diff options
author | ed <ed@FreeBSD.org> | 2009-06-02 17:52:33 +0000 |
---|---|---|
committer | ed <ed@FreeBSD.org> | 2009-06-02 17:52:33 +0000 |
commit | 3277b69d734b9c90b44ebde4ede005717e2c3b2e (patch) | |
tree | 64ba909838c23261cace781ece27d106134ea451 /lib/Bitcode | |
download | FreeBSD-src-3277b69d734b9c90b44ebde4ede005717e2c3b2e.zip FreeBSD-src-3277b69d734b9c90b44ebde4ede005717e2c3b2e.tar.gz |
Import LLVM, at r72732.
Diffstat (limited to 'lib/Bitcode')
-rw-r--r-- | lib/Bitcode/Makefile | 14 | ||||
-rw-r--r-- | lib/Bitcode/Reader/BitReader.cpp | 51 | ||||
-rw-r--r-- | lib/Bitcode/Reader/BitcodeReader.cpp | 2126 | ||||
-rw-r--r-- | lib/Bitcode/Reader/BitcodeReader.h | 214 | ||||
-rw-r--r-- | lib/Bitcode/Reader/CMakeLists.txt | 7 | ||||
-rw-r--r-- | lib/Bitcode/Reader/Deserialize.cpp | 454 | ||||
-rw-r--r-- | lib/Bitcode/Reader/DeserializeAPFloat.cpp | 24 | ||||
-rw-r--r-- | lib/Bitcode/Reader/DeserializeAPInt.cpp | 33 | ||||
-rw-r--r-- | lib/Bitcode/Reader/Makefile | 15 | ||||
-rw-r--r-- | lib/Bitcode/Writer/BitWriter.cpp | 58 | ||||
-rw-r--r-- | lib/Bitcode/Writer/BitcodeWriter.cpp | 1449 | ||||
-rw-r--r-- | lib/Bitcode/Writer/BitcodeWriterPass.cpp | 56 | ||||
-rw-r--r-- | lib/Bitcode/Writer/CMakeLists.txt | 9 | ||||
-rw-r--r-- | lib/Bitcode/Writer/Makefile | 15 | ||||
-rw-r--r-- | lib/Bitcode/Writer/Serialize.cpp | 118 | ||||
-rw-r--r-- | lib/Bitcode/Writer/SerializeAPFloat.cpp | 21 | ||||
-rw-r--r-- | lib/Bitcode/Writer/SerializeAPInt.cpp | 31 | ||||
-rw-r--r-- | lib/Bitcode/Writer/ValueEnumerator.cpp | 347 | ||||
-rw-r--r-- | lib/Bitcode/Writer/ValueEnumerator.h | 127 |
19 files changed, 5169 insertions, 0 deletions
diff --git a/lib/Bitcode/Makefile b/lib/Bitcode/Makefile new file mode 100644 index 0000000..2d6b5ad --- /dev/null +++ b/lib/Bitcode/Makefile @@ -0,0 +1,14 @@ +##===- lib/Bitcode/Makefile --------------------------------*- Makefile -*-===## +# +# The LLVM Compiler Infrastructure +# +# This file is distributed under the University of Illinois Open Source +# License. See LICENSE.TXT for details. +# +##===----------------------------------------------------------------------===## + +LEVEL = ../.. +PARALLEL_DIRS = Reader Writer + +include $(LEVEL)/Makefile.common + diff --git a/lib/Bitcode/Reader/BitReader.cpp b/lib/Bitcode/Reader/BitReader.cpp new file mode 100644 index 0000000..52851cd --- /dev/null +++ b/lib/Bitcode/Reader/BitReader.cpp @@ -0,0 +1,51 @@ +//===-- BitReader.cpp -----------------------------------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +#include "llvm-c/BitReader.h" +#include "llvm/Bitcode/ReaderWriter.h" +#include "llvm/Support/MemoryBuffer.h" +#include <string> +#include <cstring> + +using namespace llvm; + +/* Builds a module from the bitcode in the specified memory buffer, returning a + reference to the module via the OutModule parameter. Returns 0 on success. + Optionally returns a human-readable error message via OutMessage. */ +int LLVMParseBitcode(LLVMMemoryBufferRef MemBuf, + LLVMModuleRef *OutModule, char **OutMessage) { + std::string Message; + + *OutModule = wrap(ParseBitcodeFile(unwrap(MemBuf), &Message)); + if (!*OutModule) { + if (OutMessage) + *OutMessage = strdup(Message.c_str()); + return 1; + } + + return 0; +} + +/* Reads a module from the specified path, returning via the OutModule parameter + a module provider which performs lazy deserialization. Returns 0 on success. + Optionally returns a human-readable error message via OutMessage. */ +int LLVMGetBitcodeModuleProvider(LLVMMemoryBufferRef MemBuf, + LLVMModuleProviderRef *OutMP, + char **OutMessage) { + std::string Message; + + *OutMP = wrap(getBitcodeModuleProvider(unwrap(MemBuf), &Message)); + if (!*OutMP) { + if (OutMessage) + *OutMessage = strdup(Message.c_str()); + return 1; + } + + return 0; +} diff --git a/lib/Bitcode/Reader/BitcodeReader.cpp b/lib/Bitcode/Reader/BitcodeReader.cpp new file mode 100644 index 0000000..1dad04b --- /dev/null +++ b/lib/Bitcode/Reader/BitcodeReader.cpp @@ -0,0 +1,2126 @@ +//===- BitcodeReader.cpp - Internal BitcodeReader implementation ----------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This header defines the BitcodeReader class. +// +//===----------------------------------------------------------------------===// + +#include "llvm/Bitcode/ReaderWriter.h" +#include "BitcodeReader.h" +#include "llvm/Constants.h" +#include "llvm/DerivedTypes.h" +#include "llvm/InlineAsm.h" +#include "llvm/Instructions.h" +#include "llvm/MDNode.h" +#include "llvm/Module.h" +#include "llvm/AutoUpgrade.h" +#include "llvm/ADT/SmallString.h" +#include "llvm/ADT/SmallVector.h" +#include "llvm/Support/MathExtras.h" +#include "llvm/Support/MemoryBuffer.h" +#include "llvm/OperandTraits.h" +using namespace llvm; + +void BitcodeReader::FreeState() { + delete Buffer; + Buffer = 0; + std::vector<PATypeHolder>().swap(TypeList); + ValueList.clear(); + + std::vector<AttrListPtr>().swap(MAttributes); + std::vector<BasicBlock*>().swap(FunctionBBs); + std::vector<Function*>().swap(FunctionsWithBodies); + DeferredFunctionInfo.clear(); +} + +//===----------------------------------------------------------------------===// +// Helper functions to implement forward reference resolution, etc. +//===----------------------------------------------------------------------===// + +/// ConvertToString - Convert a string from a record into an std::string, return +/// true on failure. +template<typename StrTy> +static bool ConvertToString(SmallVector<uint64_t, 64> &Record, unsigned Idx, + StrTy &Result) { + if (Idx > Record.size()) + return true; + + for (unsigned i = Idx, e = Record.size(); i != e; ++i) + Result += (char)Record[i]; + return false; +} + +static GlobalValue::LinkageTypes GetDecodedLinkage(unsigned Val) { + switch (Val) { + default: // Map unknown/new linkages to external + case 0: return GlobalValue::ExternalLinkage; + case 1: return GlobalValue::WeakAnyLinkage; + case 2: return GlobalValue::AppendingLinkage; + case 3: return GlobalValue::InternalLinkage; + case 4: return GlobalValue::LinkOnceAnyLinkage; + case 5: return GlobalValue::DLLImportLinkage; + case 6: return GlobalValue::DLLExportLinkage; + case 7: return GlobalValue::ExternalWeakLinkage; + case 8: return GlobalValue::CommonLinkage; + case 9: return GlobalValue::PrivateLinkage; + case 10: return GlobalValue::WeakODRLinkage; + case 11: return GlobalValue::LinkOnceODRLinkage; + case 12: return GlobalValue::AvailableExternallyLinkage; + } +} + +static GlobalValue::VisibilityTypes GetDecodedVisibility(unsigned Val) { + switch (Val) { + default: // Map unknown visibilities to default. + case 0: return GlobalValue::DefaultVisibility; + case 1: return GlobalValue::HiddenVisibility; + case 2: return GlobalValue::ProtectedVisibility; + } +} + +static int GetDecodedCastOpcode(unsigned Val) { + switch (Val) { + default: return -1; + case bitc::CAST_TRUNC : return Instruction::Trunc; + case bitc::CAST_ZEXT : return Instruction::ZExt; + case bitc::CAST_SEXT : return Instruction::SExt; + case bitc::CAST_FPTOUI : return Instruction::FPToUI; + case bitc::CAST_FPTOSI : return Instruction::FPToSI; + case bitc::CAST_UITOFP : return Instruction::UIToFP; + case bitc::CAST_SITOFP : return Instruction::SIToFP; + case bitc::CAST_FPTRUNC : return Instruction::FPTrunc; + case bitc::CAST_FPEXT : return Instruction::FPExt; + case bitc::CAST_PTRTOINT: return Instruction::PtrToInt; + case bitc::CAST_INTTOPTR: return Instruction::IntToPtr; + case bitc::CAST_BITCAST : return Instruction::BitCast; + } +} +static int GetDecodedBinaryOpcode(unsigned Val, const Type *Ty) { + switch (Val) { + default: return -1; + case bitc::BINOP_ADD: return Instruction::Add; + case bitc::BINOP_SUB: return Instruction::Sub; + case bitc::BINOP_MUL: return Instruction::Mul; + case bitc::BINOP_UDIV: return Instruction::UDiv; + case bitc::BINOP_SDIV: + return Ty->isFPOrFPVector() ? Instruction::FDiv : Instruction::SDiv; + case bitc::BINOP_UREM: return Instruction::URem; + case bitc::BINOP_SREM: + return Ty->isFPOrFPVector() ? Instruction::FRem : Instruction::SRem; + case bitc::BINOP_SHL: return Instruction::Shl; + case bitc::BINOP_LSHR: return Instruction::LShr; + case bitc::BINOP_ASHR: return Instruction::AShr; + case bitc::BINOP_AND: return Instruction::And; + case bitc::BINOP_OR: return Instruction::Or; + case bitc::BINOP_XOR: return Instruction::Xor; + } +} + +namespace llvm { +namespace { + /// @brief A class for maintaining the slot number definition + /// as a placeholder for the actual definition for forward constants defs. + class ConstantPlaceHolder : public ConstantExpr { + ConstantPlaceHolder(); // DO NOT IMPLEMENT + void operator=(const ConstantPlaceHolder &); // DO NOT IMPLEMENT + public: + // allocate space for exactly one operand + void *operator new(size_t s) { + return User::operator new(s, 1); + } + explicit ConstantPlaceHolder(const Type *Ty) + : ConstantExpr(Ty, Instruction::UserOp1, &Op<0>(), 1) { + Op<0>() = UndefValue::get(Type::Int32Ty); + } + + /// @brief Methods to support type inquiry through isa, cast, and dyn_cast. + static inline bool classof(const ConstantPlaceHolder *) { return true; } + static bool classof(const Value *V) { + return isa<ConstantExpr>(V) && + cast<ConstantExpr>(V)->getOpcode() == Instruction::UserOp1; + } + + + /// Provide fast operand accessors + //DECLARE_TRANSPARENT_OPERAND_ACCESSORS(Value); + }; +} + +// FIXME: can we inherit this from ConstantExpr? +template <> +struct OperandTraits<ConstantPlaceHolder> : FixedNumOperandTraits<1> { +}; +} + + +void BitcodeReaderValueList::AssignValue(Value *V, unsigned Idx) { + if (Idx == size()) { + push_back(V); + return; + } + + if (Idx >= size()) + resize(Idx+1); + + WeakVH &OldV = ValuePtrs[Idx]; + if (OldV == 0) { + OldV = V; + return; + } + + // Handle constants and non-constants (e.g. instrs) differently for + // efficiency. + if (Constant *PHC = dyn_cast<Constant>(&*OldV)) { + ResolveConstants.push_back(std::make_pair(PHC, Idx)); + OldV = V; + } else { + // If there was a forward reference to this value, replace it. + Value *PrevVal = OldV; + OldV->replaceAllUsesWith(V); + delete PrevVal; + } +} + + +Constant *BitcodeReaderValueList::getConstantFwdRef(unsigned Idx, + const Type *Ty) { + if (Idx >= size()) + resize(Idx + 1); + + if (Value *V = ValuePtrs[Idx]) { + assert(Ty == V->getType() && "Type mismatch in constant table!"); + return cast<Constant>(V); + } + + // Create and return a placeholder, which will later be RAUW'd. + Constant *C = new ConstantPlaceHolder(Ty); + ValuePtrs[Idx] = C; + return C; +} + +Value *BitcodeReaderValueList::getValueFwdRef(unsigned Idx, const Type *Ty) { + if (Idx >= size()) + resize(Idx + 1); + + if (Value *V = ValuePtrs[Idx]) { + assert((Ty == 0 || Ty == V->getType()) && "Type mismatch in value table!"); + return V; + } + + // No type specified, must be invalid reference. + if (Ty == 0) return 0; + + // Create and return a placeholder, which will later be RAUW'd. + Value *V = new Argument(Ty); + ValuePtrs[Idx] = V; + return V; +} + +/// ResolveConstantForwardRefs - Once all constants are read, this method bulk +/// resolves any forward references. The idea behind this is that we sometimes +/// get constants (such as large arrays) which reference *many* forward ref +/// constants. Replacing each of these causes a lot of thrashing when +/// building/reuniquing the constant. Instead of doing this, we look at all the +/// uses and rewrite all the place holders at once for any constant that uses +/// a placeholder. +void BitcodeReaderValueList::ResolveConstantForwardRefs() { + // Sort the values by-pointer so that they are efficient to look up with a + // binary search. + std::sort(ResolveConstants.begin(), ResolveConstants.end()); + + SmallVector<Constant*, 64> NewOps; + + while (!ResolveConstants.empty()) { + Value *RealVal = operator[](ResolveConstants.back().second); + Constant *Placeholder = ResolveConstants.back().first; + ResolveConstants.pop_back(); + + // Loop over all users of the placeholder, updating them to reference the + // new value. If they reference more than one placeholder, update them all + // at once. + while (!Placeholder->use_empty()) { + Value::use_iterator UI = Placeholder->use_begin(); + + // If the using object isn't uniqued, just update the operands. This + // handles instructions and initializers for global variables. + if (!isa<Constant>(*UI) || isa<GlobalValue>(*UI)) { + UI.getUse().set(RealVal); + continue; + } + + // Otherwise, we have a constant that uses the placeholder. Replace that + // constant with a new constant that has *all* placeholder uses updated. + Constant *UserC = cast<Constant>(*UI); + for (User::op_iterator I = UserC->op_begin(), E = UserC->op_end(); + I != E; ++I) { + Value *NewOp; + if (!isa<ConstantPlaceHolder>(*I)) { + // Not a placeholder reference. + NewOp = *I; + } else if (*I == Placeholder) { + // Common case is that it just references this one placeholder. + NewOp = RealVal; + } else { + // Otherwise, look up the placeholder in ResolveConstants. + ResolveConstantsTy::iterator It = + std::lower_bound(ResolveConstants.begin(), ResolveConstants.end(), + std::pair<Constant*, unsigned>(cast<Constant>(*I), + 0)); + assert(It != ResolveConstants.end() && It->first == *I); + NewOp = operator[](It->second); + } + + NewOps.push_back(cast<Constant>(NewOp)); + } + + // Make the new constant. + Constant *NewC; + if (ConstantArray *UserCA = dyn_cast<ConstantArray>(UserC)) { + NewC = ConstantArray::get(UserCA->getType(), &NewOps[0], NewOps.size()); + } else if (ConstantStruct *UserCS = dyn_cast<ConstantStruct>(UserC)) { + NewC = ConstantStruct::get(&NewOps[0], NewOps.size(), + UserCS->getType()->isPacked()); + } else if (isa<ConstantVector>(UserC)) { + NewC = ConstantVector::get(&NewOps[0], NewOps.size()); + } else { + assert(isa<ConstantExpr>(UserC) && "Must be a ConstantExpr."); + NewC = cast<ConstantExpr>(UserC)->getWithOperands(&NewOps[0], + NewOps.size()); + } + + UserC->replaceAllUsesWith(NewC); + UserC->destroyConstant(); + NewOps.clear(); + } + + // Update all ValueHandles, they should be the only users at this point. + Placeholder->replaceAllUsesWith(RealVal); + delete Placeholder; + } +} + + +const Type *BitcodeReader::getTypeByID(unsigned ID, bool isTypeTable) { + // If the TypeID is in range, return it. + if (ID < TypeList.size()) + return TypeList[ID].get(); + if (!isTypeTable) return 0; + + // The type table allows forward references. Push as many Opaque types as + // needed to get up to ID. + while (TypeList.size() <= ID) + TypeList.push_back(OpaqueType::get()); + return TypeList.back().get(); +} + +//===----------------------------------------------------------------------===// +// Functions for parsing blocks from the bitcode file +//===----------------------------------------------------------------------===// + +bool BitcodeReader::ParseAttributeBlock() { + if (Stream.EnterSubBlock(bitc::PARAMATTR_BLOCK_ID)) + return Error("Malformed block record"); + + if (!MAttributes.empty()) + return Error("Multiple PARAMATTR blocks found!"); + + SmallVector<uint64_t, 64> Record; + + SmallVector<AttributeWithIndex, 8> Attrs; + + // Read all the records. + while (1) { + unsigned Code = Stream.ReadCode(); + if (Code == bitc::END_BLOCK) { + if (Stream.ReadBlockEnd()) + return Error("Error at end of PARAMATTR block"); + return false; + } + + if (Code == bitc::ENTER_SUBBLOCK) { + // No known subblocks, always skip them. + Stream.ReadSubBlockID(); + if (Stream.SkipBlock()) + return Error("Malformed block record"); + continue; + } + + if (Code == bitc::DEFINE_ABBREV) { + Stream.ReadAbbrevRecord(); + continue; + } + + // Read a record. + Record.clear(); + switch (Stream.ReadRecord(Code, Record)) { + default: // Default behavior: ignore. + break; + case bitc::PARAMATTR_CODE_ENTRY: { // ENTRY: [paramidx0, attr0, ...] + if (Record.size() & 1) + return Error("Invalid ENTRY record"); + + // FIXME : Remove this autoupgrade code in LLVM 3.0. + // If Function attributes are using index 0 then transfer them + // to index ~0. Index 0 is used for return value attributes but used to be + // used for function attributes. + Attributes RetAttribute = Attribute::None; + Attributes FnAttribute = Attribute::None; + for (unsigned i = 0, e = Record.size(); i != e; i += 2) { + // FIXME: remove in LLVM 3.0 + // The alignment is stored as a 16-bit raw value from bits 31--16. + // We shift the bits above 31 down by 11 bits. + + unsigned Alignment = (Record[i+1] & (0xffffull << 16)) >> 16; + if (Alignment && !isPowerOf2_32(Alignment)) + return Error("Alignment is not a power of two."); + + Attributes ReconstitutedAttr = Record[i+1] & 0xffff; + if (Alignment) + ReconstitutedAttr |= Attribute::constructAlignmentFromInt(Alignment); + ReconstitutedAttr |= (Record[i+1] & (0xffffull << 32)) >> 11; + Record[i+1] = ReconstitutedAttr; + + if (Record[i] == 0) + RetAttribute = Record[i+1]; + else if (Record[i] == ~0U) + FnAttribute = Record[i+1]; + } + + unsigned OldRetAttrs = (Attribute::NoUnwind|Attribute::NoReturn| + Attribute::ReadOnly|Attribute::ReadNone); + + if (FnAttribute == Attribute::None && RetAttribute != Attribute::None && + (RetAttribute & OldRetAttrs) != 0) { + if (FnAttribute == Attribute::None) { // add a slot so they get added. + Record.push_back(~0U); + Record.push_back(0); + } + + FnAttribute |= RetAttribute & OldRetAttrs; + RetAttribute &= ~OldRetAttrs; + } + + for (unsigned i = 0, e = Record.size(); i != e; i += 2) { + if (Record[i] == 0) { + if (RetAttribute != Attribute::None) + Attrs.push_back(AttributeWithIndex::get(0, RetAttribute)); + } else if (Record[i] == ~0U) { + if (FnAttribute != Attribute::None) + Attrs.push_back(AttributeWithIndex::get(~0U, FnAttribute)); + } else if (Record[i+1] != Attribute::None) + Attrs.push_back(AttributeWithIndex::get(Record[i], Record[i+1])); + } + + MAttributes.push_back(AttrListPtr::get(Attrs.begin(), Attrs.end())); + Attrs.clear(); + break; + } + } + } +} + + +bool BitcodeReader::ParseTypeTable() { + if (Stream.EnterSubBlock(bitc::TYPE_BLOCK_ID)) + return Error("Malformed block record"); + + if (!TypeList.empty()) + return Error("Multiple TYPE_BLOCKs found!"); + + SmallVector<uint64_t, 64> Record; + unsigned NumRecords = 0; + + // Read all the records for this type table. + while (1) { + unsigned Code = Stream.ReadCode(); + if (Code == bitc::END_BLOCK) { + if (NumRecords != TypeList.size()) + return Error("Invalid type forward reference in TYPE_BLOCK"); + if (Stream.ReadBlockEnd()) + return Error("Error at end of type table block"); + return false; + } + + if (Code == bitc::ENTER_SUBBLOCK) { + // No known subblocks, always skip them. + Stream.ReadSubBlockID(); + if (Stream.SkipBlock()) + return Error("Malformed block record"); + continue; + } + + if (Code == bitc::DEFINE_ABBREV) { + Stream.ReadAbbrevRecord(); + continue; + } + + // Read a record. + Record.clear(); + const Type *ResultTy = 0; + switch (Stream.ReadRecord(Code, Record)) { + default: // Default behavior: unknown type. + ResultTy = 0; + break; + case bitc::TYPE_CODE_NUMENTRY: // TYPE_CODE_NUMENTRY: [numentries] + // TYPE_CODE_NUMENTRY contains a count of the number of types in the + // type list. This allows us to reserve space. + if (Record.size() < 1) + return Error("Invalid TYPE_CODE_NUMENTRY record"); + TypeList.reserve(Record[0]); + continue; + case bitc::TYPE_CODE_VOID: // VOID + ResultTy = Type::VoidTy; + break; + case bitc::TYPE_CODE_FLOAT: // FLOAT + ResultTy = Type::FloatTy; + break; + case bitc::TYPE_CODE_DOUBLE: // DOUBLE + ResultTy = Type::DoubleTy; + break; + case bitc::TYPE_CODE_X86_FP80: // X86_FP80 + ResultTy = Type::X86_FP80Ty; + break; + case bitc::TYPE_CODE_FP128: // FP128 + ResultTy = Type::FP128Ty; + break; + case bitc::TYPE_CODE_PPC_FP128: // PPC_FP128 + ResultTy = Type::PPC_FP128Ty; + break; + case bitc::TYPE_CODE_LABEL: // LABEL + ResultTy = Type::LabelTy; + break; + case bitc::TYPE_CODE_OPAQUE: // OPAQUE + ResultTy = 0; + break; + case bitc::TYPE_CODE_METADATA: // METADATA + ResultTy = Type::MetadataTy; + break; + case bitc::TYPE_CODE_INTEGER: // INTEGER: [width] + if (Record.size() < 1) + return Error("Invalid Integer type record"); + + ResultTy = IntegerType::get(Record[0]); + break; + case bitc::TYPE_CODE_POINTER: { // POINTER: [pointee type] or + // [pointee type, address space] + if (Record.size() < 1) + return Error("Invalid POINTER type record"); + unsigned AddressSpace = 0; + if (Record.size() == 2) + AddressSpace = Record[1]; + ResultTy = PointerType::get(getTypeByID(Record[0], true), AddressSpace); + break; + } + case bitc::TYPE_CODE_FUNCTION: { + // FIXME: attrid is dead, remove it in LLVM 3.0 + // FUNCTION: [vararg, attrid, retty, paramty x N] + if (Record.size() < 3) + return Error("Invalid FUNCTION type record"); + std::vector<const Type*> ArgTys; + for (unsigned i = 3, e = Record.size(); i != e; ++i) + ArgTys.push_back(getTypeByID(Record[i], true)); + + ResultTy = FunctionType::get(getTypeByID(Record[2], true), ArgTys, + Record[0]); + break; + } + case