = Layout.getSectionOrder()[i]; SD->setLayoutOrder(i); + unsigned FragmentIndex = 0; for (MCSectionData::iterator it2 = SD->begin(), ie2 = SD->end(); it2 != ie2; ++it2) it2->setLayoutOrder(FragmentIndex++); } - llvm::OwningPtr<MCObjectWriter> OwnWriter(0); - if (Writer == 0) { - //no custom Writer_ : create the default one life-managed by OwningPtr - OwnWriter.reset(getBackend().createObjectWriter(OS)); - Writer = OwnWriter.get(); - if (!Writer) - report_fatal_error("unable to create object writer!"); - } - // Layout until everything fits. - while (LayoutOnce(*Writer, Layout)) + while (LayoutOnce(Layout)) continue; DEBUG_WITH_TYPE("mc-dump", { @@ -691,43 +584,42 @@ void MCAssembler::Finish(MCObjectWriter *Writer) { // Allow the object writer a chance to perform post-layout binding (for // example, to set the index fields in the symbol data). - Writer->ExecutePostLayoutBinding(*this); + getWriter().ExecutePostLayoutBinding(*this, Layout); // Evaluate and apply the fixups, generating relocation entries as necessary. for (MCAssembler::iterator it = begin(), ie = end(); it != ie; ++it) { for (MCSectionData::iterator it2 = it->begin(), ie2 = it->end(); it2 != ie2; ++it2) { MCDataFragment *DF = dyn_cast<MCDataFragment>(it2); - if (!DF) - continue; - - for (MCDataFragment::fixup_iterator it3 = DF->fixup_begin(), - ie3 = DF->fixup_end(); it3 != ie3; ++it3) { - MCFixup &Fixup = *it3; - - // Evaluate the fixup. - MCValue Target; - uint64_t FixedValue; - if (!EvaluateFixup(*Writer, Layout, Fixup, DF, Target, FixedValue)) { - // The fixup was unresolved, we need a relocation. Inform the object - // writer of the relocation, and give it an opportunity to adjust the - // fixup value if need be. - Writer->RecordRelocation(*this, Layout, DF, Fixup, Target,FixedValue); + if (DF) { + for (MCDataFragment::fixup_iterator it3 = DF->fixup_begin(), + ie3 = DF->fixup_end(); it3 != ie3; ++it3) { + MCFixup &Fixup = *it3; + uint64_t FixedValue = HandleFixup(Layout, *DF, Fixup); + getBackend().ApplyFixup(Fixup, DF->getContents().data(), + DF->getContents().size(), FixedValue); + } + } + MCInstFragment *IF = dyn_cast<MCInstFragment>(it2); + if (IF) { + for (MCInstFragment::fixup_iterator it3 = IF->fixup_begin(), + ie3 = IF->fixup_end(); it3 != ie3; ++it3) { + MCFixup &Fixup = *it3; + uint64_t FixedValue = HandleFixup(Layout, *IF, Fixup); + getBackend().ApplyFixup(Fixup, IF->getCode().data(), + IF->getCode().size(), FixedValue); } - - getBackend().ApplyFixup(Fixup, *DF, FixedValue); } } } // Write the object file. - Writer->WriteObject(*this, Layout); + getWriter().WriteObject(*this, Layout); stats::ObjectBytes += OS.tell() - StartOffset; } -bool MCAssembler::FixupNeedsRelaxation(const MCObjectWriter &Writer, - const MCFixup &Fixup, +bool MCAssembler::FixupNeedsRelaxation(const MCFixup &Fixup, const MCFragment *DF, const MCAsmLayout &Layout) const { if (getRelaxAll()) @@ -736,7 +628,7 @@ bool MCAssembler::FixupNeedsRelaxation(const MCObjectWriter &Writer, // If we cannot resolve the fixup value, it requires relaxation. MCValue Target; uint64_t Value; - if (!EvaluateFixup(Writer, Layout, Fixup, DF, Target, Value)) + if (!EvaluateFixup(Layout, Fixup, DF, Target, Value)) return true; // Otherwise, relax if the value is too big for a (signed) i8. @@ -745,8 +637,7 @@ bool MCAssembler::FixupNeedsRelaxation(const MCObjectWriter &Writer, return int64_t(Value) != int64_t(int8_t(Value)); } -bool MCAssembler::FragmentNeedsRelaxation(const MCObjectWriter &Writer, - const MCInstFragment *IF, +bool MCAssembler::FragmentNeedsRelaxation(const MCInstFragment *IF, const MCAsmLayout &Layout) const { // If this inst doesn't ever need relaxation, ignore it. This occurs when we // are intentionally pushing out inst fragments, or because we relaxed a @@ -756,16 +647,15 @@ bool MCAssembler::FragmentNeedsRelaxation(const MCObjectWriter &Writer, for (MCInstFragment::const_fixup_iterator it = IF->fixup_begin(), ie = IF->fixup_end(); it != ie; ++it) - if (FixupNeedsRelaxation(Writer, *it, IF, Layout)) + if (FixupNeedsRelaxation(*it, IF, Layout)) return true; return false; } -bool MCAssembler::RelaxInstruction(const MCObjectWriter &Writer, - MCAsmLayout &Layout, +bool MCAssembler::RelaxInstruction(MCAsmLayout &Layout, MCInstFragment &IF) { - if (!FragmentNeedsRelaxation(Writer, &IF, Layout)) + if (!FragmentNeedsRelaxation(&IF, Layout)) return false; ++stats::RelaxedInstructions; @@ -789,7 +679,6 @@ bool MCAssembler::RelaxInstruction(const MCObjectWriter &Writer, VecOS.flush(); // Update the instruction fragment. - int SlideAmount = Code.size() - IF.getInstSize(); IF.setInst(Relaxed); IF.getCode() = Code; IF.getFixups().clear(); @@ -797,135 +686,111 @@ bool MCAssembler::RelaxInstruction(const MCObjectWriter &Writer, for (unsigned i = 0, e = Fixups.size(); i != e; ++i) IF.getFixups().push_back(Fixups[i]); - // Update the layout, and remember that we relaxed. - Layout.UpdateForSlide(&IF, SlideAmount); return true; } -bool MCAssembler::RelaxOrg(const MCObjectWriter &Writer, - MCAsmLayout &Layout, - MCOrgFragment &OF) { - int64_t TargetLocation; - if (!OF.getOffset().EvaluateAsAbsolute(TargetLocation, &Layout)) - report_fatal_error("expected assembly-time absolute expression"); - - // FIXME: We need a way to communicate this error. - uint64_t FragmentOffset = Layout.getFragmentOffset(&OF); - int64_t Offset = TargetLocation - FragmentOffset; - if (Offset < 0 || Offset >= 0x40000000) - report_fatal_error("invalid .org offset '" + Twine(TargetLocation) + - "' (at offset '" + Twine(FragmentOffset) + "')"); - - unsigned OldSize = OF.getSize(); - OF.setSize(Offset); - return OldSize != OF.getSize(); -} - -bool MCAssembler::RelaxLEB(const MCObjectWriter &Writer, - MCAsmLayout &Layout, - MCLEBFragment &LF) { - int64_t Value; - LF.getValue().EvaluateAsAbsolute(Value, &Layout); - SmallString<32> Tmp; - raw_svector_ostream OSE(Tmp); +bool MCAssembler::RelaxLEB(MCAsmLayout &Layout, MCLEBFragment &LF) { + int64_t Value = 0; + uint64_t OldSize = LF.getContents().size(); + LF.getValue().EvaluateAsAbsolute(Value, Layout); + SmallString<8> &Data = LF.getContents(); + Data.clear(); + raw_svector_ostream OSE(Data); if (LF.isSigned()) MCObjectWriter::EncodeSLEB128(Value, OSE); else MCObjectWriter::EncodeULEB128(Value, OSE); - uint64_t OldSize = LF.getSize(); - LF.setSize(OSE.GetNumBytesInBuffer()); - return OldSize != LF.getSize(); + OSE.flush(); + return OldSize != LF.getContents().size(); } -bool MCAssembler::RelaxDwarfLineAddr(const MCObjectWriter &Writer, - MCAsmLayout &Layout, +bool MCAssembler::RelaxDwarfLineAddr(MCAsmLayout &Layout, MCDwarfLineAddrFragment &DF) { - int64_t AddrDelta; - DF.getAddrDelta().EvaluateAsAbsolute(AddrDelta, &Layout); + int64_t AddrDelta = 0; + uint64_t OldSize = DF.getContents().size(); + bool IsAbs = DF.getAddrDelta().EvaluateAsAbsolute(AddrDelta, Layout); + (void)IsAbs; + assert(IsAbs); int64_t LineDelta; LineDelta = DF.getLineDelta(); - uint64_t OldSize = DF.getSize(); - DF.setSize(MCDwarfLineAddr::ComputeSize(LineDelta, AddrDelta)); - return OldSize != DF.getSize(); + SmallString<8> &Data = DF.getContents(); + Data.clear(); + raw_svector_ostream OSE(Data); + MCDwarfLineAddr::Encode(LineDelta, AddrDelta, OSE); + OSE.flush(); + return OldSize != Data.size(); +} + +bool MCAssembler::RelaxDwarfCallFrameFragment(MCAsmLayout &Layout, + MCDwarfCallFrameFragment &DF) { + int64_t AddrDelta = 0; + uint64_t OldSize = DF.getContents().size(); + bool IsAbs = DF.getAddrDelta().EvaluateAsAbsolute(AddrDelta, Layout); + (void)IsAbs; + assert(IsAbs); + SmallString<8> &Data = DF.getContents(); + Data.clear(); + raw_svector_ostream OSE(Data); + MCDwarfFrameEmitter::EncodeAdvanceLoc(AddrDelta, OSE); + OSE.flush(); + return OldSize != Data.size(); +} + +bool MCAssembler::LayoutSectionOnce(MCAsmLayout &Layout, + MCSectionData &SD) { + MCFragment *FirstInvalidFragment = NULL; + // Scan for fragments that need relaxation. + for (MCSectionData::iterator it2 = SD.begin(), + ie2 = SD.end(); it2 != ie2; ++it2) { + // Check if this is an fragment that needs relaxation. + bool relaxedFrag = false; + switch(it2->getKind()) { + default: + break; + case MCFragment::FT_Inst: + relaxedFrag = RelaxInstruction(Layout, *cast<MCInstFragment>(it2)); + break; + case MCFragment::FT_Dwarf: + relaxedFrag = RelaxDwarfLineAddr(Layout, + *cast<MCDwarfLineAddrFragment>(it2)); + break; + case MCFragment::FT_DwarfFrame: + relaxedFrag = + RelaxDwarfCallFrameFragment(Layout, + *cast<MCDwarfCallFrameFragment>(it2)); + break; + case MCFragment::FT_LEB: + relaxedFrag = RelaxLEB(Layout, *cast<MCLEBFragment>(it2)); + break; + } + // Update the layout, and remember that we relaxed. + if (relaxedFrag && !FirstInvalidFragment) + FirstInvalidFragment = it2; + } + if (FirstInvalidFragment) { + Layout.Invalidate(FirstInvalidFragment); + return true; + } + return false; } -bool MCAssembler::LayoutOnce(const MCObjectWriter &Writer, - MCAsmLayout &Layout) { +bool MCAssembler::LayoutOnce(MCAsmLayout &Layout) { ++stats::RelaxationSteps; - // Layout the sections in order. - Layout.LayoutFile(); - - // Scan for fragments that need relaxation. bool WasRelaxed = false; for (iterator it = begin(), ie = end(); it != ie; ++it) { MCSectionData &SD = *it; - - for (MCSectionData::iterator it2 = SD.begin(), - ie2 = SD.end(); it2 != ie2; ++it2) { - // Check if this is an fragment that needs relaxation. - switch(it2->getKind()) { - default: - break; - case MCFragment::FT_Inst: - WasRelaxed |= RelaxInstruction(Writer, Layout, - *cast<MCInstFragment>(it2)); - break; - case MCFragment::FT_Org: - WasRelaxed |= RelaxOrg(Writer, Layout, *cast<MCOrgFragment>(it2)); - break; - case MCFragment::FT_Dwarf: - WasRelaxed |= RelaxDwarfLineAddr(Writer, Layout, - *cast<MCDwarfLineAddrFragment>(it2)); - break; - case MCFragment::FT_LEB: - WasRelaxed |= RelaxLEB(Writer, Layout, *cast<MCLEBFragment>(it2)); - break; - } - } + while(LayoutSectionOnce(Layout, SD)) + WasRelaxed = true; } return WasRelaxed; } void MCAssembler::FinishLayout(MCAsmLayout &Layout) { - // Lower out any instruction fragments, to simplify the fixup application and - // output. - // - // FIXME-PERF: We don't have to do this, but the assumption is that it is - // cheap (we will mostly end up eliminating fragments and appending on to data - // fragments), so the extra complexity downstream isn't worth it. Evaluate - // this assumption. - for (iterator it = begin(), ie = end(); it != ie; ++it) { - MCSectionData &SD = *it; - - for (MCSectionData::iterator it2 = SD.begin(), - ie2 = SD.end(); it2 != ie2; ++it2) { - MCInstFragment *IF = dyn_cast<MCInstFragment>(it2); - if (!IF) - continue; - - // Create a new data fragment for the instruction. - // - // FIXME-PERF: Reuse previous data fragment if possible. - MCDataFragment *DF = new MCDataFragment(); - SD.getFragmentList().insert(it2, DF); - - // Update the data fragments layout data. - DF->setParent(IF->getParent()); - DF->setAtom(IF->getAtom()); - DF->setLayoutOrder(IF->getLayoutOrder()); - Layout.FragmentReplaced(IF, DF); - - // Copy in the data and the fixups. - DF->getContents().append(IF->getCode().begin(), IF->getCode().end()); - for (unsigned i = 0, e = IF->getFixups().size(); i != e; ++i) - DF->getFixups().push_back(IF->getFixups()[i]); - - // Delete the instruction fragment and update the iterator. - SD.getFragmentList().erase(IF); - it2 = DF; - } + // The layout is done. Mark every fragment as valid. + for (unsigned int i = 0, n = Layout.getSectionOrder().size(); i != n; ++i) { + Layout.getFragmentOffset(&*Layout.getSectionOrder()[i]->rbegin()); } } @@ -953,19 +818,18 @@ void MCFragment::dump() { case MCFragment::FT_Inst: OS << "MCInstFragment"; break; case MCFragment::FT_Org: OS << "MCOrgFragment"; break; case MCFragment::FT_Dwarf: OS << "MCDwarfFragment"; break; + case MCFragment::FT_DwarfFrame: OS << "MCDwarfCallFrameFragment"; break; case MCFragment::FT_LEB: OS << "MCLEBFragment"; break; } OS << "<MCFragment " << (void*) this << " LayoutOrder:" << LayoutOrder - << " Offset:" << Offset << " EffectiveSize:" << EffectiveSize << ">"; + << " Offset:" << Offset << ">"; switch (getKind()) { case MCFragment::FT_Align: { const MCAlignFragment *AF = cast<MCAlignFragment>(this); if (AF->hasEmitNops()) OS << " (emit nops)"; - if (AF->hasOnlyAlignAddress()) - OS << " (only align section)"; OS << "\n "; OS << " Alignment:" << AF->getAlignment() << " Value:" << AF->getValue() << " ValueSize:" << AF->getValueSize() @@ -1021,6 +885,12 @@ void MCFragment::dump() { << " LineDelta:" << OF->getLineDelta(); break; } + case MCFragment::FT_DwarfFrame: { + const MCDwarfCallFrameFragment *CF = cast<MCDwarfCallFrameFragment>(this); + OS << "\n "; + OS << " AddrDelta:" << CF->getAddrDelta(); + break; + } case MCFragment::FT_LEB: { const MCLEBFragment *LF = cast<MCLEBFragment>(this); OS << "\n "; @@ -1035,8 +905,7 @@ void MCSectionData::dump() { raw_ostream &OS = llvm::errs(); OS << "<MCSectionData"; - OS << " Alignment:" << getAlignment() << " Address:" << Address - << " Fragments:[\n "; + OS << " Alignment:" << getAlignment() << " Fragments:[\n "; for (iterator it = begin(), ie = end(); it != ie; ++it) { if (it != begin()) OS << ",\n "; it->dump(); diff --git a/lib/MC/MCCodeEmitter.cpp b/lib/MC/MCCodeEmitter.cpp index d513237..c122763 100644 --- a/lib/MC/MCCodeEmitter.cpp +++ b/lib/MC/MCCodeEmitter.cpp @@ -16,15 +16,3 @@ MCCodeEmitter::MCCodeEmitter() { MCCodeEmitter::~MCCodeEmitter() { } - -const MCFixupKindInfo &MCCodeEmitter::getFixupKindInfo(MCFixupKind Kind) const { - static const MCFixupKindInfo Builtins[] = { - { "FK_Data_1", 0, 8, 0 }, - { "FK_Data_2", 0, 16, 0 }, - { "FK_Data_4", 0, 32, 0 }, - { "FK_Data_8", 0, 64, 0 } - }; - - assert(Kind <= 3 && "Unknown fixup kind"); - return Builtins[Kind]; -} diff --git a/lib/MC/MCContext.cpp b/lib/MC/MCContext.cpp index 87619d5..018f00c 100644 --- a/lib/MC/MCContext.cpp +++ b/lib/MC/MCContext.cpp @@ -15,8 +15,10 @@ #include "llvm/MC/MCSymbol.h" #include "llvm/MC/MCLabel.h" #include "llvm/MC/MCDwarf.h" +#include "llvm/Target/TargetAsmInfo.h" #include "llvm/ADT/SmallString.h" #include "llvm/ADT/Twine.h" +#include "llvm/Support/ELF.h" using namespace llvm; typedef StringMap<const MCSectionMachO*> MachOUniqueMapTy; @@ -24,8 +26,9 @@ typedef StringMap<const MCSectionELF*> ELFUniqueMapTy; typedef StringMap<const MCSectionCOFF*> COFFUniqueMapTy; -MCContext::MCContext(const MCAsmInfo &mai) : MAI(mai), NextUniqueID(0), - CurrentDwarfLoc(0,0,0,0,0,0) { +MCContext::MCContext(const MCAsmInfo &mai, const TargetAsmInfo *tai) : + MAI(mai), TAI(tai), NextUniqueID(0), + CurrentDwarfLoc(0,0,0,DWARF2_FLAG_IS_STMT,0,0) { MachOUniquingMap = 0; ELFUniquingMap = 0; COFFUniquingMap = 0; @@ -40,7 +43,7 @@ MCContext::MCContext(const MCAsmInfo &mai) : MAI(mai), NextUniqueID(0), MCContext::~MCContext() { // NOTE: The symbols are all allocated out of a bump pointer allocator, // we don't need to free them here. - + // If we have the MachO uniquing map, free it. delete (MachOUniqueMapTy*)MachOUniquingMap; delete (ELFUniqueMapTy*)ELFUniquingMap; @@ -48,6 +51,8 @@ MCContext::~MCContext() { // If the stream for the .secure_log_unique directive was created free it. delete (raw_ostream*)SecureLog; + + delete TAI; } //===----------------------------------------------------------------------===// @@ -56,20 +61,42 @@ MCContext::~MCContext() { MCSymbol *MCContext::GetOrCreateSymbol(StringRef Name) { assert(!Name.empty() && "Normal symbols cannot be unnamed!"); - - // Determine whether this is an assembler temporary or normal label. - bool isTemporary = Name.startswith(MAI.getPrivateGlobalPrefix()); - + // Do the lookup and get the entire StringMapEntry. We want access to the // key if we are creating the entry. StringMapEntry<MCSymbol*> &Entry = Symbols.GetOrCreateValue(Name); - if (Entry.getValue()) return Entry.getValue(); + MCSymbol *Sym = Entry.getValue(); + + if (Sym) + return Sym; + + Sym = CreateSymbol(Name); + Entry.setValue(Sym); + return Sym; +} + +MCSymbol *MCContext::CreateSymbol(StringRef Name) { + // Determine whether this is an assembler temporary or normal label. + bool isTemporary = Name.startswith(MAI.getPrivateGlobalPrefix()); + + StringMapEntry<bool> *NameEntry = &UsedNames.GetOrCreateValue(Name); + if (NameEntry->getValue()) { + assert(isTemporary && "Cannot rename non temporary symbols"); + SmallString<128> NewName; + do { + Twine T = Name + Twine(NextUniqueID++); + T.toVector(NewName); + StringRef foo = NewName; + NameEntry = &UsedNames.GetOrCreateValue(foo); + } while (NameEntry->getValue()); + } + NameEntry->setValue(true); // Ok, the entry doesn't already exist. Have the MCSymbol object itself refer - // to the copy of the string that is embedded in the StringMapEntry. - MCSymbol *Result = new (*this) MCSymbol(Entry.getKey(), isTemporary); - Entry.setValue(Result); - return Result; + // to the copy of the string that is embedded in the UsedNames entry. + MCSymbol *Result = new (*this) MCSymbol(NameEntry->getKey(), isTemporary); + + return Result; } MCSymbol *MCContext::GetOrCreateSymbol(const Twine &Name) { @@ -79,8 +106,11 @@ MCSymbol *MCContext::GetOrCreateSymbol(const Twine &Name) { } MCSymbol *MCContext::CreateTempSymbol() { - return GetOrCreateSymbol(Twine(MAI.getPrivateGlobalPrefix()) + - "tmp" + Twine(NextUniqueID++)); + SmallString<128> NameSV; + Twine Name = Twine(MAI.getPrivateGlobalPrefix()) + "tmp" + + Twine(NextUniqueID++); + Name.toVector(NameSV); + return CreateSymbol(NameSV); } unsigned MCContext::NextInstance(int64_t LocalLabelVal) { @@ -123,26 +153,26 @@ const MCSectionMachO *MCContext:: getMachOSection(StringRef Segment, StringRef Section, unsigned TypeAndAttributes, unsigned Reserved2, SectionKind Kind) { - + // We unique sections by their segment/section pair. The returned section // may not have the same flags as the requested section, if so this should be // diagnosed by the client as an error. - + // Create the map if it doesn't already exist. if (MachOUniquingMap == 0) MachOUniquingMap = new MachOUniqueMapTy(); MachOUniqueMapTy &Map = *(MachOUniqueMapTy*)MachOUniquingMap; - + // Form the name to look up. SmallString<64> Name; Name += Segment; Name.push_back(','); Name += Section; - + // Do the lookup, if we have a hit, return it. const MCSectionMachO *&Entry = Map[Name.str()]; if (Entry) return Entry; - + // Otherwise, return a new section. return Entry = new (*this) MCSectionMachO(Segment, Section, TypeAndAttributes, Reserved2, Kind); @@ -160,11 +190,11 @@ getELFSection(StringRef Section, unsigned Type, unsigned Flags, if (ELFUniquingMap == 0) ELFUniquingMap = new ELFUniqueMapTy(); ELFUniqueMapTy &Map = *(ELFUniqueMapTy*)ELFUniquingMap; - + // Do the lookup, if we have a hit, return it. StringMapEntry<const MCSectionELF*> &Entry = Map.GetOrCreateValue(Section); if (Entry.getValue()) return Entry.getValue(); - + // Possibly refine the entry size first. if (!EntrySize) { EntrySize = MCSectionELF::DetermineEntrySize(Kind); @@ -182,7 +212,7 @@ getELFSection(StringRef Section, unsigned Type, unsigned Flags, const MCSectionELF *MCContext::CreateELFGroupSection() { MCSectionELF *Result = - new (*this) MCSectionELF(".group", MCSectionELF::SHT_GROUP, 0, + new (*this) MCSectionELF(".group", ELF::SHT_GROUP, 0, SectionKind::getReadOnly(), 4, NULL); return Result; } @@ -194,15 +224,15 @@ const MCSection *MCContext::getCOFFSection(StringRef Section, if (COFFUniquingMap == 0) COFFUniquingMap = new COFFUniqueMapTy(); COFFUniqueMapTy &Map = *(COFFUniqueMapTy*)COFFUniquingMap; - + // Do the lookup, if we have a hit, return it. StringMapEntry<const MCSectionCOFF*> &Entry = Map.GetOrCreateValue(Section); if (Entry.getValue()) return Entry.getValue(); - + MCSectionCOFF *Result = new (*this) MCSectionCOFF(Entry.getKey(), Characteristics, Selection, Kind); - + Entry.setValue(Result); return Result; } @@ -261,7 +291,7 @@ unsigned MCContext::GetDwarfFile(StringRef FileName, unsigned FileNumber) { // stored at MCDwarfFiles[FileNumber].Name . DirIndex++; } - + // Now make the MCDwarfFile entry and place it in the slot in the MCDwarfFiles // vector. char *Buf = static_cast<char *>(Allocate(Name.size())); diff --git a/lib/MC/MCDisassembler/EDDisassembler.cpp b/lib/MC/MCDisassembler/EDDisassembler.cpp index 697b3d9..2fd14db 100644 --- a/lib/MC/MCDisassembler/EDDisassembler.cpp +++ b/lib/MC/MCDisassembler/EDDisassembler.cpp @@ -354,7 +354,7 @@ int EDDisassembler::parseInst(SmallVectorImpl<MCParsedAsmOperand*> &operands, SourceMgr sourceMgr; sourceMgr.AddNewSourceBuffer(buf, SMLoc()); // ownership of buf handed over - MCContext context(*AsmInfo); + MCContext context(*AsmInfo, NULL); OwningPtr<MCStreamer> streamer(createNullStreamer(context)); OwningPtr<MCAsmParser> genericParser(createMCAsmParser(*Tgt, sourceMgr, context, *streamer, diff --git a/lib/MC/MCDisassembler/EDDisassembler.h b/lib/MC/MCDisassembler/EDDisassembler.h index e2f850b..71e45f0 100644 --- a/lib/MC/MCDisassembler/EDDisassembler.h +++ b/lib/MC/MCDisassembler/EDDisassembler.h @@ -21,7 +21,7 @@ #include "llvm/ADT/OwningPtr.h" #include "llvm/ADT/Triple.h" #include "llvm/Support/raw_ostream.h" -#include "llvm/System/Mutex.h" +#include "llvm/Support/Mutex.h" #include <map> #include <set> @@ -89,8 +89,10 @@ struct EDDisassembler { bool operator<(const CPUKey &key) const { if(Arch > key.Arch) return false; - if(Syntax >= key.Syntax) - return false; + else if (Arch == key.Arch) { + if(Syntax > key.Syntax) + return false; + } return true; } }; diff --git a/lib/MC/MCDisassembler/EDInst.h b/lib/MC/MCDisassembler/EDInst.h index 39d264f..ceb9505 100644 --- a/lib/MC/MCDisassembler/EDInst.h +++ b/lib/MC/MCDisassembler/EDInst.h @@ -16,7 +16,7 @@ #ifndef LLVM_EDINST_H #define LLVM_EDINST_H -#include "llvm/System/DataTypes.h" +#include "llvm/Support/DataTypes.h" #include "llvm/ADT/SmallVector.h" #include <string> #include <vector> diff --git a/lib/MC/MCDisassembler/EDOperand.cpp b/lib/MC/MCDisassembler/EDOperand.cpp index cfeb56f..2b0c73e 100644 --- a/lib/MC/MCDisassembler/EDOperand.cpp +++ b/lib/MC/MCDisassembler/EDOperand.cpp @@ -152,10 +152,23 @@ int EDOperand::evaluate(uint64_t &result, uint64_t scaleAmount = Inst.Inst->getOperand(MCOpIndex+1).getImm(); unsigned indexReg = Inst.Inst->getOperand(MCOpIndex+2).getReg(); int64_t displacement = Inst.Inst->getOperand(MCOpIndex+3).getImm(); - //unsigned segmentReg = Inst.Inst->getOperand(MCOpIndex+4).getReg(); - + uint64_t addr = 0; + unsigned segmentReg = Inst.Inst->getOperand(MCOpIndex+4).getReg(); + + if (segmentReg != 0 && Disassembler.Key.Arch == Triple::x86_64) { + unsigned fsID = Disassembler.registerIDWithName("FS"); + unsigned gsID = Disassembler.registerIDWithName("GS"); + + if (segmentReg == fsID || + segmentReg == gsID) { + uint64_t segmentBase; + if (!callback(&segmentBase, segmentReg, arg)) + addr += segmentBase; + } + } + if (baseReg) { uint64_t baseVal; if (callback(&baseVal, baseReg, arg)) @@ -175,7 +188,7 @@ int EDOperand::evaluate(uint64_t &result, result = addr; return 0; } - } + } // switch (operandType) break; case Triple::arm: case Triple::thumb: @@ -203,6 +216,7 @@ int EDOperand::evaluate(uint64_t &result, return 0; } } + break; } return -1; diff --git a/lib/MC/MCDisassembler/EDOperand.h b/lib/MC/MCDisassembler/EDOperand.h index 6e69522..50260ec 100644 --- a/lib/MC/MCDisassembler/EDOperand.h +++ b/lib/MC/MCDisassembler/EDOperand.h @@ -16,7 +16,7 @@ #ifndef LLVM_EDOPERAND_H #define LLVM_EDOPERAND_H -#include "llvm/System/DataTypes.h" +#include "llvm/Support/DataTypes.h" namespace llvm { diff --git a/lib/MC/MCDisassembler/EDToken.cpp b/lib/MC/MCDisassembler/EDToken.cpp index 400e164..de770b4 100644 --- a/lib/MC/MCDisassembler/EDToken.cpp +++ b/lib/MC/MCDisassembler/EDToken.cpp @@ -194,6 +194,10 @@ int EDToken::tokenize(std::vector<EDToken*> &tokens, tokens.push_back(token); } + // Free any parsed operands. + for (unsigned i = 0, e = parsedOperands.size(); i != e; ++i) + delete parsedOperands[i]; + return 0; } diff --git a/lib/MC/MCDisassembler/EDToken.h b/lib/MC/MCDisassembler/EDToken.h index 6b2aeac..ba46707 100644 --- a/lib/MC/MCDisassembler/EDToken.h +++ b/lib/MC/MCDisassembler/EDToken.h @@ -17,7 +17,7 @@ #define LLVM_EDTOKEN_H #include "llvm/ADT/StringRef.h" -#include "llvm/System/DataTypes.h" +#include "llvm/Support/DataTypes.h" #include <string> #include <vector> diff --git a/lib/MC/MCDwarf.cpp b/lib/MC/MCDwarf.cpp index 679f4ee..112d7d8 100644 --- a/lib/MC/MCDwarf.cpp +++ b/lib/MC/MCDwarf.cpp @@ -7,16 +7,21 @@ // //===----------------------------------------------------------------------===// +#include "llvm/ADT/FoldingSet.h" +#include "llvm/MC/MCAsmInfo.h" #include "llvm/MC/MCDwarf.h" #include "llvm/MC/MCAssembler.h" +#include "llvm/MC/MCStreamer.h" #include "llvm/MC/MCSymbol.h" #include "llvm/MC/MCExpr.h" #include "llvm/MC/MCContext.h" #include "llvm/MC/MCObjectWriter.h" #include "llvm/ADT/SmallString.h" #include "llvm/Support/Debug.h" +#include "llvm/Support/ErrorHandling.h" #include "llvm/Support/raw_ostream.h" #include "llvm/Target/TargetAsmBackend.h" +#include "llvm/Target/TargetAsmInfo.h" using namespace llvm; // Given a special op, return the address skip amount (in units of @@ -60,7 +65,7 @@ static inline uint64_t ScaleAddrDelta(uint64_t AddrDelta) // and if there is information from the last .loc directive that has yet to have // a line entry made for it is made. // -void MCLineEntry::Make(MCObjectStreamer *MCOS, const MCSection *Section) { +void MCLineEntry::Make(MCStreamer *MCOS, const MCSection *Section) { if (!MCOS->getContext().getDwarfLocSeen()) return; @@ -80,16 +85,16 @@ void MCLineEntry::Make(MCObjectStreamer *MCOS, const MCSection *Section) { // Get the MCLineSection for this section, if one does not exist for this // section create it. - DenseMap<const MCSection *, MCLineSection *> &MCLineSections = + const DenseMap<const MCSection *, MCLineSection *> &MCLineSections = MCOS->getContext().getMCLineSections(); - MCLineSection *LineSection = MCLineSections[Section]; + MCLineSection *LineSection = MCLineSections.lookup(Section); if (!LineSection) { // Create a new MCLineSection. This will be deleted after the dwarf line // table is created using it by iterating through the MCLineSections // DenseMap. LineSection = new MCLineSection; // Save a pointer to the new LineSection into the MCLineSections DenseMap. - MCLineSections[Section] = LineSection; + MCOS->getContext().addMCLineSection(Section, LineSection); } // Add the line entry to this section's entries. @@ -99,57 +104,40 @@ void MCLineEntry::Make(MCObjectStreamer *MCOS, const MCSection *Section) { // // This helper routine returns an expression of End - Start + IntVal . // -static inline const MCExpr *MakeStartMinusEndExpr(MCObjectStreamer *MCOS, - MCSymbol *Start, - MCSymbol *End, int IntVal) { +static inline const MCExpr *MakeStartMinusEndExpr(const MCStreamer &MCOS, + const MCSymbol &Start, + const MCSymbol &End, + int IntVal) { MCSymbolRefExpr::VariantKind Variant = MCSymbolRefExpr::VK_None; const MCExpr *Res = - MCSymbolRefExpr::Create(End, Variant, MCOS->getContext()); + MCSymbolRefExpr::Create(&End, Variant, MCOS.getContext()); const MCExpr *RHS = - MCSymbolRefExpr::Create(Start, Variant, MCOS->getContext()); + MCSymbolRefExpr::Create(&Start, Variant, MCOS.getContext()); const MCExpr *Res1 = - MCBinaryExpr::Create(MCBinaryExpr::Sub, Res, RHS, MCOS->getContext()); + MCBinaryExpr::Create(MCBinaryExpr::Sub, Res, RHS, MCOS.getContext()); const MCExpr *Res2 = - MCConstantExpr::Create(IntVal, MCOS->getContext()); + MCConstantExpr::Create(IntVal, MCOS.getContext()); const MCExpr *Res3 = - MCBinaryExpr::Create(MCBinaryExpr::Sub, Res1, Res2, MCOS->getContext()); + MCBinaryExpr::Create(MCBinaryExpr::Sub, Res1, Res2, MCOS.getContext()); return Res3; } -// -// This emits an "absolute" address used in the start of a dwarf line number -// table. This will result in a relocatation entry for the address. -// -static inline void EmitDwarfSetAddress(MCObjectStreamer *MCOS, - MCSymbol *Symbol) { - MCOS->EmitIntValue(dwarf::DW_LNS_extended_op, 1); - - int sizeof_address = MCOS->getAssembler().getBackend().getPointerSize(); - MCOS->EmitULEB128IntValue(sizeof_address + 1); - - MCOS->EmitIntValue(dwarf::DW_LNE_set_address, 1); - MCOS->EmitSymbolValue(Symbol, sizeof_address); -} - // // This emits the Dwarf line table for the specified section from the entries // in the LineSection. // -static inline void EmitDwarfLineTable(MCObjectStreamer *MCOS, +static inline void EmitDwarfLineTable(MCStreamer *MCOS, const MCSection *Section, - MCLineSection *LineSection, - const MCSection *DwarfLineSection) { + const MCLineSection *LineSection) { unsigned FileNum = 1; unsigned LastLine = 1; unsigned Column = 0; unsigned Flags = DWARF2_LINE_DEFAULT_IS_STMT ? DWARF2_FLAG_IS_STMT : 0; unsigned Isa = 0; MCSymbol *LastLabel = NULL; - MCSectionData &DLS = - MCOS->getAssembler().getOrCreateSectionData(*DwarfLineSection); // Loop through each MCLineEntry and encode the dwarf line number table. - for (MCLineSection::iterator + for (MCLineSection::const_iterator