bitc::TYPE_CODE_STRUCT: { // STRUCT: [ispacked, eltty x N] + if (Record.size() < 1) + return Error("Invalid STRUCT type record"); + std::vector<const Type*> EltTys; + for (unsigned i = 1, e = Record.size(); i != e; ++i) + EltTys.push_back(getTypeByID(Record[i], true)); + ResultTy = StructType::get(EltTys, Record[0]); + break; + } + case bitc::TYPE_CODE_ARRAY: // ARRAY: [numelts, eltty] + if (Record.size() < 2) + return Error("Invalid ARRAY type record"); + ResultTy = ArrayType::get(getTypeByID(Record[1], true), Record[0]); + break; + case bitc::TYPE_CODE_VECTOR: // VECTOR: [numelts, eltty] + if (Record.size() < 2) + return Error("Invalid VECTOR type record"); + ResultTy = VectorType::get(getTypeByID(Record[1], true), Record[0]); + break; + } + + if (NumRecords == TypeList.size()) { + // If this is a new type slot, just append it. + TypeList.push_back(ResultTy ? ResultTy : OpaqueType::get()); + ++NumRecords; + } else if (ResultTy == 0) { + // Otherwise, this was forward referenced, so an opaque type was created, + // but the result type is actually just an opaque. Leave the one we + // created previously. + ++NumRecords; + } else { + // Otherwise, this was forward referenced, so an opaque type was created. + // Resolve the opaque type to the real type now. + assert(NumRecords < TypeList.size() && "Typelist imbalance"); + const OpaqueType *OldTy = cast<OpaqueType>(TypeList[NumRecords++].get()); + + // Don't directly push the new type on the Tab. Instead we want to replace + // the opaque type we previously inserted with the new concrete value. The + // refinement from the abstract (opaque) type to the new type causes all + // uses of the abstract type to use the concrete type (NewTy). This will + // also cause the opaque type to be deleted. + const_cast<OpaqueType*>(OldTy)->refineAbstractTypeTo(ResultTy); + + // This should have replaced the old opaque type with the new type in the + // value table... or with a preexisting type that was already in the + // system. Let's just make sure it did. + assert(TypeList[NumRecords-1].get() != OldTy && + "refineAbstractType didn't work!"); + } + } +} + + +bool BitcodeReader::ParseTypeSymbolTable() { + if (Stream.EnterSubBlock(bitc::TYPE_SYMTAB_BLOCK_ID)) + return Error("Malformed block record"); + + SmallVector<uint64_t, 64> Record; + + // Read all the records for this type table. + std::string TypeName; + while (1) { + unsigned Code = Stream.ReadCode(); + if (Code == bitc::END_BLOCK) { + if (Stream.ReadBlockEnd()) + return Error("Error at end of type symbol table block"); + return false; + } + + if (Code == bitc::ENTER_SUBBLOCK) { + // No known subblocks, always skip them. + Stream.ReadSubBlockID(); + if (Stream.SkipBlock()) + return Error("Malformed block record"); + continue; + } + + if (Code == bitc::DEFINE_ABBREV) { + Stream.ReadAbbrevRecord(); + continue; + } + + // Read a record. + Record.clear(); + switch (Stream.ReadRecord(Code, Record)) { + default: // Default behavior: unknown type. + break; + case bitc::TST_CODE_ENTRY: // TST_ENTRY: [typeid, namechar x N] + if (ConvertToString(Record, 1, TypeName)) + return Error("Invalid TST_ENTRY record"); + unsigned TypeID = Record[0]; + if (TypeID >= TypeList.size()) + return Error("Invalid Type ID in TST_ENTRY record"); + + TheModule->addTypeName(TypeName, TypeList[TypeID].get()); + TypeName.clear(); + break; + } + } +} + +bool BitcodeReader::ParseValueSymbolTable() { + if (Stream.EnterSubBlock(bitc::VALUE_SYMTAB_BLOCK_ID)) + return Error("Malformed block record"); + + SmallVector<uint64_t, 64> Record; + + // Read all the records for this value table. + SmallString<128> ValueName; + while (1) { + unsigned Code = Stream.ReadCode(); + if (Code == bitc::END_BLOCK) { + if (Stream.ReadBlockEnd()) + return Error("Error at end of value symbol table block"); + return false; + } + if (Code == bitc::ENTER_SUBBLOCK) { + // No known subblocks, always skip them. + Stream.ReadSubBlockID(); + if (Stream.SkipBlock()) + return Error("Malformed block record"); + continue; + } + + if (Code == bitc::DEFINE_ABBREV) { + Stream.ReadAbbrevRecord(); + continue; + } + + // Read a record. + Record.clear(); + switch (Stream.ReadRecord(Code, Record)) { + default: // Default behavior: unknown type. + break; + case bitc::VST_CODE_ENTRY: { // VST_ENTRY: [valueid, namechar x N] + if (ConvertToString(Record, 1, ValueName)) + return Error("Invalid VST_ENTRY record"); + unsigned ValueID = Record[0]; + if (ValueID >= ValueList.size()) + return Error("Invalid Value ID in VST_ENTRY record"); + Value *V = ValueList[ValueID]; + + V->setName(&ValueName[0], ValueName.size()); + ValueName.clear(); + break; + } + case bitc::VST_CODE_BBENTRY: { + if (ConvertToString(Record, 1, ValueName)) + return Error("Invalid VST_BBENTRY record"); + BasicBlock *BB = getBasicBlock(Record[0]); + if (BB == 0) + return Error("Invalid BB ID in VST_BBENTRY record"); + + BB->setName(&ValueName[0], ValueName.size()); + ValueName.clear(); + break; + } + } + } +} + +/// DecodeSignRotatedValue - Decode a signed value stored with the sign bit in +/// the LSB for dense VBR encoding. +static uint64_t DecodeSignRotatedValue(uint64_t V) { + if ((V & 1) == 0) + return V >> 1; + if (V != 1) + return -(V >> 1); + // There is no such thing as -0 with integers. "-0" really means MININT. + return 1ULL << 63; +} + +/// ResolveGlobalAndAliasInits - Resolve all of the initializers for global +/// values and aliases that we can. +bool BitcodeReader::ResolveGlobalAndAliasInits() { + std::vector<std::pair<GlobalVariable*, unsigned> > GlobalInitWorklist; + std::vector<std::pair<GlobalAlias*, unsigned> > AliasInitWorklist; + + GlobalInitWorklist.swap(GlobalInits); + AliasInitWorklist.swap(AliasInits); + + while (!GlobalInitWorklist.empty()) { + unsigned ValID = GlobalInitWorklist.back().second; + if (ValID >= ValueList.size()) { + // Not ready to resolve this yet, it requires something later in the file. + GlobalInits.push_back(GlobalInitWorklist.back()); + } else { + if (Constant *C = dyn_cast<Constant>(ValueList[ValID])) + GlobalInitWorklist.back().first->setInitializer(C); + else + return Error("Global variable initializer is not a constant!"); + } + GlobalInitWorklist.pop_back(); + } + + while (!AliasInitWorklist.empty()) { + unsigned ValID = AliasInitWorklist.back().second; + if (ValID >= ValueList.size()) { + AliasInits.push_back(AliasInitWorklist.back()); + } else { + if (Constant *C = dyn_cast<Constant>(ValueList[ValID])) + AliasInitWorklist.back().first->setAliasee(C); + else + return Error("Alias initializer is not a constant!"); + } + AliasInitWorklist.pop_back(); + } + return false; +} + + +bool BitcodeReader::ParseConstants() { + if (Stream.EnterSubBlock(bitc::CONSTANTS_BLOCK_ID)) + return Error("Malformed block record"); + + SmallVector<uint64_t, 64> Record; + + // Read all the records for this value table. + const Type *CurTy = Type::Int32Ty; + unsigned NextCstNo = ValueList.size(); + while (1) { + unsigned Code = Stream.ReadCode(); + if (Code == bitc::END_BLOCK) + break; + + if (Code == bitc::ENTER_SUBBLOCK) { + // No known subblocks, always skip them. + Stream.ReadSubBlockID(); + if (Stream.SkipBlock()) + return Error("Malformed block record"); + continue; + } + + if (Code == bitc::DEFINE_ABBREV) { + Stream.ReadAbbrevRecord(); + continue; + } + + // Read a record. + Record.clear(); + Value *V = 0; + switch (Stream.ReadRecord(Code, Record)) { + default: // Default behavior: unknown constant + case bitc::CST_CODE_UNDEF: // UNDEF + V = UndefValue::get(CurTy); + break; + case bitc::CST_CODE_SETTYPE: // SETTYPE: [typeid] + if (Record.empty()) + return Error("Malformed CST_SETTYPE record"); + if (Record[0] >= TypeList.size()) + return Error("Invalid Type ID in CST_SETTYPE record"); + CurTy = TypeList[Record[0]]; + continue; // Skip the ValueList manipulation. + case bitc::CST_CODE_NULL: // NULL + V = Constant::getNullValue(CurTy); + break; + case bitc::CST_CODE_INTEGER: // INTEGER: [intval] + if (!isa<IntegerType>(CurTy) || Record.empty()) + return Error("Invalid CST_INTEGER record"); + V = ConstantInt::get(CurTy, DecodeSignRotatedValue(Record[0])); + break; + case bitc::CST_CODE_WIDE_INTEGER: {// WIDE_INTEGER: [n x intval] + if (!isa<IntegerType>(CurTy) || Record.empty()) + return Error("Invalid WIDE_INTEGER record"); + + unsigned NumWords = Record.size(); + SmallVector<uint64_t, 8> Words; + Words.resize(NumWords); + for (unsigned i = 0; i != NumWords; ++i) + Words[i] = DecodeSignRotatedValue(Record[i]); + V = ConstantInt::get(APInt(cast<IntegerType>(CurTy)->getBitWidth(), + NumWords, &Words[0])); + break; + } + case bitc::CST_CODE_FLOAT: { // FLOAT: [fpval] + if (Record.empty()) + return Error("Invalid FLOAT record"); + if (CurTy == Type::FloatTy) + V = ConstantFP::get(APFloat(APInt(32, (uint32_t)Record[0]))); + else if (CurTy == Type::DoubleTy) + V = ConstantFP::get(APFloat(APInt(64, Record[0]))); + else if (CurTy == Type::X86_FP80Ty) { + // Bits are not stored the same way as a normal i80 APInt, compensate. + uint64_t Rearrange[2]; + Rearrange[0] = (Record[1] & 0xffffLL) | (Record[0] << 16); + Rearrange[1] = Record[0] >> 48; + V = ConstantFP::get(APFloat(APInt(80, 2, Rearrange))); + } else if (CurTy == Type::FP128Ty) + V = ConstantFP::get(APFloat(APInt(128, 2, &Record[0]), true)); + else if (CurTy == Type::PPC_FP128Ty) + V = ConstantFP::get(APFloat(APInt(128, 2, &Record[0]))); + else + V = UndefValue::get(CurTy); + break; + } + + case bitc::CST_CODE_AGGREGATE: {// AGGREGATE: [n x value number] + if (Record.empty()) + return Error("Invalid CST_AGGREGATE record"); + + unsigned Size = Record.size(); + std::vector<Constant*> Elts; + + if (const StructType *STy = dyn_cast<StructType>(CurTy)) { + for (unsigned i = 0; i != Size; ++i) + Elts.push_back(ValueList.getConstantFwdRef(Record[i], + STy->getElementType(i))); + V = ConstantStruct::get(STy, Elts); + } else if (const ArrayType *ATy = dyn_cast<ArrayType>(CurTy)) { + const Type *EltTy = ATy->getElementType(); + for (unsigned i = 0; i != Size; ++i) + Elts.push_back(ValueList.getConstantFwdRef(Record[i], EltTy)); + V = ConstantArray::get(ATy, Elts); + } else if (const VectorType *VTy = dyn_cast<VectorType>(CurTy)) { + const Type *EltTy = VTy->getElementType(); + for (unsigned i = 0; i != Size; ++i) + Elts.push_back(ValueList.getConstantFwdRef(Record[i], EltTy)); + V = ConstantVector::get(Elts); + } else { + V = UndefValue::get(CurTy); + } + break; + } + case bitc::CST_CODE_STRING: { // STRING: [values] + if (Record.empty()) + return Error("Invalid CST_AGGREGATE record"); + + const ArrayType *ATy = cast<ArrayType>(CurTy); + const Type *EltTy = ATy->getElementType(); + + unsigned Size = Record.size(); + std::vector<Constant*> Elts; + for (unsigned i = 0; i != Size; ++i) + Elts.push_back(ConstantInt::get(EltTy, Record[i])); + V = ConstantArray::get(ATy, Elts); + break; + } + case bitc::CST_CODE_CSTRING: { // CSTRING: [values] + if (Record.empty()) + return Error("Invalid CST_AGGREGATE record"); + + const ArrayType *ATy = cast<ArrayType>(CurTy); + const Type *EltTy = ATy->getElementType(); + + unsigned Size = Record.size(); + std::vector<Constant*> Elts; + for (unsigned i = 0; i != Size; ++i) + Elts.push_back(ConstantInt::get(EltTy, Record[i])); + Elts.push_back(Constant::getNullValue(EltTy)); + V = ConstantArray::get(ATy, Elts); + break; + } + case bitc::CST_CODE_CE_BINOP: { // CE_BINOP: [opcode, opval, opval] + if (Record.size() < 3) return Error("Invalid CE_BINOP record"); + int Opc = GetDecodedBinaryOpcode(Record[0], CurTy); + if (Opc < 0) { + V = UndefValue::get(CurTy); // Unknown binop. + } else { + Constant *LHS = ValueList.getConstantFwdRef(Record[1], CurTy); + Constant *RHS = ValueList.getConstantFwdRef(Record[2], CurTy); + V = ConstantExpr::get(Opc, LHS, RHS); + } + break; + } + case bitc::CST_CODE_CE_CAST: { // CE_CAST: [opcode, opty, opval] + if (Record.size() < 3) return Error("Invalid CE_CAST record"); + int Opc = GetDecodedCastOpcode(Record[0]); + if (Opc < 0) { + V = UndefValue::get(CurTy); // Unknown cast. + } else { + const Type *OpTy = getTypeByID(Record[1]); + if (!OpTy) return Error("Invalid CE_CAST record"); + Constant *Op = ValueList.getConstantFwdRef(Record[2], OpTy); + V = ConstantExpr::getCast(Opc, Op, CurTy); + } + break; + } + case bitc::CST_CODE_CE_GEP: { // CE_GEP: [n x operands] + if (Record.size() & 1) return Error("Invalid CE_GEP record"); + SmallVector<Constant*, 16> Elts; + for (unsigned i = 0, e = Record.size(); i != e; i += 2) { + const Type *ElTy = getTypeByID(Record[i]); + if (!ElTy) return Error("Invalid CE_GEP record"); + Elts.push_back(ValueList.getConstantFwdRef(Record[i+1], ElTy)); + } + V = ConstantExpr::getGetElementPtr(Elts[0], &Elts[1], Elts.size()-1); + break; + } + case bitc::CST_CODE_CE_SELECT: // CE_SELECT: [opval#, opval#, opval#] + if (Record.size() < 3) return Error("Invalid CE_SELECT record"); + V = ConstantExpr::getSelect(ValueList.getConstantFwdRef(Record[0], + Type::Int1Ty), + ValueList.getConstantFwdRef(Record[1],CurTy), + ValueList.getConstantFwdRef(Record[2],CurTy)); + break; + case bitc::CST_CODE_CE_EXTRACTELT: { // CE_EXTRACTELT: [opty, opval, opval] + if (Record.size() < 3) return Error("Invalid CE_EXTRACTELT record"); + const VectorType *OpTy = + dyn_cast_or_null<VectorType>(getTypeByID(Record[0])); + if (OpTy == 0) return Error("Invalid CE_EXTRACTELT record"); + Constant *Op0 = ValueList.getConstantFwdRef(Record[1], OpTy); + Constant *Op1 = ValueList.getConstantFwdRef(Record[2], Type::Int32Ty); + V = ConstantExpr::getExtractElement(Op0, Op1); + break; + } + case bitc::CST_CODE_CE_INSERTELT: { // CE_INSERTELT: [opval, opval, opval] + const VectorType *OpTy = dyn_cast<VectorType>(CurTy); + if (Record.size() < 3 || OpTy == 0) + return Error("Invalid CE_INSERTELT record"); + Constant *Op0 = ValueList.getConstantFwdRef(Record[0], OpTy); + Constant *Op1 = ValueList.getConstantFwdRef(Record[1], + OpTy->getElementType()); + Constant *Op2 = ValueList.getConstantFwdRef(Record[2], Type::Int32Ty); + V = ConstantExpr::getInsertElement(Op0, Op1, Op2); + break; + } + case bitc::CST_CODE_CE_SHUFFLEVEC: { // CE_SHUFFLEVEC: [opval, opval, opval] + const VectorType *OpTy = dyn_cast<VectorType>(CurTy); + if (Record.size() < 3 || OpTy == 0) + return Error("Invalid CE_SHUFFLEVEC record"); + Constant *Op0 = ValueList.getConstantFwdRef(Record[0], OpTy); + Constant *Op1 = ValueList.getConstantFwdRef(Record[1], OpTy); + const Type *ShufTy=VectorType::get(Type::Int32Ty, OpTy->getNumElements()); + Constant *Op2 = ValueList.getConstantFwdRef(Record[2], ShufTy); + V = ConstantExpr::getShuffleVector(Op0, Op1, Op2); + break; + } + case bitc::CST_CODE_CE_SHUFVEC_EX: { // [opty, opval, opval, opval] + const VectorType *RTy = dyn_cast<VectorType>(CurTy); + const VectorType *OpTy = dyn_cast<VectorType>(getTypeByID(Record[0])); + if (Record.size() < 4 || RTy == 0 || OpTy == 0) + return Error("Invalid CE_SHUFVEC_EX record"); + Constant *Op0 = ValueList.getConstantFwdRef(Record[1], OpTy); + Constant *Op1 = ValueList.getConstantFwdRef(Record[2], OpTy); + const Type *ShufTy=VectorType::get(Type::Int32Ty, RTy->getNumElements()); + Constant *Op2 = ValueList.getConstantFwdRef(Record[3], ShufTy); + V = ConstantExpr::getShuffleVector(Op0, Op1, Op2); + break; + } + case bitc::CST_CODE_CE_CMP: { // CE_CMP: [opty, opval, opval, pred] + if (Record.size() < 4) return Error("Invalid CE_CMP record"); + const Type *OpTy = getTypeByID(Record[0]); + if (OpTy == 0) return Error("Invalid CE_CMP record"); + Constant *Op0 = ValueList.getConstantFwdRef(Record[1], OpTy); + Constant *Op1 = ValueList.getConstantFwdRef(Record[2], OpTy); + + if (OpTy->isFloatingPoint()) + V = ConstantExpr::getFCmp(Record[3], Op0, Op1); + else if (!isa<VectorType>(OpTy)) + V = ConstantExpr::getICmp(Record[3], Op0, Op1); + else if (OpTy->isFPOrFPVector()) + V = ConstantExpr::getVFCmp(Record[3], Op0, Op1); + else + V = ConstantExpr::getVICmp(Record[3], Op0, Op1); + break; + } + case bitc::CST_CODE_INLINEASM: { + if (Record.size() < 2) return Error("Invalid INLINEASM record"); + std::string AsmStr, ConstrStr; + bool HasSideEffects = Record[0]; + unsigned AsmStrSize = Record[1]; + if (2+AsmStrSize >= Record.size()) + return Error("Invalid INLINEASM record"); + unsigned ConstStrSize = Record[2+AsmStrSize]; + if (3+AsmStrSize+ConstStrSize > Record.size()) + return Error("Invalid INLINEASM record"); + + for (unsigned i = 0; i != AsmStrSize; ++i) + AsmStr += (char)Record[2+i]; + for (unsigned i = 0; i != ConstStrSize; ++i) + ConstrStr += (char)Record[3+AsmStrSize+i]; + const PointerType *PTy = cast<PointerType>(CurTy); + V = InlineAsm::get(cast<FunctionType>(PTy->getElementType()), + AsmStr, ConstrStr, HasSideEffects); + break; + } + case bitc::CST_CODE_MDSTRING: { + if (Record.size() < 2) return Error("Invalid MDSTRING record"); + unsigned MDStringLength = Record.size(); + SmallString<8> String; + String.resize(MDStringLength); + for (unsigned i = 0; i != MDStringLength; ++i) + String[i] = Record[i]; + V = MDString::get(String.c_str(), String.c_str() + MDStringLength); + break; + } + case bitc::CST_CODE_MDNODE: { + if (Record.empty() || Record.size() % 2 == 1) + return Error("Invalid CST_MDNODE record"); + + unsigned Size = Record.size(); + SmallVector<Value*, 8> Elts; + for (unsigned i = 0; i != Size; i += 2) { + const Type *Ty = getTypeByID(Record[i], false); + if (Ty != Type::VoidTy) + Elts.push_back(ValueList.getValueFwdRef(Record[i+1], Ty)); + else + Elts.push_back(NULL); + } + V = MDNode::get(&Elts[0], Elts.size()); + break; + } + } + + ValueList.AssignValue(V, NextCstNo); + ++NextCstNo; + } + + if (NextCstNo != ValueList.size()) + return Error("Invalid constant reference!"); + + if (Stream.ReadBlockEnd()) + return Error("Error at end of constants block"); + + // Once all the constants have been read, go through and resolve forward + // references. + ValueList.ResolveConstantForwardRefs(); + return false; +} + +/// RememberAndSkipFunctionBody - When we see the block for a function body, +/// remember where it is and then skip it. This lets us lazily deserialize the +/// functions. +bool BitcodeReader::RememberAndSkipFunctionBody() { + // Get the function we are talking about. + if (FunctionsWithBodies.empty()) + return Error("Insufficient function protos"); + + Function *Fn = FunctionsWithBodies.back(); + FunctionsWithBodies.pop_back(); + + // Save the current stream state. + uint64_t CurBit = Stream.GetCurrentBitNo(); + DeferredFunctionInfo[Fn] = std::make_pair(CurBit, Fn->getLinkage()); + + // Set the functions linkage to GhostLinkage so we know it is lazily + // deserialized. + Fn->setLinkage(GlobalValue::GhostLinkage); + + // Skip over the function block for now. + if (Stream.SkipBlock()) + return Error("Malformed block record"); + return false; +} + +bool BitcodeReader::ParseModule(const std::string &ModuleID) { + // Reject multiple MODULE_BLOCK's in a single bitstream. + if (TheModule) + return Error("Multiple MODULE_BLOCKs in same stream"); + + if (Stream.EnterSubBlock(bitc::MODULE_BLOCK_ID)) + return Error("Malformed block record"); + + // Otherwise, create the module. + TheModule = new Module(ModuleID); + + SmallVector<uint64_t, 64> Record; + std::vector<std::string> SectionTable; + std::vector<std::string> GCTable; + + // Read all the records for this module. + while (!Stream.AtEndOfStream()) { + unsigned Code = Stream.ReadCode(); + if (Code == bitc::END_BLOCK) { + if (Stream.ReadBlockEnd()) + return Error("Error at end of module block"); + + // Patch the initializers for globals and aliases up. + ResolveGlobalAndAliasInits(); + if (!GlobalInits.empty() || !AliasInits.empty()) + return Error("Malformed global initializer set"); + if (!FunctionsWithBodies.empty()) + return Error("Too few function bodies found"); + + // Look for intrinsic functions which need to be upgraded at some point + for (Module::iterator FI = TheModule->begin(), FE = TheModule->end(); + FI != FE; ++FI) { + Function* NewFn; + if (UpgradeIntrinsicFunction(FI, NewFn)) + UpgradedIntrinsics.push_back(std::make_pair(FI, NewFn)); + } + + // Force deallocation of memory for these vectors to favor the client that + // want lazy deserialization. + std::vector<std::pair<GlobalVariable*, unsigned> >().swap(GlobalInits); + std::vector<std::pair<GlobalAlias*, unsigned> >().swap(AliasInits); + std::vector<Function*>().swap(FunctionsWithBodies); + return false; + } + + if (Code == bitc::ENTER_SUBBLOCK) { + switch (Stream.ReadSubBlockID()) { + default: // Skip unknown content. + if (Stream.SkipBlock()) + return Error("Malformed block record"); + break; + case bitc::BLOCKINFO_BLOCK_ID: + if (Stream.ReadBlockInfoBlock()) + return Error("Malformed BlockInfoBlock"); + break; + case bitc::PARAMATTR_BLOCK_ID: + if (ParseAttributeBlock()) + return true; + break; + case bitc::TYPE_BLOCK_ID: + if (ParseTypeTable()) + return true; + break; + case bitc::TYPE_SYMTAB_BLOCK_ID: + if (ParseTypeSymbolTable()) + return true; + break; + case bitc::VALUE_SYMTAB_BLOCK_ID: + if (ParseValueSymbolTable()) + return true; + break; + case bitc::CONSTANTS_BLOCK_ID: + if (ParseConstants() || ResolveGlobalAndAliasInits()) + return true; + break; + case bitc::FUNCTION_BLOCK_ID: + // If this is the first function body we've seen, reverse the + // FunctionsWithBodies list. + if (!HasReversedFunctionsWithBodies) { + std::reverse(FunctionsWithBodies.begin(), FunctionsWithBodies.end()); + HasReversedFunctionsWithBodies = true; + } + + if (RememberAndSkipFunctionBody()) + return true; + break; + } + continue; + } + + if (Code == bitc::DEFINE_ABBREV) { + Stream.ReadAbbrevRecord(); + continue; + } + + // Read a record. + switch (Stream.ReadRecord(Code, Record)) { + default: break; // Default behavior, ignore unknown content. + case bitc::MODULE_CODE_VERSION: // VERSION: [version#] + if (Record.size() < 1) + return Error("Malformed MODULE_CODE_VERSION"); + // Only version #0 is supported so far. + if (Record[0] != 0) + return Error("Unknown bitstream version!"); + break; + case bitc::MODULE_CODE_TRIPLE: { // TRIPLE: [strchr x N] + std::string S; + if (ConvertToString(Record, 0, S)) + return Error("Invalid MODULE_CODE_TRIPLE record"); + TheModule->setTargetTriple(S); + break; + } + case bitc::MODULE_CODE_DATALAYOUT: { // DATALAYOUT: [strchr x N] + std::string S; + if (ConvertToString(Record, 0, S)) + return Error("Invalid MODULE_CODE_DATALAYOUT record"); + TheModule->setDataLayout(S); + break; + } + case bitc::MODULE_CODE_ASM: { // ASM: [strchr x N] + std::string S; + if (ConvertToString(Record, 0, S)) + return Error("Invalid MODULE_CODE_ASM record"); + TheModule->setModuleInlineAsm(S); + break; + } + case bitc::MODULE_CODE_DEPLIB: { // DEPLIB: [strchr x N] + std::string S; + if (ConvertToString(Record, 0, S)) + return Error("Invalid MODULE_CODE_DEPLIB record"); + TheModule->addLibrary(S); + break; + } + case bitc::MODULE_CODE_SECTIONNAME: { // SECTIONNAME: [strchr x N] + std::string S; + if (ConvertToString(Record, 0, S)) + return Error("Invalid MODULE_CODE_SECTIONNAME record"); + SectionTable.push_back(S); + break; + } + case bitc::MODULE_CODE_GCNAME: { // SECTIONNAME: [strchr x N] + std::string S; + if (ConvertToString(Record, 0, S)) + return Error("Invalid MODULE_CODE_GCNAME record"); + GCTable.push_back(S); + break; + } + // GLOBALVAR: [pointer type, isconst, initid, + // linkage, alignment, section, visibility, threadlocal] + case bitc::MODULE_CODE_GLOBALVAR: { + if (Record.size() < 6) + return Error("Invalid MODULE_CODE_GLOBALVAR record"); + const Type *Ty = getTypeByID(Record[0]); + if (!isa<PointerType>(Ty)) + return Error("Global not a pointer type!"); + unsigned AddressSpace = cast<PointerType>(Ty)->getAddressSpace(); + Ty = cast<PointerType>(Ty)->getElementType(); + + bool isConstant = Record[1]; + GlobalValue::LinkageTypes Linkage = GetDecodedLinkage(Record[3]); + unsigned Alignment = (1 << Record[4]) >> 1; + std::string Section; + if (Record[5]) { + if (Record[5]-1 >= SectionTable.size()) + return Error("Invalid section ID"); + Section = SectionTable[Record[5]-1]; + } + GlobalValue::VisibilityTypes Visibility = GlobalValue::DefaultVisibility; + if (Record.size() > 6) + Visibility = GetDecodedVisibility(Record[6]); + bool isThreadLocal = false; + if (Record.size() > 7) + isThreadLocal = Record[7]; + + GlobalVariable *NewGV = + new GlobalVariable(Ty, isConstant, Linkage, 0, "", TheModule, + isThreadLocal, AddressSpace); + NewGV->setAlignment(Alignment); + if (!Section.empty()) + NewGV->setSection(Section); + NewGV->setVisibility(Visibility); + NewGV->setThreadLocal(isThreadLocal); + + ValueList.push_back(NewGV); + + // Remember which value to use for the global initializer. + if (unsigned InitID = Record[2]) + GlobalInits.push_back(std::make_pair(NewGV, InitID-1)); + break; + } + // FUNCTION: [type, callingconv, isproto, linkage, paramattr, + // alignment, section, visibility, gc] + case bitc::MODULE_CODE_FUNCTION: { + if (Record.size() < 8) + return Error("Invalid MODULE_CODE_FUNCTION record"); + const Type *Ty = getTypeByID(Record[0]); + if (!isa<PointerType>(Ty)) + return Error("Function not a pointer type!"); + const FunctionType *FTy = + dyn_cast<FunctionType>(cast<PointerType>(Ty)->getElementType()); + if (!FTy) + return Error("Function not a pointer to function type!"); + + Function *Func = Function::Create(FTy, GlobalValue::ExternalLinkage, + "", TheModule); + + Func->setCallingConv(Record[1]); + bool isProto = Record[2]; + Func->setLinkage(GetDecodedLinkage(Record[3])); + Func->setAttributes(getAttributes(Record[4])); + + Func->setAlignment((1 << Record[5]) >> 1); + if (Record[6]) { + if (Record[6]-1 >= SectionTable.size()) + return Error("Invalid section ID"); + Func->setSection(SectionTable[Record[6]-1]); + } + Func->setVisibility(GetDecodedVisibility(Record[7])); + if (Record.size() > 8 && Record[8]) { + if (Record[8]-1 > GCTable.size()) + return Error("Invalid GC ID"); + Func->setGC(GCTable[Record[8]-1].c_str()); + } + ValueList.push_back(Func); + + // If this is a function with a body, remember the prototype we are + // creating now, so that we can match up the body with them later. + if (!isProto) + FunctionsWithBodies.push_back(Func); + break; + } + // ALIAS: [alias type, aliasee val#, linkage] + // ALIAS: [alias type, aliasee val#, linkage, visibility] + case bitc::MODULE_CODE_ALIAS: { + if (Record.size() < 3) + return Error("Invalid MODULE_ALIAS record"); + const Type *Ty = getTypeByID(Record[0]); + if (!isa<PointerType>(Ty)) + return Error("Function not a pointer type!"); + + GlobalAlias *NewGA = new GlobalAlias(Ty, GetDecodedLinkage(Record[2]), + "", 0, TheModule); + // Old bitcode files didn't have visibility field. + if (Record.size() > 3) + NewGA->setVisibility(GetDecodedVisibility(Record[3])); + ValueList.push_back(NewGA); + AliasInits.push_back(std::make_pair(NewGA, Record[1])); + break; + } + /// MODULE_CODE_PURGEVALS: [numvals] + case bitc::MODULE_CODE_PURGEVALS: + // Trim down the value list to the specified size. + if (Record.size() < 1 || Record[0] > ValueList.size()) + return Error("Invalid MODULE_PURGEVALS record"); + ValueList.shrinkTo(Record[0]); + break; + } + Record.clear(); + } + + return Error("Premature end of bitstream"); +} + +bool BitcodeReader::ParseBitcode() { + TheModule = 0; + + if (Buffer->getBufferSize() & 3) + return Error("Bitcode stream should be a multiple of 4 bytes in length"); + + unsigned char *BufPtr = (unsigned char *)Buffer->getBufferStart(); + unsigned char *BufEnd = BufPtr+Buffer->getBufferSize(); + + // If we have a wrapper header, parse it and ignore the non-bc file contents. + // The magic number is 0x0B17C0DE stored in little endian. + if (isBitcodeWrapper(BufPtr, BufEnd)) + if (SkipBitcodeWrapperHeader(BufPtr, BufEnd)) + return Error("Invalid bitcode wrapper header"); + + StreamFile.init(BufPtr, BufEnd); + Stream.init(StreamFile); + + // Sniff for the signature. + if (Stream.Read(8) != 'B' || + Stream.Read(8) != 'C' || + Stream.Read(4) != 0x0 || + Stream.Read(4) != 0xC || + Stream.Read(4) != 0xE || + Stream.Read(4) != 0xD) + return Error("Invalid bitcode signature"); + + // We expect a number of well-defined blocks, though we don't necessarily + // need to understand them all. + while (!Stream.AtEndOfStream()) { + unsigned Code = Stream.ReadCode(); + + if (Code != bitc::ENTER_SUBBLOCK) + return Error("Invalid record at top-level"); + + unsigned BlockID = Stream.ReadSubBlockID(); + + // We only know the MODULE subblock ID. + switch (BlockID) { + case bitc::BLOCKINFO_BLOCK_ID: + if (Stream.ReadBlockInfoBlock()) + return Error("Malformed BlockInfoBlock"); + break; + case bitc::MODULE_BLOCK_ID: + if (ParseModule(Buffer->getBufferIdentifier())) + return true; + break; + default: + if (Stream.SkipBlock()) + return Error("Malformed block record"); + break; + } + } + + return false; +} + + +/// ParseFunctionBody - Lazily parse the specified function body block. +bool BitcodeReader::ParseFunctionBody(Function *F) { + if (Stream.EnterSubBlock(bitc::FUNCTION_BLOCK_ID)) + return Error("Malformed block record"); + + unsigned ModuleValueListSize = ValueList.size(); + + // Add all the function arguments to the value table. + for(Function::arg_iterator I = F->arg_begin(), E = F->arg_end(); I != E; ++I) + ValueList.push_back(I); + + unsigned NextValueNo = ValueList.size(); + BasicBlock *CurBB = 0; + unsigned CurBBNo = 0; + + // Read all the records. + SmallVector<uint64_t, 64> Record; + while (1) { + unsigned Code = Stream.ReadCode(); + if (Code == bitc::END_BLOCK) { + if (Stream.ReadBlockEnd()) + return Error("Error at end of function block"); + break; + } + + if (Code == bitc::ENTER_SUBBLOCK) { + switch (Stream.ReadSubBlockID()) { + default: // Skip unknown content. + if (Stream.SkipBlock()) + return Error("Malformed block record"); + break; + case bitc::CONSTANTS_BLOCK_ID: + if (ParseConstants()) return true; + NextValueNo = ValueList.size(); + break; + case bitc::VALUE_SYMTAB_BLOCK_ID: + if (ParseValueSymbolTable()) return true; + break; + } + continue; + } + + if (Code == bitc::DEFINE_ABBREV) { + Stream.ReadAbbrevRecord(); + continue; + } + + // Read a record. + Record.clear(); + Instruction *I = 0; + switch (Stream.ReadRecord(Code, Record)) { + default: // Default behavior: reject + return Error("Unknown instruction"); + case bitc::FUNC_CODE_DECLAREBLOCKS: // DECLAREBLOCKS: [nblocks] + if (Record.size() < 1 || Record[0] == 0) + return Error("Invalid DECLAREBLOCKS record"); + // Create all the basic blocks for the function. + FunctionBBs.resize(Record[0]); + for (unsigned i = 0, e = FunctionBBs.size(); i != e; ++i) + FunctionBBs[i] = BasicBlock::Create("", F); + CurBB = FunctionBBs[0]; + continue; + + case bitc::FUNC_CODE_INST_BINOP: { // BINOP: [opval, ty, opval, opcode] + unsigned OpNum = 0; + Value *LHS, *RHS; + if (getValueTypePair(Record, OpNum, NextValueNo, LHS) || + getValue(Record, OpNum, LHS->getType(), RHS) || + OpNum+1 != Record.size()) + return Error("Invalid BINOP record"); + + int Opc = GetDecodedBinaryOpcode(Record[OpNum], LHS->getType()); + if (Opc == -1) return Error("Invalid BINOP record"); + I = BinaryOperator::Create((Instruction::BinaryOps)Opc, LHS, RHS); + break; + } + case bitc::FUNC_CODE_INST_CAST: { // CAST: [opval, opty, destty, castopc] + unsigned OpNum = 0; + Value *Op; + if (getValueTypePair(Record, OpNum, NextValueNo, Op) || + OpNum+2 != Record.size()) + return Error("Invalid CAST record"); + + const Type *ResTy = getTypeByID(Record[OpNum]); + int Opc = GetDecodedCastOpcode(Record[OpNum+1]); + if (Opc == -1 || ResTy == 0) + return Error("Invalid CAST record"); + I = CastInst::Create((Instruction::CastOps)Opc, Op, ResTy); + break; + } + case bitc::FUNC_CODE_INST_GEP: { // GEP: [n x operands] + unsigned OpNum = 0; + Value *BasePtr; + if (getValueTypePair(Record, OpNum, NextValueNo, BasePtr)) + return Error("Invalid GEP record"); + + SmallVector<Value*, 16> GEPIdx; + while (OpNum != Record.size()) { + Value *Op; + if (getValueTypePair(Record, OpNum, NextValueNo, Op)) + return Error("Invalid GEP record"); + GEPIdx.push_back(Op); + } + + I = GetElementPtrInst::Create(BasePtr, GEPIdx.begin(), GEPIdx.end()); + break; + } + + case bitc::FUNC_CODE_INST_EXTRACTVAL: { + // EXTRACTVAL: [opty, opval, n x indices] + unsigned OpNum = 0; + Value *Agg; + if (getValueTypePair(Record, OpNum, NextValueNo, Agg)) + return Error("Invalid EXTRACTVAL record"); + + SmallVector<unsigned, 4> EXTRACTVALIdx; + for (unsigned RecSize = Record.size(); + OpNum != RecSize; ++OpNum) { + uint64_t Index = Record[OpNum]; + if ((unsigned)Index != Index) + return Error("Invalid EXTRACTVAL index"); + EXTRACTVALIdx.push_back((unsigned)Index); + } + + I = ExtractValueInst::Create(Agg, + EXTRACTVALIdx.begin(), EXTRACTVALIdx.end()); + break; + } + + case bitc::FUNC_CODE_INST_INSERTVAL: { + // INSERTVAL: [opty, opval, opty, opval, n x indices] + unsigned OpNum = 0; + Value *Agg; + if (getValueTypePair(Record, OpNum, NextValueNo, Agg)) + return Error("Invalid INSERTVAL record"); + Value *Val; + if (getValueTypePair(Record, OpNum, NextValueNo, Val)) + return Error("Invalid INSERTVAL record"); + + SmallVector<unsigned, 4> INSERTVALIdx; + for (unsigned RecSize = Record.size(); + OpNum != RecSize; ++OpNum) { + uint64_t Index = Record[OpNum]; + if ((unsigned)Index != Index) + return Error("Invalid INSERTVAL index"); + INSERTVALIdx.push_back((unsigned)Index); + } + + I = InsertValueInst::Create(Agg, Val, + INSERTVALIdx.begin(), INSERTVALIdx.end()); + break; + } + + case bitc::FUNC_CODE_INST_SELECT: { // SELECT: [opval, ty, opval, opval] + // obsolete form of select + // handles select i1 ... in old bitcode + unsigned OpNum = 0; + Value *TrueVal, *FalseVal, *Cond; + if (getValueTypePair(Record, OpNum, NextValueNo, TrueVal) || + getValue(Record, OpNum, TrueVal->getType(), FalseVal) || + getValue(Record, OpNum, Type::Int1Ty, Cond)) + return Error("Invalid SELECT record"); + + I = SelectInst::Create(Cond, TrueVal, FalseVal); + break; + } + + case bitc::FUNC_CODE_INST_VSELECT: {// VSELECT: [ty,opval,opval,predty,pred] + // new form of select + // handles select i1 or select [N x i1] + unsigned OpNum = 0; + Value *TrueVal, *FalseVal, *Cond; + if (getValueTypePair(Record, OpNum, NextValueNo, TrueVal) || + getValue(Record, OpNum, TrueVal->getType(), FalseVal) || + getValueTypePair(Record, OpNum, NextValueNo, Cond)) + return Error("Invalid SELECT record"); + + // select condition can be either i1 or [N x i1] + if (const VectorType* vector_type = + dyn_cast<const VectorType>(Cond->getType())) { + // expect <n x i1> + if (vector_type->getElementType() != Type::Int1Ty) + return Error("Invalid SELECT condition type"); + } else { + // expect i1 + if (Cond->getType() != Type::Int1Ty) + return Error("Invalid SELECT condition type"); + } + + I = SelectInst::Create(Cond, TrueVal, FalseVal); + break; + } + + case bitc::FUNC_CODE_INST_EXTRACTELT: { // EXTRACTELT: [opty, opval, opval] + unsigned OpNum = 0; + Value *Vec, *Idx; + if (getValueTypePair(Record, OpNum, NextValueNo, Vec) || + getValue(Record, OpNum, Type::Int32Ty, Idx)) + return Error("Invalid EXTRACTELT record"); + I = new ExtractElementInst(Vec, Idx); + break; + } + + case bitc::FUNC_CODE_INST_INSERTELT: { // INSERTELT: [ty, opval,opval,opval] + unsigned OpNum = 0; + Value *Vec, *Elt, *Idx; + if (getValueTypePair(Record, OpNum, NextValueNo, Vec) || + getValue(Record, OpNum, + cast<VectorType>(Vec->getType())->getElementType(), Elt) || + getValue(Record, OpNum, Type::Int32Ty, Idx)) + return Error("Invalid INSERTELT record"); + I = InsertElementInst::Create(Vec, Elt, Idx); + break; + } + + case bitc::FUNC_CODE_INST_SHUFFLEVEC: {// SHUFFLEVEC: [opval,ty,opval,opval] + unsigned OpNum = 0; + Value *Vec1, *Vec2, *Mask; + if (getValueTypePair(Record, OpNum, NextValueNo, Vec1) || + getValue(Record, OpNum, Vec1->getType(), Vec2)) + return Error("Invalid SHUFFLEVEC record"); + + if (getValueTypePair(Record, OpNum, NextValueNo, Mask)) + return Error("Invalid SHUFFLEVEC record"); + I = new ShuffleVectorInst(Vec1, Vec2, Mask); + break; + } + + case bitc::FUNC_CODE_INST_CMP: { // CMP: [opty, opval, opval, pred] + // VFCmp/VICmp + // or old form of ICmp/FCmp returning bool + unsigned OpNum = 0; + Value *LHS, *RHS; + if (getValueTypePair(Record, OpNum, NextValueNo, LHS) || + getValue(Record, OpNum, LHS->getType(), RHS) || + OpNum+1 != Record.size()) + return Error("Invalid CMP record"); + + if (LHS->getType()->isFloatingPoint()) + I = new FCmpInst((FCmpInst::Predicate)Record[OpNum], LHS, RHS); + else if (!isa<VectorType>(LHS->getType())) + I = new ICmpInst((ICmpInst::Predicate)Record[OpNum], LHS, RHS); + else if (LHS->getType()->isFPOrFPVector()) + I = new VFCmpInst((FCmpInst::Predicate)Record[OpNum], LHS, RHS); + else + I = new VICmpInst((ICmpInst::Predicate)Record[OpNum], LHS, RHS); + break; + } + case bitc::FUNC_CODE_INST_CMP2: { // CMP2: [opty, opval, opval, pred] + // Fcmp/ICmp returning bool or vector of bool + unsigned OpNum = 0; + Value *LHS, *RHS; + if (getValueTypePair(Record, OpNum, NextValueNo, LHS) || + getValue(Record, OpNum, LHS->getType(), RHS) || + OpNum+1 != Record.size()) + return Error("Invalid CMP2 record"); + + if (LHS->getType()->isFPOrFPVector()) + I = new FCmpInst((FCmpInst::Predicate)Record[OpNum], LHS, RHS); + else + I = new ICmpInst((ICmpInst::Predicate)Record[OpNum], LHS, RHS); + break; + } + case bitc::FUNC_CODE_INST_GETRESULT: { // GETRESULT: [ty, val, n] + if (Record.size() != 2) + return Error("Invalid GETRESULT record"); + unsigned OpNum = 0; + Value *Op; + getValueTypePair(Record, OpNum, NextValueNo, Op); + unsigned Index = Record[1]; + I = ExtractValueInst::Create(Op, Index); + break; + } + + case bitc::FUNC_CODE_INST_RET: // RET: [opty,opval<optional>] + { + unsigned Size = Record.size(); + if (Size == 0) { + I = ReturnInst::Create(); + break; + } + + unsigned OpNum = 0; + SmallVector<Value *,4> Vs; + do { + Value *Op = NULL; + if (getValueTypePair(Record, OpNum, NextValueNo, Op)) + return Error("Invalid RET record"); + Vs.push_back(Op); + } while(OpNum != Record.size()); + + const Type *ReturnType = F->getReturnType(); + if (Vs.size() > 1 || + (isa<StructType>(ReturnType) && + (Vs.empty() || Vs[0]->getType() != ReturnType))) { + Value *RV = UndefValue::get(ReturnType); + for (unsigned i = 0, e = Vs.size(); i != e; ++i) { + I = InsertValueInst::Create(RV, Vs[i], i, "mrv"); + CurBB->getInstList().push_back(I); + ValueList.AssignValue(I, NextValueNo++); + RV = I; + } + I = ReturnInst::Create(RV); + break; + } + + I = ReturnInst::Create(Vs[0]); + break; + } + case bitc::FUNC_CODE_INST_BR: { // BR: [bb#, bb#, opval] or [bb#] + if (Record.size() != 1 && Record.size() != 3) + return Error("Invalid BR record"); + BasicBlock *TrueDest = getBasicBlock(Record[0]); + if (TrueDest == 0) + return Error("Invalid BR record"); + + if (Record.size() == 1) + I = BranchInst::Create(TrueDest); + else { + BasicBlock *FalseDest = getBasicBlock(Record[1]); + Value *Cond = getFnValueByID(Record[2], Type::Int1Ty); + if (FalseDest == 0 || Cond == 0) + return Error("Invalid BR record"); + I = BranchInst::Create(TrueDest, FalseDest, Cond); + } + break; + } + case bitc::FUNC_CODE_INST_SWITCH: { // SWITCH: [opty, opval, n, n x ops] + if (Record.size() < 3 || (Record.size() & 1) == 0) + return Error("Invalid SWITCH record"); + const Type *OpTy = getTypeByID(Record[0]); + Value *Cond = getFnValueByID(Record[1], OpTy); + BasicBlock *Default = getBasicBlock(Record[2]); + if (OpTy == 0 || Cond == 0 || Default == 0) + return Error("Invalid SWITCH record"); + unsigned NumCases = (Record.size()-3)/2; + SwitchInst *SI = SwitchInst::Create(Cond, Default, NumCases); + for (unsigned i = 0, e = NumCases; i != e; ++i) { + ConstantInt *CaseVal = + dyn_cast_or_null<ConstantInt>(getFnValueByID(Record[3+i*2], OpTy)); + BasicBlock *DestBB = getBasicBlock(Record[1+3+i*2]); + if (CaseVal == 0 || DestBB == 0) { + delete SI; + return Error("Invalid SWITCH record!"); + } + SI->addCase(CaseVal, DestBB); + } + I = SI; + break; + } + + case bitc::FUNC_CODE_INST_INVOKE: { + // INVOKE: [attrs, cc, normBB, unwindBB, fnty, op0,op1,op2, ...] + if (Record.size() < 4) return Error("Invalid INVOKE record"); + AttrListPtr PAL = getAttributes(Record[0]); + unsigned CCInfo = Record[1]; + BasicBlock *NormalBB = getBasicBlock(Record[2]); + BasicBlock *UnwindBB = getBasicBlock(Record[3]); + + unsigned OpNum = 4; + Value *Callee; + if (getValueTypePair(Record, OpNum, NextValueNo, Callee)) + return Error("Invalid INVOKE record"); + + const PointerType *CalleeTy = dyn_cast<PointerType>(Callee->getType()); + const FunctionType *FTy = !CalleeTy ? 