it = LineSection->getMCLineEntries()->begin(), ie = LineSection->getMCLineEntries()->end(); it != ie; ++it) { @@ -185,19 +173,7 @@ static inline void EmitDwarfLineTable(MCObjectStreamer *MCOS, // At this point we want to emit/create the sequence to encode the delta in // line numbers and the increment of the address from the previous Label // and the current Label. - if (LastLabel == NULL) { - // emit the sequence to set the address - EmitDwarfSetAddress(MCOS, Label); - // emit the sequence for the LineDelta (from 1) and a zero address delta. - MCDwarfLineAddr::Emit(MCOS, LineDelta, 0); - } - else { - // Create an expression for the address delta from the LastLabel and - // this Label (plus 0). - const MCExpr *AddrDelta = MakeStartMinusEndExpr(MCOS, LastLabel, Label,0); - // Create a Dwarf Line fragment for the LineDelta and AddrDelta. - new MCDwarfLineAddrFragment(LineDelta, *AddrDelta, &DLS); - } + MCOS->EmitDwarfAdvanceLineAddr(LineDelta, LastLabel, Label); LastLine = it->getLine(); LastLabel = Label; @@ -211,53 +187,51 @@ static inline void EmitDwarfLineTable(MCObjectStreamer *MCOS, // Switch to the section to be able to create a symbol at its end. MCOS->SwitchSection(Section); + + MCContext &context = MCOS->getContext(); // Create a symbol at the end of the section. - MCSymbol *SectionEnd = MCOS->getContext().CreateTempSymbol(); + MCSymbol *SectionEnd = context.CreateTempSymbol(); // Set the value of the symbol, as we are at the end of the section. MCOS->EmitLabel(SectionEnd); // Switch back the the dwarf line section. - MCOS->SwitchSection(DwarfLineSection); - // Create an expression for the address delta from the LastLabel and this - // SectionEnd label. - const MCExpr *AddrDelta = MakeStartMinusEndExpr(MCOS, LastLabel, SectionEnd, - 0); - // Create a Dwarf Line fragment for the LineDelta and AddrDelta. - new MCDwarfLineAddrFragment(INT64_MAX, *AddrDelta, &DLS); + MCOS->SwitchSection(context.getTargetAsmInfo().getDwarfLineSection()); + + MCOS->EmitDwarfAdvanceLineAddr(INT64_MAX, LastLabel, SectionEnd); } // // This emits the Dwarf file and the line tables. // -void MCDwarfFileTable::Emit(MCObjectStreamer *MCOS, - const MCSection *DwarfLineSection) { +void MCDwarfFileTable::Emit(MCStreamer *MCOS) { + MCContext &context = MCOS->getContext(); // Switch to the section where the table will be emitted into. - MCOS->SwitchSection(DwarfLineSection); + MCOS->SwitchSection(context.getTargetAsmInfo().getDwarfLineSection()); // Create a symbol at the beginning of this section. - MCSymbol *LineStartSym = MCOS->getContext().CreateTempSymbol(); + MCSymbol *LineStartSym = context.CreateTempSymbol(); // Set the value of the symbol, as we are at the start of the section. MCOS->EmitLabel(LineStartSym); // Create a symbol for the end of the section (to be set when we get there). - MCSymbol *LineEndSym = MCOS->getContext().CreateTempSymbol(); + MCSymbol *LineEndSym = context.CreateTempSymbol(); // The first 4 bytes is the total length of the information for this // compilation unit (not including these 4 bytes for the length). - MCOS->EmitValue(MakeStartMinusEndExpr(MCOS, LineStartSym, LineEndSym, 4), - 4, 0); + MCOS->EmitAbsValue(MakeStartMinusEndExpr(*MCOS, *LineStartSym, *LineEndSym,4), + 4); // Next 2 bytes is the Version, which is Dwarf 2. MCOS->EmitIntValue(2, 2); // Create a symbol for the end of the prologue (to be set when we get there). - MCSymbol *ProEndSym = MCOS->getContext().CreateTempSymbol(); // Lprologue_end + MCSymbol *ProEndSym = context.CreateTempSymbol(); // Lprologue_end // Length of the prologue, is the next 4 bytes. Which is the start of the // section to the end of the prologue. Not including the 4 bytes for the // total length, the 2 bytes for the version, and these 4 bytes for the // length of the prologue. - MCOS->EmitValue(MakeStartMinusEndExpr(MCOS, LineStartSym, ProEndSym, + MCOS->EmitAbsValue(MakeStartMinusEndExpr(*MCOS, *LineStartSym, *ProEndSym, (4 + 2 + 4)), 4, 0); @@ -286,7 +260,7 @@ void MCDwarfFileTable::Emit(MCObjectStreamer *MCOS, // First the directory table. const std::vector<StringRef> &MCDwarfDirs = - MCOS->getContext().getMCDwarfDirs(); + context.getMCDwarfDirs(); for (unsigned i = 0; i < MCDwarfDirs.size(); i++) { MCOS->EmitBytes(MCDwarfDirs[i], 0); // the DirectoryName MCOS->EmitBytes(StringRef("\0", 1), 0); // the null term. of the string @@ -311,15 +285,32 @@ void MCDwarfFileTable::Emit(MCObjectStreamer *MCOS, MCOS->EmitLabel(ProEndSym); // Put out the line tables. - DenseMap<const MCSection *, MCLineSection *> &MCLineSections = + const DenseMap<const MCSection *, MCLineSection *> &MCLineSections = MCOS->getContext().getMCLineSections(); - for (DenseMap<const MCSection *, MCLineSection *>::iterator it = - MCLineSections.begin(), ie = MCLineSections.end(); it != ie; ++it) { - EmitDwarfLineTable(MCOS, it->first, it->second, DwarfLineSection); + const std::vector<const MCSection *> &MCLineSectionOrder = + MCOS->getContext().getMCLineSectionOrder(); + for (std::vector<const MCSection*>::const_iterator it = + MCLineSectionOrder.begin(), ie = MCLineSectionOrder.end(); it != ie; + ++it) { + const MCSection *Sec = *it; + const MCLineSection *Line = MCLineSections.lookup(Sec); + EmitDwarfLineTable(MCOS, Sec, Line); // Now delete the MCLineSections that were created in MCLineEntry::Make() // and used to emit the line table. - delete it->second; + delete Line; + } + + if (MCOS->getContext().getAsmInfo().getLinkerRequiresNonEmptyDwarfLines() + && MCLineSectionOrder.begin() == MCLineSectionOrder.end()) { + // The darwin9 linker has a bug (see PR8715). For for 32-bit architectures + // it requires: + // total_length >= prologue_length + 10 + // We are 4 bytes short, since we have total_length = 51 and + // prologue_length = 45 + + // The regular end_sequence should be sufficient. + MCDwarfLineAddr::Emit(MCOS, INT64_MAX, 0); } // This is the end of the section, so set the value of the symbol at the end @@ -327,14 +318,6 @@ void MCDwarfFileTable::Emit(MCObjectStreamer *MCOS, MCOS->EmitLabel(LineEndSym); } -/// Utility function to compute the size of the encoding. -uint64_t MCDwarfLineAddr::ComputeSize(int64_t LineDelta, uint64_t AddrDelta) { - SmallString<256> Tmp; - raw_svector_ostream OS(Tmp); - MCDwarfLineAddr::Encode(LineDelta, AddrDelta, OS); - return OS.GetNumBytesInBuffer(); -} - /// Utility function to write the encoding to an object writer. void MCDwarfLineAddr::Write(MCObjectWriter *OW, int64_t LineDelta, uint64_t AddrDelta) { @@ -345,7 +328,7 @@ void MCDwarfLineAddr::Write(MCObjectWriter *OW, int64_t LineDelta, } /// Utility function to emit the encoding to a streamer. -void MCDwarfLineAddr::Emit(MCObjectStreamer *MCOS, int64_t LineDelta, +void MCDwarfLineAddr::Emit(MCStreamer *MCOS, int64_t LineDelta, uint64_t AddrDelta) { SmallString<256> Tmp; raw_svector_ostream OS(Tmp); @@ -446,3 +429,386 @@ void MCDwarfFile::dump() const { print(dbgs()); } +static int getDataAlignmentFactor(MCStreamer &streamer) { + MCContext &context = streamer.getContext(); + const TargetAsmInfo &asmInfo = context.getTargetAsmInfo(); + int size = asmInfo.getPointerSize(); + if (asmInfo.getStackGrowthDirection() == TargetFrameLowering::StackGrowsUp) + return size; + else + return -size; +} + +static void EmitCFIInstruction(MCStreamer &Streamer, + const MCCFIInstruction &Instr) { + int dataAlignmentFactor = getDataAlignmentFactor(Streamer); + + switch (Instr.getOperation()) { + case MCCFIInstruction::Move: { + const MachineLocation &Dst = Instr.getDestination(); + const MachineLocation &Src = Instr.getSource(); + + // If advancing cfa. + if (Dst.isReg() && Dst.getReg() == MachineLocation::VirtualFP) { + assert(!Src.isReg() && "Machine move not supported yet."); + + if (Src.getReg() == MachineLocation::VirtualFP) { + Streamer.EmitIntValue(dwarf::DW_CFA_def_cfa_offset, 1); + } else { + Streamer.EmitIntValue(dwarf::DW_CFA_def_cfa, 1); + Streamer.EmitULEB128IntValue(Src.getReg()); + } + + Streamer.EmitULEB128IntValue(-Src.getOffset(), 1); + return; + } + + if (Src.isReg() && Src.getReg() == MachineLocation::VirtualFP) { + assert(Dst.isReg() && "Machine move not supported yet."); + Streamer.EmitIntValue(dwarf::DW_CFA_def_cfa_register, 1); + Streamer.EmitULEB128IntValue(Dst.getReg()); + return; + } + + unsigned Reg = Src.getReg(); + int Offset = Dst.getOffset() / dataAlignmentFactor; + + if (Offset < 0) { + Streamer.EmitIntValue(dwarf::DW_CFA_offset_extended_sf, 1); + Streamer.EmitULEB128IntValue(Reg); + Streamer.EmitSLEB128IntValue(Offset); + } else if (Reg < 64) { + Streamer.EmitIntValue(dwarf::DW_CFA_offset + Reg, 1); + Streamer.EmitULEB128IntValue(Offset, 1); + } else { + Streamer.EmitIntValue(dwarf::DW_CFA_offset_extended, 1); + Streamer.EmitULEB128IntValue(Reg, 1); + Streamer.EmitULEB128IntValue(Offset, 1); + } + return; + } + case MCCFIInstruction::Remember: + Streamer.EmitIntValue(dwarf::DW_CFA_remember_state, 1); + return; + case MCCFIInstruction::Restore: + Streamer.EmitIntValue(dwarf::DW_CFA_restore_state, 1); + return; + } + llvm_unreachable("Unhandled case in switch"); +} + +/// EmitFrameMoves - Emit frame instructions to describe the layout of the +/// frame. +static void EmitCFIInstructions(MCStreamer &streamer, + const std::vector<MCCFIInstruction> &Instrs, + MCSymbol *BaseLabel) { + for (unsigned i = 0, N = Instrs.size(); i < N; ++i) { + const MCCFIInstruction &Instr = Instrs[i]; + MCSymbol *Label = Instr.getLabel(); + // Throw out move if the label is invalid. + if (Label && !Label->isDefined()) continue; // Not emitted, in dead code. + + // Advance row if new location. + if (BaseLabel && Label) { + MCSymbol *ThisSym = Label; + if (ThisSym != BaseLabel) { + streamer.EmitDwarfAdvanceFrameAddr(BaseLabel, ThisSym); + BaseLabel = ThisSym; + } + } + + EmitCFIInstruction(streamer, Instr); + } +} + +static void EmitSymbol(MCStreamer &streamer, const MCSymbol &symbol, + unsigned symbolEncoding) { + MCContext &context = streamer.getContext(); + const TargetAsmInfo &asmInfo = context.getTargetAsmInfo(); + unsigned format = symbolEncoding & 0x0f; + unsigned application = symbolEncoding & 0x70; + unsigned size; + switch (format) { + default: + assert(0 && "Unknown Encoding"); + case dwarf::DW_EH_PE_absptr: + case dwarf::DW_EH_PE_signed: + size = asmInfo.getPointerSize(); + break; + case dwarf::DW_EH_PE_udata2: + case dwarf::DW_EH_PE_sdata2: + size = 2; + break; + case dwarf::DW_EH_PE_udata4: + case dwarf::DW_EH_PE_sdata4: + size = 4; + break; + case dwarf::DW_EH_PE_udata8: + case dwarf::DW_EH_PE_sdata8: + size = 8; + break; + } + switch (application) { + default: + assert(0 && "Unknown Encoding"); + break; + case 0: + streamer.EmitSymbolValue(&symbol, size); + break; + case dwarf::DW_EH_PE_pcrel: + streamer.EmitPCRelSymbolValue(&symbol, size); + break; + } +} + +static const MachineLocation TranslateMachineLocation( + const TargetAsmInfo &AsmInfo, + const MachineLocation &Loc) { + unsigned Reg = Loc.getReg() == MachineLocation::VirtualFP ? + MachineLocation::VirtualFP : + unsigned(AsmInfo.getDwarfRegNum(Loc.getReg(), true)); + const MachineLocation &NewLoc = Loc.isReg() ? + MachineLocation(Reg) : MachineLocation(Reg, Loc.getOffset()); + return NewLoc; +} + +static const MCSymbol &EmitCIE(MCStreamer &streamer, + const MCSymbol *personality, + unsigned personalityEncoding, + const MCSymbol *lsda, + unsigned lsdaEncoding) { + MCContext &context = streamer.getContext(); + const TargetAsmInfo &asmInfo = context.getTargetAsmInfo(); + const MCSection §ion = *asmInfo.getEHFrameSection(); + streamer.SwitchSection(§ion); + MCSymbol *sectionStart = streamer.getContext().CreateTempSymbol(); + MCSymbol *sectionEnd = streamer.getContext().CreateTempSymbol(); + + // Length + const MCExpr *Length = MakeStartMinusEndExpr(streamer, *sectionStart, + *sectionEnd, 4); + streamer.EmitLabel(sectionStart); + streamer.EmitValue(Length, 4); + + // CIE ID + streamer.EmitIntValue(0, 4); + + // Version + streamer.EmitIntValue(dwarf::DW_CIE_VERSION, 1); + + // Augmentation String + SmallString<8> Augmentation; + Augmentation += "z"; + if (personality) + Augmentation += "P"; + if (lsda) + Augmentation += "L"; + Augmentation += "R"; + streamer.EmitBytes(Augmentation.str(), 0); + streamer.EmitIntValue(0, 1); + + // Code Alignment Factor + streamer.EmitULEB128IntValue(1); + + // Data Alignment Factor + streamer.EmitSLEB128IntValue(getDataAlignmentFactor(streamer)); + + // Return Address Register + streamer.EmitULEB128IntValue(asmInfo.getDwarfRARegNum(true)); + + // Augmentation Data Length (optional) + MCSymbol *augmentationStart = streamer.getContext().CreateTempSymbol(); + MCSymbol *augmentationEnd = streamer.getContext().CreateTempSymbol(); + const MCExpr *augmentationLength = MakeStartMinusEndExpr(streamer, + *augmentationStart, + *augmentationEnd, 0); + streamer.EmitULEB128Value(augmentationLength); + + // Augmentation Data (optional) + streamer.EmitLabel(augmentationStart); + if (personality) { + // Personality Encoding + streamer.EmitIntValue(personalityEncoding, 1); + // Personality + EmitSymbol(streamer, *personality, personalityEncoding); + } + if (lsda) { + // LSDA Encoding + streamer.EmitIntValue(lsdaEncoding, 1); + } + // Encoding of the FDE pointers + streamer.EmitIntValue(dwarf::DW_EH_PE_pcrel | dwarf::DW_EH_PE_sdata4, 1); + streamer.EmitLabel(augmentationEnd); + + // Initial Instructions + + const std::vector<MachineMove> Moves = asmInfo.getInitialFrameState(); + std::vector<MCCFIInstruction> Instructions; + + for (int i = 0, n = Moves.size(); i != n; ++i) { + MCSymbol *Label = Moves[i].getLabel(); + const MachineLocation &Dst = + TranslateMachineLocation(asmInfo, Moves[i].getDestination()); + const MachineLocation &Src = + TranslateMachineLocation(asmInfo, Moves[i].getSource()); + MCCFIInstruction Inst(Label, Dst, Src); + Instructions.push_back(Inst); + } + + EmitCFIInstructions(streamer, Instructions, NULL); + + // Padding + streamer.EmitValueToAlignment(4); + + streamer.EmitLabel(sectionEnd); + return *sectionStart; +} + +static MCSymbol *EmitFDE(MCStreamer &streamer, + const MCSymbol &cieStart, + const MCDwarfFrameInfo &frame) { + MCContext &context = streamer.getContext(); + MCSymbol *fdeStart = context.CreateTempSymbol(); + MCSymbol *fdeEnd = context.CreateTempSymbol(); + + // Length + const MCExpr *Length = MakeStartMinusEndExpr(streamer, *fdeStart, *fdeEnd, 0); + streamer.EmitValue(Length, 4); + + streamer.EmitLabel(fdeStart); + // CIE Pointer + const MCExpr *offset = MakeStartMinusEndExpr(streamer, cieStart, *fdeStart, + 0); + streamer.EmitValue(offset, 4); + + // PC Begin + streamer.EmitPCRelSymbolValue(frame.Begin, 4); + + // PC Range + const MCExpr *Range = MakeStartMinusEndExpr(streamer, *frame.Begin, + *frame.End, 0); + streamer.EmitValue(Range, 4); + + // Augmentation Data Length + MCSymbol *augmentationStart = streamer.getContext().CreateTempSymbol(); + MCSymbol *augmentationEnd = streamer.getContext().CreateTempSymbol(); + const MCExpr *augmentationLength = MakeStartMinusEndExpr(streamer, + *augmentationStart, + *augmentationEnd, 0); + streamer.EmitULEB128Value(augmentationLength); + + // Augmentation Data + streamer.EmitLabel(augmentationStart); + if (frame.Lsda) + EmitSymbol(streamer, *frame.Lsda, frame.LsdaEncoding); + streamer.EmitLabel(augmentationEnd); + // Call Frame Instructions + + EmitCFIInstructions(streamer, frame.Instructions, frame.Begin); + + // Padding + streamer.EmitValueToAlignment(4); + + return fdeEnd; +} + +namespace { + struct CIEKey { + static const CIEKey getEmptyKey() { return CIEKey(0, 0, -1); } + static const CIEKey getTombstoneKey() { return CIEKey(0, -1, 0); } + + CIEKey(const MCSymbol* Personality_, unsigned PersonalityEncoding_, + unsigned LsdaEncoding_) : Personality(Personality_), + PersonalityEncoding(PersonalityEncoding_), + LsdaEncoding(LsdaEncoding_) { + } + const MCSymbol* Personality; + unsigned PersonalityEncoding; + unsigned LsdaEncoding; + }; +} + +namespace llvm { + template <> + struct DenseMapInfo<CIEKey> { + static CIEKey getEmptyKey() { + return CIEKey::getEmptyKey(); + } + static CIEKey getTombstoneKey() { + return CIEKey::getTombstoneKey(); + } + static unsigned getHashValue(const CIEKey &Key) { + FoldingSetNodeID ID; + ID.AddPointer(Key.Personality); + ID.AddInteger(Key.PersonalityEncoding); + ID.AddInteger(Key.LsdaEncoding); + return ID.ComputeHash(); + } + static bool isEqual(const CIEKey &LHS, + const CIEKey &RHS) { + return LHS.Personality == RHS.Personality && + LHS.PersonalityEncoding == RHS.PersonalityEncoding && + LHS.LsdaEncoding == RHS.LsdaEncoding; + } + }; +} + +void MCDwarfFrameEmitter::Emit(MCStreamer &streamer) { + const MCContext &context = streamer.getContext(); + const TargetAsmInfo &asmInfo = context.getTargetAsmInfo(); + MCSymbol *fdeEnd = NULL; + DenseMap<CIEKey, const MCSymbol*> CIEStarts; + + for (unsigned i = 0, n = streamer.getNumFrameInfos(); i < n; ++i) { + const MCDwarfFrameInfo &frame = streamer.getFrameInfo(i); + CIEKey key(frame.Personality, frame.PersonalityEncoding, + frame.LsdaEncoding); + const MCSymbol *&cieStart = CIEStarts[key]; + if (!cieStart) + cieStart = &EmitCIE(streamer, frame.Personality, + frame.PersonalityEncoding, frame.Lsda, + frame.LsdaEncoding); + fdeEnd = EmitFDE(streamer, *cieStart, frame); + if (i != n - 1) + streamer.EmitLabel(fdeEnd); + } + + streamer.EmitValueToAlignment(asmInfo.getPointerSize()); + if (fdeEnd) + streamer.EmitLabel(fdeEnd); +} + +void MCDwarfFrameEmitter::EmitAdvanceLoc(MCStreamer &Streamer, + uint64_t AddrDelta) { + SmallString<256> Tmp; + raw_svector_ostream OS(Tmp); + MCDwarfFrameEmitter::EncodeAdvanceLoc(AddrDelta, OS); + Streamer.EmitBytes(OS.str(), /*AddrSpace=*/0); +} + +void MCDwarfFrameEmitter::EncodeAdvanceLoc(uint64_t AddrDelta, + raw_ostream &OS) { + // FIXME: Assumes the code alignment factor is 1. + if (AddrDelta == 0) { + } else if (isUIntN(6, AddrDelta)) { + uint8_t Opcode = dwarf::DW_CFA_advance_loc | AddrDelta; + OS << Opcode; + } else if (isUInt<8>(AddrDelta)) { + OS << uint8_t(dwarf::DW_CFA_advance_loc1); + OS << uint8_t(AddrDelta); + } else if (isUInt<16>(AddrDelta)) { + // FIXME: check what is the correct behavior on a big endian machine. + OS << uint8_t(dwarf::DW_CFA_advance_loc2); + OS << uint8_t( AddrDelta & 0xff); + OS << uint8_t((AddrDelta >> 8) & 0xff); + } else { + // FIXME: check what is the correct behavior on a big endian machine. + assert(isUInt<32>(AddrDelta)); + OS << uint8_t(dwarf::DW_CFA_advance_loc4); + OS << uint8_t( AddrDelta & 0xff); + OS << uint8_t((AddrDelta >> 8) & 0xff); + OS << uint8_t((AddrDelta >> 16) & 0xff); + OS << uint8_t((AddrDelta >> 24) & 0xff); + + } +} diff --git a/lib/MC/MCELF.cpp b/lib/MC/MCELF.cpp new file mode 100644 index 0000000..ce7783e --- /dev/null +++ b/lib/MC/MCELF.cpp @@ -0,0 +1,72 @@ +//===- lib/MC/MCELF.cpp - MC ELF ------------------------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file implements ELF object file writer information. +// +//===----------------------------------------------------------------------===// + +#include "MCELF.h" +#include "llvm/MC/MCAssembler.h" +#include "llvm/MC/MCELFSymbolFlags.h" +#include "llvm/MC/MCFixupKindInfo.h" +#include "llvm/Support/ELF.h" +#include "llvm/Target/TargetAsmBackend.h" + +namespace llvm { + +void MCELF::SetBinding(MCSymbolData &SD, unsigned Binding) { + assert(Binding == ELF::STB_LOCAL || Binding == ELF::STB_GLOBAL || + Binding == ELF::STB_WEAK); + uint32_t OtherFlags = SD.getFlags() & ~(0xf << ELF_STB_Shift); + SD.setFlags(OtherFlags | (Binding << ELF_STB_Shift)); +} + +unsigned MCELF::GetBinding(const MCSymbolData &SD) { + uint32_t Binding = (SD.getFlags() & (0xf << ELF_STB_Shift)) >> ELF_STB_Shift; + assert(Binding == ELF::STB_LOCAL || Binding == ELF::STB_GLOBAL || + Binding == ELF::STB_WEAK); + return Binding; +} + +void MCELF::SetType(MCSymbolData &SD, unsigned Type) { + assert(Type == ELF::STT_NOTYPE || Type == ELF::STT_OBJECT || + Type == ELF::STT_FUNC || Type == ELF::STT_SECTION || + Type == ELF::STT_FILE || Type == ELF::STT_COMMON || + Type == ELF::STT_TLS); + + uint32_t OtherFlags = SD.getFlags() & ~(0xf << ELF_STT_Shift); + SD.setFlags(OtherFlags | (Type << ELF_STT_Shift)); +} + +unsigned MCELF::GetType(const MCSymbolData &SD) { + uint32_t Type = (SD.getFlags() & (0xf << ELF_STT_Shift)) >> ELF_STT_Shift; + assert(Type == ELF::STT_NOTYPE || Type == ELF::STT_OBJECT || + Type == ELF::STT_FUNC || Type == ELF::STT_SECTION || + Type == ELF::STT_FILE || Type == ELF::STT_COMMON || + Type == ELF::STT_TLS); + return Type; +} + +void MCELF::SetVisibility(MCSymbolData &SD, unsigned Visibility) { + assert(Visibility == ELF::STV_DEFAULT || Visibility == ELF::STV_INTERNAL || + Visibility == ELF::STV_HIDDEN || Visibility == ELF::STV_PROTECTED); + + uint32_t OtherFlags = SD.getFlags() & ~(0xf << ELF_STV_Shift); + SD.setFlags(OtherFlags | (Visibility << ELF_STV_Shift)); +} + +unsigned MCELF::GetVisibility(MCSymbolData &SD) { + unsigned Visibility = + (SD.getFlags() & (0xf << ELF_STV_Shift)) >> ELF_STV_Shift; + assert(Visibility == ELF::STV_DEFAULT || Visibility == ELF::STV_INTERNAL || + Visibility == ELF::STV_HIDDEN || Visibility == ELF::STV_PROTECTED); + return Visibility; +} + +} diff --git a/lib/MC/MCELF.h b/lib/MC/MCELF.h new file mode 100644 index 0000000..e08f1e6 --- /dev/null +++ b/lib/MC/MCELF.h @@ -0,0 +1,35 @@ +//===- lib/MC/MCELF.h - ELF MC --------------------------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file contains some support functions used by the ELF Streamer and +// ObjectWriter. +// +//===----------------------------------------------------------------------===// + +#ifndef LLVM_MC_MCELF_H +#define LLVM_MC_MCELF_H + +#include "llvm/MC/MCExpr.h" + +namespace llvm { +class MCSymbolData; + +class MCELF { + public: + static void SetBinding(MCSymbolData &SD, unsigned Binding); + static unsigned GetBinding(const MCSymbolData &SD); + static void SetType(MCSymbolData &SD, unsigned Type); + static unsigned GetType(const MCSymbolData &SD); + static void SetVisibility(MCSymbolData &SD, unsigned Visibility); + static unsigned GetVisibility(MCSymbolData &SD); +}; + +} + +#endif diff --git a/lib/MC/MCELFObjectTargetWriter.cpp b/lib/MC/MCELFObjectTargetWriter.cpp new file mode 100644 index 0000000..12a02a9 --- /dev/null +++ b/lib/MC/MCELFObjectTargetWriter.cpp @@ -0,0 +1,23 @@ +//===-- MCELFObjectTargetWriter.cpp - ELF Target Writer Subclass ----------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +#include "llvm/MC/MCELFObjectWriter.h" + +using namespace llvm; + +MCELFObjectTargetWriter::MCELFObjectTargetWriter(bool Is64Bit_, + Triple::OSType OSType_, + uint16_t EMachine_, + bool HasRelocationAddend_) + : OSType(OSType_), EMachine(EMachine_), + HasRelocationAddend(HasRelocationAddend_), Is64Bit(Is64Bit_) { +} + +MCELFObjectTargetWriter::~MCELFObjectTargetWriter() { +} diff --git a/lib/MC/MCELFStreamer.cpp b/lib/MC/MCELFStreamer.cpp index 157c0c0..464c136 100644 --- a/lib/MC/MCELFStreamer.cpp +++ b/lib/MC/MCELFStreamer.cpp @@ -11,18 +11,14 @@ // //===----------------------------------------------------------------------===// +#include "MCELFStreamer.h" +#include "MCELF.h" #include "llvm/MC/MCStreamer.h" - -#include "llvm/ADT/SmallPtrSet.h" -#include "llvm/MC/MCAssembler.h" -#include "llvm/MC/MCContext.h" #include "llvm/MC/MCCodeEmitter.h" #include "llvm/MC/MCELFSymbolFlags.h" #include "llvm/MC/MCExpr.h" #include "llvm/MC/MCInst.h" -#include "llvm/MC/MCObjectStreamer.h" #include "llvm/MC/MCSection.h" -#include "llvm/MC/MCSectionELF.h" #include "llvm/MC/MCSymbol.h" #include "llvm/MC/MCValue.h" #include "llvm/Support/Debug.h" @@ -30,159 +26,10 @@ #include "llvm/Support/ErrorHandling.h" #include "llvm/Support/raw_ostream.h" #include "llvm/Target/TargetAsmBackend.h" +#include "llvm/Target/TargetAsmInfo.h" using namespace llvm; -namespace { - -static void SetBinding(MCSymbolData &SD, unsigned Binding) { - assert(Binding == ELF::STB_LOCAL || Binding == ELF::STB_GLOBAL || - Binding == ELF::STB_WEAK); - uint32_t OtherFlags = SD.getFlags() & ~(0xf << ELF_STB_Shift); - SD.setFlags(OtherFlags | (Binding << ELF_STB_Shift)); -} - -static unsigned GetBinding(const MCSymbolData &SD) { - uint32_t Binding = (SD.getFlags() & (0xf << ELF_STB_Shift)) >> ELF_STB_Shift; - assert(Binding == ELF::STB_LOCAL || Binding == ELF::STB_GLOBAL || - Binding == ELF::STB_WEAK); - return Binding; -} - -static void SetType(MCSymbolData &SD, unsigned Type) { - assert(Type == ELF::STT_NOTYPE || Type == ELF::STT_OBJECT || - Type == ELF::STT_FUNC || Type == ELF::STT_SECTION || - Type == ELF::STT_FILE || Type == ELF::STT_COMMON || - Type == ELF::STT_TLS); - - uint32_t OtherFlags = SD.getFlags() & ~(0xf << ELF_STT_Shift); - SD.setFlags(OtherFlags | (Type << ELF_STT_Shift)); -} - -static void SetVisibility(MCSymbolData &SD, unsigned Visibility) { - assert(Visibility == ELF::STV_DEFAULT || Visibility == ELF::STV_INTERNAL || - Visibility == ELF::STV_HIDDEN || Visibility == ELF::STV_PROTECTED); - - uint32_t OtherFlags = SD.getFlags() & ~(0xf << ELF_STV_Shift); - SD.setFlags(OtherFlags | (Visibility << ELF_STV_Shift)); -} - -class MCELFStreamer : public MCObjectStreamer { -public: - MCELFStreamer(MCContext &Context, TargetAsmBackend &TAB, - raw_ostream &OS, MCCodeEmitter *Emitter) - : MCObjectStreamer(Context, TAB, OS, Emitter, false) {} - - ~MCELFStreamer() {} - - /// @name MCStreamer Interface - /// @{ - - virtual void InitSections(); - virtual void EmitLabel(MCSymbol *Symbol); - virtual void EmitAssemblerFlag(MCAssemblerFlag Flag); - virtual void EmitThumbFunc(MCSymbol *Func); - virtual void EmitAssignment(MCSymbol *Symbol, const MCExpr *Value); - virtual void EmitWeakReference(MCSymbol *Alias, const MCSymbol *Symbol); - virtual void SwitchSection(const MCSection *Section); - virtual void EmitSymbolAttribute(MCSymbol *Symbol, MCSymbolAttr Attribute); - virtual void EmitSymbolDesc(MCSymbol *Symbol, unsigned DescValue) { - assert(0 && "ELF doesn't support this directive"); - } - virtual void EmitCommonSymbol(MCSymbol *Symbol, uint64_t Size, - unsigned ByteAlignment); - virtual void BeginCOFFSymbolDef(const MCSymbol *Symbol) { - assert(0 && "ELF doesn't support this directive"); - } - - virtual void EmitCOFFSymbolStorageClass(int StorageClass) { - assert(0 && "ELF doesn't support this directive"); - } - - virtual void EmitCOFFSymbolType(int Type) { - assert(0 && "ELF doesn't support this directive"); - } - - virtual void EndCOFFSymbolDef() { - assert(0 && "ELF doesn't support this directive"); - } - - virtual void EmitELFSize(MCSymbol *Symbol, const MCExpr *Value) { - MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol); - SD.setSize(Value); - } - - virtual void EmitLocalCommonSymbol(MCSymbol *Symbol, uint64_t Size) { - assert(0 && "ELF doesn't support this directive"); - } - virtual void EmitZerofill(const MCSection *Section, MCSymbol *Symbol = 0, - unsigned Size = 0, unsigned ByteAlignment = 0) { - assert(0 && "ELF doesn't support this directive"); - } - virtual void EmitTBSSSymbol(const MCSection *Section, MCSymbol *Symbol, - uint64_t Size, unsigned ByteAlignment = 0) { - assert(0 && "ELF doesn't support this directive"); - } - virtual void EmitBytes(StringRef Data, unsigned AddrSpace); - virtual void EmitValue(const MCExpr *Value, unsigned Size,unsigned AddrSpace); - virtual void EmitGPRel32Value(const MCExpr *Value) { - assert(0 && "ELF doesn't support this directive"); - } - virtual void EmitValueToAlignment(unsigned ByteAlignment, int64_t Value = 0, - unsigned ValueSize = 1, - unsigned MaxBytesToEmit = 0); - virtual void EmitCodeAlignment(unsigned ByteAlignment, - unsigned MaxBytesToEmit = 0); - virtual void EmitValueToOffset(const MCExpr *Offset, - unsigned char Value = 0); - - virtual void EmitFileDirective(StringRef Filename); - virtual void EmitDwarfFileDirective(unsigned FileNo, StringRef Filename) { - DEBUG(dbgs() << "FIXME: MCELFStreamer:EmitDwarfFileDirective not implemented\n"); - } - - virtual void Finish(); - -private: - virtual void EmitInstToFragment(const MCInst &Inst); - virtual void EmitInstToData(const MCInst &Inst); - - struct LocalCommon { - MCSymbolData *SD; - uint64_t Size; - unsigned ByteAlignment; - }; - std::vector<LocalCommon> LocalCommons; - - SmallPtrSet<MCSymbol *, 16> BindingExplicitlySet; - /// @} - void SetSection(StringRef Section, unsigned Type, unsigned Flags, - SectionKind Kind) { - SwitchSection(getContext().getELFSection(Section, Type, Flags, Kind)); - } - - void SetSectionData() { - SetSection(".data", MCSectionELF::SHT_PROGBITS, - MCSectionELF::SHF_WRITE |MCSectionELF::SHF_ALLOC, - SectionKind::getDataRel()); - EmitCodeAlignment(4, 0); - } - void SetSectionText() { - SetSection(".text", MCSectionELF::SHT_PROGBITS, - MCSectionELF::SHF_EXECINSTR | - MCSectionELF::SHF_ALLOC, SectionKind::getText()); - EmitCodeAlignment(4, 0); - } - void SetSectionBss() { - SetSection(".bss", MCSectionELF::SHT_NOBITS, - MCSectionELF::SHF_WRITE | - MCSectionELF::SHF_ALLOC, SectionKind::getBSS()); - EmitCodeAlignment(4, 