0 : + dyn_cast<FunctionType>(CalleeTy->getElementType()); + + // Check that the right number of fixed parameters are here. + if (FTy == 0 || NormalBB == 0 || UnwindBB == 0 || + Record.size() < OpNum+FTy->getNumParams()) + return Error("Invalid INVOKE record"); + + SmallVector<Value*, 16> Ops; + for (unsigned i = 0, e = FTy->getNumParams(); i != e; ++i, ++OpNum) { + Ops.push_back(getFnValueByID(Record[OpNum], FTy->getParamType(i))); + if (Ops.back() == 0) return Error("Invalid INVOKE record"); + } + + if (!FTy->isVarArg()) { + if (Record.size() != OpNum) + return Error("Invalid INVOKE record"); + } else { + // Read type/value pairs for varargs params. + while (OpNum != Record.size()) { + Value *Op; + if (getValueTypePair(Record, OpNum, NextValueNo, Op)) + return Error("Invalid INVOKE record"); + Ops.push_back(Op); + } + } + + I = InvokeInst::Create(Callee, NormalBB, UnwindBB, + Ops.begin(), Ops.end()); + cast<InvokeInst>(I)->setCallingConv(CCInfo); + cast<InvokeInst>(I)->setAttributes(PAL); + break; + } + case bitc::FUNC_CODE_INST_UNWIND: // UNWIND + I = new UnwindInst(); + break; + case bitc::FUNC_CODE_INST_UNREACHABLE: // UNREACHABLE + I = new UnreachableInst(); + break; + case bitc::FUNC_CODE_INST_PHI: { // PHI: [ty, val0,bb0, ...] + if (Record.size() < 1 || ((Record.size()-1)&1)) + return Error("Invalid PHI record"); + const Type *Ty = getTypeByID(Record[0]); + if (!Ty) return Error("Invalid PHI record"); + + PHINode *PN = PHINode::Create(Ty); + PN->reserveOperandSpace((Record.size()-1)/2); + + for (unsigned i = 0, e = Record.size()-1; i != e; i += 2) { + Value *V = getFnValueByID(Record[1+i], Ty); + BasicBlock *BB = getBasicBlock(Record[2+i]); + if (!V || !BB) return Error("Invalid PHI record"); + PN->addIncoming(V, BB); + } + I = PN; + break; + } + + case bitc::FUNC_CODE_INST_MALLOC: { // MALLOC: [instty, op, align] + if (Record.size() < 3) + return Error("Invalid MALLOC record"); + const PointerType *Ty = + dyn_cast_or_null<PointerType>(getTypeByID(Record[0])); + Value *Size = getFnValueByID(Record[1], Type::Int32Ty); + unsigned Align = Record[2]; + if (!Ty || !Size) return Error("Invalid MALLOC record"); + I = new MallocInst(Ty->getElementType(), Size, (1 << Align) >> 1); + break; + } + case bitc::FUNC_CODE_INST_FREE: { // FREE: [op, opty] + unsigned OpNum = 0; + Value *Op; + if (getValueTypePair(Record, OpNum, NextValueNo, Op) || + OpNum != Record.size()) + return Error("Invalid FREE record"); + I = new FreeInst(Op); + break; + } + case bitc::FUNC_CODE_INST_ALLOCA: { // ALLOCA: [instty, op, align] + if (Record.size() < 3) + return Error("Invalid ALLOCA record"); + const PointerType *Ty = + dyn_cast_or_null<PointerType>(getTypeByID(Record[0])); + Value *Size = getFnValueByID(Record[1], Type::Int32Ty); + unsigned Align = Record[2]; + if (!Ty || !Size) return Error("Invalid ALLOCA record"); + I = new AllocaInst(Ty->getElementType(), Size, (1 << Align) >> 1); + break; + } + case bitc::FUNC_CODE_INST_LOAD: { // LOAD: [opty, op, align, vol] + unsigned OpNum = 0; + Value *Op; + if (getValueTypePair(Record, OpNum, NextValueNo, Op) || + OpNum+2 != Record.size()) + return Error("Invalid LOAD record"); + + I = new LoadInst(Op, "", Record[OpNum+1], (1 << Record[OpNum]) >> 1); + break; + } + case bitc::FUNC_CODE_INST_STORE2: { // STORE2:[ptrty, ptr, val, align, vol] + unsigned OpNum = 0; + Value *Val, *Ptr; + if (getValueTypePair(Record, OpNum, NextValueNo, Ptr) || + getValue(Record, OpNum, + cast<PointerType>(Ptr->getType())->getElementType(), Val) || + OpNum+2 != Record.size()) + return Error("Invalid STORE record"); + + I = new StoreInst(Val, Ptr, Record[OpNum+1], (1 << Record[OpNum]) >> 1); + break; + } + case bitc::FUNC_CODE_INST_STORE: { // STORE:[val, valty, ptr, align, vol] + // FIXME: Legacy form of store instruction. Should be removed in LLVM 3.0. + unsigned OpNum = 0; + Value *Val, *Ptr; + if (getValueTypePair(Record, OpNum, NextValueNo, Val) || + getValue(Record, OpNum, PointerType::getUnqual(Val->getType()), Ptr)|| + OpNum+2 != Record.size()) + return Error("Invalid STORE record"); + + I = new StoreInst(Val, Ptr, Record[OpNum+1], (1 << Record[OpNum]) >> 1); + break; + } + case bitc::FUNC_CODE_INST_CALL: { + // CALL: [paramattrs, cc, fnty, fnid, arg0, arg1...] + if (Record.size() < 3) + return Error("Invalid CALL record"); + + AttrListPtr PAL = getAttributes(Record[0]); + unsigned CCInfo = Record[1]; + + unsigned OpNum = 2; + Value *Callee; + if (getValueTypePair(Record, OpNum, NextValueNo, Callee)) + return Error("Invalid CALL record"); + + const PointerType *OpTy = dyn_cast<PointerType>(Callee->getType()); + const FunctionType *FTy = 0; + if (OpTy) FTy = dyn_cast<FunctionType>(OpTy->getElementType()); + if (!FTy || Record.size() < FTy->getNumParams()+OpNum) + return Error("Invalid CALL record"); + + SmallVector<Value*, 16> Args; + // Read the fixed params. + for (unsigned i = 0, e = FTy->getNumParams(); i != e; ++i, ++OpNum) { + if (FTy->getParamType(i)->getTypeID()==Type::LabelTyID) + Args.push_back(getBasicBlock(Record[OpNum])); + else + Args.push_back(getFnValueByID(Record[OpNum], FTy->getParamType(i))); + if (Args.back() == 0) return Error("Invalid CALL record"); + } + + // Read type/value pairs for varargs params. + if (!FTy->isVarArg()) { + if (OpNum != Record.size()) + return Error("Invalid CALL record"); + } else { + while (OpNum != Record.size()) { + Value *Op; + if (getValueTypePair(Record, OpNum, NextValueNo, Op)) + return Error("Invalid CALL record"); + Args.push_back(Op); + } + } + + I = CallInst::Create(Callee, Args.begin(), Args.end()); + cast<CallInst>(I)->setCallingConv(CCInfo>>1); + cast<CallInst>(I)->setTailCall(CCInfo & 1); + cast<CallInst>(I)->setAttributes(PAL); + break; + } + case bitc::FUNC_CODE_INST_VAARG: { // VAARG: [valistty, valist, instty] + if (Record.size() < 3) + return Error("Invalid VAARG record"); + const Type *OpTy = getTypeByID(Record[0]); + Value *Op = getFnValueByID(Record[1], OpTy); + const Type *ResTy = getTypeByID(Record[2]); + if (!OpTy || !Op || !ResTy) + return Error("Invalid VAARG record"); + I = new VAArgInst(Op, ResTy); + break; + } + } + + // Add instruction to end of current BB. If there is no current BB, reject + // this file. + if (CurBB == 0) { + delete I; + return Error("Invalid instruction with no BB"); + } + CurBB->getInstList().push_back(I); + + // If this was a terminator instruction, move to the next block. + if (isa<TerminatorInst>(I)) { + ++CurBBNo; + CurBB = CurBBNo < FunctionBBs.size() ? FunctionBBs[CurBBNo] : 0; + } + + // Non-void values get registered in the value table for future use. + if (I && I->getType() != Type::VoidTy) + ValueList.AssignValue(I, NextValueNo++); + } + + // Check the function list for unresolved values. + if (Argument *A = dyn_cast<Argument>(ValueList.back())) { + if (A->getParent() == 0) { + // We found at least one unresolved value. Nuke them all to avoid leaks. + for (unsigned i = ModuleValueListSize, e = ValueList.size(); i != e; ++i){ + if ((A = dyn_cast<Argument>(ValueList.back())) && A->getParent() == 0) { + A->replaceAllUsesWith(UndefValue::get(A->getType())); + delete A; + } + } + return Error("Never resolved value found in function!"); + } + } + + // Trim the value list down to the size it was before we parsed this function. + ValueList.shrinkTo(ModuleValueListSize); + std::vector<BasicBlock*>().swap(FunctionBBs); + + return false; +} + +//===----------------------------------------------------------------------===// +// ModuleProvider implementation +//===----------------------------------------------------------------------===// + + +bool BitcodeReader::materializeFunction(Function *F, std::string *ErrInfo) { + // If it already is material, ignore the request. + if (!F->hasNotBeenReadFromBitcode()) return false; + + DenseMap<Function*, std::pair<uint64_t, unsigned> >::iterator DFII = + DeferredFunctionInfo.find(F); + assert(DFII != DeferredFunctionInfo.end() && "Deferred function not found!"); + + // Move the bit stream to the saved position of the deferred function body and + // restore the real linkage type for the function. + Stream.JumpToBit(DFII->second.first); + F->setLinkage((GlobalValue::LinkageTypes)DFII->second.second); + + if (ParseFunctionBody(F)) { + if (ErrInfo) *ErrInfo = ErrorString; + return true; + } + + // Upgrade any old intrinsic calls in the function. + for (UpgradedIntrinsicMap::iterator I = UpgradedIntrinsics.begin(), + E = UpgradedIntrinsics.end(); I != E; ++I) { + if (I->first != I->second) { + for (Value::use_iterator UI = I->first->use_begin(), + UE = I->first->use_end(); UI != UE; ) { + if (CallInst* CI = dyn_cast<CallInst>(*UI++)) + UpgradeIntrinsicCall(CI, I->second); + } + } + } + + return false; +} + +void BitcodeReader::dematerializeFunction(Function *F) { + // If this function isn't materialized, or if it is a proto, this is a noop. + if (F->hasNotBeenReadFromBitcode() || F->isDeclaration()) + return; + + assert(DeferredFunctionInfo.count(F) && "No info to read function later?"); + + // Just forget the function body, we can remat it later. + F->deleteBody(); + F->setLinkage(GlobalValue::GhostLinkage); +} + + +Module *BitcodeReader::materializeModule(std::string *ErrInfo) { + for (DenseMap<Function*, std::pair<uint64_t, unsigned> >::iterator I = + DeferredFunctionInfo.begin(), E = DeferredFunctionInfo.end(); I != E; + ++I) { + Function *F = I->first; + if (F->hasNotBeenReadFromBitcode() && + materializeFunction(F, ErrInfo)) + return 0; + } + + // Upgrade any intrinsic calls that slipped through (should not happen!) and + // delete the old functions to clean up. We can't do this unless the entire + // module is materialized because there could always be another function body + // with calls to the old function. + for (std::vector<std::pair<Function*, Function*> >::iterator I = + UpgradedIntrinsics.begin(), E = UpgradedIntrinsics.end(); I != E; ++I) { + if (I->first != I->second) { + for (Value::use_iterator UI = I->first->use_begin(), + UE = I->first->use_end(); UI != UE; ) { + if (CallInst* CI = dyn_cast<CallInst>(*UI++)) + UpgradeIntrinsicCall(CI, I->second); + } + if (!I->first->use_empty()) + I->first->replaceAllUsesWith(I->second); + I->first->eraseFromParent(); + } + } + std::vector<std::pair<Function*, Function*> >().swap(UpgradedIntrinsics); + + return TheModule; +} + + +/// This method is provided by the parent ModuleProvde class and overriden +/// here. It simply releases the module from its provided and frees up our +/// state. +/// @brief Release our hold on the generated module +Module *BitcodeReader::releaseModule(std::string *ErrInfo) { + // Since we're losing control of this Module, we must hand it back complete + Module *M = ModuleProvider::releaseModule(ErrInfo); + FreeState(); + return M; +} + + +//===----------------------------------------------------------------------===// +// External interface +//===----------------------------------------------------------------------===// + +/// getBitcodeModuleProvider - lazy function-at-a-time loading from a file. +/// +ModuleProvider *llvm::getBitcodeModuleProvider(MemoryBuffer *Buffer, + std::string *ErrMsg) { + BitcodeReader *R = new BitcodeReader(Buffer); + if (R->ParseBitcode()) { + if (ErrMsg) + *ErrMsg = R->getErrorString(); + + // Don't let the BitcodeReader dtor delete 'Buffer'. + R->releaseMemoryBuffer(); + delete R; + return 0; + } + return R; +} + +/// ParseBitcodeFile - Read the specified bitcode file, returning the module. +/// If an error occurs, return null and fill in *ErrMsg if non-null. +Module *llvm::ParseBitcodeFile(MemoryBuffer *Buffer, std::string *ErrMsg){ + BitcodeReader *R; + R = static_cast<BitcodeReader*>(getBitcodeModuleProvider(Buffer, ErrMsg)); + if (!R) return 0; + + // Read in the entire module. + Module *M = R->materializeModule(ErrMsg); + + // Don't let the BitcodeReader dtor delete 'Buffer', regardless of whether + // there was an error. + R->releaseMemoryBuffer(); + + // If there was no error, tell ModuleProvider not to delete it when its dtor + // is run. + if (M) + M = R->releaseModule(ErrMsg); + + delete R; + return M; +} diff --git a/lib/Bitcode/Reader/BitcodeReader.h b/lib/Bitcode/Reader/BitcodeReader.h new file mode 100644 index 0000000..0dc470b --- /dev/null +++ b/lib/Bitcode/Reader/BitcodeReader.h @@ -0,0 +1,214 @@ +//===- BitcodeReader.h - Internal BitcodeReader impl ------------*- C++ -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This header defines the BitcodeReader class. +// +//===----------------------------------------------------------------------===// + +#ifndef BITCODE_READER_H +#define BITCODE_READER_H + +#include "llvm/ModuleProvider.h" +#include "llvm/Attributes.h" +#include "llvm/Type.h" +#include "llvm/OperandTraits.h" +#include "llvm/Bitcode/BitstreamReader.h" +#include "llvm/Bitcode/LLVMBitCodes.h" +#include "llvm/Support/ValueHandle.h" +#include "llvm/ADT/DenseMap.h" +#include <vector> + +namespace llvm { + class MemoryBuffer; + +//===----------------------------------------------------------------------===// +// BitcodeReaderValueList Class +//===----------------------------------------------------------------------===// + +class BitcodeReaderValueList { + std::vector<WeakVH> ValuePtrs; + + /// ResolveConstants - As we resolve forward-referenced constants, we add + /// information about them to this vector. This allows us to resolve them in + /// bulk instead of resolving each reference at a time. See the code in + /// ResolveConstantForwardRefs for more information about this. + /// + /// The key of this vector is the placeholder constant, the value is the slot + /// number that holds the resolved value. + typedef std::vector<std::pair<Constant*, unsigned> > ResolveConstantsTy; + ResolveConstantsTy ResolveConstants; +public: + BitcodeReaderValueList() {} + ~BitcodeReaderValueList() { + assert(ResolveConstants.empty() && "Constants not resolved?"); + } + + // vector compatibility methods + unsigned size() const { return ValuePtrs.size(); } + void resize(unsigned N) { ValuePtrs.resize(N); } + void push_back(Value *V) { + ValuePtrs.push_back(V); + } + + void clear() { + assert(ResolveConstants.empty() && "Constants not resolved?"); + ValuePtrs.clear(); + } + + Value *operator[](unsigned i) const { + assert(i < ValuePtrs.size()); + return ValuePtrs[i]; + } + + Value *back() const { return ValuePtrs.back(); } + void pop_back() { ValuePtrs.pop_back(); } + bool empty() const { return ValuePtrs.empty(); } + void shrinkTo(unsigned N) { + assert(N <= size() && "Invalid shrinkTo request!"); + ValuePtrs.resize(N); + } + + Constant *getConstantFwdRef(unsigned Idx, const Type *Ty); + Value *getValueFwdRef(unsigned Idx, const Type *Ty); + + void AssignValue(Value *V, unsigned Idx); + + /// ResolveConstantForwardRefs - Once all constants are read, this method bulk + /// resolves any forward references. + void ResolveConstantForwardRefs(); +}; + +class BitcodeReader : public ModuleProvider { + MemoryBuffer *Buffer; + BitstreamReader StreamFile; + BitstreamCursor Stream; + + const char *ErrorString; + + std::vector<PATypeHolder> TypeList; + BitcodeReaderValueList ValueList; + std::vector<std::pair<GlobalVariable*, unsigned> > GlobalInits; + std::vector<std::pair<GlobalAlias*, unsigned> > AliasInits; + + /// MAttributes - The set of attributes by index. Index zero in the + /// file is for null, and is thus not represented here. As such all indices + /// are off by one. + std::vector<AttrListPtr> MAttributes; + + /// FunctionBBs - While parsing a function body, this is a list of the basic + /// blocks for the function. + std::vector<BasicBlock*> FunctionBBs; + + // When reading the module header, this list is populated with functions that + // have bodies later in the file. + std::vector<Function*> FunctionsWithBodies; + + // When intrinsic functions are encountered which require upgrading they are + // stored here with their replacement function. + typedef std::vector<std::pair<Function*, Function*> > UpgradedIntrinsicMap; + UpgradedIntrinsicMap UpgradedIntrinsics; + + // After the module header has been read, the FunctionsWithBodies list is + // reversed. This keeps track of whether we've done this yet. + bool HasReversedFunctionsWithBodies; + + /// DeferredFunctionInfo - When function bodies are initially scanned, this + /// map contains info about where to find deferred function body (in the + /// stream) and what linkage the original function had. + DenseMap<Function*, std::pair<uint64_t, unsigned> > DeferredFunctionInfo; +public: + explicit BitcodeReader(MemoryBuffer *buffer) + : Buffer(buffer), ErrorString(0) { + HasReversedFunctionsWithBodies = false; + } + ~BitcodeReader() { + FreeState(); + } + + void FreeState(); + + /// releaseMemoryBuffer - This causes the reader to completely forget about + /// the memory buffer it contains, which prevents the buffer from being + /// destroyed when it is deleted. + void releaseMemoryBuffer() { + Buffer = 0; + } + + virtual bool materializeFunction(Function *F, std::string *ErrInfo = 0); + virtual Module *materializeModule(std::string *ErrInfo = 0); + virtual void dematerializeFunction(Function *F); + virtual Module *releaseModule(std::string *ErrInfo = 0); + + bool Error(const char *Str) { + ErrorString = Str; + return true; + } + const char *getErrorString() const { return ErrorString; } + + /// @brief Main interface to parsing a bitcode buffer. + /// @returns true if an error occurred. + bool ParseBitcode(); +private: + const Type *getTypeByID(unsigned ID, bool isTypeTable = false); + Value *getFnValueByID(unsigned ID, const Type *Ty) { + return ValueList.getValueFwdRef(ID, Ty); + } + BasicBlock *getBasicBlock(unsigned ID) const { + if (ID >= FunctionBBs.size()) return 0; // Invalid ID + return FunctionBBs[ID]; + } + AttrListPtr getAttributes(unsigned i) const { + if (i-1 < MAttributes.size()) + return MAttributes[i-1]; + return AttrListPtr(); + } + + /// getValueTypePair - Read a value/type pair out of the specified record from + /// slot 'Slot'. Increment Slot past the number of slots used in the record. + /// Return true on failure. + bool getValueTypePair(SmallVector<uint64_t, 64> &Record, unsigned &Slot, + unsigned InstNum, Value *&ResVal) { + if (Slot == Record.size()) return true; + unsigned ValNo = (unsigned)Record[Slot++]; + if (ValNo < InstNum) { + // If this is not a forward reference, just return the value we already + // have. + ResVal = getFnValueByID(ValNo, 0); + return ResVal == 0; + } else if (Slot == Record.size()) { + return true; + } + + unsigned TypeNo = (unsigned)Record[Slot++]; + ResVal = getFnValueByID(ValNo, getTypeByID(TypeNo)); + return ResVal == 0; + } + bool getValue(SmallVector<uint64_t, 64> &Record, unsigned &Slot, + const Type *Ty, Value *&ResVal) { + if (Slot == Record.size()) return true; + unsigned ValNo = (unsigned)Record[Slot++]; + ResVal = getFnValueByID(ValNo, Ty); + return ResVal == 0; + } + + + bool ParseModule(const std::string &ModuleID); + bool ParseAttributeBlock(); + bool ParseTypeTable(); + bool ParseTypeSymbolTable(); + bool ParseValueSymbolTable(); + bool ParseConstants(); + bool RememberAndSkipFunctionBody(); + bool ParseFunctionBody(Function *F); + bool ResolveGlobalAndAliasInits(); +}; + +} // End llvm namespace + +#endif diff --git a/lib/Bitcode/Reader/CMakeLists.txt b/lib/Bitcode/Reader/CMakeLists.txt new file mode 100644 index 0000000..a19c79a --- /dev/null +++ b/lib/Bitcode/Reader/CMakeLists.txt @@ -0,0 +1,7 @@ +add_llvm_library(LLVMBitReader + BitReader.cpp + BitcodeReader.cpp + Deserialize.cpp + DeserializeAPFloat.cpp + DeserializeAPInt.cpp + )