0); - } -}; - -} // end anonymous namespace. - void MCELFStreamer::InitSections() { // This emulates the same behavior of GNU as. This makes it easier // to compare the output as the major sections are in the same order. @@ -195,24 +42,13 @@ void MCELFStreamer::InitSections() { void MCELFStreamer::EmitLabel(MCSymbol *Symbol) { assert(Symbol->isUndefined() && "Cannot define a symbol twice!"); - Symbol->setSection(*CurSection); - - MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol); + MCObjectStreamer::EmitLabel(Symbol); const MCSectionELF &Section = static_cast<const MCSectionELF&>(Symbol->getSection()); - if (Section.getFlags() & MCSectionELF::SHF_TLS) - SetType(SD, ELF::STT_TLS); - - // FIXME: This is wasteful, we don't necessarily need to create a data - // fragment. Instead, we should mark the symbol as pointing into the data - // fragment if it exists, otherwise we should just queue the label and set its - // fragment pointer when we emit the next fragment. - MCDataFragment *F = getOrCreateDataFragment(); - - assert(!SD.getFragment() && "Unexpected fragment on symbol data!"); - SD.setFragment(F); - SD.setOffset(F->getContents().size()); + MCSymbolData &SD = getAssembler().getSymbolData(*Symbol); + if (Section.getFlags() & ELF::SHF_TLS) + MCELF::SetType(SD, ELF::STT_TLS); } void MCELFStreamer::EmitAssemblerFlag(MCAssemblerFlag Flag) { @@ -240,11 +76,11 @@ void MCELFStreamer::EmitAssignment(MCSymbol *Symbol, const MCExpr *Value) { Symbol->setVariableValue(AddValueSymbols(Value)); } -void MCELFStreamer::SwitchSection(const MCSection *Section) { +void MCELFStreamer::ChangeSection(const MCSection *Section) { const MCSymbol *Grp = static_cast<const MCSectionELF *>(Section)->getGroup(); if (Grp) getAssembler().getOrCreateSymbolData(*Grp); - this->MCObjectStreamer::SwitchSection(Section); + this->MCObjectStreamer::ChangeSection(Section); } void MCELFStreamer::EmitWeakReference(MCSymbol *Alias, const MCSymbol *Symbol) { @@ -284,6 +120,7 @@ void MCELFStreamer::EmitSymbolAttribute(MCSymbol *Symbol, case MCSA_LazyReference: case MCSA_Reference: case MCSA_NoDeadStrip: + case MCSA_SymbolResolver: case MCSA_PrivateExtern: case MCSA_WeakDefinition: case MCSA_WeakDefAutoPrivate: @@ -298,54 +135,54 @@ void MCELFStreamer::EmitSymbolAttribute(MCSymbol *Symbol, break; case MCSA_Global: - SetBinding(SD, ELF::STB_GLOBAL); + MCELF::SetBinding(SD, ELF::STB_GLOBAL); SD.setExternal(true); BindingExplicitlySet.insert(Symbol); break; case MCSA_WeakReference: case MCSA_Weak: - SetBinding(SD, ELF::STB_WEAK); + MCELF::SetBinding(SD, ELF::STB_WEAK); SD.setExternal(true); BindingExplicitlySet.insert(Symbol); break; case MCSA_Local: - SetBinding(SD, ELF::STB_LOCAL); + MCELF::SetBinding(SD, ELF::STB_LOCAL); SD.setExternal(false); BindingExplicitlySet.insert(Symbol); break; case MCSA_ELF_TypeFunction: - SetType(SD, ELF::STT_FUNC); + MCELF::SetType(SD, ELF::STT_FUNC); break; case MCSA_ELF_TypeObject: - SetType(SD, ELF::STT_OBJECT); + MCELF::SetType(SD, ELF::STT_OBJECT); break; case MCSA_ELF_TypeTLS: - SetType(SD, ELF::STT_TLS); + MCELF::SetType(SD, ELF::STT_TLS); break; case MCSA_ELF_TypeCommon: - SetType(SD, ELF::STT_COMMON); + MCELF::SetType(SD, ELF::STT_COMMON); break; case MCSA_ELF_TypeNoType: - SetType(SD, ELF::STT_NOTYPE); + MCELF::SetType(SD, ELF::STT_NOTYPE); break; case MCSA_Protected: - SetVisibility(SD, ELF::STV_PROTECTED); + MCELF::SetVisibility(SD, ELF::STV_PROTECTED); break; case MCSA_Hidden: - SetVisibility(SD, ELF::STV_HIDDEN); + MCELF::SetVisibility(SD, ELF::STV_HIDDEN); break; case MCSA_Internal: - SetVisibility(SD, ELF::STV_INTERNAL); + MCELF::SetVisibility(SD, ELF::STV_INTERNAL); break; } } @@ -355,17 +192,17 @@ void MCELFStreamer::EmitCommonSymbol(MCSymbol *Symbol, uint64_t Size, MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol); if (!BindingExplicitlySet.count(Symbol)) { - SetBinding(SD, ELF::STB_GLOBAL); + MCELF::SetBinding(SD, ELF::STB_GLOBAL); SD.setExternal(true); } - SetType(SD, ELF::STT_OBJECT); + MCELF::SetType(SD, ELF::STT_OBJECT); - if (GetBinding(SD) == ELF_STB_Local) { + if (MCELF::GetBinding(SD) == ELF_STB_Local) { const MCSection *Section = getAssembler().getContext().getELFSection(".bss", - MCSectionELF::SHT_NOBITS, - MCSectionELF::SHF_WRITE | - MCSectionELF::SHF_ALLOC, + ELF::SHT_NOBITS, + ELF::SHF_WRITE | + ELF::SHF_ALLOC, SectionKind::getBSS()); Symbol->setSection(*Section); @@ -378,29 +215,20 @@ void MCELFStreamer::EmitCommonSymbol(MCSymbol *Symbol, uint64_t Size, SD.setSize(MCConstantExpr::Create(Size, getContext())); } -void MCELFStreamer::EmitBytes(StringRef Data, unsigned AddrSpace) { - // TODO: This is exactly the same as WinCOFFStreamer. Consider merging into - // MCObjectStreamer. - getOrCreateDataFragment()->getContents().append(Data.begin(), Data.end()); +void MCELFStreamer::EmitLocalCommonSymbol(MCSymbol *Symbol, uint64_t Size) { + // FIXME: Should this be caught and done earlier? + MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol); + MCELF::SetBinding(SD, ELF::STB_LOCAL); + SD.setExternal(false); + BindingExplicitlySet.insert(Symbol); + // FIXME: ByteAlignment is not needed here, but is required. + EmitCommonSymbol(Symbol, Size, 1); } -void MCELFStreamer::EmitValue(const MCExpr *Value, unsigned Size, - unsigned AddrSpace) { +void MCELFStreamer::EmitBytes(StringRef Data, unsigned AddrSpace) { // TODO: This is exactly the same as WinCOFFStreamer. Consider merging into // MCObjectStreamer. - MCDataFragment *DF = getOrCreateDataFragment(); - - // Avoid fixups when possible. - int64_t AbsValue; - if (AddValueSymbols(Value)->EvaluateAsAbsolute(AbsValue)) { - // FIXME: Endianness assumption. - for (unsigned i = 0; i != Size; ++i) - DF->getContents().push_back(uint8_t(AbsValue >> (i * 8))); - } else { - DF->addFixup(MCFixup::Create(DF->getContents().size(), AddValueSymbols(Value), - MCFixup::getKindForSize(Size))); - DF->getContents().resize(DF->getContents().size() + Size, 0); - } + getOrCreateDataFragment()->getContents().append(Data.begin(), Data.end()); } void MCELFStreamer::EmitValueToAlignment(unsigned ByteAlignment, @@ -433,18 +261,11 @@ void MCELFStreamer::EmitCodeAlignment(unsigned ByteAlignment, getCurrentSectionData()->setAlignment(ByteAlignment); } -void MCELFStreamer::EmitValueToOffset(const MCExpr *Offset, - unsigned char Value) { - // TODO: This is exactly the same as MCMachOStreamer. Consider merging into - // MCObjectStreamer. - new MCOrgFragment(*Offset, Value, getCurrentSectionData()); -} - // Add a symbol for the file name of this module. This is the second // entry in the module's symbol table (the first being the null symbol). void MCELFStreamer::EmitFileDirective(StringRef Filename) { MCSymbol *Symbol = getAssembler().getContext().GetOrCreateSymbol(Filename); - Symbol->setSection(*CurSection); + Symbol->setSection(*getCurrentSection()); Symbol->setAbsolute(); MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol); @@ -452,21 +273,52 @@ void MCELFStreamer::EmitFileDirective(StringRef Filename) { SD.setFlags(ELF_STT_File | ELF_STB_Local | ELF_STV_Default); } -void MCELFStreamer::EmitInstToFragment(const MCInst &Inst) { - MCInstFragment *IF = new MCInstFragment(Inst, getCurrentSectionData()); +void MCELFStreamer::fixSymbolsInTLSFixups(const MCExpr *expr) { + switch (expr->getKind()) { + case MCExpr::Target: llvm_unreachable("Can't handle target exprs yet!"); + case MCExpr::Constant: + break; - // Add the fixups and data. - // - // FIXME: Revisit this design decision when relaxation is done, we may be - // able to get away with not storing any extra data in the MCInst. - SmallVector<MCFixup, 4> Fixups; - SmallString<256> Code; - raw_svector_ostream VecOS(Code); - getAssembler().getEmitter().EncodeInstruction(Inst, VecOS, Fixups); - VecOS.flush(); + case MCExpr::Binary: { + const MCBinaryExpr *be = cast<MCBinaryExpr>(expr); + fixSymbolsInTLSFixups(be->getLHS()); + fixSymbolsInTLSFixups(be->getRHS()); + break; + } - IF->getCode() = Code; - IF->getFixups() = Fixups; + case MCExpr::SymbolRef: { + const MCSymbolRefExpr &symRef = *cast<MCSymbolRefExpr>(expr); + switch (symRef.getKind()) { + default: + return; + case MCSymbolRefExpr::VK_NTPOFF: + case MCSymbolRefExpr::VK_GOTNTPOFF: + case MCSymbolRefExpr::VK_TLSGD: + case MCSymbolRefExpr::VK_TLSLDM: + case MCSymbolRefExpr::VK_TPOFF: + case MCSymbolRefExpr::VK_DTPOFF: + case MCSymbolRefExpr::VK_GOTTPOFF: + case MCSymbolRefExpr::VK_TLSLD: + case MCSymbolRefExpr::VK_ARM_TLSGD: + break; + } + MCSymbolData &SD = getAssembler().getOrCreateSymbolData(symRef.getSymbol()); + MCELF::SetType(SD, ELF::STT_TLS); + break; + } + + case MCExpr::Unary: + fixSymbolsInTLSFixups(cast<MCUnaryExpr>(expr)->getSubExpr()); + break; + } +} + +void MCELFStreamer::EmitInstToFragment(const MCInst &Inst) { + this->MCObjectStreamer::EmitInstToFragment(Inst); + MCInstFragment &F = *cast<MCInstFragment>(getCurrentFragment()); + + for (unsigned i = 0, e = F.getFixups().size(); i != e; ++i) + fixSymbolsInTLSFixups(F.getFixups()[i].getValue()); } void MCELFStreamer::EmitInstToData(const MCInst &Inst) { @@ -478,6 +330,9 @@ void MCELFStreamer::EmitInstToData(const MCInst &Inst) { getAssembler().getEmitter().EncodeInstruction(Inst, VecOS, Fixups); VecOS.flush(); + for (unsigned i = 0, e = Fixups.size(); i != e; ++i) + fixSymbolsInTLSFixups(Fixups[i].getValue()); + // Add the fixups and data. for (unsigned i = 0, e = Fixups.size(); i != e; ++i) { Fixups[i].setOffset(Fixups[i].getOffset() + DF->getContents().size()); @@ -487,14 +342,8 @@ void MCELFStreamer::EmitInstToData(const MCInst &Inst) { } void MCELFStreamer::Finish() { - // FIXME: duplicated code with the MachO streamer. - // Dump out the dwarf file & directory tables and line tables. - if (getContext().hasDwarfFiles()) { - const MCSection *DwarfLineSection = - getContext().getELFSection(".debug_line", 0, 0, - SectionKind::getDataRelLocal()); - MCDwarfFileTable::Emit(this, DwarfLineSection); - } + if (getNumFrameInfos()) + MCDwarfFrameEmitter::Emit(*this); for (std::vector<LocalCommon>::const_iterator i = LocalCommons.begin(), e = LocalCommons.end(); @@ -520,10 +369,12 @@ void MCELFStreamer::Finish() { } MCStreamer *llvm::createELFStreamer(MCContext &Context, TargetAsmBackend &TAB, - raw_ostream &OS, MCCodeEmitter *CE, - bool RelaxAll) { + raw_ostream &OS, MCCodeEmitter *CE, + bool RelaxAll, bool NoExecStack) { MCELFStreamer *S = new MCELFStreamer(Context, TAB, OS, CE); if (RelaxAll) S->getAssembler().setRelaxAll(true); + if (NoExecStack) + S->getAssembler().setNoExecStack(true); return S; } diff --git a/lib/MC/MCELFStreamer.h b/lib/MC/MCELFStreamer.h new file mode 100644 index 0000000..091101d --- /dev/null +++ b/lib/MC/MCELFStreamer.h @@ -0,0 +1,268 @@ +//===- lib/MC/MCELFStreamer.h - ELF Object Output -------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file assembles .s files and emits ELF .o object files. +// +//===----------------------------------------------------------------------===// + +#ifndef LLVM_MC_MCELFSTREAMER_H +#define LLVM_MC_MCELFSTREAMER_H + +#include "llvm/ADT/SmallPtrSet.h" +#include "llvm/ADT/DenseMap.h" +#include "llvm/MC/MCAssembler.h" +#include "llvm/MC/MCContext.h" +#include "llvm/MC/MCObjectStreamer.h" +#include "llvm/MC/MCSectionELF.h" + +namespace llvm { + +class MCELFStreamer : public MCObjectStreamer { +public: + MCELFStreamer(MCContext &Context, TargetAsmBackend &TAB, + raw_ostream &OS, MCCodeEmitter *Emitter) + : MCObjectStreamer(Context, TAB, OS, Emitter) {} + + ~MCELFStreamer() {} + + /// @name MCStreamer Interface + /// @{ + + virtual void InitSections(); + virtual void ChangeSection(const MCSection *Section); + virtual void EmitLabel(MCSymbol *Symbol); + virtual void EmitAssemblerFlag(MCAssemblerFlag Flag); + virtual void EmitThumbFunc(MCSymbol *Func); + virtual void EmitAssignment(MCSymbol *Symbol, const MCExpr *Value); + virtual void EmitWeakReference(MCSymbol *Alias, const MCSymbol *Symbol); + virtual void EmitSymbolAttribute(MCSymbol *Symbol, MCSymbolAttr Attribute); + virtual void EmitSymbolDesc(MCSymbol *Symbol, unsigned DescValue) { + assert(0 && "ELF doesn't support this directive"); + } + virtual void EmitCommonSymbol(MCSymbol *Symbol, uint64_t Size, + unsigned ByteAlignment); + virtual void BeginCOFFSymbolDef(const MCSymbol *Symbol) { + assert(0 && "ELF doesn't support this directive"); + } + + virtual void EmitCOFFSymbolStorageClass(int StorageClass) { + assert(0 && "ELF doesn't support this directive"); + } + + virtual void EmitCOFFSymbolType(int Type) { + assert(0 && "ELF doesn't support this directive"); + } + + virtual void EndCOFFSymbolDef() { + assert(0 && "ELF doesn't support this directive"); + } + + virtual void EmitELFSize(MCSymbol *Symbol, const MCExpr *Value) { + MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol); + SD.setSize(Value); + } + + virtual void EmitLocalCommonSymbol(MCSymbol *Symbol, uint64_t Size); + + virtual void EmitZerofill(const MCSection *Section, MCSymbol *Symbol = 0, + unsigned Size = 0, unsigned ByteAlignment = 0) { + assert(0 && "ELF doesn't support this directive"); + } + virtual void EmitTBSSSymbol(const MCSection *Section, MCSymbol *Symbol, + uint64_t Size, unsigned ByteAlignment = 0) { + assert(0 && "ELF doesn't support this directive"); + } + virtual void EmitBytes(StringRef Data, unsigned AddrSpace); + virtual void EmitValueToAlignment(unsigned ByteAlignment, int64_t Value = 0, + unsigned ValueSize = 1, + unsigned MaxBytesToEmit = 0); + virtual void EmitCodeAlignment(unsigned ByteAlignment, + unsigned MaxBytesToEmit = 0); + + virtual void EmitFileDirective(StringRef Filename); + + virtual void Finish(); + +private: + virtual void EmitInstToFragment(const MCInst &Inst); + virtual void EmitInstToData(const MCInst &Inst); + + void fixSymbolsInTLSFixups(const MCExpr *expr); + + struct LocalCommon { + MCSymbolData *SD; + uint64_t Size; + unsigned ByteAlignment; + }; + std::vector<LocalCommon> LocalCommons; + + SmallPtrSet<MCSymbol *, 16> BindingExplicitlySet; + /// @} + void SetSection(StringRef Section, unsigned Type, unsigned Flags, + SectionKind Kind) { + SwitchSection(getContext().getELFSection(Section, Type, Flags, Kind)); + } + + void SetSectionData() { + SetSection(".data", ELF::SHT_PROGBITS, + ELF::SHF_WRITE |ELF::SHF_ALLOC, + SectionKind::getDataRel()); + EmitCodeAlignment(4, 0); + } + void SetSectionText() { + SetSection(".text", ELF::SHT_PROGBITS, + ELF::SHF_EXECINSTR | + ELF::SHF_ALLOC, SectionKind::getText()); + EmitCodeAlignment(4, 0); + } + void SetSectionBss() { + SetSection(".bss", ELF::SHT_NOBITS, + ELF::SHF_WRITE | + ELF::SHF_ALLOC, SectionKind::getBSS()); + EmitCodeAlignment(4, 0); + } +}; + +} // end llvm namespace + +#endif +//===- lib/MC/MCELFStreamer.h - ELF Object Output -------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file assembles .s files and emits ELF .o object files. +// +//===----------------------------------------------------------------------===// + +#ifndef LLVM_MC_MCELFSTREAMER_H +#define LLVM_MC_MCELFSTREAMER_H + +#include "llvm/ADT/SmallPtrSet.h" +#include "llvm/ADT/DenseMap.h" +#include "llvm/MC/MCAssembler.h" +#include "llvm/MC/MCContext.h" +#include "llvm/MC/MCObjectStreamer.h" +#include "llvm/MC/MCSectionELF.h" + +namespace llvm { + +class MCELFStreamer : public MCObjectStreamer { +public: + MCELFStreamer(MCContext &Context, TargetAsmBackend &TAB, + raw_ostream &OS, MCCodeEmitter *Emitter) + : MCObjectStreamer(Context, TAB, OS, Emitter) {} + + ~MCELFStreamer() {} + + /// @name MCStreamer Interface + /// @{ + + virtual void InitSections(); + virtual void ChangeSection(const MCSection *Section); + virtual void EmitLabel(MCSymbol *Symbol); + virtual void EmitAssemblerFlag(MCAssemblerFlag Flag); + virtual void EmitThumbFunc(MCSymbol *Func); + virtual void EmitAssignment(MCSymbol *Symbol, const MCExpr *Value); + virtual void EmitWeakReference(MCSymbol *Alias, const MCSymbol *Symbol); + virtual void EmitSymbolAttribute(MCSymbol *Symbol, MCSymbolAttr Attribute); + virtual void EmitSymbolDesc(MCSymbol *Symbol, unsigned DescValue) { + assert(0 && "ELF doesn't support this directive"); + } + virtual void EmitCommonSymbol(MCSymbol *Symbol, uint64_t Size, + unsigned ByteAlignment); + virtual void BeginCOFFSymbolDef(const MCSymbol *Symbol) { + assert(0 && "ELF doesn't support this directive"); + } + + virtual void EmitCOFFSymbolStorageClass(int StorageClass) { + assert(0 && "ELF doesn't support this directive"); + } + + virtual void EmitCOFFSymbolType(int Type) { + assert(0 && "ELF doesn't support this directive"); + } + + virtual void EndCOFFSymbolDef() { + assert(0 && "ELF doesn't support this directive"); + } + + virtual void EmitELFSize(MCSymbol *Symbol, const MCExpr *Value) { + MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol); + SD.setSize(Value); + } + + virtual void EmitLocalCommonSymbol(MCSymbol *Symbol, uint64_t Size); + + virtual void EmitZerofill(const MCSection *Section, MCSymbol *Symbol = 0, + unsigned Size = 0, unsigned ByteAlignment = 0) { + assert(0 && "ELF doesn't support this directive"); + } + virtual void EmitTBSSSymbol(const MCSection *Section, MCSymbol *Symbol, + uint64_t Size, unsigned ByteAlignment = 0) { + assert(0 && "ELF doesn't support this directive"); + } + virtual void EmitBytes(StringRef Data, unsigned AddrSpace); + virtual void EmitValueToAlignment(unsigned ByteAlignment, int64_t Value = 0, + unsigned ValueSize = 1, + unsigned MaxBytesToEmit = 0); + virtual void EmitCodeAlignment(unsigned ByteAlignment, + unsigned MaxBytesToEmit = 0); + + virtual void EmitFileDirective(StringRef Filename); + + virtual void Finish(); + +private: + virtual void EmitInstToFragment(const MCInst &Inst); + virtual void EmitInstToData(const MCInst &Inst); + + void fixSymbolsInTLSFixups(const MCExpr *expr); + + struct LocalCommon { + MCSymbolData *SD; + uint64_t Size; + unsigned ByteAlignment; + }; + std::vector<LocalCommon> LocalCommons; + + SmallPtrSet<MCSymbol *, 16> BindingExplicitlySet; + /// @} + void SetSection(StringRef Section, unsigned Type, unsigned Flags, + SectionKind Kind) { + SwitchSection(getContext().getELFSection(Section, Type, Flags, Kind)); + } + + void SetSectionData() { + SetSection(".data", ELF::SHT_PROGBITS, + ELF::SHF_WRITE |ELF::SHF_ALLOC, + SectionKind::getDataRel()); + EmitCodeAlignment(4, 0); + } + void SetSectionText() { + SetSection(".text", ELF::SHT_PROGBITS, + ELF::SHF_EXECINSTR | + ELF::SHF_ALLOC, SectionKind::getText()); + EmitCodeAlignment(4, 0); + } + void SetSectionBss() { + SetSection(".bss", ELF::SHT_NOBITS, + ELF::SHF_WRITE | + ELF::SHF_ALLOC, SectionKind::getBSS()); + EmitCodeAlignment(4, 0); + } +}; + +} // end llvm namespace + +#endif diff --git a/lib/MC/MCExpr.cpp b/lib/MC/MCExpr.cpp index eea736e..54d3743 100644 --- a/lib/MC/MCExpr.cpp +++ b/lib/MC/MCExpr.cpp @@ -14,7 +14,6 @@ #include "llvm/MC/MCAsmLayout.h" #include "llvm/MC/MCAssembler.h" #include "llvm/MC/MCContext.h" -#include "llvm/MC/MCObjectFormat.h" #include "llvm/MC/MCSymbol.h" #include "llvm/MC/MCValue.h" #include "llvm/Support/Debug.h" @@ -43,10 +42,6 @@ void MCExpr::print(raw_ostream &OS) const { // absolute names. bool UseParens = Sym.getName()[0] == '$'; - if (SRE.getKind() == MCSymbolRefExpr::VK_ARM_HI16 || - SRE.getKind() == MCSymbolRefExpr::VK_ARM_LO16) - OS << MCSymbolRefExpr::getVariantKindName(SRE.getKind()); - if (SRE.getKind() == MCSymbolRefExpr::VK_PPC_HA16 || SRE.getKind() == MCSymbolRefExpr::VK_PPC_LO16) { OS << MCSymbolRefExpr::getVariantKindName(SRE.getKind()); @@ -66,8 +61,6 @@ void MCExpr::print(raw_ostream &OS) const { SRE.getKind() == MCSymbolRefExpr::VK_ARM_GOTTPOFF) OS << MCSymbolRefExpr::getVariantKindName(SRE.getKind()); else if (SRE.getKind() != MCSymbolRefExpr::VK_None && - SRE.getKind() != MCSymbolRefExpr::VK_ARM_HI16 && - SRE.getKind() != MCSymbolRefExpr::VK_ARM_LO16 && SRE.getKind() != MCSymbolRefExpr::VK_PPC_HA16 && SRE.getKind() != MCSymbolRefExpr::VK_PPC_LO16) OS << '@' << MCSymbolRefExpr::getVariantKindName(SRE.getKind()); @@ -197,14 +190,12 @@ StringRef MCSymbolRefExpr::getVariantKindName(VariantKind Kind) { case VK_TPOFF: return "TPOFF"; case VK_DTPOFF: return "DTPOFF"; case VK_TLVP: return "TLVP"; - case VK_ARM_HI16: return ":upper16:"; - case VK_ARM_LO16: return ":lower16:"; case VK_ARM_PLT: return "(PLT)"; case VK_ARM_GOT: return "(GOT)"; case VK_ARM_GOTOFF: return "(GOTOFF)"; case VK_ARM_TPOFF: return "(tpoff)"; case VK_ARM_GOTTPOFF: return "(gottpoff)"; - case VK_ARM_TLSGD: return "(tldgd)"; + case VK_ARM_TLSGD: return "(tlsgd)"; case VK_PPC_TOC: return "toc"; case VK_PPC_HA16: return "ha16"; case VK_PPC_LO16: return "lo16"; @@ -215,19 +206,33 @@ MCSymbolRefExpr::VariantKind MCSymbolRefExpr::getVariantKindForName(StringRef Name) { return StringSwitch<VariantKind>(Name) .Case("GOT", VK_GOT) + .Case("got", VK_GOT) .Case("GOTOFF", VK_GOTOFF) + .Case("gotoff", VK_GOTOFF) .Case("GOTPCREL", VK_GOTPCREL) + .Case("gotpcrel", VK_GOTPCREL) .Case("GOTTPOFF", VK_GOTTPOFF) + .Case("gottpoff", VK_GOTTPOFF) .Case("INDNTPOFF", VK_INDNTPOFF) + .Case("indntpoff", VK_INDNTPOFF) .Case("NTPOFF", VK_NTPOFF) + .Case("ntpoff", VK_NTPOFF) .Case("GOTNTPOFF", VK_GOTNTPOFF) + .Case("gotntpoff", VK_GOTNTPOFF) .Case("PLT", VK_PLT) + .Case("plt", VK_PLT) .Case("TLSGD", VK_TLSGD) + .Case("tlsgd", VK_TLSGD) .Case("TLSLD", VK_TLSLD) + .Case("tlsld", VK_TLSLD) .Case("TLSLDM", VK_TLSLDM) + .Case("tlsldm", VK_TLSLDM) .Case("TPOFF", VK_TPOFF) + .Case("tpoff", VK_TPOFF) .Case("DTPOFF", VK_DTPOFF) + .Case("dtpoff", VK_DTPOFF) .Case("TLVP", VK_TLVP) + .Case("tlvp", VK_TLVP) .Default(VK_Invalid); } @@ -237,7 +242,28 @@ void MCTargetExpr::Anchor() {} /* *** */ -bool MCExpr::EvaluateAsAbsolute(int64_t &Res, const MCAsmLayout *Layout) const { +bool MCExpr::EvaluateAsAbsolute(int64_t &Res) const { + return EvaluateAsAbsolute(Res, 0, 0, 0); +} + +bool MCExpr::EvaluateAsAbsolute(int64_t &Res, + const MCAsmLayout &Layout) const { + return EvaluateAsAbsolute(Res, &Layout.getAssembler(), &Layout, 0); +} + +bool MCExpr::EvaluateAsAbsolute(int64_t &Res, + const MCAsmLayout &Layout, + const SectionAddrMap &Addrs) const { + return EvaluateAsAbsolute(Res, &Layout.getAssembler(), &Layout, &Addrs); +} + +bool MCExpr::EvaluateAsAbsolute(int64_t &Res, const MCAssembler &Asm) const { + return EvaluateAsAbsolute(Res, &Asm, 0, 0); +} + +bool MCExpr::EvaluateAsAbsolute(int64_t &Res, const MCAssembler *Asm, + const MCAsmLayout *Layout, + const SectionAddrMap *Addrs) const { MCValue Value; // Fast path constants. @@ -246,80 +272,158 @@ bool MCExpr::EvaluateAsAbsolute(int64_t &Res, const MCAsmLayout *Layout) const { return true; } - if (!EvaluateAsRelocatable(Value, Layout) || !Value.isAbsolute()) { - // EvaluateAsAbsolute is defined to return the "current value" of - // the expression if we are given a Layout object, even in cases - // when the value is not fixed. - if (Layout) { - Res = Value.getConstant(); - if (Value.getSymA()) { - Res += Layout->getSymbolAddress( - &Layout->getAssembler().getSymbolData(Value.getSymA()->getSymbol())); - } - if (Value.getSymB()) { - Res -= Layout->getSymbolAddress( - &Layout->getAssembler().getSymbolData(Value.getSymB()->getSymbol())); - } - } - return false; - } + // FIXME: The use if InSet = Addrs is a hack. Setting InSet causes us + // absolutize differences across sections and that is what the MachO writer + // uses Addrs for. + bool IsRelocatable = + EvaluateAsRelocatableImpl(Value, Asm, Layout, Addrs, /*InSet*/ Addrs); + // Record the current value. Res = Value.getConstant(); - return true; + + return IsRelocatable && Value.isAbsolute(); } -static bool EvaluateSymbolicAdd(const MCAsmLayout *Layout, bool InSet, +/// \brief Helper method for \see EvaluateSymbolAdd(). +static void AttemptToFoldSymbolOffsetDifference(const MCAssembler *Asm, + const MCAsmLayout *Layout, + const SectionAddrMap *Addrs, + bool InSet, + const MCSymbolRefExpr *&A, + const MCSymbolRefExpr *&B, + int64_t &Addend) { + if (!A || !B) + return; + + const MCSymbol &SA = A->getSymbol(); + const MCSymbol &SB = B->getSymbol(); + + if (SA.isUndefined() || SB.isUndefined()) + return; + + if (!Asm->getWriter().IsSymbolRefDifferenceFullyResolved(*Asm, A, B, InSet)) + return; + + MCSymbolData &AD = Asm->getSymbolData(SA); + MCSymbolData &BD = Asm->getSymbolData(SB); + + if (AD.getFragment() == BD.getFragment()) { + Addend += (AD.getOffset() - BD.getOffset()); + + // Clear the symbol expr pointers to indicate we have folded these + // operands. + A = B = 0; + return; + } + + if (!Layout) + return; + + const MCSectionData &SecA = *AD.getFragment()->getParent(); + const MCSectionData &SecB = *BD.getFragment()->getParent(); + + if ((&SecA != &SecB) && !Addrs) + return; + + // Eagerly evaluate. + Addend += (Layout->getSymbolOffset(&Asm->getSymbolData(A->getSymbol())) - + Layout->getSymbolOffset(&Asm->getSymbolData(B->getSymbol()))); + if (Addrs && (&SecA != &SecB)) + Addend += (Addrs->lookup(&SecA) - Addrs->lookup(&SecB)); + + // Clear the symbol expr pointers to indicate we have folded these + // operands. + A = B = 0; +} + +/// \brief Evaluate the result of an add between (conceptually) two MCValues. +/// +/// This routine conceptually attempts to construct an MCValue: +/// Result = (Result_A - Result_B + Result_Cst) +/// from two MCValue's LHS and RHS where +/// Result = LHS + RHS +/// and +/// Result = (LHS_A - LHS_B + LHS_Cst) + (RHS_A - RHS_B + RHS_Cst). +/// +/// This routine attempts to aggresively fold the operands such that the result +/// is representable in an MCValue, but may not always succeed. +/// +/// \returns True on success, false if the result is not representable in an +/// MCValue. + +/// NOTE: It is really important to have both the Asm and Layout arguments. +/// They might look redundant, but this function can be used before layout +/// is done (see the object streamer for example) and having the Asm argument +/// lets us avoid relaxations early. +static bool EvaluateSymbolicAdd(const MCAssembler *Asm, + const MCAsmLayout *Layout, + const SectionAddrMap *Addrs, + bool InSet, const MCValue &LHS,const MCSymbolRefExpr *RHS_A, const MCSymbolRefExpr *RHS_B, int64_t RHS_Cst, MCValue &Res) { - // We can't add or subtract two symbols. - if ((LHS.getSymA() && RHS_A) || - (LHS.getSymB() && RHS_B)) - return false; - - const MCSymbolRefExpr *A = LHS.getSymA() ? LHS.getSymA() : RHS_A; - const MCSymbolRefExpr *B = LHS.getSymB() ? LHS.getSymB() : RHS_B; - if (B) { - // If we have a negated symbol, then we must have also have a non-negated - // symbol in order to encode the expression. We can do this check later to - // permit expressions which eventually fold to a representable form -- such - // as (a + (0 - b)) -- if necessary. - if (!A) - return false; + // FIXME: This routine (and other evaluation parts) are *incredibly* sloppy + // about dealing with modifiers. This will ultimately bite us, one day. + const MCSymbolRefExpr *LHS_A = LHS.getSymA(); + const MCSymbolRefExpr *LHS_B = LHS.getSymB(); + int64_t LHS_Cst = LHS.getConstant(); + + // Fold the result constant immediately. + int64_t Result_Cst = LHS_Cst + RHS_Cst; + + assert((!Layout || Asm) && + "Must have an assembler object if layout is given!"); + + // If we have a layout, we can fold resolved differences. + if (Asm) { + // First, fold out any differences which are fully resolved. By + // reassociating terms in + // Result = (LHS_A - LHS_B + LHS_Cst) + (RHS_A - RHS_B + RHS_Cst). + // we have the four possible differences: + // (LHS_A - LHS_B), + // (LHS_A - RHS_B), + // (RHS_A - LHS_B), + // (RHS_A - RHS_B). + // Since we are attempting to be as aggresive as possible about folding, we + // attempt to evaluate each possible alternative. + AttemptToFoldSymbolOffsetDifference(Asm, Layout, Addrs, InSet, LHS_A, LHS_B, + Result_Cst); + AttemptToFoldSymbolOffsetDifference(Asm, Layout, Addrs, InSet, LHS_A, RHS_B, + Result_Cst); + AttemptToFoldSymbolOffsetDifference(Asm, Layout, Addrs, InSet, RHS_A, LHS_B, + Result_Cst); + AttemptToFoldSymbolOffsetDifference(Asm, Layout, Addrs, InSet, RHS_A, RHS_B, + Result_Cst); } - // Absolutize symbol differences between defined symbols when we have a - // layout object and the target requests it. - - if (Layout && A && B) { - const MCSymbol &SA = A->getSymbol(); - const MCSymbol &SB = B->getSymbol(); - const MCObjectFormat &F = - Layout->getAssembler().getBackend().getObjectFormat(); - if (SA.isDefined() && SB.isDefined() && F.isAbsolute(InSet, SA, SB)) { - const MCAssembler &Asm = Layout->getAssembler(); - MCSymbolData &AD = Asm.getSymbolData(A->getSymbol()); - MCSymbolData &BD = Asm.getSymbolData(B->getSymbol()); - Res = MCValue::get(+ Layout->getSymbolAddress(&AD) - - Layout->getSymbolAddress(&BD) - + LHS.getConstant() - + RHS_Cst); - return true; - } - } + // We can't represent the addition or subtraction of two symbols. + if ((LHS_A && RHS_A) || (LHS_B && RHS_B)) + return false; + // At this point, we have at most one additive symbol and one subtractive + // symbol -- find them. + const MCSymbolRefExpr *A = LHS_A ? LHS_A : RHS_A; + const MCSymbolRefExpr *B = LHS_B ? LHS_B : RHS_B; + + // If we have a negated symbol, then we must have also have a non-negated + // symbol in order to encode the expression. + if (B && !A) + return false; - Res = MCValue::get(A, B, LHS.getConstant() + RHS_Cst); + Res = MCValue::get(A, B, Result_Cst); return true; } bool MCExpr::EvaluateAsRelocatable(MCValue &Res, - const MCAsmLayout *Layout) const { - return EvaluateAsRelocatableImpl(Res, Layout, false); + const MCAsmLayout &Layout) const { + return EvaluateAsRelocatableImpl(Res, &Layout.getAssembler(), &Layout, + 0, false); } bool MCExpr::EvaluateAsRelocatableImpl(MCValue &Res, + const MCAssembler *Asm, const MCAsmLayout *Layout, + const SectionAddrMap *Addrs, bool InSet) const { ++stats::MCExprEvaluate; @@ -337,7 +441,9 @@ bool MCExpr::EvaluateAsRelocatableImpl(MCValue &Res, // Evaluate recursively if this is a variable. if (Sym.isVariable() && SRE->getKind() == MCSymbolRefExpr::VK_None) { - bool Ret = Sym.getVariableValue()->EvaluateAsRelocatableImpl(Res, Layout, + bool Ret = Sym.getVariableValue()->EvaluateAsRelocatableImpl(Res, Asm, + Layout, + Addrs, true); // If we failed to simplify this to a constant, let the target // handle it. @@ -353,7 +459,8 @@ bool MCExpr::EvaluateAsRelocatableImpl(MCValue &Res, const MCUnaryExpr *AUE = cast<MCUnaryExpr>(this); MCValue Value; - if (!AUE->getSubExpr()->EvaluateAsRelocatableImpl(Value, Layout, InSet)) + if (!AUE->getSubExpr()->EvaluateAsRelocatableImpl(Value, Asm, Layout, + Addrs, InSet)) return false; switch (AUE->getOpcode()) { @@ -386,8 +493,10 @@ bool MCExpr::EvaluateAsRelocatableImpl(MCValue &Res, const MCBinaryExpr *ABE = cast<MCBinaryExpr>(this); MCValue LHSValue, RHSValue; - if (!ABE->getLHS()->EvaluateAsRelocatableImpl(LHSValue, Layout, InSet) || - !ABE->getRHS()->EvaluateAsRelocatableImpl(RHSValue, Layout, InSet)) + if (!ABE->getLHS()->EvaluateAsRelocatableImpl(LHSValue, Asm, Layout, + Addrs, InSet) || + !ABE->getRHS()->EvaluateAsRelocatableImpl(RHSValue, Asm, Layout, + Addrs, InSet)) return false; // We only support a few operations on non-constant expressions, handle @@ -398,13 +507,13 @@ bool MCExpr::EvaluateAsRelocatableImpl(MCValue &Res, return false; case MCBinaryExpr::Sub: // Negate RHS and add. - return EvaluateSymbolicAdd(Layout, InSet, LHSValue, + return EvaluateSymbolicAdd(Asm, Layout, Addrs, InSet, LHSValue, RHSValue.getSymB(), RHSValue.getSymA(), -RHSValue.getConstant(), Res); case MCBinaryExpr::Add: - return EvaluateSymbolicAdd(Layout, InSet, LHSValue, + return EvaluateSymbolicAdd(Asm, Layout, Addrs, InSet, LHSValue, RHSValue.getSymA(), RHSValue.getSymB(), RHSValue.getConstant(), Res); diff --git a/lib/MC/MCInstPrinter.cpp b/lib/MC/MCInstPrinter.cpp index 92a7154..212b85e 100644 --- a/lib/MC/MCInstPrinter.cpp +++ b/lib/MC/MCInstPrinter.cpp @@ -19,3 +19,8 @@ MCInstPrinter::~MCInstPrinter() { StringRef MCInstPrinter::getOpcodeName(unsigned Opcode) const { return ""; } + +StringRef MCInstPrinter::getRegName(unsigned RegNo) const { + assert(0 && "Target should implement this"); + return ""; +} diff --git a/lib/MC/MCLoggingStreamer.cpp b/lib/MC/MCLoggingStreamer.cpp index 9681c06..012c7f6 100644 --- a/lib/MC/MCLoggingStreamer.cpp +++ b/lib/MC/MCLoggingStreamer.cpp @@ -48,10 +48,9 @@ public: return Child->AddBlankLine(); } - virtual void SwitchSection(const MCSection *Section) { - CurSection = Section; - LogCall("SwitchSection"); - return Child->SwitchSection(Section); + virtual void ChangeSection(const MCSection *Section) { + LogCall("ChangeSection"); + return Child->ChangeSection(Section); } virtual void InitSections() { @@ -84,6 +83,13 @@ public: return Child->EmitWeakReference(Alias, Symbol); } + virtual void EmitDwarfAdvanceLineAddr(int64_t LineDelta, + const MCSymbol *LastLabel, + const MCSymbol *Label) { + LogCall("EmitDwarfAdvanceLineAddr"); + return Child->EmitDwarfAdvanceLineAddr(LineDelta, LastLabel, Label); + } + virtual void EmitSymbolAttribute(MCSymbol *Symbol, MCSymbolAttr Attribute) { LogCall("EmitSymbolAttribute"); return Child->EmitSymbolAttribute(Symbol, Attribute); @@ -147,14 +153,10 @@ public: return Child->EmitBytes(Data, AddrSpace); } - virtual void EmitValue(const MCExpr *Value, unsigned Size,unsigned AddrSpace){ + virtual void EmitValueImpl(const MCExpr *Value, unsigned Size, + bool isPCRel, unsigned AddrSpace){ LogCall("EmitValue"); - return Child->EmitValue(Value, Size, AddrSpace); - } - - virtual void EmitIntValue(uint64_t Value, unsigned Size, unsigned AddrSpace) { - LogCall("EmitIntValue"); - return Child->EmitIntValue(Value, Size, AddrSpace); + return Child->EmitValueImpl(Value, Size, isPCRel, AddrSpace); } virtual void EmitULEB128Value(const MCExpr *Value, @@ -205,12 +207,23 @@ public: return Child->EmitFileDirective(Filename); } - virtual void EmitDwarfFileDirective(unsigned FileNo, StringRef Filename) { + virtual bool EmitDwarfFileDirective(unsigned FileNo, StringRef Filename) { LogCall("EmitDwarfFileDirective", "FileNo:" + Twine(FileNo) + " Filename:" + Filename); return Child->EmitDwarfFileDirective(FileNo, Filename); } + virtual void EmitDwarfLocDirective(unsigned FileNo, unsigned Line, + unsigned Column, unsigned Flags, + unsigned Isa, unsigned Discriminator) { + LogCall("EmitDwarfLocDirective", + "FileNo:" + Twine(FileNo) + " Line:" + Twine(Line) + + " Column:" + Twine(Column) + " Flags:" + Twine(Flags) + + " Isa:" + Twine(Isa) + " Discriminator:" + Twine(Discriminator)); + return Child->EmitDwarfLocDirective(FileNo, Line, Column, Flags, + Isa, Discriminator); + } + virtual void EmitInstruction(const MCInst &Inst) { LogCall("EmitInstruction"); return Child->EmitInstruction(Inst); diff --git a/lib/MC/MCMachOStreamer.cpp b/lib/MC/MCMachOStreamer.cpp index b8640d3..d1f9f5c 100644 --- a/lib/MC/MCMachOStreamer.cpp +++ b/lib/MC/MCMachOStreamer.cpp @@ -24,6 +24,7 @@ #include "llvm/Support/ErrorHandling.h" #include "llvm/Support/raw_ostream.h" #include "llvm/Target/TargetAsmBackend.h" +#include "llvm/Target/TargetAsmInfo.h" using namespace llvm; @@ -31,13 +32,12 @@ namespace { class MCMachOStreamer : public MCObjectStreamer { private: - virtual void EmitInstToFragment(const MCInst &Inst); virtual void EmitInstToData(const MCInst &Inst); public: MCMachOStreamer(MCContext &Context, TargetAsmBackend &TAB, raw_ostream &OS, MCCodeEmitter *Emitter) - : MCObjectStreamer(Context, TAB, OS, Emitter, true) {} + : MCObjectStreamer(Context, TAB, OS, Emitter) {} /// @name MCStreamer Interface /// @{ @@ -74,17 +74,11 @@ public: virtual void EmitTBSSSymbol(const MCSection *Section, MCSymbol *Symbol, uint64_t Size, unsigned ByteAlignment = 0); virtual void EmitBytes(StringRef Data, unsigned AddrSpace); - virtual void EmitValue(const MCExpr *Value, unsigned Size,unsigned AddrSpace); - virtual void EmitGPRel32Value(const MCExpr *Value) { - assert(0 && "macho doesn't support this directive"); - } virtual void EmitValueToAlignment(unsigned ByteAlignment, int64_t Value = 0, unsigned ValueSize = 1, unsigned MaxBytesToEmit = 0); virtual void EmitCodeAlignment(unsigned ByteAlignment, unsigned MaxBytesToEmit = 0); - virtual void EmitValueToOffset(const MCExpr *Offset, - unsigned char Value = 0); virtual void EmitFileDirective(StringRef Filename) { // FIXME: Just ignore the .file; it isn't important enough to fail the @@ -92,12 +86,6 @@ public: //report_fatal_error("unsupported directive: '.file'"); } - virtual void EmitDwarfFileDirective(unsigned FileNo, StringRef Filename) { - // FIXME: Just ignore the .file; it isn't important enough to fail the - // entire assembly. - - //report_fatal_error("unsupported directive: '.file'"); - } virtual void Finish(); @@ -114,30 +102,18 @@ void MCMachOStreamer::InitSections() { } void MCMachOStreamer::EmitLabel(MCSymbol *Symbol) { - // TODO: This is almost exactly the same as WinCOFFStreamer. Consider merging - // into MCObjectStreamer. assert(Symbol->isUndefined() && "Cannot define a symbol twice!"); - assert(!Symbol->isVariable() && "Cannot emit a variable symbol!"); - assert(CurSection && "Cannot emit before setting section!"); - - Symbol->setSection(*CurSection); - - MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol); + // isSymbolLinkerVisible uses the section. + Symbol->setSection(*getCurrentSection()); // We have to create a new fragment if this is an atom defining symbol, // fragments cannot span atoms. - if (getAssembler().isSymbolLinkerVisible(SD.getSymbol())) + if (getAssembler().isSymbolLinkerVisible(*Symbol)) new MCDataFragment(getCurrentSectionData()); - // FIXME: This is wasteful, we don't necessarily need to create a data - // fragment. Instead, we should mark the symbol as pointing into the data - // fragment if it exists, otherwise we should just queue the label and set its - // fragment pointer when we emit the next fragment. - MCDataFragment *F = getOrCreateDataFragment(); - assert(!SD.getFragment() && "Unexpected fragment on symbol data!"); - SD.setFragment(F); - SD.setOffset(F->getContents().size()); + MCObjectStreamer::EmitLabel(Symbol); + MCSymbolData &SD = getAssembler().getSymbolData(*Symbol); // This causes the reference type flag to be cleared. Darwin 'as' was "trying" // to clear the weak reference and weak definition bits too, but the // implementation was buggy. For now we just try to match 'as', for @@ -149,6 +125,9 @@ void MCMachOStreamer::EmitLabel(MCSymbol *Symbol) { } void MCMachOStreamer::EmitAssemblerFlag(MCAssemblerFlag Flag) { + // Let the target do whatever target specific stuff it needs to do. + getAssembler().getBackend().HandleAssemblerFlag(Flag); + // Do any generic stuff we need to do. switch (Flag) { case MCAF_SyntaxUnified: return; // no-op here. case MCAF_Code16: return; // no-op here. @@ -161,8 +140,16 @@ void MCMachOStreamer::EmitAssemblerFlag(MCAssemblerFlag Flag) { } } -void MCMachOStreamer::EmitThumbFunc(MCSymbol *Func) { +void MCMachOStreamer::EmitThumbFunc(MCSymbol *Symbol) { // FIXME: Flag the function ISA as thumb with DW_AT_APPLE_isa. + + // Remember that the function is a thumb function. Fixup and relocation + // values will need adjusted. + getAssembler().setIsThumbFunc(Symbol); + + // Mark the thumb bit on the symbol. + MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol); + SD.setFlags(SD.getFlags() | SF_ThumbFunc); } void MCMachOStreamer::EmitAssignment(MCSymbol *Symbol, const MCExpr *Value) { @@ -241,6 +228,10 @@ void MCMachOStreamer::EmitSymbolAttribute(MCSymbol *Symbol, SD.setFlags(SD.getFlags() | SF_NoDeadStrip); break; + case MCSA_SymbolResolver: + SD.setFlags(SD.getFlags() | SF_SymbolResolver); + break; + case MCSA_PrivateExtern: SD.setExternal(true); SD.setPrivateExtern(true); @@ -324,26 +315,6 @@ void MCMachOStreamer::EmitBytes(StringRef Data, unsigned AddrSpace) { getOrCreateDataFragment()->getContents().append(Data.begin(), Data.end()); } -void MCMachOStreamer::EmitValue(const MCExpr *Value, unsigned Size, - unsigned AddrSpace) { - // TODO: This is exactly the same as WinCOFFStreamer. Consider merging into - // MCObjectStreamer. - MCDataFragment *DF = getOrCreateDataFragment(); - - // Avoid fixups when possible. - int64_t AbsValue; - if (AddValueSymbols(Value)->EvaluateAsAbsolute(AbsValue)) { - // FIXME: Endianness assumption. - for (unsigned i = 0; i != Size; ++i) - DF->getContents().push_back(uint8_t(AbsValue >> (i * 8))); - } else { - DF->addFixup(MCFixup::Create(DF->getContents().size(), - AddValueSymbols(Value), - MCFixup::getKindForSize(Size))); - DF->getContents().resize(DF->getContents().size() + Size, 0); - } -} - void MCMachOStreamer::EmitValueToAlignment(unsigned ByteAlignment, int64_t Value, unsigned ValueSize, unsigned MaxBytesToEmit) { @@ -374,28 +345,6 @@ void MCMachOStreamer::EmitCodeAlignment(unsigned ByteAlignment, getCurrentSectionData()->setAlignment(ByteAlignment); } -void MCMachOStreamer::EmitValueToOffset(const MCExpr *Offset, - unsigned char Value) { - new MCOrgFragment(*Offset, Value, getCurrentSectionData()); -} - -void MCMachOStreamer::EmitInstToFragment(const MCInst &Inst) { - MCInstFragment *IF = new MCInstFragment(Inst, getCurrentSectionData()); - - // Add the fixups and data. - // - // FIXME: Revisit this design decision when relaxation is done, we may be - // able to get away with not storing any extra data in the MCInst. - SmallVector<MCFixup, 4> Fixups; - SmallString<256> Code; - raw_svector_ostream VecOS(Code); - getAssembler().getEmitter().EncodeInstruction(Inst, VecOS, Fixups); - VecOS.flush(); - - IF->getCode() = Code; - IF->getFixups() = Fixups; -} - void MCMachOStreamer::EmitInstToData(const MCInst &Inst) { MCDataFragment *DF = getOrCreateDataFragment(); @@ -414,15 +363,6 @@ void MCMachOStreamer::EmitInstToData(const MCInst &Inst) { } void MCMachOStreamer::Finish() { - // Dump out the dwarf file & directory tables and line tables. - if (getContext().hasDwarfFiles()) { - const MCSection *DwarfLineSection = getContext().getMachOSection("__DWARF", - "__debug_line", - MCSectionMachO::S_ATTR_DEBUG, - 0, SectionKind::getDataRelLocal()); - MCDwarfFileTable::Emit(this, DwarfLineSection); - } - // We have to set the fragment atom associations so we can relax properly for // Mach-O. diff --git a/lib/MC/MCMachObjectTargetWriter.cpp b/lib/MC/MCMachObjectTargetWriter.cpp new file mode 100644 index 0000000..146cebf --- /dev/null +++ b/lib/MC/MCMachObjectTargetWriter.cpp @@ -0,0 +1,22 @@ +//===-- MCMachObjectTargetWriter.cpp - Mach-O Target Writer Subclass ------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +#include "llvm/MC/MCMachObjectWriter.h" + +using namespace llvm; + +MCMachObjectTargetWriter::MCMachObjectTargetWriter( + bool Is64Bit_, uint32_t CPUType_, uint32_t CPUSubtype_, + bool UseAggressiveSymbolFolding_) + : Is64Bit(Is64Bit_), CPUType(CPUType_), CPUSubtype(CPUSubtype_), + UseAggressiveSymbolFolding(UseAggressiveSymbolFolding_) { +} + +MCMachObjectTargetWriter::~MCMachObjectTargetWriter() { +} diff --git a/lib/MC/MCNullStreamer.cpp b/lib/MC/MCNullStreamer.cpp index b95a4f6..08ddf01 100644 --- a/lib/MC/MCNullStreamer.cpp +++ b/lib/MC/MCNullStreamer.cpp @@ -28,15 +28,13 @@ namespace { virtual void InitSections() { } - virtual void SwitchSection(const MCSection *Section) { - PrevSection = CurSection; - CurSection = Section; + virtual void ChangeSection(const MCSection *Section) { } virtual void EmitLabel(MCSymbol *Symbol) { assert(Symbol->isUndefined() && "Cannot define a symbol twice!"); - assert(CurSection && "Cannot emit before setting section!"); - Symbol->setSection(*CurSection); + assert(getCurrentSection() && "Cannot emit before setting section!"); + Symbol->setSection(*getCurrentSection()); } virtual void EmitAssemblerFlag(MCAssemblerFlag Flag) {} @@ -44,6 +42,9 @@ namespace { virtual void EmitAssignment(MCSymbol *Symbol, const MCExpr *Value) {} virtual void EmitWeakReference(MCSymbol *Alias, const MCSymbol *Symbol){} + virtual void EmitDwarfAdvanceLineAddr(int64_t LineDelta, + const MCSymbol *LastLabel, + const MCSymbol *Label) {} virtual void EmitSymbolAttribute(MCSymbol *Symbol, MCSymbolAttr Attribute){} @@ -65,8 +66,8 @@ namespace { uint64_t Size, unsigned ByteAlignment) {} virtual void EmitBytes(StringRef Data, unsigned AddrSpace) {} - virtual void EmitValue(const MCExpr *Value, unsigned Size, - unsigned AddrSpace) {} + virtual void EmitValueImpl(const MCExpr *Value, unsigned Size, + bool isPCRel, unsigned AddrSpace) {} virtual void EmitULEB128Value(const MCExpr *Value, unsigned AddrSpace = 0) {} virtual void EmitSLEB128Value(const MCExpr *Value, @@ -83,7 +84,12 @@ namespace { unsigned char Value = 0) {} virtual void EmitFileDirective(StringRef Filename) {} - virtual void EmitDwarfFileDirective(unsigned FileNo,StringRef Filename) {} + virtual bool EmitDwarfFileDirective(unsigned FileNo,StringRef Filename) { + return false; + } + virtual void EmitDwarfLocDirective(unsigned FileNo, unsigned Line, + unsigned Column, unsigned Flags, + unsigned Isa, unsigned Discriminator) {} virtual void EmitInstruction(const MCInst &Inst) {} virtual void Finish() {} diff --git a/lib/MC/MCObjectFormat.cpp b/lib/MC/MCObjectFormat.cpp deleted file mode 100644 index aeff334..0000000 --- a/lib/MC/MCObjectFormat.cpp +++ /dev/null @@ -1,34 +0,0 @@ -//===- lib/MC/MCObjectFormat.cpp - MCObjectFormat implementation ----------===// -// -// The LLVM Compiler Infrastructure -// -// This file is distributed under the University of Illinois Open Source -// License. See LICENSE.TXT for details. -// -//===----------------------------------------------------------------------===// - -#include "llvm/MC/MCObjectFormat.h" -#include "llvm/MC/MCSymbol.h" - -using namespace llvm; - -MCObjectFormat::~MCObjectFormat() { -} - -bool MCELFObjectFormat::isAbsolute(bool IsSet, const MCSymbol &A, - const MCSymbol &B) const { - // On ELF A - B is absolute if A and B are in the same section. - return &A.getSection() == &B.getSection(); -} - -bool MCMachOObjectFormat::isAbsolute(bool IsSet, const MCSymbol &A, - const MCSymbol &B) const { - // On MachO A - B is absolute only if in a set. - return IsSet; -} - -bool MCCOFFObjectFormat::isAbsolute(bool IsSet, const MCSymbol &A, - const MCSymbol &B) const { - // On COFF A - B is absolute if A and B are in the same section. - return &A.getSection() == &B.getSection(); -} diff --git a/lib/MC/MCObjectStreamer.cpp b/lib/MC/MCObjectStreamer.cpp index a40c754..e67d9b0 100644 --- a/lib/MC/MCObjectStreamer.cpp +++ b/lib/MC/MCObjectStreamer.cpp @@ -7,23 +7,26 @@ // //===----------------------------------------------------------------------===// +#include "llvm/MC/MCAsmInfo.h" #include "llvm/MC/MCObjectStreamer.h" #include "llvm/Support/ErrorHandling.h" #include "llvm/MC/MCAssembler.h" #include "llvm/MC/MCCodeEmitter.h" +#include "llvm/MC/MCContext.h" #include "llvm/MC/MCDwarf.h" #include "llvm/MC/MCExpr.h" +#include "llvm/MC/MCSymbol.h" #include "llvm/Target/TargetAsmBackend.h" +#include "llvm/Target/TargetAsmInfo.h" using namespace llvm; MCObjectStreamer::MCObjectStreamer(MCContext &Context, TargetAsmBackend &TAB, - raw_ostream &_OS, MCCodeEmitter *_Emitter, - bool _PadSectionToAlignment) - : MCStreamer(Context), Assembler(new MCAssembler(Context, TAB, - *_Emitter, - _PadSectionToAlignment, - _OS)), + raw_ostream &OS, MCCodeEmitter *Emitter_) + : MCStreamer(Context), + Assembler(new MCAssembler(Context, TAB, + *Emitter_, *TAB.createObjectWriter(OS), + OS)), CurSectionData(0) { } @@ -31,6 +34,7 @@ MCObjectStreamer::MCObjectStreamer(MCContext &Context, TargetAsmBackend &TAB, MCObjectStreamer::~MCObjectStreamer() { delete &Assembler->getBackend(); delete &Assembler->getEmitter(); + delete &Assembler->getWriter(); delete Assembler; } @@ -52,7 +56,10 @@ MCDataFragment *MCObjectStreamer::getOrCreateDataFragment() const { const MCExpr *MCObjectStreamer::AddValueSymbols(const MCExpr *Value) { switch (Value->getKind()) { - case MCExpr::Target: llvm_unreachable("Can't handle target exprs yet!"); + case MCExpr::Target: + cast<MCTargetExpr>(Value)->AddValueSymbols(Assembler); + break; + case MCExpr::Constant: break; @@ -75,13 +82,58 @@ const MCExpr *MCObjectStreamer::AddValueSymbols(const MCExpr *Value) { return Value; } +void MCObjectStreamer::EmitValueImpl(const MCExpr *Value, unsigned Size, + bool isPCRel, unsigned AddrSpace) { + assert(AddrSpace == 0 && "Address space must be 0!"); + MCDataFragment *DF = getOrCreateDataFragment(); + + // Avoid fixups when possible. + int64_t AbsValue; + if (AddValueSymbols(Value)->EvaluateAsAbsolute(AbsValue, getAssembler())) { + EmitIntValue(AbsValue, Size, AddrSpace); + return; + } + DF->addFixup(MCFixup::Create(DF->getContents().size(), + Value, + MCFixup::getKindForSize(Size, isPCRel))); + DF->getContents().resize(DF->getContents().size() + Size, 0); +} + +void MCObjectStreamer::EmitLabel(MCSymbol *Symbol) { + assert(!Symbol->isVariable() && "Cannot emit a variable symbol!"); + assert(getCurrentSection() && "Cannot emit before setting section!"); + + Symbol->setSection(*getCurrentSection()); + + MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol); + + // FIXME: This is wasteful, we don't necessarily need to create a data + // fragment. Instead, we should mark the symbol as pointing into the data + // fragment if it exists, otherwise we should just queue the label and set its + // fragment pointer when we emit the next fragment. + MCDataFragment *F = getOrCreateDataFragment(); + assert(!SD.getFragment() && "Unexpected fragment on symbol data!"); + SD.setFragment(F); + SD.setOffset(F->getContents().size()); +} + void MCObjectStreamer::EmitULEB128Value(const MCExpr *Value, unsigned AddrSpace) { + int64_t IntValue; + if (Value->EvaluateAsAbsolute(IntValue, getAssembler())) { + EmitULEB128IntValue(IntValue, AddrSpace); + return; + } new MCLEBFragment(*Value, false, getCurrentSectionData()); } void MCObjectStreamer::EmitSLEB128Value(const MCExpr *Value, unsigned AddrSpace) { + int64_t IntValue; + if (Value->EvaluateAsAbsolute(IntValue, getAssembler())) { + EmitSLEB128IntValue(IntValue, AddrSpace); + return; + } new MCLEBFragment(*Value, true, getCurrentSectionData()); } @@ -90,14 +142,9 @@ void MCObjectStreamer::EmitWeakReference(MCSymbol *Alias, report_fatal_error("This file format doesn't support weak aliases."); } -void MCObjectStreamer::SwitchSection(const MCSection *Section) { +void MCObjectStreamer::ChangeSection(const MCSection *Section) { assert(Section && "Cannot switch to a null section!"); - // If already in this section, then this is a noop. - if (Section == CurSection) return; - - PrevSection = CurSection; - CurSection = Section; CurSectionData = &getAssembler().getOrCreateSectionData(*Section); } @@ -134,6 +181,90 @@ void MCObjectStreamer::EmitInstruction(const MCInst &Inst) { EmitInstToFragment(Inst); } +void MCObjectStreamer::EmitInstToFragment(const MCInst &Inst) { + MCInstFragment *IF = new MCInstFragment(Inst, getCurrentSectionData()); + + raw_svector_ostream VecOS(IF->getCode()); + getAssembler().getEmitter().EncodeInstruction(Inst, VecOS, IF->getFixups()); +} + +static const MCExpr *BuildSymbolDiff(MCContext &Context, + const MCSymbol *A, const MCSymbol *B) { + MCSymbolRefExpr::VariantKind Variant = MCSymbolRefExpr::VK_None; + const MCExpr *ARef = + MCSymbolRefExpr::Create(A, Variant, Context); + const MCExpr *BRef = + MCSymbolRefExpr::Create(B, Variant, Context); + const MCExpr *AddrDelta = + MCBinaryExpr::Create(MCBinaryExpr::Sub, ARef, BRef, Context); + return AddrDelta; +} + +static const MCExpr *ForceExpAbs(MCObjectStreamer *Streamer, + MCContext &Context, const MCExpr* Expr) { + if (Context.getAsmInfo().hasAggressiveSymbolFolding()) + return Expr; + + MCSymbol *ABS = Context.CreateTempSymbol(); + Streamer->EmitAssignment(ABS, Expr); + return MCSymbolRefExpr::Create(ABS, Context); +} + +void MCObjectStreamer::EmitDwarfAdvanceLineAddr(int64_t LineDelta, + const MCSymbol *LastLabel, + const MCSymbol *Label) { + if (!LastLabel) { + int PointerSize = getContext().getTargetAsmInfo().getPointerSize(); + EmitDwarfSetLineAddr(LineDelta, Label, PointerSize); + return; + } + const MCExpr *AddrDelta = BuildSymbolDiff(getContext(), Label, LastLabel); + int64_t Res; + if (AddrDelta->EvaluateAsAbsolute(Res, getAssembler())) { + MCDwarfLineAddr::Emit(this, LineDelta, Res); + return; + } + AddrDelta = ForceExpAbs(this, getContext(), AddrDelta); + new MCDwarfLineAddrFragment(LineDelta, *AddrDelta, getCurrentSectionData()); +} + +void MCObjectStreamer::EmitDwarfAdvanceFrameAddr(const MCSymbol *LastLabel, + const MCSymbol *Label) { + const MCExpr *AddrDelta = BuildSymbolDiff(getContext(), Label, LastLabel); + int64_t Res; + if (AddrDelta->EvaluateAsAbsolute(Res, getAssembler())) { + MCDwarfFrameEmitter::EmitAdvanceLoc(*this, Res); + return; + } + AddrDelta = ForceExpAbs(this, getContext(), AddrDelta); + new MCDwarfCallFrameFragment(*AddrDelta, getCurrentSectionData()); +} + +void MCObjectStreamer::EmitValueToOffset(const MCExpr *Offset, + unsigned char Value) { + int64_t Res; + if (Offset->EvaluateAsAbsolute(Res, getAssembler())) { + new MCOrgFragment(*Offset, Value, getCurrentSectionData()); + return; + } + + MCSymbol *CurrentPos = getContext().CreateTempSymbol(); + EmitLabel(CurrentPos); + MCSymbolRefExpr::VariantKind Variant = MCSymbolRefExpr::VK_None; + const MCExpr *Ref = + MCSymbolRefExpr::Create(CurrentPos, Variant, getContext()); + const MCExpr *Delta = + MCBinaryExpr::Create(MCBinaryExpr::Sub, Offset, Ref, getContext()); + + if (!Delta->EvaluateAsAbsolute(Res, getAssembler())) + report_fatal_error("expected assembly-time absolute expression"); + EmitFill(Res, Value, 0); +} + void MCObjectStreamer::Finish() { + // Dump out the dwarf file & directory tables and line tables. + if (getContext().hasDwarfFiles()) + MCDwarfFileTable::Emit(this); + getAssembler().Finish(); } diff --git a/lib/MC/MCObjectWriter.cpp b/lib/MC/MCObjectWriter.cpp index 6cee76d..efe9f68 100644 --- a/lib/MC/MCObjectWriter.cpp +++ b/lib/MC/MCObjectWriter.cpp @@ -7,7 +7,10 @@ // //===----------------------------------------------------------------------===// +#include "llvm/MC/MCAssembler.h" +#include "llvm/MC/MCExpr.h" #include "llvm/MC/MCObjectWriter.h" +#include "llvm/MC/MCSymbol.h" using namespace llvm; @@ -39,3 +42,39 @@ void MCObjectWriter::EncodeULEB128(uint64_t Value, raw_ostream &OS) { OS << char(Byte); } while (Value != 0); } + +bool +MCObjectWriter::IsSymbolRefDifferenceFullyResolved(const MCAssembler &Asm, + const MCSymbolRefExpr *A, + const MCSymbolRefExpr *B, + bool InSet) const { + // Modified symbol references cannot be resolved. + if (A->getKind() != MCSymbolRefExpr::VK_None || + B->getKind() != MCSymbolRefExpr::VK_None) + return false; + + const MCSymbol &SA = A->getSymbol(); + const MCSymbol &SB = B->getSymbol(); + if (SA.AliasedSymbol().isUndefined() || SB.AliasedSymbol().isUndefined()) + return false; + + const MCSymbolData &DataA = Asm.getSymbolData(SA); + const MCSymbolData &DataB = Asm.getSymbolData(SB); + + return IsSymbolRefDifferenceFullyResolvedImpl(Asm, DataA, + *DataB.getFragment(), + InSet, + false); +} + +bool +MCObjectWriter::IsSymbolRefDifferenceFullyResolvedImpl(const MCAssembler &Asm, + const MCSymbolData &DataA, + const MCFragment &FB, + bool InSet, + bool IsPCRel) const { + const MCSection &SecA = DataA.getSymbol().AliasedSymbol().getSection(); + const MCSection &SecB = FB.getParent()->getSection(); + // On ELF and COFF A - B is absolute if A and B are in the same section. + return &SecA == &SecB; +} diff --git a/lib/MC/MCParser/AsmLexer.cpp b/lib/MC/MCParser/AsmLexer.cpp index 5c1ae2f..89374d0 100644 --- a/lib/MC/MCParser/AsmLexer.cpp +++ b/lib/MC/MCParser/AsmLexer.cpp @@ -15,7 +15,7 @@ #include "llvm/Support/SMLoc.h" #include "llvm/Support/MemoryBuffer.h" #include "llvm/MC/MCAsmInfo.h" -#include <ctype.h> +#include <cctype> #include <cerrno> #include <cstdio> #include <cstdlib> @@ -31,12 +31,12 @@ AsmLexer::~AsmLexer() { void AsmLexer::setBuffer(const MemoryBuffer *buf, const char *ptr) { CurBuf = buf; - + if (ptr) CurPtr = ptr; else CurPtr = CurBuf->getBufferStart(); - + TokStart = 0; } @@ -44,7 +44,7 @@ void AsmLexer::setBuffer(const MemoryBuffer *buf, const char *ptr) { /// location. This is defined to always return AsmToken::Error. AsmToken AsmLexer::ReturnError(const char *Loc, const std::string &Msg) { SetError(SMLoc::getFromPointer(Loc), Msg); - + return AsmToken(AsmToken::Error, StringRef(Loc, 0)); } @@ -58,9 +58,9 @@ int AsmLexer::getNextChar() { // a random nul in the file. Disambiguate that here. if (CurPtr-1 != CurBuf->getBufferEnd()) return 0; // Just whitespace. - + // Otherwise, return end of file. - --CurPtr; // Another call to lex will return EOF again. + --CurPtr; // Another call to lex will return EOF again. return EOF; } } @@ -106,11 +106,11 @@ AsmToken AsmLexer::LexIdentifier() { while (IsIdentifierChar(*CurPtr)) ++CurPtr; - + // Handle . as a special case. if (CurPtr == TokStart+1 && TokStart[0] == '.') return AsmToken(AsmToken::Dot, StringRef(TokStart, 1)); - + return AsmToken(AsmToken::Identifier, StringRef(TokStart, CurPtr - TokStart)); } @@ -133,7 +133,7 @@ AsmToken AsmLexer::LexSlash() { case '*': // End of the comment? if (CurPtr[0] != '/') break; - + ++CurPtr; // End the */. return LexToken(); } @@ -148,7 +148,7 @@ AsmToken AsmLexer::LexLineComment() { int CurChar = getNextChar(); while (CurChar != '\n' && CurChar != '\n' && CurChar != EOF) CurChar = getNextChar(); - + if (CurChar == EOF) return AsmToken(AsmToken::Eof, StringRef(CurPtr, 0)); return AsmToken(AsmToken::EndOfStatement, StringRef(CurPtr, 0)); @@ -184,21 +184,21 @@ AsmToken AsmLexer::LexDigit() { long long Value; if (Result.getAsInteger(10, Value)) { - // We have to handle minint_as_a_positive_value specially, because - // - minint_as_a_positive_value = minint and it is valid. - if (Result == "9223372036854775808") - Value = -9223372036854775808ULL; - else - return ReturnError(TokStart, "Invalid decimal number"); + // Allow positive values that are too large to fit into a signed 64-bit + // integer, but that do fit in an unsigned one, we just convert them over. + unsigned long long UValue; + if (Result.getAsInteger(10, UValue)) + return ReturnError(TokStart, "invalid decimal number"); + Value = (long long)UValue; } - + // The darwin/x86 (and x86-64) assembler accepts and ignores ULL and LL // suffixes on integer literals. SkipIgnoredIntegerSuffix(CurPtr); - + return AsmToken(AsmToken::Integer, Result, Value); } - + if (*CurPtr == 'b') { ++CurPtr; // See if we actually have "0b" as part of something like "jmp 0b\n" @@ -210,30 +210,30 @@ AsmToken AsmLexer::LexDigit() { const char *NumStart = CurPtr; while (CurPtr[0] == '0' || CurPtr[0] == '1') ++CurPtr; - + // Requires at least one binary digit. if (CurPtr == NumStart) return ReturnError(TokStart, "Invalid binary number"); - + StringRef Result(TokStart, CurPtr - TokStart); - + long long Value; if (Result.substr(2).getAsInteger(2, Value)) return ReturnError(TokStart, "Invalid binary number"); - + // The darwin/x86 (and x86-64) assembler accepts and ignores ULL and LL // suffixes on integer literals. SkipIgnoredIntegerSuffix(CurPtr); - + return AsmToken(AsmToken::Integer, Result, Value); } - + if (*CurPtr == 'x') { ++CurPtr; const char *NumStart = CurPtr; while (isxdigit(CurPtr[0])) ++CurPtr; - + // Requires at least one hex digit. if (CurPtr == NumStart) return ReturnError(CurPtr-2, "Invalid hexadecimal number"); @@ -241,31 +241,67 @@ AsmToken AsmLexer::LexDigit() { unsigned long long Result; if (StringRef(TokStart, CurPtr - TokStart).getAsInteger(0, Result)) return ReturnError(TokStart, "Invalid hexadecimal number"); - + // The darwin/x86 (and x86-64) assembler accepts and ignores ULL and LL // suffixes on integer literals. SkipIgnoredIntegerSuffix(CurPtr); - + return AsmToken(AsmToken::Integer, StringRef(TokStart, CurPtr - TokStart), (int64_t)Result); } - + // Must be an octal number, it starts with 0. while (*CurPtr >= '0' && *CurPtr <= '7') ++CurPtr; - + StringRef Result(TokStart, CurPtr - TokStart); long long Value; if (Result.getAsInteger(8, Value)) return ReturnError(TokStart, "Invalid octal number"); - + // The darwin/x86 (and x86-64) assembler accepts and ignores ULL and LL // suffixes on integer literals. SkipIgnoredIntegerSuffix(CurPtr); - + return AsmToken(AsmToken::Integer, Result, Value); } +/// LexSingleQuote: Integer: 'b' +AsmToken AsmLexer::LexSingleQuote() { + int CurChar = getNextChar(); + + if (CurChar == '\\') + CurChar = getNextChar(); + + if (CurChar == EOF) + return ReturnError(TokStart, "unterminated single quote"); + + CurChar = getNextChar(); + + if (CurChar != '\'') + return ReturnError(TokStart, "single quote way too long"); + + // The idea here being that 'c' is basically just an integral + // constant. + StringRef Res = StringRef(TokStart,CurPtr - TokStart); + long long Value; + + if (Res.startswith("\'\\")) { + char theChar = Res[2]; + switch (theChar) { + default: Value = theChar; break; + case '\'': Value = '\''; break; + case 't': Value = '\t'; break; + case 'n': Value = '\n'; break; + case 'b': Value = '\b'; break; + } + } else + Value = TokStart[1]; + + return AsmToken(AsmToken::Integer, Res, Value); +} + + /// LexQuote: String: "..." AsmToken AsmLexer::LexQuote() { int CurChar = getNextChar(); @@ -275,13 +311,13 @@ AsmToken AsmLexer::LexQuote() { // Allow \", etc. CurChar = getNextChar(); } - + if (CurChar == EOF) return ReturnError(TokStart, "unterminated string constant"); CurChar = getNextChar(); } - + return AsmToken(AsmToken::String, StringRef(TokStart, CurPtr - TokStart)); } @@ -307,7 +343,7 @@ AsmToken AsmLexer::LexToken() { TokStart = CurPtr; // This always consumes at least one character. int CurChar = getNextChar(); - + if (isAtStartOfComment(CurChar)) return LexLineComment(); @@ -316,7 +352,7 @@ AsmToken AsmLexer::LexToken() { // Handle identifier: [a-zA-Z_.][a-zA-Z0-9_$.