\ No newline at end of file diff --git a/lib/Bitcode/Reader/Deserialize.cpp b/lib/Bitcode/Reader/Deserialize.cpp new file mode 100644 index 0000000..06da6ce --- /dev/null +++ b/lib/Bitcode/Reader/Deserialize.cpp @@ -0,0 +1,454 @@ +//==- Deserialize.cpp - Generic Object Serialization to Bitcode --*- C++ -*-==// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file defines the internal methods used for object serialization. +// +//===----------------------------------------------------------------------===// + +#include "llvm/Bitcode/Deserialize.h" + +#ifdef DEBUG_BACKPATCH +#include "llvm/Support/Streams.h" +#endif + +using namespace llvm; + +Deserializer::Deserializer(BitstreamReader& stream) + : Stream(stream), RecIdx(0), FreeList(NULL), AbbrevNo(0), RecordCode(0) { + + StreamStart = Stream.GetCurrentBitNo(); +} + +Deserializer::~Deserializer() { + assert (RecIdx >= Record.size() && + "Still scanning bitcode record when deserialization completed."); + +#ifdef DEBUG_BACKPATCH + for (MapTy::iterator I=BPatchMap.begin(), E=BPatchMap.end(); I!=E; ++I) + assert (I->first.hasFinalPtr() && + "Some pointers were not backpatched."); +#endif +} + + +bool Deserializer::inRecord() { + if (Record.size() > 0) { + if (RecIdx >= Record.size()) { + RecIdx = 0; + Record.clear(); + AbbrevNo = 0; + return false; + } + else + return true; + } + + return false; +} + +bool Deserializer::AdvanceStream() { + assert (!inRecord() && + "Cannot advance stream. Still processing a record."); + + if (AbbrevNo == bitc::ENTER_SUBBLOCK || + AbbrevNo >= bitc::UNABBREV_RECORD) + return true; + + while (!Stream.AtEndOfStream()) { + + uint64_t Pos = Stream.GetCurrentBitNo(); + AbbrevNo = Stream.ReadCode(); + + switch (AbbrevNo) { + case bitc::ENTER_SUBBLOCK: { + unsigned id = Stream.ReadSubBlockID(); + + // Determine the extent of the block. This is useful for jumping around + // the stream. This is hack: we read the header of the block, save + // the length, and then revert the bitstream to a location just before + // the block is entered. + uint64_t BPos = Stream.GetCurrentBitNo(); + Stream.ReadVBR(bitc::CodeLenWidth); // Skip the code size. + Stream.SkipToWord(); + unsigned NumWords = Stream.Read(bitc::BlockSizeWidth); + Stream.JumpToBit(BPos); + + BlockStack.push_back(Location(Pos,id,NumWords)); + break; + } + + case bitc::END_BLOCK: { + bool x = Stream.ReadBlockEnd(); + assert(!x && "Error at block end."); x=x; + BlockStack.pop_back(); + continue; + } + + case bitc::DEFINE_ABBREV: + Stream.ReadAbbrevRecord(); + continue; + + default: + break; + } + + return true; + } + + return false; +} + +void Deserializer::ReadRecord() { + + while (AdvanceStream() && AbbrevNo == bitc::ENTER_SUBBLOCK) { + assert (!BlockStack.empty()); + Stream.EnterSubBlock(BlockStack.back().BlockID); + AbbrevNo = 0; + } + + if (Stream.AtEndOfStream()) + return; + + assert (Record.empty()); + assert (AbbrevNo >= bitc::UNABBREV_RECORD); + RecordCode = Stream.ReadRecord(AbbrevNo,Record); + assert (Record.size() > 0); +} + +void Deserializer::SkipBlock() { + assert (!inRecord()); + + if (AtEnd()) + return; + + AdvanceStream(); + + assert (AbbrevNo == bitc::ENTER_SUBBLOCK); + BlockStack.pop_back(); + Stream.SkipBlock(); + + AbbrevNo = 0; + AdvanceStream(); +} + +bool Deserializer::SkipToBlock(unsigned BlockID) { + assert (!inRecord()); + + AdvanceStream(); + assert (AbbrevNo == bitc::ENTER_SUBBLOCK); + + unsigned BlockLevel = BlockStack.size(); + + while (!AtEnd() && + BlockLevel == BlockStack.size() && + getCurrentBlockID() != BlockID) + SkipBlock(); + + return !(AtEnd() || BlockLevel != BlockStack.size()); +} + +Deserializer::Location Deserializer::getCurrentBlockLocation() { + if (!inRecord()) + AdvanceStream(); + + return BlockStack.back(); +} + +bool Deserializer::JumpTo(const Location& Loc) { + + assert (!inRecord()); + + AdvanceStream(); + + assert (!BlockStack.empty() || AtEnd()); + + uint64_t LastBPos = StreamStart; + + while (!BlockStack.empty()) { + + LastBPos = BlockStack.back().BitNo; + + // Determine of the current block contains the location of the block + // we are looking for. + if (BlockStack.back().contains(Loc)) { + // We found the enclosing block. We must first POP it off to + // destroy any accumulated context within the block scope. We then + // jump to the position of the block and enter it. + Stream.JumpToBit(LastBPos); + + if (BlockStack.size() == Stream.BlockScope.size()) + Stream.PopBlockScope(); + + BlockStack.pop_back(); + + AbbrevNo = 0; + AdvanceStream(); + assert (AbbrevNo == bitc::ENTER_SUBBLOCK); + + Stream.EnterSubBlock(BlockStack.back().BlockID); + break; + } + + // This block does not contain the block we are looking for. Pop it. + if (BlockStack.size() == Stream.BlockScope.size()) + Stream.PopBlockScope(); + + BlockStack.pop_back(); + + } + + // Check if we have popped our way to the outermost scope. If so, + // we need to adjust our position. + if (BlockStack.empty()) { + assert (Stream.BlockScope.empty()); + + Stream.JumpToBit(Loc.BitNo < LastBPos ? StreamStart : LastBPos); + AbbrevNo = 0; + AdvanceStream(); + } + + assert (AbbrevNo == bitc::ENTER_SUBBLOCK); + assert (!BlockStack.empty()); + + while (!AtEnd() && BlockStack.back() != Loc) { + if (BlockStack.back().contains(Loc)) { + Stream.EnterSubBlock(BlockStack.back().BlockID); + AbbrevNo = 0; + AdvanceStream(); + continue; + } + else + SkipBlock(); + } + + if (AtEnd()) + return false; + + assert (BlockStack.back() == Loc); + + return true; +} + +void Deserializer::Rewind() { + while (!Stream.BlockScope.empty()) + Stream.PopBlockScope(); + + while (!BlockStack.empty()) + BlockStack.pop_back(); + + Stream.JumpToBit(StreamStart); + AbbrevNo = 0; +} + + +unsigned Deserializer::getCurrentBlockID() { + if (!inRecord()) + AdvanceStream(); + + return BlockStack.back().BlockID; +} + +unsigned Deserializer::getRecordCode() { + if (!inRecord()) { + AdvanceStream(); + assert (AbbrevNo >= bitc::UNABBREV_RECORD); + ReadRecord(); + } + + return RecordCode; +} + +bool Deserializer::FinishedBlock(Location BlockLoc) { + if (!inRecord()) + AdvanceStream(); + + for (llvm::SmallVector<Location,8>::reverse_iterator + I=BlockStack.rbegin(), E=BlockStack.rend(); I!=E; ++I) + if (*I == BlockLoc) + return false; + + return true; +} + +unsigned Deserializer::getAbbrevNo() { + if (!inRecord()) + AdvanceStream(); + + return AbbrevNo; +} + +bool Deserializer::AtEnd() { + if (inRecord()) + return false; + + if (!AdvanceStream()) + return true; + + return false; +} + +uint64_t Deserializer::ReadInt() { + // FIXME: Any error recovery/handling with incomplete or bad files? + if (!inRecord()) + ReadRecord(); + + return Record[RecIdx++]; +} + +int64_t Deserializer::ReadSInt() { + uint64_t x = ReadInt(); + int64_t magnitude = x >> 1; + return x & 0x1 ? -magnitude : magnitude; +} + +char* Deserializer::ReadCStr(char* cstr, unsigned MaxLen, bool isNullTerm) { + if (cstr == NULL) + MaxLen = 0; // Zero this just in case someone does something funny. + + unsigned len = ReadInt(); + + assert (MaxLen == 0 || (len + (isNullTerm ? 1 : 0)) <= MaxLen); + + if (!cstr) + cstr = new char[len + (isNullTerm ? 1 : 0)]; + + assert (cstr != NULL); + + for (unsigned i = 0; i < len; ++i) + cstr[i] = (char) ReadInt(); + + if (isNullTerm) + cstr[len] = '\0'; + + return cstr; +} + +void Deserializer::ReadCStr(std::vector<char>& buff, bool isNullTerm, + unsigned Idx) { + + unsigned len = ReadInt(); + + // If Idx is beyond the current before size, reduce Idx to refer to the + // element after the last element. + if (Idx > buff.size()) + Idx = buff.size(); + + buff.reserve(len+Idx); + buff.resize(Idx); + + for (unsigned i = 0; i < len; ++i) + buff.push_back((char) ReadInt()); + + if (isNullTerm) + buff.push_back('\0'); +} + +void Deserializer::RegisterPtr(const SerializedPtrID& PtrId, + const void* Ptr) { + + MapTy::value_type& E = BPatchMap.FindAndConstruct(BPKey(PtrId)); + + assert (!HasFinalPtr(E) && "Pointer already registered."); + +#ifdef DEBUG_BACKPATCH + llvm::cerr << "RegisterPtr: " << PtrId << " => " << Ptr << "\n"; +#endif + + SetPtr(E,Ptr); +} + +void Deserializer::ReadUIntPtr(uintptr_t& PtrRef, + const SerializedPtrID& PtrId, + bool AllowBackpatch) { + if (PtrId == 0) { + PtrRef = 0; + return; + } + + MapTy::value_type& E = BPatchMap.FindAndConstruct(BPKey(PtrId)); + + if (HasFinalPtr(E)) { + PtrRef = GetFinalPtr(E); + +#ifdef DEBUG_BACKPATCH + llvm::cerr << "ReadUintPtr: " << PtrId + << " <-- " << (void*) GetFinalPtr(E) << '\n'; +#endif + } + else { + assert (AllowBackpatch && + "Client forbids backpatching for this pointer."); + +#ifdef DEBUG_BACKPATCH + llvm::cerr << "ReadUintPtr: " << PtrId << " (NO PTR YET)\n"; +#endif + + // Register backpatch. Check the freelist for a BPNode. + BPNode* N; + + if (FreeList) { + N = FreeList; + FreeList = FreeList->Next; + } + else // No available BPNode. Allocate one. + N = (BPNode*) Allocator.Allocate<BPNode>(); + + new (N) BPNode(GetBPNode(E),PtrRef); + SetBPNode(E,N); + } +} + +uintptr_t Deserializer::ReadInternalRefPtr() { + SerializedPtrID PtrId = ReadPtrID(); + + assert (PtrId != 0 && "References cannot refer the NULL address."); + + MapTy::value_type& E = BPatchMap.FindAndConstruct(BPKey(PtrId)); + + assert (HasFinalPtr(E) && + "Cannot backpatch references. Object must be already deserialized."); + + return GetFinalPtr(E); +} + +void Deserializer::BPEntry::SetPtr(BPNode*& FreeList, void* P) { + BPNode* Last = NULL; + + for (BPNode* N = Head; N != NULL; N=N->Next) { + Last = N; + N->PtrRef |= reinterpret_cast<uintptr_t>(P); + } + + if (Last) { + Last->Next = FreeList; + FreeList = Head; + } + + Ptr = const_cast<void*>(P); +} + + +#define INT_READ(TYPE)\ +void SerializeTrait<TYPE>::Read(Deserializer& D, TYPE& X) {\ + X = (TYPE) D.ReadInt(); } + +INT_READ(bool) +INT_READ(unsigned char) +INT_READ(unsigned short) +INT_READ(unsigned int) +INT_READ(unsigned long) + +#define SINT_READ(TYPE)\ +void SerializeTrait<TYPE>::Read(Deserializer& D, TYPE& X) {\ + X = (TYPE) D.ReadSInt(); } + +INT_READ(signed char) +INT_READ(signed short) +INT_READ(signed int) +INT_READ(signed long) diff --git a/lib/Bitcode/Reader/DeserializeAPFloat.cpp b/lib/Bitcode/Reader/DeserializeAPFloat.cpp new file mode 100644 index 0000000..ee24b68 --- /dev/null +++ b/lib/Bitcode/Reader/DeserializeAPFloat.cpp @@ -0,0 +1,24 @@ +//===-- SerializeAPInt.cpp - Serialization for APFloat ---------*- C++ -*--===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file implements deserialization of APFloat. +// +//===----------------------------------------------------------------------===// + +#include "llvm/ADT/APFloat.h" +#include "llvm/Bitcode/Deserialize.h" + +using namespace llvm; + +APFloat APFloat::ReadVal(Deserializer& D) { + APInt x; + D.Read(x); + return APFloat(x); +} + diff --git a/lib/Bitcode/Reader/DeserializeAPInt.cpp b/lib/Bitcode/Reader/DeserializeAPInt.cpp new file mode 100644 index 0000000..1b5b2bf --- /dev/null +++ b/lib/Bitcode/Reader/DeserializeAPInt.cpp @@ -0,0 +1,33 @@ +//===-- DeserializeAPInt.cpp - Deserialization for APInts ------*- C++ -*--===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file implements deserialization of APInts. +// +//===----------------------------------------------------------------------===// + +#include "llvm/ADT/APInt.h" +#include "llvm/Bitcode/Deserialize.h" +#include <cassert> + +using namespace llvm; + +void APInt::Read(Deserializer& D) { + BitWidth = D.ReadInt(); + + if (isSingleWord()) + VAL = D.ReadInt(); + else { + uint32_t NumWords = D.ReadInt(); + assert (NumWords > 1); + pVal = new uint64_t[NumWords]; + assert (pVal && "Allocation in deserialization of APInt failed."); + for (unsigned i = 0; i < NumWords; ++i) + pVal[i] = D.ReadInt(); + } +} diff --git a/lib/Bitcode/Reader/Makefile b/lib/Bitcode/Reader/Makefile new file mode 100644 index 0000000..59af8d53 --- /dev/null +++ b/lib/Bitcode/Reader/Makefile @@ -0,0 +1,15 @@ +##===- lib/Bitcode/Reader/Makefile -------------------------*- Makefile -*-===## +# +# The LLVM Compiler Infrastructure +# +# This file is distributed under the University of Illinois Open Source +# License. See LICENSE.TXT for details. +# +##===----------------------------------------------------------------------===## + +LEVEL = ../../.. +LIBRARYNAME = LLVMBitReader +BUILD_ARCHIVE = 1 + +include $(LEVEL)/Makefile.common + diff --git a/lib/Bitcode/Writer/BitWriter.cpp b/lib/Bitcode/Writer/BitWriter.cpp new file mode 100644 index 0000000..8834964 --- /dev/null +++ b/lib/Bitcode/Writer/BitWriter.cpp @@ -0,0 +1,58 @@ +//===-- BitWriter.cpp -----------------------------------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +#include "llvm-c/BitWriter.h" +#include "llvm/Bitcode/ReaderWriter.h" +#include <fstream> + +using namespace llvm; + + +/*===-- Operations on modules ---------------------------------------------===*/ + +int LLVMWriteBitcodeToFile(LLVMModuleRef M, const char *Path) { + std::ofstream OS(Path, std::ios_base::out|std::ios::trunc|std::ios::binary); + + if (!OS.fail()) + WriteBitcodeToFile(unwrap(M), OS); + + if (OS.fail()) + return -1; + + return 0; +} + +#if defined(__GNUC__) && (__GNUC__ > 3 || __GNUC__ == 3 && __GNUC_MINOR >= 4) +#include <ext/stdio_filebuf.h> + +// FIXME: Control this with configure? Provide some portable abstraction in +// libSystem? As is, the user will just get a linker error if they use this on +// non-GCC. Some C++ stdlibs even have ofstream::ofstream(int fd). +int LLVMWriteBitcodeToFileHandle(LLVMModuleRef M, int FileHandle) { + __gnu_cxx::stdio_filebuf<char> Buffer(FileHandle, std::ios_base::out | + std::ios::trunc | + std::ios::binary); + std::ostream OS(&Buffer); + + if (!OS.fail()) + WriteBitcodeToFile(unwrap(M), OS); + + if (OS.fail()) + return -1; + + return 0; +} + +#else + +int LLVMWriteBitcodeToFileHandle(LLVMModuleRef M, int FileHandle) { + return -1; // Not supported. +} + +#endif diff --git a/lib/Bitcode/Writer/BitcodeWriter.cpp b/lib/Bitcode/Writer/BitcodeWriter.cpp new file mode 100644 index 0000000..bfc029c --- /dev/null +++ b/lib/Bitcode/Writer/BitcodeWriter.cpp @@ -0,0 +1,1449 @@ +//===--- Bitcode/Writer/BitcodeWriter.cpp - Bitcode Writer ----------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// Bitcode writer implementation. +// +//===----------------------------------------------------------------------===// + +#include "llvm/Bitcode/ReaderWriter.h" +#include "llvm/Bitcode/BitstreamWriter.h" +#include "llvm/Bitcode/LLVMBitCodes.h" +#include "ValueEnumerator.h" +#include "llvm/Constants.h" +#include "llvm/DerivedTypes.h" +#include "llvm/InlineAsm.h" +#include "llvm/Instructions.h" +#include "llvm/MDNode.h" +#include "llvm/Module.h" +#include "llvm/TypeSymbolTable.h" +#include "llvm/ValueSymbolTable.h" +#include "llvm/Support/MathExtras.h" +#include "llvm/Support/Streams.h" +#include "llvm/Support/raw_ostream.h" +#include "llvm/System/Program.h" +using namespace llvm; + +/// These are manifest constants used by the bitcode writer. They do not need to +/// be kept in sync with the reader, but need to be consistent within this file. +enum { + CurVersion = 0, + + // VALUE_SYMTAB_BLOCK abbrev id's. + VST_ENTRY_8_ABBREV = bitc::FIRST_APPLICATION_ABBREV, + VST_ENTRY_7_ABBREV, + VST_ENTRY_6_ABBREV, + VST_BBENTRY_6_ABBREV, + + // CONSTANTS_BLOCK abbrev id's. + CONSTANTS_SETTYPE_ABBREV = bitc::FIRST_APPLICATION_ABBREV, + CONSTANTS_INTEGER_ABBREV, + CONSTANTS_CE_CAST_Abbrev, + CONSTANTS_NULL_Abbrev, + + // FUNCTION_BLOCK abbrev id's. + FUNCTION_INST_LOAD_ABBREV = bitc::FIRST_APPLICATION_ABBREV, + FUNCTION_INST_BINOP_ABBREV, + FUNCTION_INST_CAST_ABBREV, + FUNCTION_INST_RET_VOID_ABBREV, + FUNCTION_INST_RET_VAL_ABBREV, + FUNCTION_INST_UNREACHABLE_ABBREV +}; + + +static unsigned GetEncodedCastOpcode(unsigned Opcode) { + switch (Opcode) { + default: assert(0 && "Unknown cast instruction!"); + case Instruction::Trunc : return bitc::CAST_TRUNC; + case Instruction::ZExt : return bitc::CAST_ZEXT; + case Instruction::SExt : return bitc::CAST_SEXT; + case Instruction::FPToUI : return bitc::CAST_FPTOUI; + case Instruction::FPToSI : return bitc::CAST_FPTOSI; + case Instruction::UIToFP : return bitc::CAST_UITOFP; + case Instruction::SIToFP : return bitc::CAST_SITOFP; + case Instruction::FPTrunc : return bitc::CAST_FPTRUNC; + case Instruction::FPExt : return bitc::CAST_FPEXT; + case Instruction::PtrToInt: return bitc::CAST_PTRTOINT; + case Instruction::IntToPtr: return bitc::CAST_INTTOPTR; + case Instruction::BitCast : return bitc::CAST_BITCAST; + } +} + +static unsigned GetEncodedBinaryOpcode(unsigned Opcode) { + switch (Opcode) { + default: assert(0 && "Unknown binary instruction!"); + case Instruction::Add: return bitc::BINOP_ADD; + case Instruction::Sub: return bitc::BINOP_SUB; + case Instruction::Mul: return bitc::BINOP_MUL; + case Instruction::UDiv: return bitc::BINOP_UDIV; + case Instruction::FDiv: + case Instruction::SDiv: return bitc::BINOP_SDIV; + case Instruction::URem: return bitc::BINOP_UREM; + case Instruction::FRem: + case Instruction::SRem: return bitc::BINOP_SREM; + case Instruction::Shl: return bitc::BINOP_SHL; + case Instruction::LShr: return bitc::BINOP_LSHR; + case Instruction::AShr: return bitc::BINOP_ASHR; + case Instruction::And: return bitc::BINOP_AND; + case Instruction::Or: return bitc::BINOP_OR; + case Instruction::Xor: return bitc::BINOP_XOR; + } +} + + + +static void WriteStringRecord(unsigned Code, const std::string &Str, + unsigned AbbrevToUse, BitstreamWriter &Stream) { + SmallVector<unsigned, 64> Vals; + + // Code: [strchar x N] + for (unsigned i = 0, e = Str.size(); i != e; ++i) + Vals.push_back(Str[i]); + + // Emit the finished record. + Stream.EmitRecord(Code, Vals, AbbrevToUse); +} + +// Emit information about parameter attributes. +static void WriteAttributeTable(const ValueEnumerator &VE, + BitstreamWriter &Stream) { + const std::vector<AttrListPtr> &Attrs = VE.getAttributes(); + if (Attrs.empty()) return; + + Stream.EnterSubblock(bitc::PARAMATTR_BLOCK_ID, 3); + + SmallVector<uint64_t, 64> Record; + for (unsigned i = 0, e = Attrs.size(); i != e; ++i) { + const AttrListPtr &A = Attrs[i]; + for (unsigned i = 0, e = A.getNumSlots(); i != e; ++i) { + const AttributeWithIndex &PAWI = A.getSlot(i); + Record.push_back(PAWI.Index); + + // FIXME: remove in LLVM 3.0 + // Store the alignment in the bitcode as a 16-bit raw value instead of a + // 5-bit log2 encoded value. Shift the bits above the alignment up by + // 11 bits. + uint64_t FauxAttr = PAWI.Attrs & 0xffff; + if (PAWI.Attrs & Attribute::Alignment) + FauxAttr |= (1ull<<16)<<(((PAWI.Attrs & Attribute::Alignment)-1) >> 16); + FauxAttr |= (PAWI.Attrs & (0x3FFull << 21)) << 11; + + Record.push_back(FauxAttr); + } + + Stream.EmitRecord(bitc::PARAMATTR_CODE_ENTRY, Record); + Record.clear(); + } + + Stream.ExitBlock(); +} + +/// WriteTypeTable - Write out the type table for a module. +static void WriteTypeTable(const ValueEnumerator &VE, BitstreamWriter &Stream) { + const ValueEnumerator::TypeList &TypeList = VE.getTypes(); + + Stream.EnterSubblock(bitc::TYPE_BLOCK_ID, 4 /*count from # abbrevs */); + SmallVector<uint64_t, 64> TypeVals; + + // Abbrev for TYPE_CODE_POINTER. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::TYPE_CODE_POINTER)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, + Log2_32_Ceil(VE.getTypes().size()+1))); + Abbv->Add(BitCodeAbbrevOp(0)); // Addrspace = 0 + unsigned PtrAbbrev = Stream.EmitAbbrev(Abbv); + + // Abbrev for TYPE_CODE_FUNCTION. + Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::TYPE_CODE_FUNCTION)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 1)); // isvararg + Abbv->Add(BitCodeAbbrevOp(0)); // FIXME: DEAD value, remove in LLVM 3.0 + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Array)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, + Log2_32_Ceil(VE.getTypes().size()+1))); + unsigned FunctionAbbrev = Stream.EmitAbbrev(Abbv); + + // Abbrev for TYPE_CODE_STRUCT. + Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::TYPE_CODE_STRUCT)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 1)); // ispacked + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Array)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, + Log2_32_Ceil(VE.getTypes().size()+1))); + unsigned StructAbbrev = Stream.EmitAbbrev(Abbv); + + // Abbrev for TYPE_CODE_ARRAY. + Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::TYPE_CODE_ARRAY)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 8)); // size + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, + Log2_32_Ceil(VE.getTypes().size()+1))); + unsigned ArrayAbbrev = Stream.EmitAbbrev(Abbv); + + // Emit an entry count so the reader can reserve space. + TypeVals.push_back(TypeList.size()); + Stream.EmitRecord(bitc::TYPE_CODE_NUMENTRY, TypeVals); + TypeVals.clear(); + + // Loop over all of the types, emitting each in turn. + for (unsigned i = 0, e = TypeList.size(); i != e; ++i) { + const Type *T = TypeList[i].first; + int AbbrevToUse = 0; + unsigned Code = 0; + + switch (T->getTypeID()) { + default: assert(0 && "Unknown type!"); + case Type::VoidTyID: Code = bitc::TYPE_CODE_VOID; break; + case Type::FloatTyID: Code = bitc::TYPE_CODE_FLOAT; break; + case Type::DoubleTyID: Code = bitc::TYPE_CODE_DOUBLE; break; + case Type::X86_FP80TyID: Code = bitc::TYPE_CODE_X86_FP80; break; + case Type::FP128TyID: Code = bitc::TYPE_CODE_FP128; break; + case Type::PPC_FP128TyID: Code = bitc::TYPE_CODE_PPC_FP128; break; + case Type::LabelTyID: Code = bitc::TYPE_CODE_LABEL; break; + case Type::OpaqueTyID: Code = bitc::TYPE_CODE_OPAQUE; break; + case Type::MetadataTyID: Code = bitc::TYPE_CODE_METADATA; break; + case Type::IntegerTyID: + // INTEGER: [width] + Code = bitc::TYPE_CODE_INTEGER; + TypeVals.push_back(cast<IntegerType>(T)->getBitWidth()); + break; + case Type::PointerTyID: { + const PointerType *PTy = cast<PointerType>(T); + // POINTER: [pointee type, address space] + Code = bitc::TYPE_CODE_POINTER; + TypeVals.push_back(VE.getTypeID(PTy->getElementType())); + unsigned AddressSpace = PTy->getAddressSpace(); + TypeVals.push_back(AddressSpace); + if (AddressSpace == 0) AbbrevToUse = PtrAbbrev; + break; + } + case Type::FunctionTyID: { + const FunctionType *FT = cast<FunctionType>(T); + // FUNCTION: [isvararg, attrid, retty, paramty x N] + Code = bitc::TYPE_CODE_FUNCTION; + TypeVals.push_back(FT->isVarArg()); + TypeVals.push_back(0); // FIXME: DEAD: remove in llvm 3.0 + TypeVals.push_back(VE.getTypeID(FT->getReturnType())); + for (unsigned i = 0, e = FT->getNumParams(); i != e; ++i) + TypeVals.push_back(VE.getTypeID(FT->getParamType(i))); + AbbrevToUse = FunctionAbbrev; + break; + } + case Type::StructTyID: { + const StructType *ST = cast<StructType>(T); + // STRUCT: [ispacked, eltty x N] + Code = bitc::TYPE_CODE_STRUCT; + TypeVals.push_back(ST->isPacked()); + // Output all of the element types. + for (StructType::element_iterator I = ST->element_begin(), + E = ST->element_end(); I != E; ++I) + TypeVals.push_back(VE.getTypeID(*I)); + AbbrevToUse = StructAbbrev; + break; + } + case Type::ArrayTyID: { + const ArrayType *AT = cast<ArrayType>(T); + // ARRAY: [numelts, eltty] + Code = bitc::TYPE_CODE_ARRAY; + TypeVals.push_back(AT->getNumElements()); + TypeVals.push_back(VE.getTypeID(AT->getElementType())); + AbbrevToUse = ArrayAbbrev; + break; + } + case Type::VectorTyID: { + const VectorType *VT = cast<VectorType>(T); + // VECTOR [numelts, eltty] + Code = bitc::TYPE_CODE_VECTOR; + TypeVals.push_back(VT->getNumElements()); + TypeVals.push_back(VE.getTypeID(VT->getElementType())); + break; + } + } + + // Emit the finished record. + Stream.EmitRecord(Code, TypeVals, AbbrevToUse); + TypeVals.clear(); + } + + Stream.ExitBlock(); +} + +static unsigned getEncodedLinkage(const GlobalValue *GV) { + switch (GV->getLinkage()) { + default: assert(0 && "Invalid linkage!"); + case GlobalValue::GhostLinkage: // Map ghost linkage onto external. + case GlobalValue::ExternalLinkage: return 0; + case GlobalValue::WeakAnyLinkage: return 1; + case GlobalValue::AppendingLinkage: return 2; + case GlobalValue::InternalLinkage: return 3; + case GlobalValue::LinkOnceAnyLinkage: return 4; + case GlobalValue::DLLImportLinkage: return 5; + case GlobalValue::DLLExportLinkage: return 6; + case GlobalValue::ExternalWeakLinkage: return 7; + case GlobalValue::CommonLinkage: return 8; + case GlobalValue::PrivateLinkage: return 9; + case GlobalValue::WeakODRLinkage: return 10; + case GlobalValue::LinkOnceODRLinkage: return 11; + case GlobalValue::AvailableExternallyLinkage: return 12; + } +} + +static unsigned getEncodedVisibility(const GlobalValue *GV) { + switch (GV->getVisibility()) { + default: assert(0 && "Invalid visibility!"); + case GlobalValue::DefaultVisibility: return 0; + case GlobalValue::HiddenVisibility: return 1; + case GlobalValue::ProtectedVisibility: return 2; + } +} + +// Emit top-level description of module, including target triple, inline asm, +// descriptors for global variables, and function prototype info. +static void WriteModuleInfo(const Module *M, const ValueEnumerator &VE, + BitstreamWriter &Stream) { + // Emit the list of dependent libraries for the Module. + for (Module::lib_iterator I = M->lib_begin(), E = M->lib_end(); I != E; ++I) + WriteStringRecord(bitc::MODULE_CODE_DEPLIB, *I, 0/*TODO*/, Stream); + + // Emit various pieces of data attached to a module. + if (!M->getTargetTriple().empty()) + WriteStringRecord(bitc::MODULE_CODE_TRIPLE, M->getTargetTriple(), + 0/*TODO*/, Stream); + if (!M->getDataLayout().empty()) + WriteStringRecord(bitc::MODULE_CODE_DATALAYOUT, M->getDataLayout(), + 0/*TODO*/, Stream); + if (!M->getModuleInlineAsm().empty()) + WriteStringRecord(bitc::MODULE_CODE_ASM, M->getModuleInlineAsm(), + 0/*TODO*/, Stream); + + // Emit information about sections and GC, computing how many there are. Also + // compute the maximum alignment value. + std::map<std::string, unsigned> SectionMap; + std::map<std::string, unsigned> GCMap; + unsigned MaxAlignment = 0; + unsigned MaxGlobalType = 0; + for (Module::const_global_iterator GV = M->global_begin(),E = M->global_end(); + GV != E; ++GV) { + MaxAlignment = std::max(MaxAlignment, GV->getAlignment()); + MaxGlobalType = std::max(MaxGlobalType, VE.getTypeID(GV->getType())); + + if (!GV->hasSection()) continue; + // Give section names unique ID's. + unsigned &Entry = SectionMap[GV->getSection()]; + if (Entry != 0) continue; + WriteStringRecord(bitc::MODULE_CODE_SECTIONNAME, GV->getSection(), + 0/*TODO*/, Stream); + Entry = SectionMap.size(); + } + for (Module::const_iterator F = M->begin(), E = M->end(); F != E; ++F) { + MaxAlignment = std::max(MaxAlignment, F->getAlignment()); + if (F->hasSection()) { + // Give section names unique ID's. + unsigned &Entry = SectionMap[F->getSection()]; + if (!Entry) { + WriteStringRecord(bitc::MODULE_CODE_SECTIONNAME, F->getSection(), + 0/*TODO*/, Stream); + Entry = SectionMap.size(); + } + } + if (F->hasGC()) { + // Same for GC names. + unsigned &Entry = GCMap[F->getGC()]; + if (!Entry) { + WriteStringRecord(bitc::MODULE_CODE_GCNAME, F->getGC(), + 0/*TODO*/, Stream); + Entry = GCMap.size(); + } + } + } + + // Emit abbrev for globals, now that we know # sections and max alignment. + unsigned SimpleGVarAbbrev = 0; + if (!M->global_empty()) { + // Add an abbrev for common globals with no visibility or thread localness. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::MODULE_CODE_GLOBALVAR)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, + Log2_32_Ceil(MaxGlobalType+1))); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 1)); // Constant. + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 6)); // Initializer. + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 4)); // Linkage. + if (MaxAlignment == 0) // Alignment. + Abbv->Add(BitCodeAbbrevOp(0)); + else { + unsigned MaxEncAlignment = Log2_32(MaxAlignment)+1; + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, + Log2_32_Ceil(MaxEncAlignment+1))); + } + if (SectionMap.empty()) // Section. + Abbv->Add(BitCodeAbbrevOp(0)); + else + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, + Log2_32_Ceil(SectionMap.size()+1))); + // Don't bother emitting vis + thread local. + SimpleGVarAbbrev = Stream.EmitAbbrev(Abbv); + } + + // Emit the global variable information. + SmallVector<unsigned, 64> Vals; + for (Module::const_global_iterator GV = M->global_begin(),E = M->global_end(); + GV != E; ++GV) { + unsigned AbbrevToUse = 0; + + // GLOBALVAR: [type, isconst, initid, + // linkage, alignment, section, visibility, threadlocal] + Vals.push_back(VE.getTypeID(GV->getType())); + Vals.push_back(GV->isConstant()); + Vals.push_back(GV->isDeclaration() ? 0 : + (VE.getValueID(GV->getInitializer()) + 1)); + Vals.push_back(getEncodedLinkage(GV)); + Vals.push_back(Log2_32(GV->getAlignment())+1); + Vals.push_back(GV->hasSection() ? SectionMap[GV->getSection()] : 0); + if (GV->isThreadLocal() || + GV->getVisibility() != GlobalValue::DefaultVisibility) { + Vals.push_back(getEncodedVisibility(GV)); + Vals.push_back(GV->isThreadLocal()); + } else { + AbbrevToUse = SimpleGVarAbbrev; + } + + Stream.EmitRecord(bitc::MODULE_CODE_GLOBALVAR, Vals, AbbrevToUse); + Vals.clear(); + } + + // Emit the function proto information. + for (Module::const_iterator F = M->begin(), E = M->end(); F != E; ++F) { + // FUNCTION: [type, callingconv, isproto, paramattr, + // linkage, alignment, section, visibility, gc] + Vals.push_back(VE.getTypeID(F->getType())); + Vals.push_back(F->getCallingConv()); + Vals.push_back(F->isDeclaration()); + Vals.push_back(getEncodedLinkage(F)); + Vals.push_back(VE.getAttributeID(F->getAttributes())); + Vals.push_back(Log2_32(F->getAlignment())+1); + Vals.push_back(F->hasSection() ? SectionMap[F->getSection()] : 0); + Vals.push_back(getEncodedVisibility(F)); + Vals.push_back(F->hasGC() ? GCMap[F->getGC()] : 0); + + unsigned AbbrevToUse = 0; + Stream.EmitRecord(bitc::MODULE_CODE_FUNCTION, Vals, AbbrevToUse); + Vals.clear(); + } + + + // Emit the alias information. + for (Module::const_alias_iterator AI = M->alias_begin(), E = M->alias_end(); + AI != E; ++AI) { + Vals.push_back(VE.getTypeID(AI->getType())); + Vals.push_back(VE.getValueID(AI->getAliasee())); + Vals.push_back(getEncodedLinkage(AI)); + Vals.push_back(getEncodedVisibility(AI)); + unsigned AbbrevToUse = 0; + Stream.EmitRecord(bitc::MODULE_CODE_ALIAS, Vals, AbbrevToUse); + Vals.clear(); + } +} + + +static void WriteConstants(unsigned FirstVal, unsigned LastVal, + const ValueEnumerator &VE, + BitstreamWriter &Stream, bool isGlobal) { + if (FirstVal == LastVal) return; + + Stream.EnterSubblock(bitc::CONSTANTS_BLOCK_ID, 4); + + unsigned AggregateAbbrev = 0; + unsigned String8Abbrev = 0; + unsigned CString7Abbrev = 0; + unsigned CString6Abbrev = 0; + unsigned MDString8Abbrev = 0; + unsigned MDString6Abbrev = 0; + // If this is a constant pool for the module, emit module-specific abbrevs. + if (isGlobal) { + // Abbrev for CST_CODE_AGGREGATE. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::CST_CODE_AGGREGATE)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Array)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, Log2_32_Ceil(LastVal+1))); + AggregateAbbrev = Stream.EmitAbbrev(Abbv); + + // Abbrev for CST_CODE_STRING. + Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::CST_CODE_STRING)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Array)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 8)); + String8Abbrev = Stream.EmitAbbrev(Abbv); + // Abbrev for CST_CODE_CSTRING. + Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::CST_CODE_CSTRING)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Array)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 7)); + CString7Abbrev = Stream.EmitAbbrev(Abbv); + // Abbrev for CST_CODE_CSTRING. + Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::CST_CODE_CSTRING)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Array)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Char6)); + CString6Abbrev = Stream.EmitAbbrev(Abbv); + + // Abbrev for CST_CODE_MDSTRING. + Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::CST_CODE_MDSTRING)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Array)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 8)); + MDString8Abbrev = Stream.EmitAbbrev(Abbv); + // Abbrev for CST_CODE_MDSTRING. + Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::CST_CODE_MDSTRING)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Array)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Char6)); + MDString6Abbrev = Stream.EmitAbbrev(Abbv); + } + + SmallVector<uint64_t, 64> Record; + + const ValueEnumerator::ValueList &Vals = VE.getValues(); + const Type *LastTy = 0; + for (unsigned i = FirstVal; i != LastVal; ++i) { + const Value *V = Vals[i].first; + // If we need to switch types, do so now. + if (V->getType() != LastTy) { + LastTy = V->getType(); + Record.push_back(VE.getTypeID(LastTy)); + Stream.EmitRecord(bitc::CST_CODE_SETTYPE, Record, + CONSTANTS_SETTYPE_ABBREV); + Record.clear(); + } + + if (const InlineAsm *IA = dyn_cast<InlineAsm>(V)) { + Record.push_back(unsigned(IA->hasSideEffects())); + + // Add the asm string. + const std::string &AsmStr = IA->getAsmString(); + Record.push_back(AsmStr.size()); + for (unsigned i = 0, e = AsmStr.size(); i != e; ++i) + Record.push_back(AsmStr[i]); + + // Add the constraint string. + const std::string &ConstraintStr = IA->getConstraintString(); + Record.push_back(ConstraintStr.size()); + for (unsigned i = 0, e = ConstraintStr.size(); i != e; ++i) + Record.push_back(ConstraintStr[i]); + Stream.EmitRecord(bitc::CST_CODE_INLINEASM, Record); + Record.clear(); + continue; + } + const Constant *C = cast<Constant>(V); + unsigned Code = -1U; + unsigned AbbrevToUse = 0; + if (C->isNullValue()) { + Code = bitc::CST_CODE_NULL; + } else if (isa<UndefValue>(C)) { + Code = bitc::CST_CODE_UNDEF; + } else if (const ConstantInt *IV = dyn_cast<ConstantInt>(C)) { + if (IV->getBitWidth() <= 64) { + int64_t V = IV->getSExtValue(); + if (V >= 0) + Record.push_back(V << 1); + else + Record.push_back((-V << 1) | 1); + Code = bitc::CST_CODE_INTEGER; + AbbrevToUse = CONSTANTS_INTEGER_ABBREV; + } else { // Wide integers, > 64 bits in size. + // We have an arbitrary precision integer value to write whose + // bit width is > 64. However, in canonical unsigned integer + // format it is likely that the high bits are going to be zero. + // So, we only write the number of active words. + unsigned NWords = IV->getValue().getActiveWords(); + const uint64_t *RawWords = IV->getValue().getRawData(); + for (unsigned i = 0; i != NWords; ++i) { + int64_t V = RawWords[i]; + if (V >= 0) + Record.push_back(V << 1); + else + Record.push_back((-V << 1) | 1); + } + Code = bitc::CST_CODE_WIDE_INTEGER; + } + } else if (const ConstantFP *CFP = dyn_cast<ConstantFP>(C)) { + Code = bitc::CST_CODE_FLOAT; + const Type *Ty = CFP->getType(); + if (Ty == Type::FloatTy || Ty == Type::DoubleTy) { + Record.push_back(CFP->getValueAPF().bitcastToAPInt().getZExtValue()); + } else if (Ty == Type::X86_FP80Ty) { + // api needed to prevent premature destruction + // bits are not in the same order as a normal i80 APInt, compensate. + APInt api = CFP->getValueAPF().bitcastToAPInt(); + const uint64_t *p = api.getRawData(); + Record.push_back((p[1] << 48) | (p[0] >> 16)); + Record.push_back(p[0] & 0xffffLL); + } else if (Ty == Type::FP128Ty || Ty == Type::PPC_FP128Ty) { + APInt api = CFP->getValueAPF().bitcastToAPInt(); + const uint64_t *p = api.getRawData(); + Record.push_back(p[0]); + Record.push_back(p[1]); + } else { + assert (0 && "Unknown FP type!"); + } + } else if (isa<ConstantArray>(C) && cast<ConstantArray>(C)->isString()) { + // Emit constant strings specially. + unsigned NumOps = C->getNumOperands(); + // If this is a null-terminated string, use the denser CSTRING encoding. + if (C->getOperand(NumOps-1)->isNullValue()) { + Code = bitc::CST_CODE_CSTRING; + --NumOps; // Don't encode the null, which isn't allowed by char6. + } else { + Code = bitc::CST_CODE_STRING; + AbbrevToUse = String8Abbrev; + } + bool isCStr7 = Code == bitc::CST_CODE_CSTRING; + bool isCStrChar6 = Code == bitc::CST_CODE_CSTRING; + for (unsigned i = 0; i != NumOps; ++i) { + unsigned char V = cast<ConstantInt>(C->getOperand(i))->getZExtValue(); + Record.push_back(V); + isCStr7 &= (V & 128) == 0; + if (isCStrChar6) + isCStrChar6 = BitCodeAbbrevOp::isChar6(V); + } + + if (isCStrChar6) + AbbrevToUse = CString6Abbrev; + else if (isCStr7) + AbbrevToUse = CString7Abbrev; + } else if (isa<ConstantArray>(C) || isa<ConstantStruct>(V) || + isa<ConstantVector>(V)) { + Code = bitc::CST_CODE_AGGREGATE; + for (unsigned i = 0, e = C->getNumOperands(); i != e; ++i) + Record.push_back(VE.getValueID(C->getOperand(i))); + AbbrevToUse = AggregateAbbrev; + } else if (const ConstantExpr *CE = dyn_cast<ConstantExpr>(C)) { + switch (CE->getOpcode()) { + default: + if (Instruction::isCast(CE->getOpcode())) { + Code = bitc::CST_CODE_CE_CAST; + Record.push_back(GetEncodedCastOpcode(CE->getOpcode())); + Record.push_back(VE.getTypeID(C->getOperand(0)->getType())); + Record.push_back(VE.getValueID(C->getOperand(0))); + AbbrevToUse = CONSTANTS_CE_CAST_Abbrev; + } else { + assert(CE->getNumOperands() == 2 && "Unknown constant expr!"); + Code = bitc::CST_CODE_CE_BINOP; + Record.push_back(GetEncodedBinaryOpcode(CE->getOpcode())); + Record.push_back(VE.getValueID(C->getOperand(0))); + Record.push_back(VE.getValueID(C->getOperand(1))); + } + break; + case Instruction::GetElementPtr: + Code = bitc::CST_CODE_CE_GEP; + for (unsigned i = 0, e = CE->getNumOperands(); i != e; ++i) { + Record.push_back(VE.getTypeID(C->getOperand(i)->getType())); + Record.push_back(VE.getValueID(C->getOperand(i))); + } + break; + case Instruction::Select: + Code = bitc::CST_CODE_CE_SELECT; + Record.push_back(VE.getValueID(C->getOperand(0))); + Record.push_back(VE.getValueID(C->getOperand(1))); + Record.push_back(VE.getValueID(C->getOperand(2))); + break; + case Instruction::ExtractElement: + Code = bitc::CST_CODE_CE_EXTRACTELT; + Record.push_back(VE.getTypeID(C->getOperand(0)->getType())); + Record.push_back(VE.getValueID(C->getOperand(0))); + Record.push_back(VE.getValueID(C->getOperand(1))); + break; + case Instruction::InsertElement: + Code = bitc::CST_CODE_CE_INSERTELT; + Record.push_back(VE.getValueID(C->getOperand(0))); + Record.push_back(VE.getValueID(C->getOperand(1))); + Record.push_back(VE.getValueID(C->getOperand(2))); + break; + case Instruction::ShuffleVector: + // If the return type and argument types are the same, this is a + // standard shufflevector instruction. If the types are different, + // then the shuffle is widening or truncating the input vectors, and + // the argument type must also be encoded. + if (C->getType() == C->getOperand(0)->getType()) { + Code = bitc::CST_CODE_CE_SHUFFLEVEC; + } else { + Code = bitc::CST_CODE_CE_SHUFVEC_EX; + Record.push_back(VE.getTypeID(C->getOperand(0)->getType())); + } + Record.push_back(VE.getValueID(C->getOperand(0))); + Record.push_back(VE.getValueID(C->getOperand(1))); + Record.push_back(VE.getValueID(C->getOperand(2))); + break; + case Instruction::ICmp: + case Instruction::FCmp: + case Instruction::VICmp: + case Instruction::VFCmp: + if (isa<VectorType>(C->getOperand(0)->getType()) + && (CE->getOpcode() == Instruction::ICmp + || CE->getOpcode() == Instruction::FCmp)) { + // compare returning vector of Int1Ty + assert(0 && "Unsupported constant!"); + } else { + Code = bitc::CST_CODE_CE_CMP; + } + Record.push_back(VE.getTypeID(C->getOperand(0)->getType())); + Record.push_back(VE.getValueID(C->getOperand(0))); + Record.push_back(VE.getValueID(C->getOperand(1))); + Record.push_back(CE->getPredicate()); + break; + } + } else if (const MDString *S = dyn_cast<MDString>(C)) { + Code = bitc::CST_CODE_MDSTRING; + AbbrevToUse = MDString6Abbrev; + for (unsigned i = 0, e = S->size(); i != e; ++i) { + char V = S->begin()[i]; + Record.push_back(V); + + if (!BitCodeAbbrevOp::isChar6(V)) + AbbrevToUse = MDString8Abbrev; + } + } else if (const MDNode *N = dyn_cast<MDNode>(C)) { + Code = bitc::CST_CODE_MDNODE; + for (unsigned i = 0, e = N->getNumElements(); i != e; ++i) { + if (N->getElement(i)) { + Record.push_back(VE.getTypeID(N->getElement(i)->getType())); + Record.push_back(VE.getValueID(N->getElement(i))); + } else { + Record.push_back(VE.getTypeID(Type::VoidTy)); + Record.push_back(0); + } + } + } else { + assert(0 && "Unknown constant!"); + } + Stream.EmitRecord(Code, Record, AbbrevToUse); + Record.clear(); + } + + Stream.ExitBlock(); +} + +static void WriteModuleConstants(const ValueEnumerator &VE, + BitstreamWriter &Stream) { + const ValueEnumerator::ValueList &Vals = VE.getValues(); + + // Find the first constant to emit, which is the first non-globalvalue value. + // We know globalvalues have been emitted by WriteModuleInfo. + for (unsigned i = 0, e = Vals.size(); i != e; ++i) { + if (!isa<GlobalValue>(Vals[i].first)) { + WriteConstants(i, Vals.size(), VE, Stream, true); + return; + } + } +} + +/// PushValueAndType - The file has to encode both the value and type id for +/// many values, because we need to know what type to create for forward +/// references. However, most operands are not forward references, so this type +/// field is not needed. +/// +/// This function adds V's value ID to Vals. If the value ID is higher than the +/// instruction ID, then it is a forward reference, and it also includes the +/// type ID. +static bool PushValueAndType(const Value *V, unsigned InstID, + SmallVector<unsigned, 64> &Vals, + ValueEnumerator &VE) { + unsigned ValID = VE.getValueID(V); + Vals.push_back(ValID); + if (ValID >= InstID) { + Vals.push_back(VE.getTypeID(V->getType())); + return true; + } + return false; +} + +/// WriteInstruction - Emit an instruction to the specified stream. +static void WriteInstruction(const Instruction &I, unsigned InstID, + ValueEnumerator &VE, BitstreamWriter &Stream, + SmallVector<unsigned, 64> &Vals) { + unsigned Code = 0; + unsigned AbbrevToUse = 0; + switch (I.getOpcode()) { + default: + if (Instruction::isCast(I.getOpcode())) { + Code = bitc::FUNC_CODE_INST_CAST; + if (!PushValueAndType(I.getOperand(0), InstID, Vals, VE)) + AbbrevToUse = FUNCTION_INST_CAST_ABBREV; + Vals.push_back(VE.getTypeID(I.getType())); + Vals.push_back(GetEncodedCastOpcode(I.getOpcode())); + } else { + assert(isa<BinaryOperator>(I) && "Unknown instruction!"); + Code = bitc::FUNC_CODE_INST_BINOP; + if (!PushValueAndType(I.getOperand(0), InstID, Vals, VE)) + AbbrevToUse = FUNCTION_INST_BINOP_ABBREV; + Vals.push_back(VE.getValueID(I.getOperand(1))); + Vals.push_back(GetEncodedBinaryOpcode(I.getOpcode())); + } + break; + + case Instruction::GetElementPtr: + Code = bitc::FUNC_CODE_INST_GEP; + for (unsigned i = 0, e = I.getNumOperands(); i != e; ++i) + PushValueAndType(I.getOperand(i), InstID, Vals, VE); + break; + case Instruction::ExtractValue: { + Code = bitc::FUNC_CODE_INST_EXTRACTVAL; + PushValueAndType(I.getOperand(0), InstID, Vals, VE); + const ExtractValueInst *EVI = cast<ExtractValueInst>(&I); + for (const unsigned *i = EVI->idx_begin(), *e = EVI->idx_end(); i != e; ++i) + Vals.push_back(*i); + break; + } + case Instruction::InsertValue: { + Code = bitc::FUNC_CODE_INST_INSERTVAL; + PushValueAndType(I.getOperand(0), InstID, Vals, VE); + PushValueAndType(I.getOperand(1), InstID, Vals, VE); + const InsertValueInst *IVI = cast<InsertValueInst>(&I); + for (const unsigned *i = IVI->idx_begin(), *e = IVI->idx_end(); i != e; ++i) + Vals.push_back(*i); + break; + } + case Instruction::Select: + Code = bitc::FUNC_CODE_INST_VSELECT; + PushValueAndType(I.getOperand(1), InstID, Vals, VE); + Vals.push_back(VE.getValueID(I.getOperand(2))); + PushValueAndType(I.getOperand(0), InstID, Vals, VE); + break; + case Instruction::ExtractElement: + Code = bitc::FUNC_CODE_INST_EXTRACTELT; + PushValueAndType(I.getOperand(0), InstID, Vals, VE); + Vals.push_back(VE.getValueID(I.getOperand(1))); + break; + case Instruction::InsertElement: + Code = bitc::FUNC_CODE_INST_INSERTELT; + PushValueAndType(I.getOperand(0), InstID, Vals, VE); + Vals.push_back(VE.getValueID(I.getOperand(1))); + Vals.push_back(VE.getValueID(I.getOperand(2))); + break; + case Instruction::ShuffleVector: + Code = bitc::FUNC_CODE_INST_SHUFFLEVEC; + PushValueAndType(I.getOperand(0), InstID, Vals, VE); + Vals.push_back(VE.getValueID(I.getOperand(1))); + Vals.push_back(VE.getValueID(I.getOperand(2))); + break; + case Instruction::ICmp: + case Instruction::FCmp: + case Instruction::VICmp: + case Instruction::VFCmp: + if (I.getOpcode() == Instruction::ICmp + || I.getOpcode() == Instruction::FCmp) { + // compare returning Int1Ty or vector of Int1Ty + Code = bitc::FUNC_CODE_INST_CMP2; + } else { + Code = bitc::FUNC_CODE_INST_CMP; + } + PushValueAndType(I.getOperand(0), InstID, Vals, VE); + Vals.push_back(VE.getValueID(I.getOperand(1))); + Vals.push_back(cast<CmpInst>(I).getPredicate()); + break; + + case Instruction::Ret: + { + Code = bitc::FUNC_CODE_INST_RET; + unsigned NumOperands = I.getNumOperands(); + if (NumOperands == 0) + AbbrevToUse = FUNCTION_INST_RET_VOID_ABBREV; + else if (NumOperands == 1) { + if (!PushValueAndType(I.getOperand(0), InstID, Vals, VE)) + AbbrevToUse = FUNCTION_INST_RET_VAL_ABBREV; + } else { + for (unsigned i = 0, e = NumOperands; i != e; ++i) + PushValueAndType(I.getOperand(i), InstID, Vals, VE); + } + } + break; + case Instruction::Br: + { + Code = bitc::FUNC_CODE_INST_BR; + BranchInst &II(cast<BranchInst>(I)); + Vals.push_back(VE.getValueID(II.getSuccessor(0))); + if (II.isConditional()) { + Vals.push_back(VE.getValueID(II.getSuccessor(1))); + Vals.push_back(VE.getValueID(II.getCondition())); + } + } + break; + case Instruction::Switch: + Code = bitc::FUNC_CODE_INST_SWITCH; + Vals.push_back(VE.getTypeID(I.getOperand(0)->getType())); + for (unsigned i = 0, e = I.getNumOperands(); i != e; ++i) + Vals.push_back(VE.getValueID(I.getOperand(i))); + break; + case Instruction::Invoke: { + const InvokeInst *II = cast<InvokeInst>(&I); + const Value *Callee(II->getCalledValue()); + const PointerType *PTy = cast<PointerType>(Callee->getType()); + const FunctionType *FTy = cast<FunctionType>(PTy->getElementType()); + Code = bitc::FUNC_CODE_INST_INVOKE; + + Vals.push_back(VE.getAttributeID(II->getAttributes())); + Vals.push_back(II->getCallingConv()); + Vals.push_back(VE.getValueID(II->getNormalDest())); + Vals.push_back(VE.getValueID(II->getUnwindDest())); + PushValueAndType(Callee, InstID, Vals, VE); + + // Emit value #'s for the fixed parameters. + for (unsigned i = 0, e = FTy->getNumParams(); i != e; ++i) + Vals.push_back(VE.getValueID(I.getOperand(i+3))); // fixed param. + + // Emit type/value pairs for varargs params. + if (FTy->isVarArg()) { + for (unsigned i = 3+FTy->getNumParams(), e = I.getNumOperands(); + i != e; ++i) + PushValueAndType(I.getOperand(i), InstID, Vals, VE); // vararg + } + break; + } + case Instruction::Unwind: + Code = bitc::FUNC_CODE_INST_UNWIND; + break; + case Instruction::Unreachable: + Code = bitc::FUNC_CODE_INST_UNREACHABLE; + AbbrevToUse = FUNCTION_INST_UNREACHABLE_ABBREV; + break; + + case Instruction::PHI: + Code = bitc::FUNC_CODE_INST_PHI; + Vals.push_back(VE.getTypeID(I.getType())); + for (unsigned i = 0, e = I.getNumOperands(); i != e; ++i) + Vals.push_back(VE.getValueID(I.getOperand(i))); + break; + + case Instruction::Malloc: + Code = bitc::FUNC_CODE_INST_MALLOC; + Vals.push_back(VE.getTypeID(I.getType())); + Vals.push_back(VE.getValueID(I.getOperand(0))); // size. + Vals.push_back(Log2_32(cast<MallocInst>(I).getAlignment())+1); + break; + + case Instruction::Free: + Code = bitc::FUNC_CODE_INST_FREE; + PushValueAndType(I.getOperand(0), InstID, Vals, VE); + break; + + case Instruction::Alloca: + Code = bitc::FUNC_CODE_INST_ALLOCA; + Vals.push_back(VE.getTypeID(I.getType())); + Vals.push_back(VE.getValueID(I.getOperand(0))); // size. + Vals.push_back(Log2_32(cast<AllocaInst>(I).getAlignment())+1); + break; + + case Instruction::Load: + Code = bitc::FUNC_CODE_INST_LOAD; + if (!PushValueAndType(I.getOperand(0), InstID, Vals, VE)) // ptr + AbbrevToUse = FUNCTION_INST_LOAD_ABBREV; + + Vals.push_back(Log2_32(cast<LoadInst>(I).getAlignment())+1); + Vals.push_back(cast<LoadInst>(I).isVolatile()); + break; + case Instruction::Store: + Code = bitc::FUNC_CODE_INST_STORE2; + PushValueAndType(I.getOperand(1), InstID, Vals, VE); // ptrty + ptr + Vals.push_back(VE.getValueID(I.getOperand(0))); // val. + Vals.push_back(Log2_32(cast<StoreInst>(I).getAlignment())+1); + Vals.push_back(cast<StoreInst>(I).isVolatile()); + break; + case Instruction::Call: { + const PointerType *PTy = cast<PointerType>(I.getOperand(0)->getType()); + const FunctionType *FTy = cast<FunctionType>(PTy->getElementType()); + + Code = bitc::FUNC_CODE_INST_CALL; + + const CallInst *CI = cast<CallInst>(&I); + Vals.push_back(VE.getAttributeID(CI->getAttributes())); + Vals.push_back((CI->getCallingConv() << 1) | unsigned(CI->isTailCall())); + PushValueAndType(CI->getOperand(0), InstID, Vals, VE); // Callee + + // Emit value #'s for the fixed parameters. + for (unsigned i = 0, e = FTy->getNumParams(); i != e; ++i) + Vals.push_back(VE.getValueID(I.getOperand(i+1))); // fixed param. + + // Emit type/value pairs for varargs params. + if (FTy->isVarArg()) { + unsigned NumVarargs = I.getNumOperands()-1-FTy->getNumParams(); + for (unsigned i = I.getNumOperands()-NumVarargs, e = I.getNumOperands(); + i != e; ++i) + PushValueAndType(I.getOperand(i), InstID, Vals, VE); // varargs + } + break; + } + case Instruction::VAArg: + Code = bitc::FUNC_CODE_INST_VAARG; + Vals.push_back(VE.getTypeID(I.getOperand(0)->getType())); // valistty + Vals.push_back(VE.getValueID(I.getOperand(0))); // valist. + Vals.push_back(VE.getTypeID(I.getType())); // restype. + break; + } + + Stream.EmitRecord(Code, Vals, AbbrevToUse); + Vals.clear(); +} + +// Emit names for globals/functions etc. +static void WriteValueSymbolTable(const ValueSymbolTable &VST, + const ValueEnumerator &VE, + BitstreamWriter &Stream) { + if (VST.empty()) return; + Stream.EnterSubblock(bitc::VALUE_SYMTAB_BLOCK_ID, 4); + + // FIXME: Set up the abbrev, we know how many values there are! + // FIXME: We know if the type names can use 7-bit ascii. + SmallVector<unsigned, 64> NameVals; + + for (ValueSymbolTable::const_iterator SI = VST.begin(), SE = VST.end(); + SI != SE; ++SI) { + + const ValueName &Name = *SI; + + // Figure out the encoding to use for the name. + bool is7Bit = true; + bool isChar6 = true; + for (const char *C = Name.getKeyData(), *E = C+Name.getKeyLength(); + C != E; ++C) { + if (isChar6) + isChar6 = BitCodeAbbrevOp::isChar6(*C); + if ((unsigned char)*C & 128) { + is7Bit = false; + break; // don't bother scanning the rest. + } + } + + unsigned AbbrevToUse = VST_ENTRY_8_ABBREV; + + // VST_ENTRY: [valueid, namechar x N] + // VST_BBENTRY: [bbid, namechar x N] + unsigned Code; + if (isa<BasicBlock>(SI->getValue())) { + Code = bitc::VST_CODE_BBENTRY; + if (isChar6) + AbbrevToUse = VST_BBENTRY_6_ABBREV; + } else { + Code = bitc::VST_CODE_ENTRY; + if (isChar6) + AbbrevToUse = VST_ENTRY_6_ABBREV; + else if (is7Bit) + AbbrevToUse = VST_ENTRY_7_ABBREV; + } + + NameVals.push_back(VE.getValueID(SI->getValue())); + for (const char *P = Name.getKeyData(), + *E = Name.getKeyData()+Name.getKeyLength(); P != E; ++P) + NameVals.push_back((unsigned char)*P); + + // Emit the finished record. + Stream.EmitRecord(Code, NameVals, AbbrevToUse); + NameVals.clear(); + } + Stream.ExitBlock(); +} + +/// WriteFunction - Emit a function body to the module stream. +static void WriteFunction(const Function &F, ValueEnumerator &VE, + BitstreamWriter &Stream) { + Stream.EnterSubblock(bitc::FUNCTION_BLOCK_ID, 4); + VE.incorporateFunction(F); + + SmallVector<unsigned, 64> Vals; + + // Emit the number of basic blocks, so the reader can create them ahead of + // time. + Vals.push_back(VE.getBasicBlocks().size()); + Stream.EmitRecord(bitc::FUNC_CODE_DECLAREBLOCKS, Vals); + Vals.clear(); + + // If there are function-local constants, emit them now. + unsigned CstStart, CstEnd; + VE.getFunctionConstantRange(CstStart, CstEnd); + WriteConstants(CstStart, CstEnd, VE, Stream, false); + + // Keep a running idea of what the instruction ID is. + unsigned InstID = CstEnd; + + // Finally, emit all the instructions, in order. + for (Function::const_iterator BB = F.begin(), E = F.end(); BB != E; ++BB) + for (BasicBlock::const_iterator I = BB->begin(), E = BB->end(); + I != E; ++I) { + WriteInstruction(*I, InstID, VE, Stream, Vals); + if (I->getType() != Type::VoidTy) + ++InstID; + } + + // Emit names for all the instructions etc. + WriteValueSymbolTable(F.getValueSymbolTable(), VE, Stream); + + VE.purgeFunction(); + Stream.ExitBlock(); +} + +/// WriteTypeSymbolTable - Emit a block for the specified type symtab. +static void WriteTypeSymbolTable(const TypeSymbolTable &TST, + const ValueEnumerator &VE, + BitstreamWriter &Stream) { + if (TST.empty()) return; + + Stream.EnterSubblock(bitc::TYPE_SYMTAB_BLOCK_ID, 3); + + // 7-bit fixed width VST_CODE_ENTRY strings. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::VST_CODE_ENTRY)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, + Log2_32_Ceil(VE.getTypes().size()+1))); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Array)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 7)); + unsigned V7Abbrev = Stream.EmitAbbrev(Abbv); + + SmallVector<unsigned, 64> NameVals; + + for (TypeSymbolTable::const_iterator TI = TST.begin(), TE = TST.end(); + TI != TE; ++TI) { + // TST_ENTRY: [typeid, namechar x N] + NameVals.push_back(VE.getTypeID(TI->second)); + + const std::string &Str = TI->first; + bool is7Bit = true; + for (unsigned i = 0, e = Str.size(); i != e; ++i) { + NameVals.push_back((unsigned char)Str[i]); + if (Str[i] & 128) + is7Bit = false; + } + + // Emit the finished record. + Stream.EmitRecord(bitc::VST_CODE_ENTRY, NameVals, is7Bit ? V7Abbrev : 0); + NameVals.clear(); + } + + Stream.ExitBlock(); +} + +// Emit blockinfo, which defines the standard abbreviations etc. +static void WriteBlockInfo(const ValueEnumerator &VE, BitstreamWriter &Stream) { + // We only want to emit block info records for blocks that have multiple + // instances: CONSTANTS_BLOCK, FUNCTION_BLOCK and VALUE_SYMTAB_BLOCK. Other + // blocks can defined their abbrevs inline. + Stream.EnterBlockInfoBlock(2); + + { // 8-bit fixed-width VST_ENTRY/VST_BBENTRY strings. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 3)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 8)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Array)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 8)); + if (Stream.EmitBlockInfoAbbrev(bitc::VALUE_SYMTAB_BLOCK_ID, + Abbv) != VST_ENTRY_8_ABBREV) + assert(0 && "Unexpected abbrev ordering!"); + } + + { // 7-bit fixed width VST_ENTRY strings. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::VST_CODE_ENTRY)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 8)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Array)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 7)); + if (Stream.EmitBlockInfoAbbrev(bitc::VALUE_SYMTAB_BLOCK_ID, + Abbv) != VST_ENTRY_7_ABBREV) + assert(0 && "Unexpected abbrev ordering!"); + } + { // 6-bit char6 VST_ENTRY strings. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::VST_CODE_ENTRY)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 8)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Array)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Char6)); + if (Stream.EmitBlockInfoAbbrev(bitc::VALUE_SYMTAB_BLOCK_ID, + Abbv) != VST_ENTRY_6_ABBREV) + assert(0 && "Unexpected abbrev ordering!"); + } + { // 6-bit char6 VST_BBENTRY strings. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::VST_CODE_BBENTRY)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 8)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Array)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Char6)); + if (Stream.EmitBlockInfoAbbrev(bitc::VALUE_SYMTAB_BLOCK_ID, + Abbv) != VST_BBENTRY_6_ABBREV) + assert(0 && "Unexpected abbrev ordering!"); + } + + + + { // SETTYPE abbrev for CONSTANTS_BLOCK. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::CST_CODE_SETTYPE)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, + Log2_32_Ceil(VE.getTypes().size()+1))); + if (Stream.EmitBlockInfoAbbrev(bitc::CONSTANTS_BLOCK_ID, + Abbv) != CONSTANTS_SETTYPE_ABBREV) + assert(0 && "Unexpected abbrev ordering!"); + } + + { // INTEGER abbrev for CONSTANTS_BLOCK. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::CST_CODE_INTEGER)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 8)); + if (Stream.EmitBlockInfoAbbrev(bitc::CONSTANTS_BLOCK_ID, + Abbv) != CONSTANTS_INTEGER_ABBREV) + assert(0 && "Unexpected abbrev ordering!"); + } + + { // CE_CAST abbrev for CONSTANTS_BLOCK. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::CST_CODE_CE_CAST)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 4)); // cast opc + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, // typeid + Log2_32_Ceil(VE.getTypes().size()+1))); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 8)); // value id + + if (Stream.EmitBlockInfoAbbrev(bitc::CONSTANTS_BLOCK_ID, + Abbv) != CONSTANTS_CE_CAST_Abbrev) + assert(0 && "Unexpected abbrev ordering!"); + } + { // NULL abbrev for CONSTANTS_BLOCK. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::CST_CODE_NULL)); + if (Stream.EmitBlockInfoAbbrev(bitc::CONSTANTS_BLOCK_ID, + Abbv) != CONSTANTS_NULL_Abbrev) + assert(0 && "Unexpected abbrev ordering!"); + } + + // FIXME: This should only use space for first class types! + + { // INST_LOAD abbrev for FUNCTION_BLOCK. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::FUNC_CODE_INST_LOAD)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 6)); // Ptr + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 4)); // Align + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 1)); // volatile + if (Stream.EmitBlockInfoAbbrev(bitc::FUNCTION_BLOCK_ID, + Abbv) != FUNCTION_INST_LOAD_ABBREV) + assert(0 && "Unexpected abbrev ordering!"); + } + { // INST_BINOP abbrev for FUNCTION_BLOCK. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::FUNC_CODE_INST_BINOP)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 6)); // LHS + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 6)); // RHS + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 4)); // opc + if (Stream.EmitBlockInfoAbbrev(bitc::FUNCTION_BLOCK_ID, + Abbv) != FUNCTION_INST_BINOP_ABBREV) + assert(0 && "Unexpected abbrev ordering!"); + } + { // INST_CAST abbrev for FUNCTION_BLOCK. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::FUNC_CODE_INST_CAST)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 6)); // OpVal + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, // dest ty + Log2_32_Ceil(VE.getTypes().size()+1))); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 4)); // opc + if (Stream.EmitBlockInfoAbbrev(bitc::FUNCTION_BLOCK_ID, + Abbv) != FUNCTION_INST_CAST_ABBREV) + assert(0 && "Unexpected abbrev ordering!"); + } + + { // INST_RET abbrev for FUNCTION_BLOCK. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::FUNC_CODE_INST_RET)); + if (Stream.EmitBlockInfoAbbrev(bitc::FUNCTION_BLOCK_ID, + Abbv) != FUNCTION_INST_RET_VOID_ABBREV) + assert(0 && "Unexpected abbrev ordering!"); + } + { // INST_RET abbrev for FUNCTION_BLOCK. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::FUNC_CODE_INST_RET)); + Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 6)); // ValID + if (Stream.EmitBlockInfoAbbrev(bitc::FUNCTION_BLOCK_ID, + Abbv) != FUNCTION_INST_RET_VAL_ABBREV) + assert(0 && "Unexpected abbrev ordering!"); + } + { // INST_UNREACHABLE abbrev for FUNCTION_BLOCK. + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + Abbv->Add(BitCodeAbbrevOp(bitc::FUNC_CODE_INST_UNREACHABLE)); + if (Stream.EmitBlockInfoAbbrev(bitc::FUNCTION_BLOCK_ID, + Abbv) != FUNCTION_INST_UNREACHABLE_ABBREV) + assert(0 && "Unexpected abbrev ordering!"); + } + + Stream.ExitBlock(); +} + + +/// WriteModule - Emit the specified module to the bitstream. +static void WriteModule(const Module *M, BitstreamWriter &Stream) { + Stream.EnterSubblock(bitc::MODULE_BLOCK_ID, 3); + + // Emit the version number if it is non-zero. + if (CurVersion) { + SmallVector<unsigned, 1> Vals; + Vals.push_back(CurVersion); + Stream.EmitRecord(bitc::MODULE_CODE_VERSION, Vals); + } + + // Analyze the module, enumerating globals, functions, etc. + ValueEnumerator VE(M); + + // Emit blockinfo, which defines the standard abbreviations etc. + WriteBlockInfo(VE, Stream); + + // Emit information about parameter attributes. + WriteAttributeTable(VE, Stream); + + // Emit information describing all of the types in the module. + WriteTypeTable(VE, Stream); + + // Emit top-level description of module, including target triple, inline asm, + // descriptors for global variables, and function prototype info. + WriteModuleInfo(M, VE, Stream); + + // Emit constants. + WriteModuleConstants(VE, Stream); + + // If we have any aggregate values in the value table, purge them - these can + // only be used to initialize global variables. Doing so makes the value + // namespace smaller for code in functions. + int NumNonAggregates = VE.PurgeAggregateValues(); + if (NumNonAggregates != -1) { + SmallVector<unsigned, 1> Vals; + Vals.push_back(NumNonAggregates); + Stream.EmitRecord(bitc::MODULE_CODE_PURGEVALS, Vals); + } + + // Emit function bodies. + for (Module::const_iterator I = M->begin(), E = M->end(); I != E; ++I) + if (!I->isDeclaration()) + WriteFunction(*I, VE, Stream); + + // Emit the type symbol table information. + WriteTypeSymbolTable(M->getTypeSymbolTable(), VE, Stream); + + // Emit names for globals/functions etc. + WriteValueSymbolTable(M->getValueSymbolTable(), VE, Stream); + + Stream.ExitBlock(); +} + +/// EmitDarwinBCHeader - If generating a bc file on darwin, we have to emit a +/// header and trailer to make it compatible with the system archiver. To do +/// this we emit the following header, and then emit a trailer that pads the +/// file out to be a multiple of 16 bytes. +/// +/// struct bc_header { +/// uint32_t Magic; // 0x0B17C0DE +/// uint32_t Version; // Version, currently always 0. +/// uint32_t BitcodeOffset; // Offset to traditional bitcode file. +/// uint32_t BitcodeSize; // Size of traditional bitcode file. +/// uint32_t CPUType; // CPU specifier. +/// ... potentially more later ... +/// }; +enum { + DarwinBCSizeFieldOffset = 3*4, // Offset to bitcode_size. + DarwinBCHeaderSize = 5*4 +}; + +static void EmitDarwinBCHeader(BitstreamWriter &Stream, + const std::string &TT) { + unsigned CPUType = ~0U; + + // Match x86_64-*, i[3-9]86-*, powerpc-*, powerpc64-*. The CPUType is a + // magic number from /usr/include/mach/machine.h. It is ok to reproduce the + // specific constants here because they are implicitly part of the Darwin ABI. + enum { + DARWIN_CPU_ARCH_ABI64 = 0x01000000, + DARWIN_CPU_TYPE_X86 = 7, + DARWIN_CPU_TYPE_POWERPC = 18 + }; + + if (TT.find("x86_64-") == 0) + CPUType = DARWIN_CPU_TYPE_X86 | DARWIN_CPU_ARCH_ABI64; + else if (TT.size() >= 5 && TT[0] == 'i' && TT[2] == '8' && TT[3] == '6' && + TT[4] == '-' && TT[1] - '3' < 6) + CPUType = DARWIN_CPU_TYPE_X86; + else if (TT.find("powerpc-") == 0) + CPUType = DARWIN_CPU_TYPE_POWERPC; + else if (TT.find("powerpc64-") == 0) + CPUType = DARWIN_CPU_TYPE_POWERPC | DARWIN_CPU_ARCH_ABI64; + + // Traditional Bitcode starts after header. + unsigned BCOffset = DarwinBCHeaderSize; + + Stream.Emit(0x0B17C0DE, 32); + Stream.Emit(0 , 32); // Version. + Stream.Emit(BCOffset , 32); + Stream.Emit(0 , 32); // Filled in later. + Stream.Emit(CPUType , 32); +} + +/// EmitDarwinBCTrailer - Emit the darwin epilog after the bitcode file and +/// finalize the header. +static void EmitDarwinBCTrailer(BitstreamWriter &Stream, unsigned BufferSize) { + // Update the size field in the header. + Stream.BackpatchWord(DarwinBCSizeFieldOffset, BufferSize-DarwinBCHeaderSize); + + // If the file is not a multiple of 16 bytes, insert dummy padding. + while (BufferSize & 15) { + Stream.Emit(0, 8); + ++BufferSize; + } +} + + +/// WriteBitcodeToFile - Write the specified module to the specified output +/// stream. +void llvm::WriteBitcodeToFile(const Module *M, std::ostream &Out) { + raw_os_ostream RawOut(Out); + // If writing to stdout, set binary mode. + if (llvm::cout == Out) + sys::Program::ChangeStdoutToBinary(); + WriteBitcodeToFile(M, RawOut); +} + +/// WriteBitcodeToFile - Write the specified module to the specified output +/// stream. +void llvm::WriteBitcodeToFile(const Module *M, raw_ostream &Out) { + std::vector<unsigned char> Buffer; + BitstreamWriter Stream(Buffer); + + Buffer.reserve(256*1024); + + WriteBitcodeToStream( M, Stream ); + + // If writing to stdout, set binary mode. + if (&llvm::outs() == &Out) + sys::Program::ChangeStdoutToBinary(); + + // Write the generated bitstream to "Out". + Out.write((char*)&Buffer.front(), Buffer.size()); + + // Make sure it hits disk now. + Out.flush(); +} + +/// WriteBitcodeToStream - Write the specified module to the specified output +/// stream. +void llvm::WriteBitcodeToStream(const Module *M, BitstreamWriter &Stream) { + // If this is darwin, emit a file header and trailer if needed. + bool isDarwin = M->getTargetTriple().find("-darwin") != std::string::npos; + if (isDarwin) + EmitDarwinBCHeader(Stream, M->getTargetTriple()); + + // Emit the file header. + Stream.Emit((unsigned)'B', 8); + Stream.Emit((unsigned)'C', 8); + Stream.Emit(0x0, 4); + Stream.Emit(0xC, 4); + Stream.Emit(0xE, 