@]* if (isalpha(CurChar) || CurChar == '_' || CurChar == '.') return LexIdentifier(); - + // Unknown character, emit an error. return ReturnError(TokStart, "invalid character in input"); case EOF: return AsmToken(AsmToken::Eof, StringRef(TokStart, 0)); @@ -342,49 +378,50 @@ AsmToken AsmLexer::LexToken() { case ',': return AsmToken(AsmToken::Comma, StringRef(TokStart, 1)); case '$': return AsmToken(AsmToken::Dollar, StringRef(TokStart, 1)); case '@': return AsmToken(AsmToken::At, StringRef(TokStart, 1)); - case '=': + case '=': if (*CurPtr == '=') return ++CurPtr, AsmToken(AsmToken::EqualEqual, StringRef(TokStart, 2)); return AsmToken(AsmToken::Equal, StringRef(TokStart, 1)); - case '|': + case '|': if (*CurPtr == '|') return ++CurPtr, AsmToken(AsmToken::PipePipe, StringRef(TokStart, 2)); return AsmToken(AsmToken::Pipe, StringRef(TokStart, 1)); case '^': return AsmToken(AsmToken::Caret, StringRef(TokStart, 1)); - case '&': + case '&': if (*CurPtr == '&') return ++CurPtr, AsmToken(AsmToken::AmpAmp, StringRef(TokStart, 2)); return AsmToken(AsmToken::Amp, StringRef(TokStart, 1)); - case '!': + case '!': if (*CurPtr == '=') return ++CurPtr, AsmToken(AsmToken::ExclaimEqual, StringRef(TokStart, 2)); return AsmToken(AsmToken::Exclaim, StringRef(TokStart, 1)); case '%': return AsmToken(AsmToken::Percent, StringRef(TokStart, 1)); case '/': return LexSlash(); case '#': return AsmToken(AsmToken::Hash, StringRef(TokStart, 1)); + case '\'': return LexSingleQuote(); case '"': return LexQuote(); case '0': case '1': case '2': case '3': case '4': case '5': case '6': case '7': case '8': case '9': return LexDigit(); case '<': switch (*CurPtr) { - case '<': return ++CurPtr, AsmToken(AsmToken::LessLess, + case '<': return ++CurPtr, AsmToken(AsmToken::LessLess, StringRef(TokStart, 2)); - case '=': return ++CurPtr, AsmToken(AsmToken::LessEqual, + case '=': return ++CurPtr, AsmToken(AsmToken::LessEqual, StringRef(TokStart, 2)); - case '>': return ++CurPtr, AsmToken(AsmToken::LessGreater, + case '>': return ++CurPtr, AsmToken(AsmToken::LessGreater, StringRef(TokStart, 2)); default: return AsmToken(AsmToken::Less, StringRef(TokStart, 1)); } case '>': switch (*CurPtr) { - case '>': return ++CurPtr, AsmToken(AsmToken::GreaterGreater, + case '>': return ++CurPtr, AsmToken(AsmToken::GreaterGreater, StringRef(TokStart, 2)); - case '=': return ++CurPtr, AsmToken(AsmToken::GreaterEqual, + case '=': return ++CurPtr, AsmToken(AsmToken::GreaterEqual, StringRef(TokStart, 2)); default: return AsmToken(AsmToken::Greater, StringRef(TokStart, 1)); } - + // TODO: Quoted identifiers (objc methods etc) // local labels: [0-9][:] // Forward/backward labels: [0-9][fb] diff --git a/lib/MC/MCParser/AsmParser.cpp b/lib/MC/MCParser/AsmParser.cpp index fa7a785..a84917f 100644 --- a/lib/MC/MCParser/AsmParser.cpp +++ b/lib/MC/MCParser/AsmParser.cpp @@ -30,6 +30,7 @@ #include "llvm/Support/MemoryBuffer.h" #include "llvm/Support/SourceMgr.h" #include "llvm/Support/raw_ostream.h" +#include "llvm/Target/TargetAsmInfo.h" #include "llvm/Target/TargetAsmParser.h" #include <cctype> #include <vector> @@ -167,11 +168,12 @@ private: /// will be either the EndOfStatement or EOF. StringRef ParseStringToEndOfStatement(); - bool ParseAssignment(StringRef Name); + bool ParseAssignment(StringRef Name, bool allow_redef); bool ParsePrimaryExpr(const MCExpr *&Res, SMLoc &EndLoc); bool ParseBinOpRHS(unsigned Precedence, const MCExpr *&Res, SMLoc &EndLoc); bool ParseParenExpr(const MCExpr *&Res, SMLoc &EndLoc); + bool ParseBracketExpr(const MCExpr *&Res, SMLoc &EndLoc); /// ParseIdentifier - Parse an identifier or string (as a quoted identifier) /// and set \arg Res to the identifier contents. @@ -186,7 +188,7 @@ private: bool ParseDirectiveFill(); // ".fill" bool ParseDirectiveSpace(); // ".space" bool ParseDirectiveZero(); // ".zero" - bool ParseDirectiveSet(StringRef IDVal); // ".set" or ".equ" + bool ParseDirectiveSet(StringRef IDVal, bool allow_redef); // ".set", ".equ", ".equiv" bool ParseDirectiveOrg(); // ".org" // ".align{,32}", ".p2align{,w,l}" bool ParseDirectiveAlign(bool IsPow2, unsigned ValueSize); @@ -201,6 +203,8 @@ private: bool ParseDirectiveInclude(); // ".include" bool ParseDirectiveIf(SMLoc DirectiveLoc); // ".if" + // ".ifdef" or ".ifndef", depending on expect_defined + bool ParseDirectiveIfdef(SMLoc DirectiveLoc, bool expect_defined); bool ParseDirectiveElseIf(SMLoc DirectiveLoc); // ".elseif" bool ParseDirectiveElse(SMLoc DirectiveLoc); // ".else" bool ParseDirectiveEndIf(SMLoc DirectiveLoc); // .endif @@ -221,7 +225,6 @@ class GenericAsmParser : public MCAsmParserExtension { getParser().AddDirectiveHandler(this, Directive, HandleDirective<GenericAsmParser, Handler>); } - public: GenericAsmParser() {} @@ -239,6 +242,28 @@ public: AddDirectiveHandler<&GenericAsmParser::ParseDirectiveLoc>(".loc"); AddDirectiveHandler<&GenericAsmParser::ParseDirectiveStabs>(".stabs"); + // CFI directives. + AddDirectiveHandler<&GenericAsmParser::ParseDirectiveCFIStartProc>( + ".cfi_startproc"); + AddDirectiveHandler<&GenericAsmParser::ParseDirectiveCFIEndProc>( + ".cfi_endproc"); + AddDirectiveHandler<&GenericAsmParser::ParseDirectiveCFIDefCfa>( + ".cfi_def_cfa"); + AddDirectiveHandler<&GenericAsmParser::ParseDirectiveCFIDefCfaOffset>( + ".cfi_def_cfa_offset"); + AddDirectiveHandler<&GenericAsmParser::ParseDirectiveCFIDefCfaRegister>( + ".cfi_def_cfa_register"); + AddDirectiveHandler<&GenericAsmParser::ParseDirectiveCFIOffset>( + ".cfi_offset"); + AddDirectiveHandler< + &GenericAsmParser::ParseDirectiveCFIPersonalityOrLsda>(".cfi_personality"); + AddDirectiveHandler< + &GenericAsmParser::ParseDirectiveCFIPersonalityOrLsda>(".cfi_lsda"); + AddDirectiveHandler< + &GenericAsmParser::ParseDirectiveCFIRememberState>(".cfi_remember_state"); + AddDirectiveHandler< + &GenericAsmParser::ParseDirectiveCFIRestoreState>(".cfi_restore_state"); + // Macro directives. AddDirectiveHandler<&GenericAsmParser::ParseDirectiveMacrosOnOff>( ".macros_on"); @@ -252,10 +277,21 @@ public: AddDirectiveHandler<&GenericAsmParser::ParseDirectiveLEB128>(".uleb128"); } + bool ParseRegisterOrRegisterNumber(int64_t &Register, SMLoc DirectiveLoc); + bool ParseDirectiveFile(StringRef, SMLoc DirectiveLoc); bool ParseDirectiveLine(StringRef, SMLoc DirectiveLoc); bool ParseDirectiveLoc(StringRef, SMLoc DirectiveLoc); bool ParseDirectiveStabs(StringRef, SMLoc DirectiveLoc); + bool ParseDirectiveCFIStartProc(StringRef, SMLoc DirectiveLoc); + bool ParseDirectiveCFIEndProc(StringRef, SMLoc DirectiveLoc); + bool ParseDirectiveCFIDefCfa(StringRef, SMLoc DirectiveLoc); + bool ParseDirectiveCFIDefCfaOffset(StringRef, SMLoc DirectiveLoc); + bool ParseDirectiveCFIDefCfaRegister(StringRef, SMLoc DirectiveLoc); + bool ParseDirectiveCFIOffset(StringRef, SMLoc DirectiveLoc); + bool ParseDirectiveCFIPersonalityOrLsda(StringRef, SMLoc DirectiveLoc); + bool ParseDirectiveCFIRememberState(StringRef, SMLoc DirectiveLoc); + bool ParseDirectiveCFIRestoreState(StringRef, SMLoc DirectiveLoc); bool ParseDirectiveMacrosOnOff(StringRef, SMLoc DirectiveLoc); bool ParseDirectiveMacro(StringRef, SMLoc DirectiveLoc); @@ -457,6 +493,20 @@ bool AsmParser::ParseParenExpr(const MCExpr *&Res, SMLoc &EndLoc) { return false; } +/// ParseBracketExpr - Parse a bracket expression and return it. +/// NOTE: This assumes the leading '[' has already been consumed. +/// +/// bracketexpr ::= expr] +/// +bool AsmParser::ParseBracketExpr(const MCExpr *&Res, SMLoc &EndLoc) { + if (ParseExpression(Res)) return true; + if (Lexer.isNot(AsmToken::RBrac)) + return TokError("expected ']' in brackets expression"); + EndLoc = Lexer.getLoc(); + Lex(); + return false; +} + /// ParsePrimaryExpr - Parse a primary expression and return it. /// primaryexpr ::= (parenexpr /// primaryexpr ::= symbol @@ -492,7 +542,7 @@ bool AsmParser::ParsePrimaryExpr(const MCExpr *&Res, SMLoc &EndLoc) { Variant = MCSymbolRefExpr::getVariantKindForName(Split.second); if (Variant == MCSymbolRefExpr::VK_Invalid) { Variant = MCSymbolRefExpr::VK_None; - TokError("invalid variant '" + Split.second + "'"); + return TokError("invalid variant '" + Split.second + "'"); } } @@ -532,6 +582,13 @@ bool AsmParser::ParsePrimaryExpr(const MCExpr *&Res, SMLoc &EndLoc) { } return false; } + case AsmToken::Real: { + APFloat RealVal(APFloat::IEEEdouble, getTok().getString()); + uint64_t IntVal = RealVal.bitcastToAPInt().getZExtValue(); + Res = MCConstantExpr::Create(IntVal, getContext()); + Lex(); // Eat token. + return false; + } case AsmToken::Dot: { // This is a '.' reference, which references the current PC. Emit a // temporary label to the streamer and refer to it. @@ -542,10 +599,14 @@ bool AsmParser::ParsePrimaryExpr(const MCExpr *&Res, SMLoc &EndLoc) { Lex(); // Eat identifier. return false; } - case AsmToken::LParen: Lex(); // Eat the '('. return ParseParenExpr(Res, EndLoc); + case AsmToken::LBrac: + if (!PlatformParser->HasBracketExpressions()) + return TokError("brackets expression not supported on this target"); + Lex(); // Eat the '['. + return ParseBracketExpr(Res, EndLoc); case AsmToken::Minus: Lex(); // Eat the operator. if (ParsePrimaryExpr(Res, EndLoc)) @@ -809,12 +870,17 @@ bool AsmParser::ParseStatement() { return false; } - // Statements always start with an identifier. + // Statements always start with an identifier or are a full line comment. AsmToken ID = getTok(); SMLoc IDLoc = ID.getLoc(); StringRef IDVal; int64_t LocalLabelVal = -1; - // GUESS allow an integer followed by a ':' as a directional local label + // A full line comment is a '#' as the first token. + if (Lexer.is(AsmToken::Hash)) { + EatToEndOfStatement(); + return false; + } + // Allow an integer followed by a ':' as a directional local label. if (Lexer.is(AsmToken::Integer)) { LocalLabelVal = getTok().getIntVal(); if (LocalLabelVal < 0) { @@ -842,6 +908,10 @@ bool AsmParser::ParseStatement() { // example. if (IDVal == ".if") return ParseDirectiveIf(IDLoc); + if (IDVal == ".ifdef") + return ParseDirectiveIfdef(IDLoc, true); + if (IDVal == ".ifndef" || IDVal == ".ifnotdef") + return ParseDirectiveIfdef(IDLoc, false); if (IDVal == ".elseif") return ParseDirectiveElseIf(IDLoc); if (IDVal == ".else") @@ -896,7 +966,7 @@ bool AsmParser::ParseStatement() { // identifier '=' ... -> assignment statement Lex(); - return ParseAssignment(IDVal); + return ParseAssignment(IDVal, true); default: // Normal instruction or directive. break; @@ -911,7 +981,9 @@ bool AsmParser::ParseStatement() { if (IDVal[0] == '.') { // Assembler features if (IDVal == ".set" || IDVal == ".equ") - return ParseDirectiveSet(IDVal); + return ParseDirectiveSet(IDVal, true); + if (IDVal == ".equiv") + return ParseDirectiveSet(IDVal, false); // Data directives @@ -926,11 +998,19 @@ bool AsmParser::ParseStatement() { return ParseDirectiveValue(2); if (IDVal == ".value") return ParseDirectiveValue(2); + if (IDVal == ".2byte") + return ParseDirectiveValue(2); if (IDVal == ".long") return ParseDirectiveValue(4); + if (IDVal == ".int") + return ParseDirectiveValue(4); + if (IDVal == ".4byte") + return ParseDirectiveValue(4); if (IDVal == ".quad") return ParseDirectiveValue(8); - if (IDVal == ".single") + if (IDVal == ".8byte") + return ParseDirectiveValue(8); + if (IDVal == ".single" || IDVal == ".float") return ParseDirectiveRealValue(APFloat::IEEEsingle); if (IDVal == ".double") return ParseDirectiveRealValue(APFloat::IEEEdouble); @@ -983,6 +1063,8 @@ bool AsmParser::ParseStatement() { return ParseDirectiveSymbolAttribute(MCSA_LazyReference); if (IDVal == ".no_dead_strip") return ParseDirectiveSymbolAttribute(MCSA_NoDeadStrip); + if (IDVal == ".symbol_resolver") + return ParseDirectiveSymbolAttribute(MCSA_SymbolResolver); if (IDVal == ".private_extern") return ParseDirectiveSymbolAttribute(MCSA_PrivateExtern); if (IDVal == ".protected") @@ -1008,6 +1090,9 @@ bool AsmParser::ParseStatement() { if (IDVal == ".include") return ParseDirectiveInclude(); + if (IDVal == ".code16" || IDVal == ".code32" || IDVal == ".code64") + return TokError(Twine(IDVal) + " not supported yet"); + // Look up the handler in the handler table. std::pair<MCAsmParserExtension*, DirectiveHandler> Handler = DirectiveMap.lookup(IDVal); @@ -1208,7 +1293,7 @@ static void MarkUsed(const MCExpr *Value) { } } -bool AsmParser::ParseAssignment(StringRef Name) { +bool AsmParser::ParseAssignment(StringRef Name, bool allow_redef) { // FIXME: Use better location, we should use proper tokens. SMLoc EqualLoc = Lexer.getLoc(); @@ -1234,7 +1319,7 @@ bool AsmParser::ParseAssignment(StringRef Name) { // FIXME: Diagnose assignment to protected identifier (e.g., register name). if (Sym->isUndefined() && !Sym->isUsed() && !Sym->isVariable()) ; // Allow redefinitions of undefined symbols only used in directives. - else if (!Sym->isUndefined() && !Sym->isAbsolute()) + else if (!Sym->isUndefined() && (!Sym->isAbsolute() || !allow_redef)) return Error(EqualLoc, "redefinition of '" + Name + "'"); else if (!Sym->isVariable()) return Error(EqualLoc, "invalid assignment to '" + Name + "'"); @@ -1295,8 +1380,10 @@ bool AsmParser::ParseIdentifier(StringRef &Res) { } /// ParseDirectiveSet: +/// ::= .equ identifier ',' expression +/// ::= .equiv identifier ',' expression /// ::= .set identifier ',' expression -bool AsmParser::ParseDirectiveSet(StringRef IDVal) { +bool AsmParser::ParseDirectiveSet(StringRef IDVal, bool allow_redef) { StringRef Name; if (ParseIdentifier(Name)) @@ -1306,7 +1393,7 @@ bool AsmParser::ParseDirectiveSet(StringRef IDVal) { return TokError("unexpected token in '" + Twine(IDVal) + "'"); Lex(); - return ParseAssignment(Name); + return ParseAssignment(Name, allow_redef); } bool AsmParser::ParseEscapedString(std::string &Data) { @@ -1871,6 +1958,31 @@ bool AsmParser::ParseDirectiveIf(SMLoc DirectiveLoc) { return false; } +bool AsmParser::ParseDirectiveIfdef(SMLoc DirectiveLoc, bool expect_defined) { + StringRef Name; + TheCondStack.push_back(TheCondState); + TheCondState.TheCond = AsmCond::IfCond; + + if (TheCondState.Ignore) { + EatToEndOfStatement(); + } else { + if (ParseIdentifier(Name)) + return TokError("expected identifier after '.ifdef'"); + + Lex(); + + MCSymbol *Sym = getContext().LookupSymbol(Name); + + if (expect_defined) + TheCondState.CondMet = (Sym != NULL && !Sym->isUndefined()); + else + TheCondState.CondMet = (Sym == NULL || Sym->isUndefined()); + TheCondState.Ignore = !TheCondState.CondMet; + } + + return false; +} + /// ParseDirectiveElseIf /// ::= .elseif expression bool AsmParser::ParseDirectiveElseIf(SMLoc DirectiveLoc) { @@ -1974,9 +2086,8 @@ bool GenericAsmParser::ParseDirectiveFile(StringRef, SMLoc DirectiveLoc) { if (FileNumber == -1) getStreamer().EmitFileDirective(Filename); else { - if (getContext().GetDwarfFile(Filename, FileNumber) == 0) + if (getStreamer().EmitDwarfFileDirective(FileNumber, Filename)) Error(FileNumberLoc, "file number already allocated"); - getStreamer().EmitDwarfFileDirective(FileNumber, Filename); } return false; @@ -2104,8 +2215,8 @@ bool GenericAsmParser::ParseDirectiveLoc(StringRef, SMLoc DirectiveLoc) { } } - getContext().setCurrentDwarfLoc(FileNumber, LineNumber, ColumnPos, Flags, - Isa, Discriminator); + getStreamer().EmitDwarfLocDirective(FileNumber, LineNumber, ColumnPos, Flags, + Isa, Discriminator); return false; } @@ -2117,6 +2228,163 @@ bool GenericAsmParser::ParseDirectiveStabs(StringRef Directive, return TokError("unsupported directive '" + Directive + "'"); } +/// ParseDirectiveCFIStartProc +/// ::= .cfi_startproc +bool GenericAsmParser::ParseDirectiveCFIStartProc(StringRef, + SMLoc DirectiveLoc) { + return getStreamer().EmitCFIStartProc(); +} + +/// ParseDirectiveCFIEndProc +/// ::= .cfi_endproc +bool GenericAsmParser::ParseDirectiveCFIEndProc(StringRef, SMLoc DirectiveLoc) { + return getStreamer().EmitCFIEndProc(); +} + +/// ParseRegisterOrRegisterNumber - parse register name or number. +bool GenericAsmParser::ParseRegisterOrRegisterNumber(int64_t &Register, + SMLoc DirectiveLoc) { + unsigned RegNo; + + if (getLexer().is(AsmToken::Percent)) { + if (getParser().getTargetParser().ParseRegister(RegNo, DirectiveLoc, + DirectiveLoc)) + return true; + Register = getContext().getTargetAsmInfo().getDwarfRegNum(RegNo, true); + } else + return getParser().ParseAbsoluteExpression(Register); + + return false; +} + +/// ParseDirectiveCFIDefCfa +/// ::= .cfi_def_cfa register, offset +bool GenericAsmParser::ParseDirectiveCFIDefCfa(StringRef, + SMLoc DirectiveLoc) { + int64_t Register = 0; + if (ParseRegisterOrRegisterNumber(Register, DirectiveLoc)) + return true; + + if (getLexer().isNot(AsmToken::Comma)) + return TokError("unexpected token in directive"); + Lex(); + + int64_t Offset = 0; + if (getParser().ParseAbsoluteExpression(Offset)) + return true; + + return getStreamer().EmitCFIDefCfa(Register, Offset); +} + +/// ParseDirectiveCFIDefCfaOffset +/// ::= .cfi_def_cfa_offset offset +bool GenericAsmParser::ParseDirectiveCFIDefCfaOffset(StringRef, + SMLoc DirectiveLoc) { + int64_t Offset = 0; + if (getParser().ParseAbsoluteExpression(Offset)) + return true; + + return getStreamer().EmitCFIDefCfaOffset(Offset); +} + +/// ParseDirectiveCFIDefCfaRegister +/// ::= .cfi_def_cfa_register register +bool GenericAsmParser::ParseDirectiveCFIDefCfaRegister(StringRef, + SMLoc DirectiveLoc) { + int64_t Register = 0; + if (ParseRegisterOrRegisterNumber(Register, DirectiveLoc)) + return true; + + return getStreamer().EmitCFIDefCfaRegister(Register); +} + +/// ParseDirectiveCFIOffset +/// ::= .cfi_off register, offset +bool GenericAsmParser::ParseDirectiveCFIOffset(StringRef, SMLoc DirectiveLoc) { + int64_t Register = 0; + int64_t Offset = 0; + + if (ParseRegisterOrRegisterNumber(Register, DirectiveLoc)) + return true; + + if (getLexer().isNot(AsmToken::Comma)) + return TokError("unexpected token in directive"); + Lex(); + + if (getParser().ParseAbsoluteExpression(Offset)) + return true; + + return getStreamer().EmitCFIOffset(Register, Offset); +} + +static bool isValidEncoding(int64_t Encoding) { + if (Encoding & ~0xff) + return false; + + if (Encoding == dwarf::DW_EH_PE_omit) + return true; + + const unsigned Format = Encoding & 0xf; + if (Format != dwarf::DW_EH_PE_absptr && Format != dwarf::DW_EH_PE_udata2 && + Format != dwarf::DW_EH_PE_udata4 && Format != dwarf::DW_EH_PE_udata8 && + Format != dwarf::DW_EH_PE_sdata2 && Format != dwarf::DW_EH_PE_sdata4 && + Format != dwarf::DW_EH_PE_sdata8 && Format != dwarf::DW_EH_PE_signed) + return false; + + const unsigned Application = Encoding & 0x70; + if (Application != dwarf::DW_EH_PE_absptr && + Application != dwarf::DW_EH_PE_pcrel) + return false; + + return true; +} + +/// ParseDirectiveCFIPersonalityOrLsda +/// ::= .cfi_personality encoding, [symbol_name] +/// ::= .cfi_lsda encoding, [symbol_name] +bool GenericAsmParser::ParseDirectiveCFIPersonalityOrLsda(StringRef IDVal, + SMLoc DirectiveLoc) { + int64_t Encoding = 0; + if (getParser().ParseAbsoluteExpression(Encoding)) + return true; + if (Encoding == dwarf::DW_EH_PE_omit) + return false; + + if (!isValidEncoding(Encoding)) + return TokError("unsupported encoding."); + + if (getLexer().isNot(AsmToken::Comma)) + return TokError("unexpected token in directive"); + Lex(); + + StringRef Name; + if (getParser().ParseIdentifier(Name)) + return TokError("expected identifier in directive"); + + MCSymbol *Sym = getContext().GetOrCreateSymbol(Name); + + if (IDVal == ".cfi_personality") + return getStreamer().EmitCFIPersonality(Sym, Encoding); + else { + assert(IDVal == ".cfi_lsda"); + return getStreamer().EmitCFILsda(Sym, Encoding); + } +} + +/// ParseDirectiveCFIRememberState +/// ::= .cfi_remember_state +bool GenericAsmParser::ParseDirectiveCFIRememberState(StringRef IDVal, + SMLoc DirectiveLoc) { + return getStreamer().EmitCFIRememberState(); +} + +/// ParseDirectiveCFIRestoreState +/// ::= .cfi_remember_state +bool GenericAsmParser::ParseDirectiveCFIRestoreState(StringRef IDVal, + SMLoc DirectiveLoc) { + return getStreamer().EmitCFIRestoreState(); +} + /// ParseDirectiveMacrosOnOff /// ::= .macros_on /// ::= .macros_off diff --git a/lib/MC/MCParser/ELFAsmParser.cpp b/lib/MC/MCParser/ELFAsmParser.cpp index d074ea9..dcf689a 100644 --- a/lib/MC/MCParser/ELFAsmParser.cpp +++ b/lib/MC/MCParser/ELFAsmParser.cpp @@ -16,6 +16,7 @@ #include "llvm/MC/MCParser/MCAsmLexer.h" #include "llvm/MC/MCSectionELF.h" #include "llvm/MC/MCStreamer.h" +#include "llvm/Support/ELF.h" using namespace llvm; namespace { @@ -29,9 +30,12 @@ class ELFAsmParser : public MCAsmParserExtension { bool ParseSectionSwitch(StringRef Section, unsigned Type, unsigned Flags, SectionKind Kind); + bool SeenIdent; public: - ELFAsmParser() {} + ELFAsmParser() : SeenIdent(false) { + BracketExpressionsSupported = true; + } virtual void Initialize(MCAsmParser &Parser) { // Call the base implementation. @@ -48,6 +52,8 @@ public: AddDirectiveHandler<&ELFAsmParser::ParseSectionDirectiveDataRelRoLocal>(".data.rel.ro.local"); AddDirectiveHandler<&ELFAsmParser::ParseSectionDirectiveEhFrame>(".eh_frame"); AddDirectiveHandler<&ELFAsmParser::ParseDirectiveSection>(".section"); + AddDirectiveHandler<&ELFAsmParser::ParseDirectivePushSection>(".pushsection"); + AddDirectiveHandler<&ELFAsmParser::ParseDirectivePopSection>(".popsection"); AddDirectiveHandler<&ELFAsmParser::ParseDirectiveSize>(".size"); AddDirectiveHandler<&ELFAsmParser::ParseDirectivePrevious>(".previous"); AddDirectiveHandler<&ELFAsmParser::ParseDirectiveType>(".type"); @@ -59,61 +65,63 @@ public: // FIXME: Part of this logic is duplicated in the MCELFStreamer. What is // the best way for us to get access to it? bool ParseSectionDirectiveData(StringRef, SMLoc) { - return ParseSectionSwitch(".data", MCSectionELF::SHT_PROGBITS, - MCSectionELF::SHF_WRITE |MCSectionELF::SHF_ALLOC, + return ParseSectionSwitch(".data", ELF::SHT_PROGBITS, + ELF::SHF_WRITE |ELF::SHF_ALLOC, SectionKind::getDataRel()); } bool ParseSectionDirectiveText(StringRef, SMLoc) { - return ParseSectionSwitch(".text", MCSectionELF::SHT_PROGBITS, - MCSectionELF::SHF_EXECINSTR | - MCSectionELF::SHF_ALLOC, SectionKind::getText()); + return ParseSectionSwitch(".text", ELF::SHT_PROGBITS, + ELF::SHF_EXECINSTR | + ELF::SHF_ALLOC, SectionKind::getText()); } bool ParseSectionDirectiveBSS(StringRef, SMLoc) { - return ParseSectionSwitch(".bss", MCSectionELF::SHT_NOBITS, - MCSectionELF::SHF_WRITE | - MCSectionELF::SHF_ALLOC, SectionKind::getBSS()); + return ParseSectionSwitch(".bss", ELF::SHT_NOBITS, + ELF::SHF_WRITE | + ELF::SHF_ALLOC, SectionKind::getBSS()); } bool ParseSectionDirectiveRoData(StringRef, SMLoc) { - return ParseSectionSwitch(".rodata", MCSectionELF::SHT_PROGBITS, - MCSectionELF::SHF_ALLOC, + return ParseSectionSwitch(".rodata", ELF::SHT_PROGBITS, + ELF::SHF_ALLOC, SectionKind::getReadOnly()); } bool ParseSectionDirectiveTData(StringRef, SMLoc) { - return ParseSectionSwitch(".tdata", MCSectionELF::SHT_PROGBITS, - MCSectionELF::SHF_ALLOC | - MCSectionELF::SHF_TLS | MCSectionELF::SHF_WRITE, + return ParseSectionSwitch(".tdata", ELF::SHT_PROGBITS, + ELF::SHF_ALLOC | + ELF::SHF_TLS | ELF::SHF_WRITE, SectionKind::getThreadData()); } bool ParseSectionDirectiveTBSS(StringRef, SMLoc) { - return ParseSectionSwitch(".tbss", MCSectionELF::SHT_NOBITS, - MCSectionELF::SHF_ALLOC | - MCSectionELF::SHF_TLS | MCSectionELF::SHF_WRITE, + return ParseSectionSwitch(".tbss", ELF::SHT_NOBITS, + ELF::SHF_ALLOC | + ELF::SHF_TLS | ELF::SHF_WRITE, SectionKind::getThreadBSS()); } bool ParseSectionDirectiveDataRel(StringRef, SMLoc) { - return ParseSectionSwitch(".data.rel", MCSectionELF::SHT_PROGBITS, - MCSectionELF::SHF_ALLOC | - MCSectionELF::SHF_WRITE, + return ParseSectionSwitch(".data.rel", ELF::SHT_PROGBITS, + ELF::SHF_ALLOC | + ELF::SHF_WRITE, SectionKind::getDataRel()); } bool ParseSectionDirectiveDataRelRo(StringRef, SMLoc) { - return ParseSectionSwitch(".data.rel.ro", MCSectionELF::SHT_PROGBITS, - MCSectionELF::SHF_ALLOC | - MCSectionELF::SHF_WRITE, + return ParseSectionSwitch(".data.rel.ro", ELF::SHT_PROGBITS, + ELF::SHF_ALLOC | + ELF::SHF_WRITE, SectionKind::getReadOnlyWithRel()); } bool ParseSectionDirectiveDataRelRoLocal(StringRef, SMLoc) { - return ParseSectionSwitch(".data.rel.ro.local", MCSectionELF::SHT_PROGBITS, - MCSectionELF::SHF_ALLOC | - MCSectionELF::SHF_WRITE, + return ParseSectionSwitch(".data.rel.ro.local", ELF::SHT_PROGBITS, + ELF::SHF_ALLOC | + ELF::SHF_WRITE, SectionKind::getReadOnlyWithRelLocal()); } bool ParseSectionDirectiveEhFrame(StringRef, SMLoc) { - return ParseSectionSwitch(".eh_frame", MCSectionELF::SHT_PROGBITS, - MCSectionELF::SHF_ALLOC | - MCSectionELF::SHF_WRITE, + return ParseSectionSwitch(".eh_frame", ELF::SHT_PROGBITS, + ELF::SHF_ALLOC | + ELF::SHF_WRITE, SectionKind::getDataRel()); } + bool ParseDirectivePushSection(StringRef, SMLoc); + bool ParseDirectivePopSection(StringRef, SMLoc); bool ParseDirectiveSection(StringRef, SMLoc); bool ParseDirectiveSize(StringRef, SMLoc); bool ParseDirectivePrevious(StringRef, SMLoc); @@ -167,6 +175,12 @@ bool ELFAsmParser::ParseSectionName(StringRef &SectionName) { SMLoc FirstLoc = getLexer().getLoc(); unsigned Size = 0; + if (getLexer().is(AsmToken::String)) { + SectionName = getTok().getIdentifier(); + Lex(); + return false; + } + for (;;) { StringRef Tmp; unsigned CurSize; @@ -175,10 +189,15 @@ bool ELFAsmParser::ParseSectionName(StringRef &SectionName) { if (getLexer().is(AsmToken::Minus)) { CurSize = 1; Lex(); // Consume the "-". - } else if (!getParser().ParseIdentifier(Tmp)) - CurSize = Tmp.size(); - else + } else if (getLexer().is(AsmToken::String)) { + CurSize = getTok().getIdentifier().size() + 2; + Lex(); + } else if (getLexer().is(AsmToken::Identifier)) { + CurSize = getTok().getIdentifier().size(); + Lex(); + } else { break; + } Size += CurSize; SectionName = StringRef(FirstLoc.getPointer(), Size); @@ -193,6 +212,71 @@ bool ELFAsmParser::ParseSectionName(StringRef &SectionName) { return false; } +static SectionKind computeSectionKind(unsigned Flags) { + if (Flags & ELF::SHF_EXECINSTR) + return SectionKind::getText(); + if (Flags & ELF::SHF_TLS) + return SectionKind::getThreadData(); + return SectionKind::getDataRel(); +} + +static int parseSectionFlags(StringRef flagsStr) { + int flags = 0; + + for (unsigned i = 0; i < flagsStr.size(); i++) { + switch (flagsStr[i]) { + case 'a': + flags |= ELF::SHF_ALLOC; + break; + case 'x': + flags |= ELF::SHF_EXECINSTR; + break; + case 'w': + flags |= ELF::SHF_WRITE; + break; + case 'M': + flags |= ELF::SHF_MERGE; + break; + case 'S': + flags |= ELF::SHF_STRINGS; + break; + case 'T': + flags |= ELF::SHF_TLS; + break; + case 'c': + flags |= ELF::XCORE_SHF_CP_SECTION; + break; + case 'd': + flags |= ELF::XCORE_SHF_DP_SECTION; + break; + case 'G': + flags |= ELF::SHF_GROUP; + break; + default: + return -1; + } + } + + return flags; +} + +bool ELFAsmParser::ParseDirectivePushSection(StringRef s, SMLoc loc) { + getStreamer().PushSection(); + + if (ParseDirectiveSection(s, loc)) { + getStreamer().PopSection(); + return true; + } + + return false; +} + +bool ELFAsmParser::ParseDirectivePopSection(StringRef, SMLoc) { + if (!getStreamer().PopSection()) + return TokError(".popsection without corresponding .pushsection"); + return false; +} + // FIXME: This is a work in progress. bool ELFAsmParser::ParseDirectiveSection(StringRef, SMLoc) { StringRef SectionName; @@ -200,21 +284,34 @@ bool ELFAsmParser::ParseDirectiveSection(StringRef, SMLoc) { if (ParseSectionName(SectionName)) return TokError("expected identifier in directive"); - StringRef FlagsStr; StringRef TypeName; int64_t Size = 0; StringRef GroupName; + unsigned Flags = 0; + + // Set the defaults first. + if (SectionName == ".fini" || SectionName == ".init" || + SectionName == ".rodata") + Flags |= ELF::SHF_ALLOC; + if (SectionName == ".fini" || SectionName == ".init") + Flags |= ELF::SHF_EXECINSTR; + if (getLexer().is(AsmToken::Comma)) { Lex(); if (getLexer().isNot(AsmToken::String)) return TokError("expected string in directive"); - FlagsStr = getTok().getStringContents(); + StringRef FlagsStr = getTok().getStringContents(); Lex(); - bool Mergeable = FlagsStr.find('M') != StringRef::npos; - bool Group = FlagsStr.find('G') != StringRef::npos; + int extraFlags = parseSectionFlags(FlagsStr); + if (extraFlags < 0) + return TokError("unknown flag"); + Flags |= extraFlags; + + bool Mergeable = Flags & ELF::SHF_MERGE; + bool Group = Flags & ELF::SHF_GROUP; if (getLexer().isNot(AsmToken::Comma)) { if (Mergeable) @@ -261,70 +358,28 @@ bool ELFAsmParser::ParseDirectiveSection(StringRef, SMLoc) { if (getLexer().isNot(AsmToken::EndOfStatement)) return TokError("unexpected token in directive"); - unsigned Flags = 0; - unsigned Type = MCSectionELF::SHT_NULL; - - // Set the defaults first. - if (SectionName == ".fini" || SectionName == ".init" || SectionName == ".rodata") { - Type = MCSectionELF::SHT_PROGBITS; - Flags |= MCSectionELF::SHF_ALLOC; - } - if (SectionName == ".fini" || SectionName == ".init") { - Flags |= MCSectionELF::SHF_EXECINSTR; - } - - for (unsigned i = 0; i < FlagsStr.size(); i++) { - switch (FlagsStr[i]) { - case 'a': - Flags |= MCSectionELF::SHF_ALLOC; - break; - case 'x': - Flags |= MCSectionELF::SHF_EXECINSTR; - break; - case 'w': - Flags |= MCSectionELF::SHF_WRITE; - break; - case 'M': - Flags |= MCSectionELF::SHF_MERGE; - break; - case 'S': - Flags |= MCSectionELF::SHF_STRINGS; - break; - case 'T': - Flags |= MCSectionELF::SHF_TLS; - break; - case 'c': - Flags |= MCSectionELF::XCORE_SHF_CP_SECTION; - break; - case 'd': - Flags |= MCSectionELF::XCORE_SHF_DP_SECTION; - break; - case 'G': - Flags |= MCSectionELF::SHF_GROUP; - break; - default: - return TokError("unknown flag"); - } - } + unsigned Type = ELF::SHT_PROGBITS; if (!TypeName.empty()) { if (TypeName == "init_array") - Type = MCSectionELF::SHT_INIT_ARRAY; + Type = ELF::SHT_INIT_ARRAY; else if (TypeName == "fini_array") - Type = MCSectionELF::SHT_FINI_ARRAY; + Type = ELF::SHT_FINI_ARRAY; else if (TypeName == "preinit_array") - Type = MCSectionELF::SHT_PREINIT_ARRAY; + Type = ELF::SHT_PREINIT_ARRAY; else if (TypeName == "nobits") - Type = MCSectionELF::SHT_NOBITS; + Type = ELF::SHT_NOBITS; else if (TypeName == "progbits") - Type = MCSectionELF::SHT_PROGBITS; + Type = ELF::SHT_PROGBITS; + else if (TypeName == "note") + Type = ELF::SHT_NOTE; + else if (TypeName == "unwind") + Type = ELF::SHT_X86_64_UNWIND; else return TokError("unknown section type"); } - SectionKind Kind = (Flags & MCSectionELF::SHF_EXECINSTR) - ? SectionKind::getText() - : SectionKind::getDataRel(); + SectionKind Kind = computeSectionKind(Flags); getStreamer().SwitchSection(getContext().getELFSection(SectionName, Type, Flags, Kind, Size, GroupName)); @@ -333,8 +388,9 @@ bool ELFAsmParser::ParseDirectiveSection(StringRef, SMLoc) { bool ELFAsmParser::ParseDirectivePrevious(StringRef DirName, SMLoc) { const MCSection *PreviousSection = getStreamer().getPreviousSection(); - if (PreviousSection != NULL) - getStreamer().SwitchSection(PreviousSection); + if (PreviousSection == NULL) + return TokError(".previous without corresponding .section"); + getStreamer().SwitchSection(PreviousSection); return false; } @@ -396,23 +452,22 @@ bool ELFAsmParser::ParseDirectiveIdent(StringRef, SMLoc) { Lex(); - const MCSection *OldSection = getStreamer().getCurrentSection(); const MCSection *Comment = - getContext().getELFSection(".comment", MCSectionELF::SHT_PROGBITS, - MCSectionELF::SHF_MERGE | - MCSectionELF::SHF_STRINGS, + getContext().getELFSection(".comment", ELF::SHT_PROGBITS, + ELF::SHF_MERGE | + ELF::SHF_STRINGS, SectionKind::getReadOnly(), 1, ""); - static bool First = true; - + getStreamer().PushSection(); getStreamer().SwitchSection(Comment); - if (First) + if (!SeenIdent) { getStreamer().EmitIntValue(0, 1); - First = false; + SeenIdent = true; + } getStreamer().EmitBytes(Data, 0); getStreamer().EmitIntValue(0, 1); - getStreamer().SwitchSection(OldSection); + getStreamer().PopSection(); return false; } diff --git a/lib/MC/MCParser/MCAsmParserExtension.cpp b/lib/MC/MCParser/MCAsmParserExtension.cpp index c30d306..3f25a14 100644 --- a/lib/MC/MCParser/MCAsmParserExtension.cpp +++ b/lib/MC/MCParser/MCAsmParserExtension.cpp @@ -10,7 +10,8 @@ #include "llvm/MC/MCParser/MCAsmParserExtension.h" using namespace llvm; -MCAsmParserExtension::MCAsmParserExtension() { +MCAsmParserExtension::MCAsmParserExtension() : + BracketExpressionsSupported(false) { } MCAsmParserExtension::~MCAsmParserExtension() { diff --git a/lib/MC/MCPureStreamer.cpp b/lib/MC/MCPureStreamer.cpp new file mode 100644 index 0000000..6098e6b --- /dev/null +++ b/lib/MC/MCPureStreamer.cpp @@ -0,0 +1,234 @@ +//===- lib/MC/MCPureStreamer.cpp - MC "Pure" Object Output ----------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +#include "llvm/MC/MCStreamer.h" +#include "llvm/MC/MCAssembler.h" +#include "llvm/MC/MCCodeEmitter.h" +#include "llvm/MC/MCContext.h" +#include "llvm/MC/MCExpr.h" +#include "llvm/MC/MCObjectStreamer.h" +// FIXME: Remove this. +#include "llvm/MC/MCSectionMachO.h" +#include "llvm/MC/MCSymbol.h" +#include "llvm/Support/ErrorHandling.h" + +using namespace llvm; + +namespace { + +class MCPureStreamer : public MCObjectStreamer { +private: + virtual void EmitInstToFragment(const MCInst &Inst); + virtual void EmitInstToData(const MCInst &Inst); + +public: + MCPureStreamer(MCContext &Context, TargetAsmBackend &TAB, + raw_ostream &OS, MCCodeEmitter *Emitter) + : MCObjectStreamer(Context, TAB, OS, Emitter) {} + + /// @name MCStreamer Interface + /// @{ + + virtual void InitSections(); + virtual void EmitLabel(MCSymbol *Symbol); + virtual void EmitAssignment(MCSymbol *Symbol, const MCExpr *Value); + virtual void EmitZerofill(const MCSection *Section, MCSymbol *Symbol = 0, + unsigned Size = 0, unsigned ByteAlignment = 0); + virtual void EmitBytes(StringRef Data, unsigned AddrSpace); + virtual void EmitValueToAlignment(unsigned ByteAlignment, int64_t Value = 0, + unsigned ValueSize = 1, + unsigned MaxBytesToEmit = 0); + virtual void EmitCodeAlignment(unsigned ByteAlignment, + unsigned MaxBytesToEmit = 0); + virtual void EmitValueToOffset(const MCExpr *Offset, + unsigned char Value = 0); + virtual void Finish(); + + + virtual void EmitSymbolAttribute(MCSymbol *Symbol, MCSymbolAttr Attribute) { + report_fatal_error("unsupported directive in pure streamer"); + } + virtual void EmitAssemblerFlag(MCAssemblerFlag Flag) { + report_fatal_error("unsupported directive in pure streamer"); + } + virtual void EmitTBSSSymbol(const MCSection *Section, MCSymbol *Symbol, + uint64_t Size, unsigned ByteAlignment = 0) { + report_fatal_error("unsupported directive in pure streamer"); + } + virtual void EmitSymbolDesc(MCSymbol *Symbol, unsigned DescValue) { + report_fatal_error("unsupported directive in pure streamer"); + } + virtual void EmitCommonSymbol(MCSymbol *Symbol, uint64_t Size, + unsigned ByteAlignment) { + report_fatal_error("unsupported directive in pure streamer"); + } + virtual void EmitThumbFunc(MCSymbol *Func) { + report_fatal_error("unsupported directive in pure streamer"); + } + virtual void BeginCOFFSymbolDef(const MCSymbol *Symbol) { + report_fatal_error("unsupported directive in pure streamer"); + } + virtual void EmitCOFFSymbolStorageClass(int StorageClass) { + report_fatal_error("unsupported directive in pure streamer"); + } + virtual void EmitCOFFSymbolType(int Type) { + report_fatal_error("unsupported directive in pure streamer"); + } + virtual void EndCOFFSymbolDef() { + report_fatal_error("unsupported directive in pure streamer"); + } + virtual void EmitELFSize(MCSymbol *Symbol, const MCExpr *Value) { + report_fatal_error("unsupported directive in pure streamer"); + } + virtual void EmitLocalCommonSymbol(MCSymbol *Symbol, uint64_t Size) { + report_fatal_error("unsupported directive in pure streamer"); + } + virtual void EmitFileDirective(StringRef Filename) { + report_fatal_error("unsupported directive in pure streamer"); + } + virtual bool EmitDwarfFileDirective(unsigned FileNo, StringRef Filename) { + report_fatal_error("unsupported directive in pure streamer"); + return false; + } + + /// @} +}; + +} // end anonymous namespace. + +void MCPureStreamer::InitSections() { + // FIMXE: To what!? + SwitchSection(getContext().getMachOSection("__TEXT", "__text", + MCSectionMachO::S_ATTR_PURE_INSTRUCTIONS, + 0, SectionKind::getText())); + +} + +void MCPureStreamer::EmitLabel(MCSymbol *Symbol) { + assert(Symbol->isUndefined() && "Cannot define a symbol twice!"); + assert(!Symbol->isVariable() && "Cannot emit a variable symbol!"); + assert(getCurrentSection() && "Cannot emit before setting section!"); + + Symbol->setSection(*getCurrentSection()); + + MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol); + + // We have to create a new fragment if this is an atom defining symbol, + // fragments cannot span atoms. + if (getAssembler().isSymbolLinkerVisible(SD.getSymbol())) + new MCDataFragment(getCurrentSectionData()); + + // FIXME: This is wasteful, we don't necessarily need to create a data + // fragment. Instead, we should mark the symbol as pointing into the data + // fragment if it exists, otherwise we should just queue the label and set its + // fragment pointer when we emit the next fragment. + MCDataFragment *F = getOrCreateDataFragment(); + assert(!SD.getFragment() && "Unexpected fragment on symbol data!"); + SD.setFragment(F); + SD.setOffset(F->getContents().size()); +} + +void MCPureStreamer::EmitAssignment(MCSymbol *Symbol, const MCExpr *Value) { + // TODO: This is exactly the same as WinCOFFStreamer. Consider merging into + // MCObjectStreamer. + // FIXME: Lift context changes into super class. + getAssembler().getOrCreateSymbolData(*Symbol); + Symbol->setVariableValue(AddValueSymbols(Value)); +} + +void MCPureStreamer::EmitZerofill(const MCSection *Section, MCSymbol *Symbol, + unsigned Size, unsigned ByteAlignment) { + report_fatal_error("not yet implemented in pure streamer"); +} + +void MCPureStreamer::EmitBytes(StringRef Data, unsigned AddrSpace) { + // TODO: This is exactly the same as WinCOFFStreamer. Consider merging into + // MCObjectStreamer. + getOrCreateDataFragment()->getContents().append(Data.begin(), Data.end()); +} + +void MCPureStreamer::EmitValueToAlignment(unsigned ByteAlignment, + int64_t Value, unsigned ValueSize, + unsigned MaxBytesToEmit) { + // TODO: This is exactly the same as WinCOFFStreamer. Consider merging into + // MCObjectStreamer. + if (MaxBytesToEmit == 0) + MaxBytesToEmit = ByteAlignment; + new MCAlignFragment(ByteAlignment, Value, ValueSize, MaxBytesToEmit, + getCurrentSectionData()); + + // Update the maximum alignment on the current section if necessary. + if (ByteAlignment > getCurrentSectionData()->getAlignment()) + getCurrentSectionData()->setAlignment(ByteAlignment); +} + +void MCPureStreamer::EmitCodeAlignment(unsigned ByteAlignment, + unsigned MaxBytesToEmit) { + // TODO: This is exactly the same as WinCOFFStreamer. Consider merging into + // MCObjectStreamer. + if (MaxBytesToEmit == 0) + MaxBytesToEmit = ByteAlignment; + MCAlignFragment *F = new MCAlignFragment(ByteAlignment, 0, 1, MaxBytesToEmit, + getCurrentSectionData()); + F->setEmitNops(true); + + // Update the maximum alignment on the current section if necessary. + if (ByteAlignment > getCurrentSectionData()->getAlignment()) + getCurrentSectionData()->setAlignment(ByteAlignment); +} + +void MCPureStreamer::EmitValueToOffset(const MCExpr *Offset, + unsigned char Value) { + new MCOrgFragment(*Offset, Value, getCurrentSectionData()); +} + +void MCPureStreamer::EmitInstToFragment(const MCInst &Inst) { + MCInstFragment *IF = new MCInstFragment(Inst, getCurrentSectionData()); + + // Add the fixups and data. + // + // FIXME: Revisit this design decision when relaxation is done, we may be + // able to get away with not storing any extra data in the MCInst. + SmallVector<MCFixup, 4> Fixups; + SmallString<256> Code; + raw_svector_ostream VecOS(Code); + getAssembler().getEmitter().EncodeInstruction(Inst, VecOS, Fixups); + VecOS.flush(); + + IF->getCode() = Code; + IF->getFixups() = Fixups; +} + +void MCPureStreamer::EmitInstToData(const MCInst &Inst) { + MCDataFragment *DF = getOrCreateDataFragment(); + + SmallVector<MCFixup, 4> Fixups; + SmallString<256> Code; + raw_svector_ostream VecOS(Code); + getAssembler().getEmitter().EncodeInstruction(Inst, VecOS, Fixups); + VecOS.flush(); + + // Add the fixups and data. + for (unsigned i = 0, e = Fixups.size(); i != e; ++i) { + Fixups[i].setOffset(Fixups[i].getOffset() + DF->getContents().size()); + DF->addFixup(Fixups[i]); + } + DF->getContents().append(Code.begin(), Code.end()); +} + +void MCPureStreamer::Finish() { + // FIXME: Handle DWARF tables? + + this->MCObjectStreamer::Finish(); +} + +MCStreamer *llvm::createPureStreamer(MCContext &Context, TargetAsmBackend &TAB, + raw_ostream &OS, MCCodeEmitter *CE) { + return new MCPureStreamer(Context, TAB, OS, CE); +} diff --git a/lib/MC/MCSectionCOFF.cpp b/lib/MC/MCSectionCOFF.cpp index 0909df4..90091f0 100644 --- a/lib/MC/MCSectionCOFF.cpp +++ b/lib/MC/MCSectionCOFF.cpp @@ -78,3 +78,7 @@ void MCSectionCOFF::PrintSwitchToSection(const MCAsmInfo &MAI, bool MCSectionCOFF::UseCodeAlign() const { return getKind().isText(); } + +bool MCSectionCOFF::isVirtualSection() const { + return getCharacteristics() & COFF::IMAGE_SCN_CNT_UNINITIALIZED_DATA; +} diff --git a/lib/MC/MCSectionELF.cpp b/lib/MC/MCSectionELF.cpp index ab72a0e..dfd77c3 100644 --- a/lib/MC/MCSectionELF.cpp +++ b/lib/MC/MCSectionELF.cpp @@ -11,7 +11,9 @@ #include "llvm/MC/MCAsmInfo.h" #include "llvm/MC/MCContext.h" #include "llvm/MC/MCSymbol.h" +#include "llvm/Support/ELF.h" #include "llvm/Support/raw_ostream.h" + using namespace llvm; MCSectionELF::~MCSectionELF() {} // anchor. @@ -37,41 +39,63 @@ void MCSectionELF::PrintSwitchToSection(const MCAsmInfo &MAI, return; } - OS << "\t.section\t" << getSectionName(); - + StringRef name = getSectionName(); + if (name.find_first_not_of("0123456789_." + "abcdefghijklmnopqrstuvwxyz" + "ABCDEFGHIJKLMNOPQRSTUVWXYZ") == name.npos) { + OS << "\t.section\t" << name; + } else { + OS << "\t.section\t\""; + for (const char *b = name.begin(), *e = name.end(); b < e; ++b) { + if (*b == '"') // Unquoted " + OS << "\\\""; + else if (*b != '\\') // Neither " or backslash + OS << *b; + else if (b + 1 == e) // Trailing backslash + OS << "\\\\"; + else { + OS << b[0] << b[1]; // Quoted character + ++b; + } + } + OS << '"'; + } + // Handle the weird solaris syntax if desired. if (MAI.usesSunStyleELFSectionSwitchSyntax() && - !(Flags & MCSectionELF::SHF_MERGE)) { - if (Flags & MCSectionELF::SHF_ALLOC) + !(Flags & ELF::SHF_MERGE)) { + if (Flags & ELF::SHF_ALLOC) OS << ",#alloc"; - if (Flags & MCSectionELF::SHF_EXECINSTR) + if (Flags & ELF::SHF_EXECINSTR) OS << ",#execinstr"; - if (Flags & MCSectionELF::SHF_WRITE) + if (Flags & ELF::SHF_WRITE) OS << ",#write"; - if (Flags & MCSectionELF::SHF_TLS) + if (Flags & ELF::SHF_TLS) OS << ",#tls"; OS << '\n'; return; } OS << ",\""; - if (Flags & MCSectionELF::SHF_ALLOC) + if (Flags & ELF::SHF_ALLOC) OS << 'a'; - if (Flags & MCSectionELF::SHF_EXECINSTR) + if (Flags & ELF::SHF_EXECINSTR) OS << 'x'; - if (Flags & MCSectionELF::SHF_WRITE) + if (Flags & ELF::SHF_GROUP) + OS << 'G'; + if (Flags & ELF::SHF_WRITE) OS << 'w'; - if (Flags & MCSectionELF::SHF_MERGE) + if (Flags & ELF::SHF_MERGE) OS << 'M'; - if (Flags & MCSectionELF::SHF_STRINGS) + if (Flags & ELF::SHF_STRINGS) OS << 'S'; - if (Flags & MCSectionELF::SHF_TLS) + if (Flags & ELF::SHF_TLS) OS << 'T'; // If there are target-specific flags, print them. - if (Flags & MCSectionELF::XCORE_SHF_CP_SECTION) + if (Flags & ELF::XCORE_SHF_CP_SECTION) OS << 'c'; - if (Flags & MCSectionELF::XCORE_SHF_DP_SECTION) + if (Flags & ELF::XCORE_SHF_DP_SECTION) OS << 'd'; OS << '"'; @@ -84,38 +108,35 @@ void MCSectionELF::PrintSwitchToSection(const MCAsmInfo &MAI, else OS << '@'; - if (Type == MCSectionELF::SHT_INIT_ARRAY) + if (Type == ELF::SHT_INIT_ARRAY) OS << "init_array"; - else if (Type == MCSectionELF::SHT_FINI_ARRAY) + else if (Type == ELF::SHT_FINI_ARRAY) OS << "fini_array"; - else if (Type == MCSectionELF::SHT_PREINIT_ARRAY) + else if (Type == ELF::SHT_PREINIT_ARRAY) OS << "preinit_array"; - else if (Type == MCSectionELF::SHT_NOBITS) + else if (Type == ELF::SHT_NOBITS) OS << "nobits"; - else if (Type == MCSectionELF::SHT_PROGBITS) + else if (Type == ELF::SHT_NOTE) + OS << "note"; + else if (Type == ELF::SHT_PROGBITS) OS << "progbits"; if (EntrySize) { - assert(Flags & MCSectionELF::SHF_MERGE); + assert(Flags & ELF::SHF_MERGE); OS << "," << EntrySize; } + if (Flags & ELF::SHF_GROUP) + OS << "," << Group->getName() << ",comdat"; OS << '\n'; } bool MCSectionELF::UseCodeAlign() const { - return getFlags() & MCSectionELF::SHF_EXECINSTR; + return getFlags() & ELF::SHF_EXECINSTR; } -// HasCommonSymbols - True if this section holds common symbols, this is -// indicated on the ELF object file by a symbol with SHN_COMMON section -// header index. -bool MCSectionELF::HasCommonSymbols() const { - - if (StringRef(SectionName).startswith(".gnu.linkonce.")) - return true; - - return false; +bool MCSectionELF::isVirtualSection() const { + return getType() == ELF::SHT_NOBITS; } unsigned MCSectionELF::DetermineEntrySize(SectionKind Kind) { diff --git a/lib/MC/MCSectionMachO.cpp b/lib/MC/MCSectionMachO.cpp index 853d990..577e93a 100644 --- a/lib/MC/MCSectionMachO.cpp +++ b/lib/MC/MCSectionMachO.cpp @@ -10,7 +10,7 @@ #include "llvm/MC/MCSectionMachO.h" #include "llvm/MC/MCContext.h" #include "llvm/Support/raw_ostream.h" -#include <ctype.h> +#include <cctype> using namespace llvm; /// SectionTypeDescriptors - These are strings that describe the various section @@ -101,16 +101,18 @@ void MCSectionMachO::PrintSwitchToSection(const MCAsmInfo &MAI, return; } - OS << ','; - unsigned SectionType = TAA & MCSectionMachO::SECTION_TYPE; assert(SectionType <= MCSectionMachO::LAST_KNOWN_SECTION_TYPE && "Invalid SectionType specified!"); - if (SectionTypeDescriptors[SectionType].AssemblerName) + if (SectionTypeDescriptors[SectionType].AssemblerName) { + OS << ','; OS << SectionTypeDescriptors[SectionType].AssemblerName; - else - OS << "<<" << SectionTypeDescriptors[SectionType].EnumName << ">>"; + } else { + // If we have no name for the attribute, stop here. + OS << '\n'; + return; + } // If we don't have any attributes, we're done. unsigned SectionAttrs = TAA & MCSectionMachO::SECTION_ATTRIBUTES; @@ -125,7 +127,9 @@ void MCSectionMachO::PrintSwitchToSection(const MCAsmInfo &MAI, // Check each attribute to see if we have it. char Separator = ','; - for (unsigned i = 0; SectionAttrDescriptors[i].AttrFlag; ++i) { + for (unsigned i = 0; + SectionAttrs != 0 && SectionAttrDescriptors[i].AttrFlag; + ++i) { // Check to see if we have this attribute. if ((SectionAttrDescriptors[i].AttrFlag & SectionAttrs) == 0) continue; @@ -153,6 +157,12 @@ bool MCSectionMachO::UseCodeAlign() const { return hasAttribute(MCSectionMachO::S_ATTR_PURE_INSTRUCTIONS); } +bool MCSectionMachO::isVirtualSection() const { + return (getType() == MCSectionMachO::S_ZEROFILL || + getType() == MCSectionMachO::S_GB_ZEROFILL || + getType() == MCSectionMachO::S_THREAD_LOCAL_ZEROFILL); +} + /// StripSpaces - This removes leading and trailing spaces from the StringRef. static void StripSpaces(StringRef &Str) { while (!Str.empty() && isspace(Str[0])) @@ -201,7 +211,6 @@ std::string MCSectionMachO::ParseSectionSpecifier(StringRef Spec, // In. "between 1 and 16 characters"; // If there is no comma after the section, we're done. - TAA = 0; StubSize = 0; if (Comma.second.empty()) return ""; diff --git a/lib/MC/MCStreamer.cpp b/lib/MC/MCStreamer.cpp index 98667f4..1bd287b 100644 --- a/lib/MC/MCStreamer.cpp +++ b/lib/MC/MCStreamer.cpp @@ -7,6 +7,8 @@ // //===----------------------------------------------------------------------===// +#include "llvm/MC/MCAsmInfo.h" +#include "llvm/MC/MCContext.h" #include "llvm/MC/MCStreamer.h" #include "llvm/MC/MCExpr.h" #include "llvm/MC/MCObjectWriter.h" @@ -17,8 +19,9 @@ #include <cstdlib> using namespace llvm; -MCStreamer::MCStreamer(MCContext &Ctx) : Context(Ctx), CurSection(0), - PrevSection(0) { +MCStreamer::MCStreamer(MCContext &Ctx) : Context(Ctx) { + const MCSection *section = NULL; + SectionStack.push_back(std::make_pair(section, section)); } MCStreamer::~MCStreamer() { @@ -29,29 +32,90 @@ raw_ostream &MCStreamer::GetCommentOS() { return nulls(); } +void MCStreamer::EmitDwarfSetLineAddr(int64_t LineDelta, + const MCSymbol *Label, int PointerSize) { + // emit the sequence to set the address + EmitIntValue(dwarf::DW_LNS_extended_op, 1); + EmitULEB128IntValue(PointerSize + 1); + EmitIntValue(dwarf::DW_LNE_set_address, 1); + EmitSymbolValue(Label, PointerSize); + + // emit the sequence for the LineDelta (from 1) and a zero address delta. + MCDwarfLineAddr::Emit(this, LineDelta, 0); +} /// EmitIntValue - Special case of EmitValue that avoids the client having to /// pass in a MCExpr for constant integers. void MCStreamer::EmitIntValue(uint64_t Value, unsigned Size, unsigned AddrSpace) { - EmitValue(MCConstantExpr::Create(Value, getContext()), Size, AddrSpace); + assert(Size <= 8 && "Invalid size"); + assert((isUIntN(8 * Size, Value) || isIntN(8 * Size, Value)) && + "Invalid size"); + char buf[8]; + // FIXME: Endianness assumption. + for (unsigned i = 0; i != Size; ++i) + buf[i] = uint8_t(Value >> (i * 8)); + EmitBytes(StringRef(buf, Size), AddrSpace); } /// EmitULEB128Value - Special case of EmitULEB128Value that avoids the /// client having to pass in a MCExpr for constant integers. void MCStreamer::EmitULEB128IntValue(uint64_t Value, unsigned AddrSpace) { - EmitULEB128Value(MCConstantExpr::Create(Value, getContext()), AddrSpace); + SmallString<32> Tmp; + raw_svector_ostream OSE(Tmp); + MCObjectWriter::EncodeULEB128(Value, OSE); + EmitBytes(OSE.str(), AddrSpace); } /// EmitSLEB128Value - Special case of EmitSLEB128Value that avoids the /// client having to pass in a MCExpr for constant integers. void MCStreamer::EmitSLEB128IntValue(int64_t Value, unsigned AddrSpace) { - EmitSLEB128Value(MCConstantExpr::Create(Value, getContext()), AddrSpace); + SmallString<32> Tmp; + raw_svector_ostream OSE(Tmp); + MCObjectWriter::EncodeSLEB128(Value, OSE); + EmitBytes(OSE.str(), AddrSpace); +} + +void MCStreamer::EmitAbsValue(const MCExpr *Value, unsigned Size, + unsigned AddrSpace) { + if (getContext().getAsmInfo().hasAggressiveSymbolFolding()) { + EmitValue(Value, Size, AddrSpace); + return; + } + MCSymbol *ABS = getContext().CreateTempSymbol(); + EmitAssignment(ABS, Value); + EmitSymbolValue(ABS, Size, AddrSpace); +} + + +void MCStreamer::EmitValue(const MCExpr *Value, unsigned Size, + unsigned AddrSpace) { + EmitValueImpl(Value, Size, false, AddrSpace); +} + +void MCStreamer::EmitPCRelValue(const MCExpr *Value, unsigned Size, + unsigned AddrSpace) { + EmitValueImpl(Value, Size, true, AddrSpace); +} + +void MCStreamer::EmitSymbolValue(const MCSymbol *Sym, unsigned Size, + bool isPCRel, unsigned AddrSpace) { + EmitValueImpl(MCSymbolRefExpr::Create(Sym, getContext()), Size, isPCRel, + AddrSpace); } void MCStreamer::EmitSymbolValue(const MCSymbol *Sym, unsigned Size, unsigned AddrSpace) { - EmitValue(MCSymbolRefExpr::Create(Sym, getContext()), Size, AddrSpace); + EmitSymbolValue(Sym, Size, false, AddrSpace); +} + +void MCStreamer::EmitPCRelSymbolValue(const MCSymbol *Sym, unsigned Size, + unsigned AddrSpace) { + EmitSymbolValue(Sym, Size, true, AddrSpace); +} + +void MCStreamer::EmitGPRel32Value(const MCExpr *Value) { + report_fatal_error("unsupported directive in streamer"); } /// EmitFill - Emit NumBytes bytes worth of the value specified by @@ -63,6 +127,178 @@ void MCStreamer::EmitFill(uint64_t NumBytes, uint8_t FillValue, EmitValue(E, 1, AddrSpace); } +bool MCStreamer::EmitDwarfFileDirective(unsigned FileNo, + StringRef Filename) { + return getContext().GetDwarfFile(Filename, FileNo) == 0; +} + +void MCStreamer::EmitDwarfLocDirective(unsigned FileNo, unsigned Line, + unsigned Column, unsigned Flags, + unsigned Isa, + unsigned Discriminator) { + getContext().setCurrentDwarfLoc(FileNo, Line, Column, Flags, Isa, + Discriminator); +} + +MCDwarfFrameInfo *MCStreamer::getCurrentFrameInfo() { + if (FrameInfos.empty()) + return NULL; + return &FrameInfos.back(); +} + +void MCStreamer::EnsureValidFrame() { + MCDwarfFrameInfo *CurFrame = getCurrentFrameInfo(); + if (!CurFrame || CurFrame->End) + report_fatal_error("No open frame"); +} + +bool MCStreamer::EmitCFIStartProc() { + MCDwarfFrameInfo *CurFrame = getCurrentFrameInfo(); + if (CurFrame && !CurFrame->End) { + report_fatal_error("Starting a frame before finishing the previous one!"); + return true; + } + MCDwarfFrameInfo Frame; + Frame.Begin = getContext().CreateTempSymbol(); + EmitLabel(Frame.Begin); + FrameInfos.push_back(Frame); + return false; +} + +bool MCStreamer::EmitCFIEndProc() { + EnsureValidFrame(); + MCDwarfFrameInfo *CurFrame = getCurrentFrameInfo(); + CurFrame->End = getContext().CreateTempSymbol(); + EmitLabel(CurFrame->End); + return false; +} + +bool MCStreamer::EmitCFIDefCfa(int64_t Register, int64_t Offset) { + EnsureValidFrame(); + MCDwarfFrameInfo *CurFrame = getCurrentFrameInfo(); + MCSymbol *Label = getContext().CreateTempSymbol(); + EmitLabel(Label); + MachineLocation Dest(MachineLocation::VirtualFP); + MachineLocation Source(Register, -Offset); + MCCFIInstruction Instruction(Label, Dest, Source); + CurFrame->Instructions.push_back(Instruction); + return false; +} + +bool MCStreamer::EmitCFIDefCfaOffset(int64_t Offset) { + EnsureValidFrame(); + MCDwarfFrameInfo *CurFrame = getCurrentFrameInfo(); + MCSymbol *Label = getContext().CreateTempSymbol(); + EmitLabel(Label); + MachineLocation Dest(MachineLocation::VirtualFP); + MachineLocation Source(MachineLocation::VirtualFP, -Offset); + MCCFIInstruction Instruction(Label, Dest, Source); + CurFrame->Instructions.push_back(Instruction); + return false; +} + +bool MCStreamer::EmitCFIDefCfaRegister(int64_t Register) { + EnsureValidFrame(); + MCDwarfFrameInfo *CurFrame = getCurrentFrameInfo(); + MCSymbol *Label = getContext().CreateTempSymbol(); + EmitLabel(Label); + MachineLocation Dest(Register); + MachineLocation Source(MachineLocation::VirtualFP); + MCCFIInstruction Instruction(Label, Dest, Source); + CurFrame->Instructions.push_back(Instruction); + return false; +} + +bool MCStreamer::EmitCFIOffset(int64_t Register, int64_t Offset) { + EnsureValidFrame(); + MCDwarfFrameInfo *CurFrame = getCurrentFrameInfo(); + MCSymbol *Label = getContext().CreateTempSymbol(); + EmitLabel(Label); + MachineLocation Dest(Register, Offset); + MachineLocation Source(Register, Offset); + MCCFIInstruction Instruction(Label, Dest, Source); + CurFrame->Instructions.push_back(Instruction); + return false; +} + +bool MCStreamer::EmitCFIPersonality(const MCSymbol *Sym, + unsigned Encoding) { + EnsureValidFrame(); + MCDwarfFrameInfo *CurFrame = getCurrentFrameInfo(); + CurFrame->Personality = Sym; + CurFrame->PersonalityEncoding = Encoding; + return false; +} + +bool MCStreamer::EmitCFILsda(const MCSymbol *Sym, unsigned Encoding) { + EnsureValidFrame(); + MCDwarfFrameInfo *CurFrame = getCurrentFrameInfo(); + CurFrame->Lsda = Sym; + CurFrame->LsdaEncoding = Encoding; + return false; +} + +bool MCStreamer::EmitCFIRememberState() { + EnsureValidFrame(); + MCDwarfFrameInfo *CurFrame = getCurrentFrameInfo(); + MCSymbol *Label = getContext().CreateTempSymbol(); + EmitLabel(Label); + MCCFIInstruction Instruction(MCCFIInstruction::Remember, Label); + CurFrame->Instructions.push_back(Instruction); + return false; +} + +bool MCStreamer::EmitCFIRestoreState() { + // FIXME: Error if there is no matching cfi_remember_state. + EnsureValidFrame(); + MCDwarfFrameInfo *CurFrame = getCurrentFrameInfo(); + MCSymbol *Label = getContext().CreateTempSymbol(); + EmitLabel(Label); + MCCFIInstruction Instruction(MCCFIInstruction::Restore, Label); + CurFrame->Instructions.push_back(Instruction); + return false; +} + +void MCStreamer::EmitFnStart() { + errs() << "Not implemented yet\n"; + abort(); +} + +void MCStreamer::EmitFnEnd() { + errs() << "Not implemented yet\n"; + abort(); +} + +void MCStreamer::EmitCantUnwind() { + errs() << "Not implemented yet\n"; + abort(); +} + +void MCStreamer::EmitHandlerData() { + errs() << "Not implemented yet\n"; + abort(); +} + +void MCStreamer::EmitPersonality(const MCSymbol *Personality) { + errs() << "Not implemented yet\n"; + abort(); +} + +void MCStreamer::EmitSetFP(unsigned FpReg, unsigned SpReg, int64_t Offset) { + errs() << "Not implemented yet\n"; + abort(); +} + +void MCStreamer::EmitPad(int64_t Offset) { + errs() << "Not implemented yet\n"; + abort(); +} + +void MCStreamer::EmitRegSave(const SmallVectorImpl<unsigned> &RegList, bool) { + errs() << "Not implemented yet\n"; + abort(); +} + /// EmitRawText - If this file is backed by an assembly streamer, this dumps /// the specified string in the output .s file. This capability is /// indicated by the hasRawTextSupport() predicate. diff --git a/lib/MC/MachObjectWriter.cpp b/lib/MC/MachObjectWriter.cpp index 41c11fb..de53494 100644 --- a/lib/MC/MachObjectWriter.cpp +++ b/lib/MC/MachObjectWriter.cpp @@ -7,6 +7,8 @@ // //===----------------------------------------------------------------------===// +#include "llvm/MC/MCMachObjectWriter.h" +#include "llvm/ADT/OwningPtr.h" #include "llvm/ADT/StringMap.h" #include "llvm/ADT/Twine.h" #include "llvm/MC/MCAssembler.h" @@ -17,25 +19,29 @@ #include "llvm/MC/MCSymbol.h" #include "llvm/MC/MCMachOSymbolFlags.h" #include "llvm/MC/MCValue.h" +#include "llvm/Object/MachOFormat.h" #include "llvm/Support/ErrorHandling.h" -#include "llvm/Support/MachO.h" #include "llvm/Target/TargetAsmBackend.h" // FIXME: Gross. +#include "../Target/ARM/ARMFixupKinds.h" #include "../Target/X86/X86FixupKinds.h" #include <vector> using namespace llvm; +using namespace llvm::object; // FIXME: this has been copied from (or to) X86AsmBackend.cpp static unsigned getFixupKindLog2Size(unsigned Kind) { switch (Kind) { - default: llvm_unreachable("invalid fixup kind!"); - case X86::reloc_pcrel_1byte: + default: + llvm_unreachable("invalid fixup kind!"); + case FK_PCRel_1: case FK_Data_1: return 0; - case X86::reloc_pcrel_2byte: + case FK_PCRel_2: case FK_Data_2: return 1; - case X86::reloc_pcrel_4byte: + case FK_PCRel_4: + // FIXME: Remove these!!! case X86::reloc_riprel_4byte: case X86::reloc_riprel_4byte_movq_load: case X86::reloc_signed_4byte: @@ -44,24 +50,6 @@ static unsigned getFixupKindLog2Size(unsigned Kind) { } } -static bool isFixupKindPCRel(unsigned Kind) { - switch (Kind) { - default: - return false; - case X86::reloc_pcrel_1byte: - case X86::reloc_pcrel_2byte: - case X86::reloc_pcrel_4byte: - case X86::reloc_riprel_4byte: - case X86::reloc_riprel_4byte_movq_load: - return true; - } -} - -static bool isFixupKindRIPRel(unsigned Kind) { - return Kind == X86::reloc_riprel_4byte || - Kind == X86::reloc_riprel_4byte_movq_load; -} - static bool doesSymbolRequireExternRelocation(MCSymbolData *SD) { // Undefined symbols are always extern. if (SD->Symbol->isUndefined()) @@ -76,176 +64,9 @@ static bool doesSymbolRequireExternRelocation(MCSymbolData *SD) { return false; } -static bool isScatteredFixupFullyResolved(const MCAssembler &Asm, - const MCValue Target, - const MCSymbolData *BaseSymbol) { - // The effective fixup address is - // addr(atom(A)) + offset(A) - // - addr(atom(B)) - offset(B) - // - addr(BaseSymbol) + <fixup offset from base symbol> - // and the offsets are not relocatable, so the fixup is fully resolved when - // addr(atom(A)) - addr(atom(B)) - addr(BaseSymbol) == 0. - // - // Note that "false" is almost always conservatively correct (it means we emit - // a relocation which is unnecessary), except when it would force us to emit a - // relocation which the target cannot encode. - - const MCSymbolData *A_Base = 0, *B_Base = 0; - if (const MCSymbolRefExpr *A = Target.getSymA()) { - // Modified symbol references cannot be resolved. - if (A->getKind() != MCSymbolRefExpr::VK_None) - return false; - - A_Base = Asm.getAtom(&Asm.getSymbolData(A->getSymbol())); - if (!A_Base) - return false; - } - - if (const MCSymbolRefExpr *B = Target.getSymB()) { - // Modified symbol references cannot be resolved. - if (B->getKind() != MCSymbolRefExpr::VK_None) - return false; - - B_Base = Asm.getAtom(&Asm.getSymbolData(B->getSymbol())); - if (!B_Base) - return false; - } - - // If there is no base, A and B have to be the same atom for this fixup to be - // fully resolved. - if (!BaseSymbol) - return A_Base == B_Base; - - // Otherwise, B must be missing and A must be the base. - return !B_Base && BaseSymbol == A_Base; -} - -static bool isScatteredFixupFullyResolvedSimple(const MCAssembler &Asm, - const MCValue Target, - const MCSection *BaseSection) { - // The effective fixup address is - // addr(atom(A)) + offset(A) - // - addr(atom(B)) - offset(B) - // - addr(<base symbol>) + <fixup offset from base symbol> - // and the offsets are not relocatable, so the fixup is fully resolved when - // addr(atom(A)) - addr(atom(B)) - addr(<base symbol>)) == 0. - // - // The simple (Darwin, except on x86_64) way of dealing with this was to - // assume that any reference to a temporary symbol *must* be a temporary - // symbol in the same atom, unless the sections differ. Therefore, any PCrel - // relocation to a temporary symbol (in the same section) is fully - // resolved. This also works in conjunction with absolutized .set, which - // requires the compiler to use .set to absolutize the differences between - // symbols which the compiler knows to be assembly time constants, so we don't - // need to worry about considering symbol differences fully resolved. - - // Non-relative fixups are only resolved if constant. - if (!BaseSection) - return Target.isAbsolute(); - - // Otherwise, relative fixups are only resolved if not a difference and the - // target is a temporary in the same section. - if (Target.isAbsolute() || Target.getSymB()) - return false; - - const MCSymbol *A = &Target.getSymA()->getSymbol(); - if (!A->isTemporary() || !A->isInSection() || - &A->getSection() != BaseSection) - return false; - - return true; -} - namespace { class MachObjectWriter : public MCObjectWriter { - // See <mach-o/loader.h>. - enum { - Header_Magic32 = 0xFEEDFACE, - Header_Magic64 = 0xFEEDFACF - }; - - enum { - Header32Size = 28, - Header64Size = 32, - SegmentLoadCommand32Size = 56, - SegmentLoadCommand64Size = 72, - Section32Size = 68, - Section64Size = 80, - SymtabLoadCommandSize = 24, - DysymtabLoadCommandSize = 80, - Nlist32Size = 12, - Nlist64Size = 16, - RelocationInfoSize = 8 - }; - - enum HeaderFileType { - HFT_Object = 0x1 - }; - - enum HeaderFlags { - HF_SubsectionsViaSymbols = 0x2000 - }; - - enum LoadCommandType { - LCT_Segment = 0x1, - LCT_Symtab = 0x2, - LCT_Dysymtab = 0xb, - LCT_Segment64 = 0x19 - }; - - // See <mach-o/nlist.h>. - enum SymbolTypeType { - STT_Undefined = 0x00, - STT_Absolute = 0x02, - STT_Section = 0x0e - }; - - enum SymbolTypeFlags { - // If any of these bits are set, then the entry is a stab entry number (see - // <mach-o/stab.h>. Otherwise the other masks apply. - STF_StabsEntryMask = 0xe0, - - STF_TypeMask = 0x0e, - STF_External = 0x01, - STF_PrivateExtern = 0x10 - }; - - /// IndirectSymbolFlags - Flags for encoding special values in the indirect - /// symbol entry. - enum IndirectSymbolFlags { - ISF_Local = 0x80000000, - ISF_Absolute = 0x40000000 - }; - - /// RelocationFlags - Special flags for addresses. - enum RelocationFlags { - RF_Scattered = 0x80000000 - }; - - enum RelocationInfoType { - RIT_Vanilla = 0, - RIT_Pair = 1, - RIT_Difference = 2, - RIT_PreboundLazyPointer = 3, - RIT_LocalDifference = 4, - RIT_TLV = 5 - }; - - /// X86_64 uses its own relocation types. - enum RelocationInfoTypeX86_64 { - RIT_X86_64_Unsigned = 0, - RIT_X86_64_Signed = 1, - RIT_X86_64_Branch = 2, - RIT_X86_64_GOTLoad = 3, - RIT_X86_64_GOT = 4, - RIT_X86_64_Subtractor = 5, - RIT_X86_64_Signed1 = 6, - RIT_X86_64_Signed2 = 7, - RIT_X86_64_Signed4 = 8, - RIT_X86_64_TLV = 9 - }; - /// MachSymbolData - Helper struct for containing some precomputed information /// on symbols. struct MachSymbolData { @@ -260,16 +81,14 @@ class MachObjectWriter : public MCObjectWriter { } }; + /// The target specific Mach-O writer instance. + llvm::OwningPtr<MCMachObjectTargetWriter> TargetObjectWriter; + /// @name Relocation Data /// @{ - struct MachRelocationEntry { - uint32_t Word0; - uint32_t Word1; - }; - llvm::DenseMap<const MCSectionData*, - std::vector<MachRelocationEntry> > Relocations; + std::vector<macho::RelocationEntry> > Relocations; llvm::DenseMap<const MCSectionData*, unsigned> IndirectSymBase; /// @} @@ -283,25 +102,70 @@ class MachObjectWriter : public MCObjectWriter { /// @} - unsigned Is64Bit : 1; +private: + /// @name Utility Methods + /// @{ + + bool isFixupKindPCRel(const MCAssembler &Asm, unsigned Kind) { + const MCFixupKindInfo &FKI = Asm.getBackend().getFixupKindInfo( + (MCFixupKind) Kind); + + return FKI.Flags & MCFixupKindInfo::FKF_IsPCRel; + } + + /// @} + + SectionAddrMap SectionAddress; + uint64_t getSectionAddress(const MCSectionData* SD) const { + return SectionAddress.lookup(SD); + } + uint64_t getSymbolAddress(const MCSymbolData* SD, + const MCAsmLayout &Layout) const { + return getSectionAddress(SD->getFragment()->getParent()) + + Layout.getSymbolOffset(SD); + } + uint64_t getFragmentAddress(const MCFragment *Fragment, + const MCAsmLayout &Layout) const { + return getSectionAddress(Fragment->getParent()) + + Layout.getFragmentOffset(Fragment); + } - uint32_t CPUType; - uint32_t CPUSubtype; + uint64_t getPaddingSize(const MCSectionData *SD, + const MCAsmLayout &Layout) const { + uint64_t EndAddr = getSectionAddress(SD) + Layout.getSectionAddressSize(SD); + unsigned Next = SD->getLayoutOrder() + 1; + if (Next >= Layout.getSectionOrder().size()) + return 0; + + const MCSectionData &NextSD = *Layout.getSectionOrder()[Next]; + if (NextSD.getSection().isVirtualSection()) + return 0; + return OffsetToAlignment(EndAddr, NextSD.getAlignment()); + } public: - MachObjectWriter(raw_ostream &_OS, - bool _Is64Bit, uint32_t _CPUType, uint32_t _CPUSubtype, + MachObjectWriter(MCMachObjectTargetWriter *MOTW, raw_ostream &_OS, bool _IsLittleEndian) - : MCObjectWriter(_OS, _IsLittleEndian), - Is64Bit(_Is64Bit), CPUType(_CPUType), CPUSubtype(_CPUSubtype) { + : MCObjectWriter(_OS, _IsLittleEndian), TargetObjectWriter(MOTW) { + } + + /// @name Target Writer Proxy Accessors + /// @{ + + bool is64Bit() const { return TargetObjectWriter->is64Bit(); } + bool isARM() const { + uint32_t CPUType = TargetObjectWriter->getCPUType() & ~mach::CTFM_ArchMask; + return CPUType == mach::CTM_ARM; } + /// @} + void WriteHeader(unsigned NumLoadCommands, unsigned LoadCommandsSize, bool SubsectionsViaSymbols) { uint32_t Flags = 0; if (SubsectionsViaSymbols) - Flags |= HF_SubsectionsViaSymbols; + Flags |= macho::HF_SubsectionsViaSymbols; // struct mach_header (28 bytes) or // struct mach_header_64 (32 bytes) @@ -309,20 +173,20 @@ public: uint64_t Start = OS.tell(); (void) Start; - Write32(Is64Bit ? Header_Magic64 : Header_Magic32); + Write32(is64Bit() ? macho::HM_Object64 : macho::HM_Object32); - Write32(CPUType); - Write32(CPUSubtype); + Write32(TargetObjectWriter->getCPUType()); + Write32(TargetObjectWriter->getCPUSubtype()); - Write32(HFT_Object); - Write32(NumLoadCommands); // Object files have a single load command, the - // segment. + Write32(macho::HFT_Object); + Write32(NumLoadCommands); Write32(LoadCommandsSize); Write32(Flags); - if (Is64Bit) + if (is64Bit()) Write32(0); // reserved - assert(OS.tell() - Start == Is64Bit ? Header64Size : Header32Size); + assert(OS.tell() - Start == + (is64Bit() ? macho::Header64Size : macho::Header32Size)); } /// WriteSegmentLoadCommand - Write a segment load command. @@ -339,14 +203,16 @@ public: uint64_t Start = OS.tell(); (void) Start; - unsigned SegmentLoadCommandSize = Is64Bit ? SegmentLoadCommand64Size : - SegmentLoadCommand32Size; - Write32(Is64Bit ? LCT_Segment64 : LCT_Segment); + unsigned SegmentLoadCommandSize = + is64Bit() ? macho::SegmentLoadCommand64Size: + macho::SegmentLoadCommand32Size; + Write32(is64Bit() ? macho::LCT_Segment64 : macho::LCT_Segment); Write32(SegmentLoadCommandSize + - NumSections * (Is64Bit ? Section64Size : Section32Size)); + NumSections * (is64Bit() ? macho::Section64Size : + macho::Section32Size)); WriteBytes("", 16); - if (Is64Bit) { + if (is64Bit()) { Write64(0); // vmaddr Write64(VMSize); // vmsize Write64(SectionDataStartOffset); // file offset @@ -368,10 +234,10 @@ public: void WriteSection(const MCAssembler &Asm, const MCAsmLayout &Layout, const MCSectionData &SD, uint64_t FileOffset, uint64_t RelocationsStart, unsigned NumRelocations) { - uint64_t SectionSize = Layout.getSectionSize(&SD); + uint64_t SectionSize = Layout.getSectionAddressSize(&SD); // The offset is unused for virtual sections. - if (Asm.getBackend().isVirtualSection(SD.getSection())) { + if (SD.getSection().isVirtualSection()) { assert(Layout.getSectionFileSize(&SD) == 0 && "Invalid file size!"); FileOffset = 0; } @@ -385,11 +251,11 @@ public: const MCSectionMachO &Section = cast<MCSectionMachO>(SD.getSection()); WriteBytes(Section.getSectionName(), 16); WriteBytes(Section.getSegmentName(), 16); - if (Is64Bit) { - Write64(Layout.getSectionAddress(&SD)); // address + if (is64Bit()) { + Write64(getSectionAddress(&SD)); // address Write64(SectionSize); // size } else { - Write32(Layout.getSectionAddress(&SD)); // address + Write32(getSectionAddress(&SD)); // address Write32(SectionSize); // size } Write32(FileOffset); @@ -405,10 +271,11 @@ public: Write32(Flags); Write32(IndirectSymBase.lookup(&SD)); // reserved1 Write32(Section.getStubSize()); // reserved2 - if (Is64Bit) + if (is64Bit()) Write32(0); // reserved3 - assert(OS.tell() - Start == Is64Bit ? Section64Size : Section32Size); + assert(OS.tell() - Start == (is64Bit() ? macho::Section64Size : + macho::Section32Size)); } void WriteSymtabLoadCommand(uint32_t SymbolOffset, uint32_t NumSymbols, @@ -419,14 +286,14 @@ public: uint64_t Start = OS.tell(); (void) Start; - Write32(LCT_Symtab); - Write32(SymtabLoadCommandSize); + Write32(macho::LCT_Symtab); + Write32(macho::SymtabLoadCommandSize); Write32(SymbolOffset); Write32(NumSymbols); Write32(StringTableOffset); Write32(StringTableSize); - assert(OS.tell() - Start == SymtabLoadCommandSize); + assert(OS.tell() - Start == macho::SymtabLoadCommandSize); } void WriteDysymtabLoadCommand(uint32_t FirstLocalSymbol, @@ -442,8 +309,8 @@ public: uint64_t Start = OS.tell(); (void) Start; - Write32(LCT_Dysymtab); - Write32(DysymtabLoadCommandSize); + Write32(macho::LCT_Dysymtab); + Write32(macho::DysymtabLoadCommandSize); Write32(FirstLocalSymbol); Write32(NumLocalSymbols); Write32(FirstExternalSymbol); @@ -463,7 +330,7 @@ public: Write32(0); // locreloff Write32(0); // nlocrel - assert(OS.tell() - Start == DysymtabLoadCommandSize); + assert(OS.tell() - Start == macho::DysymtabLoadCommandSize); } void WriteNlist(MachSymbolData &MSD, const MCAsmLayout &Layout) { @@ -477,27 +344,27 @@ public: // // FIXME: Are the prebound or indirect fields possible here? if (Symbol.isUndefined()) - Type = STT_Undefined; + Type = macho::STT_Undefined; else if (Symbol.isAbsolute()) - Type = STT_Absolute; + Type = macho::STT_Absolute; else - Type = STT_Section; + Type = macho::STT_Section; // FIXME: Set STAB bits. if (Data.isPrivateExtern()) - Type |= STF_PrivateExtern; + Type |= macho::STF_PrivateExtern; // Set external bit. if (Data.isExternal() || Symbol.isUndefined()) - Type |= STF_External; + Type |= macho::STF_External; // Compute the symbol address. if (Symbol.isDefined()) { if (Symbol.isAbsolute()) { Address = cast<MCConstantExpr>(Symbol.getVariableValue())->getValue(); } else { - Address = Layout.getSymbolAddress(&Data); + Address = getSymbolAddress(&Data, Layout); } } else if (Data.isCommon()) { // Common symbols are encoded with the size in the address @@ -525,7 +392,7 @@ public: // The Mach-O streamer uses the lowest 16-bits of the flags for the 'desc' // value. Write16(Flags); - if (Is64Bit) + if (is64Bit()) Write64(Address); else Write32(Address); @@ -545,11 +412,15 @@ public: // - Input errors, where something cannot be correctly encoded. 'as' allows // these through in many cases. + static bool isFixupKindRIPRel(unsigned Kind) { + return Kind == X86::reloc_riprel_4byte || + Kind == X86::reloc_riprel_4byte_movq_load; + } void RecordX86_64Relocation(const MCAssembler &Asm, const MCAsmLayout &Layout, const MCFragment *Fragment, const MCFixup &Fixup, MCValue Target, uint64_t &FixedValue) { - unsigned IsPCRel = isFixupKindPCRel(Fixup.getKind()); + unsigned IsPCRel = isFixupKindPCRel(Asm, Fixup.getKind()); unsigned IsRIPRel = isFixupKindRIPRel(Fixup.getKind()); unsigned Log2Size = getFixupKindLog2Size(Fixup.getKind()); @@ -557,7 +428,7 @@ public: uint32_t FixupOffset = Layout.getFragmentOffset(Fragment) + Fixup.getOffset(); uint32_t FixupAddress = - Layout.getFragmentAddress(Fragment) + Fixup.getOffset(); + getFragmentAddress(Fragment, Layout) + Fixup.getOffset(); int64_t Value = 0; unsigned Index = 0; unsigned IsExtern = 0; @@ -576,7 +447,7 @@ public: if (Target.isAbsolute()) { // constant // SymbolNum of 0 indicates the absolute section. - Type = RIT_X86_64_Unsigned; + Type = macho::RIT_X86_64_Unsigned; Index = 0; // FIXME: I believe this is broken, I don't think the linker can @@ -586,7 +457,7 @@ public: // yet). if (IsPCRel) { IsExtern = 1; - Type = RIT_X86_64_Branch; + Type = macho::RIT_X86_64_Branch; } } else if (Target.getSymB()) { // A - B + constant const MCSymbol *A = &Target.getSymA()->getSymbol(); @@ -620,10 +491,10 @@ public: if (A_Base == B_Base && A_Base) report_fatal_error("unsupported relocation with identical base"); - Value += Layout.getSymbolAddress(&A_SD) - - (A_Base == NULL ? 0 : Layout.getSymbolAddress(A_Base)); - Value -= Layout.getSymbolAddress(&B_SD) - - (B_Base == NULL ? 0 : Layout.getSymbolAddress(B_Base)); + Value += getSymbolAddress(&A_SD, Layout) - + (A_Base == NULL ? 0 : getSymbolAddress(A_Base, Layout)); + Value -= getSymbolAddress(&B_SD, Layout) - + (B_Base == NULL ? 0 : getSymbolAddress(B_Base, Layout)); if (A_Base) { Index = A_Base->getIndex(); @@ -633,9 +504,9 @@ public: Index = A_SD.getFragment()->getParent()->getOrdinal() + 1; IsExtern = 0; } - Type = RIT_X86_64_Unsigned; + Type = macho::RIT_X86_64_Unsigned; - MachRelocationEntry MRE; + macho::RelocationEntry MRE; MRE.Word0 = FixupOffset; MRE.Word1 = ((Index << 0) | (IsPCRel << 24) | @@ -652,7 +523,7 @@ public: Index = B_SD.getFragment()->getParent()->getOrdinal() + 1; IsExtern = 0; } - Type = RIT_X86_64_Subtractor; + Type = macho::RIT_X86_64_Subtractor; } else { const MCSymbol *Symbol = &Target.getSymA()->getSymbol(); MCSymbolData &SD = Asm.getSymbolData(*Symbol); @@ -678,15 +549,26 @@ public: // Add the local offset, if needed. if (Base != &SD) - Value += Layout.getSymbolAddress(&SD) - Layout.getSymbolAddress(Base); + Value += Layout.getSymbolOffset(&SD) - Layout.getSymbolOffset(Base); } else if (Symbol->isInSection()) { // The index is the section ordinal (1-based). Index = SD.getFragment()->getParent()->getOrdinal() + 1; IsExtern = 0; - Value += Layout.getSymbolAddress(&SD); + Value += getSymbolAddress(&SD, Layout); if (IsPCRel) Value -= FixupAddress + (1 << Log2Size); + } else if (Symbol->isVariable()) { + const MCExpr *Value = Symbol->getVariableValue(); + int64_t Res; + bool isAbs = Value->EvaluateAsAbsolute(Res, Layout, SectionAddress); + if (isAbs) { + FixedValue = Res; + return; + } else { + report_fatal_error("unsupported relocation of variable '" + + Symbol->getName() + "'"); + } } else { report_fatal_error("unsupported relocation of undefined symbol '" + Symbol->getName() + "'"); @@ -700,15 +582,15 @@ public: // rewrite the movq to an leaq at link time if the symbol ends up in // the same linkage unit. if (unsigned(Fixup.getKind()) == X86::reloc_riprel_4byte_movq_load) - Type = RIT_X86_64_GOTLoad; + Type = macho::RIT_X86_64_GOTLoad; else - Type = RIT_X86_64_GOT; + Type = macho::RIT_X86_64_GOT; } else if (Modifier == MCSymbolRefExpr::VK_TLVP) { - Type = RIT_X86_64_TLV; + Type = macho::RIT_X86_64_TLV; } else if (Modifier != MCSymbolRefExpr::VK_None) { report_fatal_error("unsupported symbol modifier in relocation"); } else { - Type = RIT_X86_64_Signed; + Type = macho::RIT_X86_64_Signed; // The Darwin x86_64 relocation format has a problem where it cannot // encode an address (L<foo> + <constant>) which is outside the atom @@ -725,9 +607,9 @@ public: // (the additional bias), but instead appear to just look at the // final offset. switch (-(Target.getConstant() + (1LL << Log2Size))) { - case 1: Type = RIT_X86_64_Signed1; break; - case 2: Type = RIT_X86_64_Signed2; break; - case 4: Type = RIT_X86_64_Signed4; break; + case 1: Type = macho::RIT_X86_64_Signed1; break; + case 2: Type = macho::RIT_X86_64_Signed2; break; + case 4: Type = macho::RIT_X86_64_Signed4; break; } } } else { @@ -735,24 +617,24 @@ public: report_fatal_error("unsupported symbol modifier in branch " "relocation"); - Type = RIT_X86_64_Branch; + Type = macho::RIT_X86_64_Branch; } } else { if (Modifier == MCSymbolRefExpr::VK_GOT) { - Type = RIT_X86_64_GOT; + Type = macho::RIT_X86_64_GOT; } else if (Modifier == MCSymbolRefExpr::VK_GOTPCREL) { // GOTPCREL is allowed as a modifier on non-PCrel instructions, in // which case all we do is set the PCrel bit in the relocation entry; // this is used with exception handling, for example. The source is // required to include any necessary offset directly. - Type = RIT_X86_64_GOT; + Type = macho::RIT_X86_64_GOT; IsPCRel = 1; } else if (Modifier == MCSymbolRefExpr::VK_TLVP) { report_fatal_error("TLVP symbol modifier should have been rip-rel"); } else if (Modifier != MCSymbolRefExpr::VK_None) report_fatal_error("unsupported symbol modifier in relocation"); else - Type = RIT_X86_64_Unsigned; + Type = macho::RIT_X86_64_Unsigned; } } @@ -760,7 +642,7 @@ public: FixedValue = Value; // struct relocation_info (8 bytes) - MachRelocationEntry MRE; + macho::RelocationEntry MRE; MRE.Word0 = FixupOffset; MRE.Word1 = ((Index << 0) | (IsPCRel << 24) | @@ -774,11 +656,11 @@ public: const MCAsmLayout &Layout, const MCFragment *Fragment, const MCFixup &Fixup, MCValue Target, + unsigned Log2Size, uint64_t &FixedValue) { uint32_t FixupOffset = Layout.getFragmentOffset(Fragment)+Fixup.getOffset(); - unsigned IsPCRel = isFixupKindPCRel(Fixup.getKind()); - unsigned Log2Size = getFixupKindLog2Size(Fixup.getKind()); - unsigned Type = RIT_Vanilla; + unsigned IsPCRel = isFixupKindPCRel(Asm, Fixup.getKind()); + unsigned Type = macho::RIT_Vanilla; // See <reloc.h>. const MCSymbol *A = &Target.getSymA()->getSymbol(); @@ -788,7 +670,9 @@ public: report_fatal_error("symbol '" + A->getName() + "' can not be undefined in a subtraction expression"); - uint32_t Value = Layout.getSymbolAddress(A_SD); + uint32_t Value = getSymbolAddress(A_SD, Layout); + uint64_t SecAddr = getSectionAddress(A_SD->getFragment()->getParent()); + FixedValue += SecAddr; uint32_t Value2 = 0; if (const MCSymbolRefExpr *B = Target.getSymB()) { @@ -803,28 +687,184 @@ public: // Note that there is no longer any semantic difference between these two // relocation types from the linkers point of view, this is done solely // for pedantic compatibility with 'as'. - Type = A_SD->isExternal() ? RIT_Difference : RIT_LocalDifference; - Value2 = Layout.getSymbolAddress(B_SD); + Type = A_SD->isExternal() ? (unsigned)macho::RIT_Difference : + (unsigned)macho::RIT_Generic_LocalDifference; + Value2 = getSymbolAddress(B_SD, Layout); + FixedValue -= getSectionAddress(B_SD->getFragment()->getParent()); } // Relocations are written out in reverse order, so the PAIR comes first. - if (Type == RIT_Difference || Type == RIT_LocalDifference) { - MachRelocationEntry MRE; + if (Type == macho::RIT_Difference || + Type == macho::RIT_Generic_LocalDifference) { + macho::RelocationEntry MRE; MRE.Word0 = ((0 << 0) | - (RIT_Pair << 24) | + (macho::RIT_Pair << 24) | (Log2Size << 28) | (IsPCRel << 30) | - RF_Scattered); + macho::RF_Scattered); MRE.Word1 = Value2; Relocations[Fragment->getParent()].push_back(MRE); } - MachRelocationEntry MRE; + macho::RelocationEntry MRE; MRE.Word0 = ((FixupOffset << 0) | (Type << 24) | (Log2Size << 28) | (IsPCRel << 30) | - RF_Scattered); + macho::RF_Scattered); + MRE.Word1 = Value; + Relocations[Fragment->getParent()].push_back(MRE); + } + + void RecordARMScatteredRelocation(const MCAssembler &Asm, + const MCAsmLayout &Layout, + const MCFragment *Fragment, + const MCFixup &Fixup, MCValue Target, + unsigned Log2Size, + uint64_t &FixedValue) { + uint32_t FixupOffset = Layout.getFragmentOffset(Fragment)+Fixup.getOffset(); + unsigned IsPCRel = isFixupKindPCRel(Asm, Fixup.getKind()); + unsigned Type = macho::RIT_Vanilla; + + // See <reloc.h>. + const MCSymbol *A = &Target.getSymA()->getSymbol(); + MCSymbolData *A_SD = &Asm.getSymbolData(*A); + + if (!A_SD->getFragment()) + report_fatal_error("symbol '" + A->getName() + + "' can not be undefined in a subtraction expression"); + + uint32_t Value = getSymbolAddress(A_SD, Layout); + uint64_t SecAddr = getSectionAddress(A_SD->getFragment()->getParent()); + FixedValue += SecAddr; + uint32_t Value2 = 0; + + if (const MCSymbolRefExpr *B = Target.getSymB()) { + MCSymbolData *B_SD = &Asm.getSymbolData(B->getSymbol()); + + if (!B_SD->getFragment()) + report_fatal_error("symbol '" + B->getSymbol().getName() + + "' can not be undefined in a subtraction expression"); + + // Select the appropriate difference relocation type. + Type = macho::RIT_Difference; + Value2 = getSymbolAddress(B_SD, Layout); + FixedValue -= getSectionAddress(B_SD->getFragment()->getParent()); + } + + // Relocations are written out in reverse order, so the PAIR comes first. + if (Type == macho::RIT_Difference || + Type == macho::RIT_Generic_LocalDifference) { + macho::RelocationEntry MRE; + MRE.Word0 = ((0 << 0) | + (macho::RIT_Pair << 24) | + (Log2Size << 28) | + (IsPCRel << 30) | + macho::RF_Scattered); + MRE.Word1 = Value2; + Relocations[Fragment->getParent()].push_back(MRE); + } + + macho::RelocationEntry MRE; + MRE.Word0 = ((FixupOffset << 0) | + (Type << 24) | + (Log2Size << 28) | + (IsPCRel << 30) | + macho::RF_Scattered); + MRE.Word1 = Value; + Relocations[Fragment->getParent()].push_back(MRE); + } + + void RecordARMMovwMovtRelocation(const MCAssembler &Asm, + const MCAsmLayout &Layout, + const MCFragment *Fragment, + const MCFixup &Fixup, MCValue Target, + uint64_t &FixedValue) { + uint32_t FixupOffset = Layout.getFragmentOffset(Fragment)+Fixup.getOffset(); + unsigned IsPCRel = isFixupKindPCRel(Asm, Fixup.getKind()); + unsigned Type = macho::RIT_ARM_Half; + + // See <reloc.h>. + const MCSymbol *A = &Target.getSymA()->getSymbol(); + MCSymbolData *A_SD = &Asm.getSymbolData(*A); + + if (!A_SD->getFragment()) + report_fatal_error("symbol '" + A->getName() + + "' can not be undefined in a subtraction expression"); + + uint32_t Value = getSymbolAddress(A_SD, Layout); + uint32_t Value2 = 0; + uint64_t SecAddr = getSectionAddress(A_SD->getFragment()->getParent()); + FixedValue += SecAddr; + + if (const MCSymbolRefExpr *B = Target.getSymB()) { + MCSymbolData *B_SD = &Asm.getSymbolData(B->getSymbol()); + + if (!B_SD->getFragment()) + report_fatal_error("symbol '" + B->getSymbol().getName() + + "' can not be undefined in a subtraction expression"); + + // Select the appropriate difference relocation type. + Type = macho::RIT_ARM_HalfDifference; + Value2 = getSymbolAddress(B_SD, Layout); + FixedValue -= getSectionAddress(B_SD->getFragment()->getParent()); + } + + // Relocations are written out in reverse order, so the PAIR comes first. + // ARM_RELOC_HALF and ARM_RELOC_HALF_SECTDIFF abuse the r_length field: + // + // For these two r_type relocations they always have a pair following them + // and the r_length bits are used differently. The encoding of the + // r_length is as follows: + // low bit of r_length: + // 0 - :lower16: for movw instructions + // 1 - :upper16: for movt instructions + // high bit of r_length: + // 0 - arm instructions + // 1 - thumb instructions + // the other half of the relocated expression is in the following pair + // relocation entry in the the low 16 bits of r_address field. + unsigned ThumbBit = 0; + unsigned MovtBit = 0; + switch (Fixup.getKind()) { + default: break; + case ARM::fixup_arm_movt_hi16: + case ARM::fixup_arm_movt_hi16_pcrel: + MovtBit = 1; + break; + case ARM::fixup_t2_movt_hi16: + case ARM::fixup_t2_movt_hi16_pcrel: + MovtBit = 1; + // Fallthrough + case ARM::fixup_t2_movw_lo16: + case ARM::fixup_t2_movw_lo16_pcrel: + ThumbBit = 1; + break; + } + + + if (Type == macho::RIT_ARM_HalfDifference) { + uint32_t OtherHalf = MovtBit + ? (FixedValue & 0xffff) : ((FixedValue & 0xffff0000) >> 16); + + macho::RelocationEntry MRE; + MRE.Word0 = ((OtherHalf << 0) | + (macho::RIT_Pair << 24) | + (MovtBit << 28) | + (ThumbBit << 29) | + (IsPCRel << 30) | + macho::RF_Scattered); + MRE.Word1 = Value2; + Relocations[Fragment->getParent()].push_back(MRE); + } + + macho::RelocationEntry MRE; + MRE.Word0 = ((FixupOffset << 0) | + (Type << 24) | + (MovtBit << 28) | + (ThumbBit << 29) | + (IsPCRel << 30) | + macho::RF_Scattered); MRE.Word1 = Value; Relocations[Fragment->getParent()].push_back(MRE); } @@ -835,7 +875,7 @@ public: const MCFixup &Fixup, MCValue Target, uint64_t &FixedValue) { assert(Target.getSymA()->getKind() == MCSymbolRefExpr::VK_TLVP && - !Is64Bit && + !is64Bit() && "Should only be called with a 32-bit TLVP relocation!"); unsigned Log2Size = getFixupKindLog2Size(Fixup.getKind()); @@ -853,10 +893,10 @@ public: if (Target.getSymB()) { // If this is a subtraction then we're pcrel. uint32_t FixupAddress = - Layout.getFragmentAddress(Fragment) + Fixup.getOffset(); + getFragmentAddress(Fragment, Layout) + Fixup.getOffset(); MCSymbolData *SD_B = &Asm.getSymbolData(Target.getSymB()->getSymbol()); IsPCRel = 1; - FixedValue = (FixupAddress - Layout.getSymbolAddress(SD_B) + + FixedValue = (FixupAddress - getSymbolAddress(SD_B, Layout) + Target.getConstant()); FixedValue += 1ULL << Log2Size; } else { @@ -864,25 +904,192 @@ public: } // struct relocation_info (8 bytes) - MachRelocationEntry MRE; + macho::RelocationEntry MRE; MRE.Word0 = Value; + MRE.Word1 = ((Index << 0) | + (IsPCRel << 24) | + (Log2Size << 25) | + (1 << 27) | // Extern + (macho::RIT_Generic_TLV << 28)); // Type + Relocations[Fragment->getParent()].push_back(MRE); + } + + static bool getARMFixupKindMachOInfo(unsigned Kind, unsigned &RelocType, + unsigned &Log2Size) { + RelocType = unsigned(macho::RIT_Vanilla); + Log2Size = ~0U; + + switch (Kind) { + default: + return false; + + case FK_Data_1: + Log2Size = llvm::Log2_32(1); + return true; + case FK_Data_2: + Log2Size = llvm::Log2_32(2); + return true; + case FK_Data_4: + Log2Size = llvm::Log2_32(4); + return true; + case FK_Data_8: + Log2Size = llvm::Log2_32(8); + return true; + + // Handle 24-bit branch kinds. + case ARM::fixup_arm_ldst_pcrel_12: + case ARM::fixup_arm_pcrel_10: + case ARM::fixup_arm_adr_pcrel_12: + case ARM::fixup_arm_condbranch: + case ARM::fixup_arm_uncondbranch: + RelocType = unsigned(macho::RIT_ARM_Branch24Bit); + // Report as 'long', even though that is not quite accurate. + Log2Size = llvm::Log2_32(4); + return true; + + // Handle Thumb branches. + case ARM::fixup_arm_thumb_br: + RelocType = unsigned(macho::RIT_ARM_ThumbBranch22Bit); + Log2Size = llvm::Log2_32(2); + return true; + + case ARM::fixup_arm_thumb_bl: + RelocType = unsigned(macho::RIT_ARM_ThumbBranch32Bit); + Log2Size = llvm::Log2_32(4); + return true; + + case ARM::fixup_arm_thumb_blx: + RelocType = unsigned(macho::RIT_ARM_ThumbBranch22Bit); + // Report as 'long', even though that is not quite accurate. + Log2Size = llvm::Log2_32(4); + return true; + + case ARM::fixup_arm_movt_hi16: + case ARM::fixup_arm_movt_hi16_pcrel: + case ARM::fixup_t2_movt_hi16: + case ARM::fixup_t2_movt_hi16_pcrel: + RelocType = unsigned(macho::RIT_ARM_HalfDifference); + // Report as 'long', even though that is not quite accurate. + Log2Size = llvm::Log2_32(4); + return true; + + case ARM::fixup_arm_movw_lo16: + case ARM::fixup_arm_movw_lo16_pcrel: + case ARM::fixup_t2_movw_lo16: + case ARM::fixup_t2_movw_lo16_pcrel: + RelocType = unsigned(macho::RIT_ARM_Half); + // Report as 'long', even though that is not quite accurate. + Log2Size = llvm::Log2_32(4); + return true; + } + } + void RecordARMRelocation(const MCAssembler &Asm, const MCAsmLayout &Layout, + const MCFragment *Fragment, const MCFixup &Fixup, + MCValue Target, uint64_t &FixedValue) { + unsigned IsPCRel = isFixupKindPCRel(Asm, Fixup.getKind()); + unsigned Log2Size; + unsigned RelocType = macho::RIT_Vanilla; + if (!getARMFixupKindMachOInfo(Fixup.getKind(), RelocType, Log2Size)) { + report_fatal_error("unknown ARM fixup kind!"); + return; + } + + // If this is a difference or a defined symbol plus an offset, then we need + // a scattered relocation entry. Differences always require scattered + // relocations. + if (Target.getSymB()) { + if (RelocType == macho::RIT_ARM_Half || + RelocType == macho::RIT_ARM_HalfDifference) + return RecordARMMovwMovtRelocation(Asm, Layout, Fragment, Fixup, + Target, FixedValue); + return RecordARMScatteredRelocation(Asm, Layout, Fragment, Fixup, + Target, Log2Size, FixedValue); + } + + // Get the symbol data, if any. + MCSymbolData *SD = 0; + if (Target.getSymA()) + SD = &Asm.getSymbolData(Target.getSymA()->getSymbol()); + + // FIXME: For other platforms, we need to use scattered relocations for + // internal relocations with offsets. If this is an internal relocation + // with an offset, it also needs a scattered relocation entry. + // + // Is this right for ARM? + uint32_t Offset = Target.getConstant(); + if (IsPCRel && RelocType == macho::RIT_Vanilla) + Offset += 1 << Log2Size; + if (Offset && SD && !doesSymbolRequireExternRelocation(SD)) + return RecordARMScatteredRelocation(Asm, Layout, Fragment, Fixup, Target, + Log2Size, FixedValue); + + // See <reloc.h>. + uint32_t FixupOffset = Layout.getFragmentOffset(Fragment)+Fixup.getOffset(); + unsigned Index = 0; + unsigned IsExtern = 0; + unsigned Type = 0; + + if (Target.isAbsolute()) { // constant + // FIXME! + report_fatal_error("FIXME: relocations to absolute targets " + "not yet implemented"); + } else if (SD->getSymbol().isVariable()) { + int64_t Res; + if (SD->getSymbol().getVariableValue()->EvaluateAsAbsolute( + Res, Layout, SectionAddress)) { + FixedValue = Res; + return; + } + + report_fatal_error("unsupported relocation of variable '" + + SD->getSymbol().getName() + "'"); + } else { + // Check whether we need an external or internal relocation. + if (doesSymbolRequireExternRelocation(SD)) { + IsExtern = 1; + Index = SD->getIndex(); + // For external relocations, make sure to offset the fixup value to + // compensate for the addend of the symbol address, if it was + // undefined. This occurs with weak definitions, for example. + if (!SD->Symbol->isUndefined()) + FixedValue -= Layout.getSymbolOffset(SD); + } else { + // The index is the section ordinal (1-based). + Index = SD->getFragment()->getParent()->getOrdinal() + 1; + FixedValue += getSectionAddress(SD->getFragment()->getParent()); + } + if (IsPCRel) + FixedValue -= getSectionAddress(Fragment->getParent()); + + // The type is determined by the fixup kind. + Type = RelocType; + } + + // struct relocation_info (8 bytes) + macho::RelocationEntry MRE; + MRE.Word0 = FixupOffset; MRE.Word1 = ((Index << 0) | (IsPCRel << 24) | (Log2Size << 25) | - (1 << 27) | // Extern - (RIT_TLV << 28)); // Type + (IsExtern << 27) | + (Type << 28)); Relocations[Fragment->getParent()].push_back(MRE); } void RecordRelocation(const MCAssembler &Asm, const MCAsmLayout &Layout, const MCFragment *Fragment, const MCFixup &Fixup, MCValue Target, uint64_t &FixedValue) { - if (Is64Bit) { + // FIXME: These needs to be factored into the target Mach-O writer. + if (isARM()) { + RecordARMRelocation(Asm, Layout, Fragment, Fixup, Target, FixedValue); + return; + } + if (is64Bit()) { RecordX86_64Relocation(Asm, Layout, Fragment, Fixup, Target, FixedValue); return; } - unsigned IsPCRel = isFixupKindPCRel(Fixup.getKind()); + unsigned IsPCRel = isFixupKindPCRel(Asm, Fixup.getKind()); unsigned Log2Size = getFixupKindLog2Size(Fixup.getKind()); // If this is a 32-bit TLVP reloc it's handled a bit differently. @@ -897,7 +1104,7 @@ public: // Differences always require scattered relocations. if (Target.getSymB()) return RecordScatteredRelocation(Asm, Layout, Fragment, Fixup, - Target, FixedValue); + Target, Log2Size, FixedValue); // Get the symbol data, if any. MCSymbolData *SD = 0; @@ -911,7 +1118,7 @@ public: Offset += 1 << Log2Size; if (Offset && SD && !doesSymbolRequireExternRelocation(SD)) return RecordScatteredRelocation(Asm, Layout, Fragment, Fixup, - Target, FixedValue); + Target, Log2Size, FixedValue); // See <reloc.h>. uint32_t FixupOffset = Layout.getFragmentOffset(Fragment)+Fixup.getOffset(); @@ -924,7 +1131,17 @@ public: // // FIXME: Currently, these are never generated (see code below). I cannot // find a case where they are actually emitted. - Type = RIT_Vanilla; + Type = macho::RIT_Vanilla; + } else if (SD->getSymbol().isVariable()) { + int64_t Res; + if (SD->getSymbol().getVariableValue()->EvaluateAsAbsolute( + Res, Layout, SectionAddress)) { + FixedValue = Res; + return; + } + + report_fatal_error("unsupported relocation of variable '" + + SD->getSymbol().getName() + "'"); } else { // Check whether we need an external or internal relocation. if (doesSymbolRequireExternRelocation(SD)) { @@ -934,17 +1151,20 @@ public: // compensate for the addend of the symbol address, if it was // undefined. This occurs with weak definitions, for example. if (!SD->Symbol->isUndefined()) - FixedValue -= Layout.getSymbolAddress(SD); + FixedValue -= Layout.getSymbolOffset(SD); } else { // The index is the section ordinal (1-based). Index = SD->getFragment()->getParent()->getOrdinal() + 1; + FixedValue += getSectionAddress(SD->getFragment()->getParent()); } + if (IsPCRel) + FixedValue -= getSectionAddress(Fragment->getParent()); - Type = RIT_Vanilla; + Type = macho::RIT_Vanilla; } // struct relocation_info (8 bytes) - MachRelocationEntry MRE; + macho::RelocationEntry MRE; MRE.Word0 = FixupOffset; MRE.Word1 = ((Index << 0) | (IsPCRel << 24) | @@ -1118,7 +1338,25 @@ public: StringTable += '\x00'; } - void ExecutePostLayoutBinding(MCAssembler &Asm) { + void computeSectionAddresses(const MCAssembler &Asm, + const MCAsmLayout &Layout) { + uint64_t StartAddress = 0; + const SmallVectorImpl<MCSectionData*> &Order = Layout.getSectionOrder(); + for (int i = 0, n = Order.size(); i != n ; ++i) { + const MCSectionData *SD = Order[i]; + StartAddress = RoundUpToAlignment(StartAddress, SD->getAlignment()); + SectionAddress[SD] = StartAddress; + StartAddress += Layout.getSectionAddressSize(SD); + // Explicitly pad the section to match the alignment requirements of the + // following one. This is for 'gas' compatibility, it shouldn't + /// strictly be necessary. + StartAddress += getPaddingSize(SD, Layout); + } + } + + void ExecutePostLayoutBinding(MCAssembler &Asm, const MCAsmLayout &Layout) { + computeSectionAddresses(Asm, Layout); + // Create symbol data for any indirect symbols. BindIndirectSymbols(Asm); @@ -1127,34 +1365,62 @@ public: UndefinedSymbolData); } + virtual bool IsSymbolRefDifferenceFullyResolvedImpl(const MCAssembler &Asm, + const MCSymbolData &DataA, + const MCFragment &FB, + bool InSet, + bool IsPCRel) const { + if (InSet) + return true; + + // The effective address is + // addr(atom(A)) + offset(A) + // - addr(atom(B)) - offset(B) + // and the offsets are not relocatable, so the fixup is fully resolved when + // addr(atom(A)) - addr(atom(B)) == 0. + const MCSymbolData *A_Base = 0, *B_Base = 0; + + const MCSymbol &SA = DataA.getSymbol().AliasedSymbol(); + const MCSection &SecA = SA.getSection(); + const MCSection &SecB = FB.getParent()->getSection(); - bool IsFixupFullyResolved(const MCAssembler &Asm, - const MCValue Target, - bool IsPCRel, - const MCFragment *DF) const { - // If we are using scattered symbols, determine whether this value is - // actually resolved; scattering may cause atoms to move. - if (Asm.getBackend().hasScatteredSymbols()) { - if (Asm.getBackend().hasReliableSymbolDifference()) { - // If this is a PCrel relocation, find the base atom (identified by its - // symbol) that the fixup value is relative to. - const MCSymbolData *BaseSymbol = 0; - if (IsPCRel) { - BaseSymbol = DF->getAtom(); - if (!BaseSymbol) - return false; - } - - return isScatteredFixupFullyResolved(Asm, Target, BaseSymbol); - } else { - const MCSection *BaseSection = 0; - if (IsPCRel) - BaseSection = &DF->getParent()->getSection(); - - return isScatteredFixupFullyResolvedSimple(Asm, Target, BaseSection); + if (IsPCRel) { + // The simple (Darwin, except on x86_64) way of dealing with this was to + // assume that any reference to a temporary symbol *must* be a temporary + // symbol in the same atom, unless the sections differ. Therefore, any + // PCrel relocation to a temporary symbol (in the same section) is fully + // resolved. This also works in conjunction with absolutized .set, which + // requires the compiler to use .set to absolutize the differences between + // symbols which the compiler knows to be assembly time constants, so we + // don't need to worry about considering symbol differences fully + // resolved. + + if (!Asm.getBackend().hasReliableSymbolDifference()) { + if (!SA.isTemporary() || !SA.isInSection() || &SecA != &SecB) + return false; + return true; } + } else { + if (!TargetObjectWriter->useAggressiveSymbolFolding()) + return false; } - return true; + + const MCFragment &FA = *Asm.getSymbolData(SA).getFragment(); + + A_Base = FA.getAtom(); + if (!A_Base) + return false; + + B_Base = FB.getAtom(); + if (!B_Base) + return false; + + // If the atoms are the same, they are guaranteed to have the same address. + if (A_Base == B_Base) + return true; + + // Otherwise, we can't prove this is fully resolved. + return false; } void WriteObject(MCAssembler &Asm, const MCAsmLayout &Layout) { @@ -1163,35 +1429,37 @@ public: // The section data starts after the header, the segment load command (and // section headers) and the symbol table. unsigned NumLoadCommands = 1; - uint64_t LoadCommandsSize = Is64Bit ? - SegmentLoadCommand64Size + NumSections * Section64Size : - SegmentLoadCommand32Size + NumSections * Section32Size; + uint64_t LoadCommandsSize = is64Bit() ? + macho::SegmentLoadCommand64Size + NumSections * macho::Section64Size : + macho::SegmentLoadCommand32Size + NumSections * macho::Section32Size; // Add the symbol table load command sizes, if used. unsigned NumSymbols = LocalSymbolData.size() + ExternalSymbolData.size() + UndefinedSymbolData.size(); if (NumSymbols) { NumLoadCommands += 2; - LoadCommandsSize += SymtabLoadCommandSize + DysymtabLoadCommandSize; + LoadCommandsSize += (macho::SymtabLoadCommandSize + + macho::DysymtabLoadCommandSize); } // Compute the total size of the section data, as well as its file size and // vm size. - uint64_t SectionDataStart = (Is64Bit ? Header64Size : Header32Size) - + LoadCommandsSize; + uint64_t SectionDataStart = (is64Bit() ? macho::Header64Size : + macho::Header32Size) + LoadCommandsSize; uint64_t SectionDataSize = 0; uint64_t SectionDataFileSize = 0; uint64_t VMSize = 0; for (MCAssembler::const_iterator it = Asm.begin(), ie = Asm.end(); it != ie; ++it) { const MCSectionData &SD = *it; - uint64_t Address = Layout.getSectionAddress(&SD); - uint64_t Size = Layout.getSectionSize(&SD); + uint64_t Address = getSectionAddress(&SD); + uint64_t Size = Layout.getSectionAddressSize(&SD); uint64_t FileSize = Layout.getSectionFileSize(&SD); + FileSize += getPaddingSize(&SD, Layout); VMSize = std::max(VMSize, Address + Size); - if (Asm.getBackend().isVirtualSection(SD.getSection())) + if (SD.getSection().isVirtualSection()) continue; SectionDataSize = std::max(SectionDataSize, Address + Size); @@ -1214,11 +1482,11 @@ public: uint64_t RelocTableEnd = SectionDataStart + SectionDataFileSize; for (MCAssembler::const_iterator it = Asm.begin(), ie = Asm.end(); it != ie; ++it) { - std::vector<MachRelocationEntry> &Relocs = Relocations[it]; + std::vector<macho::RelocationEntry> &Relocs = Relocations[it]; unsigned NumRelocs = Relocs.size(); - uint64_t SectionStart = SectionDataStart + Layout.getSectionAddress(it); + uint64_t SectionStart = SectionDataStart + getSectionAddress(it); WriteSection(Asm, Layout, *it, SectionStart, RelocTableEnd, NumRelocs); - RelocTableEnd += NumRelocs * RelocationInfoSize; + RelocTableEnd += NumRelocs * macho::RelocationInfoSize; } // Write the symbol table load command, if used. @@ -1244,8 +1512,8 @@ public: // The string table is written after symbol table. uint64_t StringTableOffset = - SymbolTableOffset + NumSymTabSymbols * (Is64Bit ? Nlist64Size : - Nlist32Size); + SymbolTableOffset + NumSymTabSymbols * (is64Bit() ? macho::Nlist64Size : + macho::Nlist32Size); WriteSymtabLoadCommand(SymbolTableOffset, NumSymTabSymbols, StringTableOffset, StringTable.size()); @@ -1257,8 +1525,13 @@ public: // Write the actual section data. for (MCAssembler::const_iterator it = Asm.begin(), - ie = Asm.end(); it != ie; ++it) - Asm.WriteSectionData(it, Layout, this); + ie = Asm.end(); it != ie; ++it) { + Asm.WriteSectionData(it, Layout); + + uint64_t Pad = getPaddingSize(it, Layout); + for (unsigned int i = 0; i < Pad; ++i) + Write8(0); + } // Write the extra padding. WriteZeros(SectionDataPadding); @@ -1268,7 +1541,7 @@ public: ie = Asm.end(); it != ie; ++it) { // Write the section relocation entries, in reverse order to match 'as' // (approximately, the exact algorithm is more complicated than this). - std::vector<MachRelocationEntry> &Relocs = Relocations[it]; + std::vector<macho::RelocationEntry> &Relocs = Relocations[it]; for (unsigned i = 0, e = Relocs.size(); i != e; ++i) { Write32(Relocs[e - i - 1].Word0); Write32(Relocs[e - i - 1].Word1); @@ -1289,9 +1562,9 @@ public: // If this symbol is defined and internal, mark it as such. if (it->Symbol->isDefined() && !Asm.getSymbolData(*it->Symbol).isExternal()) { - uint32_t Flags = ISF_Local; + uint32_t Flags = macho::ISF_Local; if (it->Symbol->isAbsolute()) - Flags |= ISF_Absolute; + Flags |= macho::ISF_Absolute; Write32(Flags); continue; } @@ -1318,9 +1591,8 @@ public: } -MCObjectWriter *llvm::createMachObjectWriter(raw_ostream &OS, bool is64Bit, - uint32_t CPUType, - uint32_t CPUSubtype, +MCObjectWriter *llvm::createMachObjectWriter(MCMachObjectTargetWriter *MOTW, + raw_ostream &OS, bool IsLittleEndian) { - return new MachObjectWriter(OS, is64Bit, CPUType, CPUSubtype, IsLittleEndian); + return new MachObjectWriter(MOTW, OS, IsLittleEndian); } diff --git a/lib/MC/TargetAsmBackend.cpp b/lib/MC/TargetAsmBackend.cpp index 1f10410..1927557 100644 --- a/lib/MC/TargetAsmBackend.cpp +++ b/lib/MC/TargetAsmBackend.cpp @@ -10,12 +10,28 @@ #include "llvm/Target/TargetAsmBackend.h" using namespace llvm; -TargetAsmBackend::TargetAsmBackend(const Target &T) - : TheTarget(T), - HasReliableSymbolDifference(false), - HasScatteredSymbols(false) +TargetAsmBackend::TargetAsmBackend() + : HasReliableSymbolDifference(false) { } TargetAsmBackend::~TargetAsmBackend() { } + +const MCFixupKindInfo & +TargetAsmBackend::getFixupKindInfo(MCFixupKind Kind) const { + static const MCFixupKindInfo Builtins[] = { + { "FK_Data_1", 0, 8, 0 }, + { "FK_Data_2", 0, 16, 0 }, + { "FK_Data_4", 0, 32, 0 }, + { "FK_Data_8", 0, 64, 0 }, + { "FK_PCRel_1", 0, 8, MCFixupKindInfo::FKF_IsPCRel }, + { "FK_PCRel_2", 0, 16, MCFixupKindInfo::FKF_IsPCRel }, + { "FK_PCRel_4", 0, 32, MCFixupKindInfo::FKF_IsPCRel }, + { "FK_PCRel_8", 0, 64, MCFixupKindInfo::FKF_IsPCRel } + }; + + assert((size_t)Kind <= sizeof(Builtins) / sizeof(Builtins[0]) && + "Unknown fixup kind"); + return Builtins[Kind]; +} diff --git a/lib/MC/WinCOFFObjectWriter.cpp b/lib/MC/WinCOFFObjectWriter.cpp index fd79203..6ca5d37 100644 --- a/lib/MC/WinCOFFObjectWriter.cpp +++ b/lib/MC/WinCOFFObjectWriter.cpp @@ -31,7 +31,7 @@ #include "llvm/Support/Debug.h" #include "llvm/Support/ErrorHandling.h" -#include "llvm/System/TimeValue.h" +#include "llvm/Support/TimeValue.h" #include "../Target/X86/X86FixupKinds.h" @@ -170,7 +170,7 @@ public: // MCObjectWriter interface implementation. - void ExecutePostLayoutBinding(MCAssembler &Asm); + void ExecutePostLayoutBinding(MCAssembler &Asm, const MCAsmLayout &Layout); void RecordRelocation(const MCAssembler &Asm, const MCAsmLayout &Layout, @@ -179,11 +179,6 @@ public: MCValue Target, uint64_t &FixedValue); - virtual bool IsFixupFullyResolved(const MCAssembler &Asm, - const MCValue Target, - bool IsPCRel, - const MCFragment *DF) const; - void WriteObject(MCAssembler &Asm, const MCAsmLayout &Layout); }; } @@ -616,7 +611,8 @@ void WinCOFFObjectWriter::WriteRelocation(const COFF::relocation &R) { //////////////////////////////////////////////////////////////////////////////// // MCObjectWriter interface implementations -void WinCOFFObjectWriter::ExecutePostLayoutBinding(MCAssembler &Asm) { +void WinCOFFObjectWriter::ExecutePostLayoutBinding(MCAssembler &Asm, + const MCAsmLayout &Layout) { // "Define" each section & symbol. This creates section & symbol // entries in the staging area. @@ -661,7 +657,7 @@ void WinCOFFObjectWriter::RecordRelocation(const MCAssembler &Asm, const MCSymbol *B = &Target.getSymB()->getSymbol(); MCSymbolData &B_SD = Asm.getSymbolData(*B); - FixedValue = Layout.getSymbolAddress(&A_SD) - Layout.getSymbolAddress(&B_SD); + FixedValue = Layout.getSymbolOffset(&A_SD) - Layout.getSymbolOffset(&B_SD); // In the case where we have SymbA and SymB, we just need to store the delta // between the two symbols. Update FixedValue to account for the delta, and @@ -689,7 +685,7 @@ void WinCOFFObjectWriter::RecordRelocation(const MCAssembler &Asm, Reloc.Data.VirtualAddress += Fixup.getOffset(); switch ((unsigned)Fixup.getKind()) { - case X86::reloc_pcrel_4byte: + case FK_PCRel_4: case X86::reloc_riprel_4byte: case X86::reloc_riprel_4byte_movq_load: Reloc.Data.Type = Is64Bit ? COFF::IMAGE_REL_AMD64_REL32 @@ -716,36 +712,6 @@ void WinCOFFObjectWriter::RecordRelocation(const MCAssembler &Asm, coff_section->Relocations.push_back(Reloc); } -bool WinCOFFObjectWriter::IsFixupFullyResolved(const MCAssembler &Asm, - const MCValue Target, - bool IsPCRel, - const MCFragment *DF) const { - // If this is a PCrel relocation, find the section this fixup value is - // relative to. - const MCSection *BaseSection = 0; - if (IsPCRel) { - BaseSection = &DF->getParent()->getSection(); - assert(BaseSection); - } - - const MCSection *SectionA = 0; - const MCSymbol *SymbolA = 0; - if (const MCSymbolRefExpr *A = Target.getSymA()) { - SymbolA = &A->getSymbol(); - SectionA = &SymbolA->getSection(); - } - - const MCSection *SectionB = 0; - if (const MCSymbolRefExpr *B = Target.getSymB()) { - SectionB = &B->getSymbol().getSection(); - } - - if (!BaseSection) - return SectionA == SectionB; - - return !SectionB && BaseSection == SectionA; -} - void WinCOFFObjectWriter::WriteObject(MCAssembler &Asm, const MCAsmLayout &Layout) { // Assign symbol and section indexes and offsets. @@ -753,7 +719,7 @@ void WinCOFFObjectWriter::WriteObject(MCAssembler &Asm, for (sections::iterator i = Sections.begin(), e = Sections.end(); i != e; i++) { - if (Layout.getSectionSize((*i)->MCData) > 0) { + if (Layout.getSectionAddressSize((*i)->MCData) > 0) { MakeSectionReal(**i, ++Header.NumberOfSections); } else { (*i)->Number = -1; @@ -873,7 +839,7 @@ void WinCOFFObjectWriter::WriteObject(MCAssembler &Asm, assert(OS.tell() == (*i)->Header.PointerToRawData && "Section::PointerToRawData is insane!"); - Asm.WriteSectionData(j, Layout, this); + Asm.WriteSectionData(j, Layout); } if ((*i)->Relocations.size() > 0) { diff --git a/lib/MC/WinCOFFStreamer.cpp b/lib/MC/WinCOFFStreamer.cpp index 3c5a3be..46968e6 100644 --- a/lib/MC/WinCOFFStreamer.cpp +++ b/lib/MC/WinCOFFStreamer.cpp @@ -68,16 +68,11 @@ public: virtual void EmitTBSSSymbol(const MCSection *Section, MCSymbol *Symbol, uint64_t Size, unsigned ByteAlignment); virtual void EmitBytes(StringRef Data, unsigned AddrSpace); - virtual void EmitValue(const MCExpr *Value, unsigned Size, - unsigned AddrSpace); - virtual void EmitGPRel32Value(const MCExpr *Value); virtual void EmitValueToAlignment(unsigned ByteAlignment, int64_t Value, unsigned ValueSize, unsigned MaxBytesToEmit); virtual void EmitCodeAlignment(unsigned ByteAlignment, unsigned MaxBytesToEmit); - virtual void EmitValueToOffset(const MCExpr *Offset, unsigned char Value); virtual void EmitFileDirective(StringRef Filename); - virtual void EmitDwarfFileDirective(unsigned FileNo,StringRef Filename); virtual void EmitInstruction(const MCInst &Instruction); virtual void Finish(); @@ -129,7 +124,7 @@ WinCOFFStreamer::WinCOFFStreamer(MCContext &Context, TargetAsmBackend &TAB, MCCodeEmitter &CE, raw_ostream &OS) - : MCObjectStreamer(Context, TAB, OS, &CE, true) + : MCObjectStreamer(Context, TAB, OS, &CE) , CurSymbol(NULL) { } @@ -178,25 +173,8 @@ void WinCOFFStreamer::InitSections() { } void WinCOFFStreamer::EmitLabel(MCSymbol *Symbol) { - // TODO: This is copied almost exactly from the MachOStreamer. Consider - // merging into MCObjectStreamer? assert(Symbol->isUndefined() && "Cannot define a symbol twice!"); - assert(!Symbol->isVariable() && "Cannot emit a variable symbol!"); - assert(CurSection && "Cannot emit before setting section!"); - - Symbol->setSection(*CurSection); - - MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol); - - // FIXME: This is wasteful, we don't necessarily need to create a data - // fragment. Instead, we should mark the symbol as pointing into the data - // fragment if it exists, otherwise we should just queue the label and set its - // fragment pointer when we emit the next fragment. - MCDataFragment *DF = getOrCreateDataFragment(); - - assert(!SD.getFragment() && "Unexpected fragment on symbol data!"); - SD.setFragment(DF); - SD.setOffset(DF->getContents().size()); + MCObjectStreamer::EmitLabel(Symbol); } void WinCOFFStreamer::EmitAssemblerFlag(MCAssemblerFlag Flag) { @@ -347,32 +325,6 @@ void WinCOFFStreamer::EmitBytes(StringRef Data, unsigned AddrSpace) { getOrCreateDataFragment()->getContents().append(Data.begin(), Data.end()); } -void WinCOFFStreamer::EmitValue(const MCExpr *Value, unsigned Size, - unsigned AddrSpace) { - assert(AddrSpace == 0 && "Address space must be 0!"); - - // TODO: This is copied exactly from the MachOStreamer. Consider merging into - // MCObjectStreamer? - MCDataFragment *DF = getOrCreateDataFragment(); - - // Avoid fixups when possible. - int64_t AbsValue; - if (AddValueSymbols(Value)->EvaluateAsAbsolute(AbsValue)) { - // FIXME: Endianness assumption. - for (unsigned i = 0; i != Size; ++i) - DF->getContents().push_back(uint8_t(AbsValue >> (i * 8))); - } else { - DF->addFixup(MCFixup::Create(DF->getContents().size(), - AddValueSymbols(Value), - MCFixup::getKindForSize(Size))); - DF->getContents().resize(DF->getContents().size() + Size, 0); - } -} - -void WinCOFFStreamer::EmitGPRel32Value(const MCExpr *Value) { - llvm_unreachable("not implemented"); -} - void WinCOFFStreamer::EmitValueToAlignment(unsigned ByteAlignment, int64_t Value, unsigned ValueSize, @@ -404,21 +356,11 @@ void WinCOFFStreamer::EmitCodeAlignment(unsigned ByteAlignment, getCurrentSectionData()->setAlignment(ByteAlignment); } -void WinCOFFStreamer::EmitValueToOffset(const MCExpr *Offset, - unsigned char Value) { - llvm_unreachable("not implemented"); -} - void WinCOFFStreamer::EmitFileDirective(StringRef Filename) { // Ignore for now, linkers don't care, and proper debug // info will be a much large effort. } -void WinCOFFStreamer::EmitDwarfFileDirective(unsigned FileNo, - StringRef Filename) { - llvm_unreachable("not implemented"); -} - void WinCOFFStreamer::EmitInstruction(const MCInst &Instruction) { for (unsigned i = 0, e = Instruction.getNumOperands(); i != e; ++i) if (Instruction.getOperand(i).isExpr()) |