4); + Stream.Emit(0xD, 4); + + // Emit the module. + WriteModule(M, Stream); + + if (isDarwin) + EmitDarwinBCTrailer(Stream, Stream.getBuffer().size()); +} diff --git a/lib/Bitcode/Writer/BitcodeWriterPass.cpp b/lib/Bitcode/Writer/BitcodeWriterPass.cpp new file mode 100644 index 0000000..209cf09 --- /dev/null +++ b/lib/Bitcode/Writer/BitcodeWriterPass.cpp @@ -0,0 +1,56 @@ +//===--- Bitcode/Writer/BitcodeWriterPass.cpp - Bitcode Writer ------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// BitcodeWriterPass implementation. +// +//===----------------------------------------------------------------------===// + +#include "llvm/Bitcode/ReaderWriter.h" +#include "llvm/Pass.h" +using namespace llvm; + +namespace { + class WriteBitcodePass : public ModulePass { + // FIXME: Kill off std::ostream + std::ostream *Out; + raw_ostream *RawOut; // raw_ostream to print on + public: + static char ID; // Pass identification, replacement for typeid + explicit WriteBitcodePass(std::ostream &o) + : ModulePass(&ID), Out(&o), RawOut(0) {} + explicit WriteBitcodePass(raw_ostream &o) + : ModulePass(&ID), Out(0), RawOut(&o) {} + + const char *getPassName() const { return "Bitcode Writer"; } + + bool runOnModule(Module &M) { + if (Out) { + WriteBitcodeToFile(&M, *Out); + } else { + WriteBitcodeToFile(&M, *RawOut); + } + return false; + } + }; +} + +char WriteBitcodePass::ID = 0; + +/// CreateBitcodeWriterPass - Create and return a pass that writes the module +/// to the specified ostream. +ModulePass *llvm::CreateBitcodeWriterPass(std::ostream &Str) { + return new WriteBitcodePass(Str); +} + + +/// createBitcodeWriterPass - Create and return a pass that writes the module +/// to the specified ostream. +ModulePass *llvm::createBitcodeWriterPass(raw_ostream &Str) { + return new WriteBitcodePass(Str); +} diff --git a/lib/Bitcode/Writer/CMakeLists.txt b/lib/Bitcode/Writer/CMakeLists.txt new file mode 100644 index 0000000..ac5bb99 --- /dev/null +++ b/lib/Bitcode/Writer/CMakeLists.txt @@ -0,0 +1,9 @@ +add_llvm_library(LLVMBitWriter + BitWriter.cpp + BitcodeWriter.cpp + BitcodeWriterPass.cpp + Serialize.cpp + SerializeAPFloat.cpp + SerializeAPInt.cpp + ValueEnumerator.cpp + ) diff --git a/lib/Bitcode/Writer/Makefile b/lib/Bitcode/Writer/Makefile new file mode 100644 index 0000000..7b0bd72 --- /dev/null +++ b/lib/Bitcode/Writer/Makefile @@ -0,0 +1,15 @@ +##===- lib/Bitcode/Reader/Makefile -------------------------*- Makefile -*-===## +# +# The LLVM Compiler Infrastructure +# +# This file is distributed under the University of Illinois Open Source +# License. See LICENSE.TXT for details. +# +##===----------------------------------------------------------------------===## + +LEVEL = ../../.. +LIBRARYNAME = LLVMBitWriter +BUILD_ARCHIVE = 1 + +include $(LEVEL)/Makefile.common + diff --git a/lib/Bitcode/Writer/Serialize.cpp b/lib/Bitcode/Writer/Serialize.cpp new file mode 100644 index 0000000..79464a6 --- /dev/null +++ b/lib/Bitcode/Writer/Serialize.cpp @@ -0,0 +1,118 @@ +//==- Serialize.cpp - Generic Object Serialization to Bitcode ----*- C++ -*-==// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file defines the internal methods used for object serialization. +// +//===----------------------------------------------------------------------===// + +#include "llvm/Bitcode/Serialize.h" +#include "string.h" + +#ifdef DEBUG_BACKPATCH +#include "llvm/Support/Streams.h" +#endif + +using namespace llvm; + +Serializer::Serializer(BitstreamWriter& stream) + : Stream(stream), BlockLevel(0) {} + +Serializer::~Serializer() { + if (inRecord()) + EmitRecord(); + + while (BlockLevel > 0) + Stream.ExitBlock(); + + Stream.FlushToWord(); +} + +void Serializer::EmitRecord() { + assert(Record.size() > 0 && "Cannot emit empty record."); + Stream.EmitRecord(8,Record); + Record.clear(); +} + +void Serializer::EnterBlock(unsigned BlockID,unsigned CodeLen) { + FlushRecord(); + Stream.EnterSubblock(BlockID,CodeLen); + ++BlockLevel; +} + +void Serializer::ExitBlock() { + assert (BlockLevel > 0); + --BlockLevel; + FlushRecord(); + Stream.ExitBlock(); +} + +void Serializer::EmitInt(uint64_t X) { + assert (BlockLevel > 0); + Record.push_back(X); +} + +void Serializer::EmitSInt(int64_t X) { + if (X >= 0) + EmitInt(X << 1); + else + EmitInt((-X << 1) | 1); +} + +void Serializer::EmitCStr(const char* s, const char* end) { + Record.push_back(end - s); + + while(s != end) { + Record.push_back(*s); + ++s; + } +} + +void Serializer::EmitCStr(const char* s) { + EmitCStr(s,s+strlen(s)); +} + +SerializedPtrID Serializer::getPtrId(const void* ptr) { + if (!ptr) + return 0; + + MapTy::iterator I = PtrMap.find(ptr); + + if (I == PtrMap.end()) { + unsigned id = PtrMap.size()+1; +#ifdef DEBUG_BACKPATCH + llvm::cerr << "Registered PTR: " << ptr << " => " << id << "\n"; +#endif + PtrMap[ptr] = id; + return id; + } + else return I->second; +} + +bool Serializer::isRegistered(const void* ptr) const { + MapTy::const_iterator I = PtrMap.find(ptr); + return I != PtrMap.end(); +} + + +#define INT_EMIT(TYPE)\ +void SerializeTrait<TYPE>::Emit(Serializer&S, TYPE X) { S.EmitInt(X); } + +INT_EMIT(bool) +INT_EMIT(unsigned char) +INT_EMIT(unsigned short) +INT_EMIT(unsigned int) +INT_EMIT(unsigned long) + +#define SINT_EMIT(TYPE)\ +void SerializeTrait<TYPE>::Emit(Serializer&S, TYPE X) { S.EmitSInt(X); } + +SINT_EMIT(signed char) +SINT_EMIT(signed short) +SINT_EMIT(signed int) +SINT_EMIT(signed long) diff --git a/lib/Bitcode/Writer/SerializeAPFloat.cpp b/lib/Bitcode/Writer/SerializeAPFloat.cpp new file mode 100644 index 0000000..25d954f --- /dev/null +++ b/lib/Bitcode/Writer/SerializeAPFloat.cpp @@ -0,0 +1,21 @@ +//===-- SerializeAPInt.cpp - Serialization for APFloat ---------*- C++ -*--===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file implements serialization of APFloat. +// +//===----------------------------------------------------------------------===// + +#include "llvm/ADT/APFloat.h" +#include "llvm/Bitcode/Serialize.h" + +using namespace llvm; + +void APFloat::Emit(Serializer& S) const { + S.Emit(bitcastToAPInt()); +} diff --git a/lib/Bitcode/Writer/SerializeAPInt.cpp b/lib/Bitcode/Writer/SerializeAPInt.cpp new file mode 100644 index 0000000..47792c7 --- /dev/null +++ b/lib/Bitcode/Writer/SerializeAPInt.cpp @@ -0,0 +1,31 @@ +//===-- SerializeAPInt.cpp - Serialization for APInts ----------*- C++ -*--===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file implements serialization of APInts. +// +//===----------------------------------------------------------------------===// + +#include "llvm/ADT/APInt.h" +#include "llvm/Bitcode/Serialize.h" +#include <cassert> + +using namespace llvm; + +void APInt::Emit(Serializer& S) const { + S.EmitInt(BitWidth); + + if (isSingleWord()) + S.EmitInt(VAL); + else { + uint32_t NumWords = getNumWords(); + S.EmitInt(NumWords); + for (unsigned i = 0; i < NumWords; ++i) + S.EmitInt(pVal[i]); + } +} diff --git a/lib/Bitcode/Writer/ValueEnumerator.cpp b/lib/Bitcode/Writer/ValueEnumerator.cpp new file mode 100644 index 0000000..8002a36 --- /dev/null +++ b/lib/Bitcode/Writer/ValueEnumerator.cpp @@ -0,0 +1,347 @@ +//===-- ValueEnumerator.cpp - Number values and types for bitcode 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 the ValueEnumerator class. +// +//===----------------------------------------------------------------------===// + +#include "ValueEnumerator.h" +#include "llvm/Constants.h" +#include "llvm/DerivedTypes.h" +#include "llvm/MDNode.h" +#include "llvm/Module.h" +#include "llvm/TypeSymbolTable.h" +#include "llvm/ValueSymbolTable.h" +#include "llvm/Instructions.h" +#include <algorithm> +using namespace llvm; + +static bool isSingleValueType(const std::pair<const llvm::Type*, + unsigned int> &P) { + return P.first->isSingleValueType(); +} + +static bool isIntegerValue(const std::pair<const Value*, unsigned> &V) { + return isa<IntegerType>(V.first->getType()); +} + +static bool CompareByFrequency(const std::pair<const llvm::Type*, + unsigned int> &P1, + const std::pair<const llvm::Type*, + unsigned int> &P2) { + return P1.second > P2.second; +} + +/// ValueEnumerator - Enumerate module-level information. +ValueEnumerator::ValueEnumerator(const Module *M) { + // Enumerate the global variables. + for (Module::const_global_iterator I = M->global_begin(), + E = M->global_end(); I != E; ++I) + EnumerateValue(I); + + // Enumerate the functions. + for (Module::const_iterator I = M->begin(), E = M->end(); I != E; ++I) { + EnumerateValue(I); + EnumerateAttributes(cast<Function>(I)->getAttributes()); + } + + // Enumerate the aliases. + for (Module::const_alias_iterator I = M->alias_begin(), E = M->alias_end(); + I != E; ++I) + EnumerateValue(I); + + // Remember what is the cutoff between globalvalue's and other constants. + unsigned FirstConstant = Values.size(); + + // Enumerate the global variable initializers. + for (Module::const_global_iterator I = M->global_begin(), + E = M->global_end(); I != E; ++I) + if (I->hasInitializer()) + EnumerateValue(I->getInitializer()); + + // Enumerate the aliasees. + for (Module::const_alias_iterator I = M->alias_begin(), E = M->alias_end(); + I != E; ++I) + EnumerateValue(I->getAliasee()); + + // Enumerate types used by the type symbol table. + EnumerateTypeSymbolTable(M->getTypeSymbolTable()); + + // Insert constants that are named at module level into the slot pool so that + // the module symbol table can refer to them... + EnumerateValueSymbolTable(M->getValueSymbolTable()); + + // Enumerate types used by function bodies and argument lists. + for (Module::const_iterator F = M->begin(), E = M->end(); F != E; ++F) { + + for (Function::const_arg_iterator I = F->arg_begin(), E = F->arg_end(); + I != E; ++I) + EnumerateType(I->getType()); + + for (Function::const_iterator BB = F->begin(), E = F->end(); BB != E; ++BB) + for (BasicBlock::const_iterator I = BB->begin(), E = BB->end(); I!=E;++I){ + for (User::const_op_iterator OI = I->op_begin(), E = I->op_end(); + OI != E; ++OI) + EnumerateOperandType(*OI); + EnumerateType(I->getType()); + if (const CallInst *CI = dyn_cast<CallInst>(I)) + EnumerateAttributes(CI->getAttributes()); + else if (const InvokeInst *II = dyn_cast<InvokeInst>(I)) + EnumerateAttributes(II->getAttributes()); + } + } + + // Optimize constant ordering. + OptimizeConstants(FirstConstant, Values.size()); + + // Sort the type table by frequency so that most commonly used types are early + // in the table (have low bit-width). + std::stable_sort(Types.begin(), Types.end(), CompareByFrequency); + + // Partition the Type ID's so that the single-value types occur before the + // aggregate types. This allows the aggregate types to be dropped from the + // type table after parsing the global variable initializers. + std::partition(Types.begin(), Types.end(), isSingleValueType); + + // Now that we rearranged the type table, rebuild TypeMap. + for (unsigned i = 0, e = Types.size(); i != e; ++i) + TypeMap[Types[i].first] = i+1; +} + +// Optimize constant ordering. +namespace { + struct CstSortPredicate { + ValueEnumerator &VE; + explicit CstSortPredicate(ValueEnumerator &ve) : VE(ve) {} + bool operator()(const std::pair<const Value*, unsigned> &LHS, + const std::pair<const Value*, unsigned> &RHS) { + // Sort by plane. + if (LHS.first->getType() != RHS.first->getType()) + return VE.getTypeID(LHS.first->getType()) < + VE.getTypeID(RHS.first->getType()); + // Then by frequency. + return LHS.second > RHS.second; + } + }; +} + +/// OptimizeConstants - Reorder constant pool for denser encoding. +void ValueEnumerator::OptimizeConstants(unsigned CstStart, unsigned CstEnd) { + if (CstStart == CstEnd || CstStart+1 == CstEnd) return; + + CstSortPredicate P(*this); + std::stable_sort(Values.begin()+CstStart, Values.begin()+CstEnd, P); + + // Ensure that integer constants are at the start of the constant pool. This + // is important so that GEP structure indices come before gep constant exprs. + std::partition(Values.begin()+CstStart, Values.begin()+CstEnd, + isIntegerValue); + + // Rebuild the modified portion of ValueMap. + for (; CstStart != CstEnd; ++CstStart) + ValueMap[Values[CstStart].first] = CstStart+1; +} + + +/// EnumerateTypeSymbolTable - Insert all of the types in the specified symbol +/// table. +void ValueEnumerator::EnumerateTypeSymbolTable(const TypeSymbolTable &TST) { + for (TypeSymbolTable::const_iterator TI = TST.begin(), TE = TST.end(); + TI != TE; ++TI) + EnumerateType(TI->second); +} + +/// EnumerateValueSymbolTable - Insert all of the values in the specified symbol +/// table into the values table. +void ValueEnumerator::EnumerateValueSymbolTable(const ValueSymbolTable &VST) { + for (ValueSymbolTable::const_iterator VI = VST.begin(), VE = VST.end(); + VI != VE; ++VI) + EnumerateValue(VI->getValue()); +} + +void ValueEnumerator::EnumerateValue(const Value *V) { + assert(V->getType() != Type::VoidTy && "Can't insert void values!"); + + // Check to see if it's already in! + unsigned &ValueID = ValueMap[V]; + if (ValueID) { + // Increment use count. + Values[ValueID-1].second++; + return; + } + + // Enumerate the type of this value. + EnumerateType(V->getType()); + + if (const Constant *C = dyn_cast<Constant>(V)) { + if (isa<GlobalValue>(C)) { + // Initializers for globals are handled explicitly elsewhere. + } else if (isa<ConstantArray>(C) && cast<ConstantArray>(C)->isString()) { + // Do not enumerate the initializers for an array of simple characters. + // The initializers just polute the value table, and we emit the strings + // specially. + } else if (C->getNumOperands()) { + // If a constant has operands, enumerate them. This makes sure that if a + // constant has uses (for example an array of const ints), that they are + // inserted also. + + // We prefer to enumerate them with values before we enumerate the user + // itself. This makes it more likely that we can avoid forward references + // in the reader. We know that there can be no cycles in the constants + // graph that don't go through a global variable. + for (User::const_op_iterator I = C->op_begin(), E = C->op_end(); + I != E; ++I) + EnumerateValue(*I); + + // Finally, add the value. Doing this could make the ValueID reference be + // dangling, don't reuse it. + Values.push_back(std::make_pair(V, 1U)); + ValueMap[V] = Values.size(); + return; + } else if (const MDNode *N = dyn_cast<MDNode>(C)) { + for (MDNode::const_elem_iterator I = N->elem_begin(), E = N->elem_end(); + I != E; ++I) { + if (*I) + EnumerateValue(*I); + else + EnumerateType(Type::VoidTy); + } + + Values.push_back(std::make_pair(V, 1U)); + ValueMap[V] = Values.size(); + return; + } + } + + // Add the value. + Values.push_back(std::make_pair(V, 1U)); + ValueID = Values.size(); +} + + +void ValueEnumerator::EnumerateType(const Type *Ty) { + unsigned &TypeID = TypeMap[Ty]; + + if (TypeID) { + // If we've already seen this type, just increase its occurrence count. + Types[TypeID-1].second++; + return; + } + + // First time we saw this type, add it. + Types.push_back(std::make_pair(Ty, 1U)); + TypeID = Types.size(); + + // Enumerate subtypes. + for (Type::subtype_iterator I = Ty->subtype_begin(), E = Ty->subtype_end(); + I != E; ++I) + EnumerateType(*I); +} + +// Enumerate the types for the specified value. If the value is a constant, +// walk through it, enumerating the types of the constant. +void ValueEnumerator::EnumerateOperandType(const Value *V) { + EnumerateType(V->getType()); + if (const Constant *C = dyn_cast<Constant>(V)) { + // If this constant is already enumerated, ignore it, we know its type must + // be enumerated. + if (ValueMap.count(V)) return; + + // This constant may have operands, make sure to enumerate the types in + // them. + for (unsigned i = 0, e = C->getNumOperands(); i != e; ++i) + EnumerateOperandType(C->getOperand(i)); + + if (const MDNode *N = dyn_cast<MDNode>(V)) { + for (unsigned i = 0, e = N->getNumElements(); i != e; ++i) + EnumerateOperandType(N->getElement(i)); + } + } +} + +void ValueEnumerator::EnumerateAttributes(const AttrListPtr &PAL) { + if (PAL.isEmpty()) return; // null is always 0. + // Do a lookup. + unsigned &Entry = AttributeMap[PAL.getRawPointer()]; + if (Entry == 0) { + // Never saw this before, add it. + Attributes.push_back(PAL); + Entry = Attributes.size(); + } +} + + +/// PurgeAggregateValues - If there are any aggregate values at the end of the +/// value list, remove them and return the count of the remaining values. If +/// there are none, return -1. +int ValueEnumerator::PurgeAggregateValues() { + // If there are no aggregate values at the end of the list, return -1. + if (Values.empty() || Values.back().first->getType()->isSingleValueType()) + return -1; + + // Otherwise, remove aggregate values... + while (!Values.empty() && !Values.back().first->getType()->isSingleValueType()) + Values.pop_back(); + + // ... and return the new size. + return Values.size(); +} + +void ValueEnumerator::incorporateFunction(const Function &F) { + NumModuleValues = Values.size(); + + // Adding function arguments to the value table. + for(Function::const_arg_iterator I = F.arg_begin(), E = F.arg_end(); + I != E; ++I) + EnumerateValue(I); + + FirstFuncConstantID = Values.size(); + + // Add all function-level constants to the value table. + for (Function::const_iterator BB = F.begin(), E = F.end(); BB != E; ++BB) { + for (BasicBlock::const_iterator I = BB->begin(), E = BB->end(); I!=E; ++I) + for (User::const_op_iterator OI = I->op_begin(), E = I->op_end(); + OI != E; ++OI) { + if ((isa<Constant>(*OI) && !isa<GlobalValue>(*OI)) || + isa<InlineAsm>(*OI)) + EnumerateValue(*OI); + } + BasicBlocks.push_back(BB); + ValueMap[BB] = BasicBlocks.size(); + } + + // Optimize the constant layout. + OptimizeConstants(FirstFuncConstantID, Values.size()); + + // Add the function's parameter attributes so they are available for use in + // the function's instruction. + EnumerateAttributes(F.getAttributes()); + + FirstInstID = Values.size(); + + // Add all of the instructions. + for (Function::const_iterator BB = F.begin(), E = F.end(); BB != E; ++BB) { + for (BasicBlock::const_iterator I = BB->begin(), E = BB->end(); I!=E; ++I) { + if (I->getType() != Type::VoidTy) + EnumerateValue(I); + } + } +} + +void ValueEnumerator::purgeFunction() { + /// Remove purged values from the ValueMap. + for (unsigned i = NumModuleValues, e = Values.size(); i != e; ++i) + ValueMap.erase(Values[i].first); + for (unsigned i = 0, e = BasicBlocks.size(); i != e; ++i) + ValueMap.erase(BasicBlocks[i]); + + Values.resize(NumModuleValues); + BasicBlocks.clear(); +} + diff --git a/lib/Bitcode/Writer/ValueEnumerator.h b/lib/Bitcode/Writer/ValueEnumerator.h new file mode 100644 index 0000000..bb0324b --- /dev/null +++ b/lib/Bitcode/Writer/ValueEnumerator.h @@ -0,0 +1,127 @@ +//===-- Bitcode/Writer/ValueEnumerator.h - Number values --------*- C++ -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This class gives values and types Unique ID's. +// +//===----------------------------------------------------------------------===// + +#ifndef VALUE_ENUMERATOR_H +#define VALUE_ENUMERATOR_H + +#include "llvm/ADT/DenseMap.h" +#include "llvm/Attributes.h" +#include <vector> + +namespace llvm { + +class Type; +class Value; +class BasicBlock; +class Function; +class Module; +class AttrListPtr; +class TypeSymbolTable; +class ValueSymbolTable; + +class ValueEnumerator { +public: + // For each type, we remember its Type* and occurrence frequency. + typedef std::vector<std::pair<const Type*, unsigned> > TypeList; + + // For each value, we remember its Value* and occurrence frequency. + typedef std::vector<std::pair<const Value*, unsigned> > ValueList; +private: + typedef DenseMap<const Type*, unsigned> TypeMapType; + TypeMapType TypeMap; + TypeList Types; + + typedef DenseMap<const Value*, unsigned> ValueMapType; + ValueMapType ValueMap; + ValueList Values; + + typedef DenseMap<void*, unsigned> AttributeMapType; + AttributeMapType AttributeMap; + std::vector<AttrListPtr> Attributes; + + /// BasicBlocks - This contains all the basic blocks for the currently + /// incorporated function. Their reverse mapping is stored in ValueMap. + std::vector<const BasicBlock*> BasicBlocks; + + /// When a function is incorporated, this is the size of the Values list + /// before incorporation. + unsigned NumModuleValues; + unsigned FirstFuncConstantID; + unsigned FirstInstID; + + ValueEnumerator(const ValueEnumerator &); // DO NOT IMPLEMENT + void operator=(const ValueEnumerator &); // DO NOT IMPLEMENT +public: + ValueEnumerator(const Module *M); + + unsigned getValueID(const Value *V) const { + ValueMapType::const_iterator I = ValueMap.find(V); + assert(I != ValueMap.end() && "Value not in slotcalculator!"); + return I->second-1; + } + + unsigned getTypeID(const Type *T) const { + TypeMapType::const_iterator I = TypeMap.find(T); + assert(I != TypeMap.end() && "Type not in ValueEnumerator!"); + return I->second-1; + } + + unsigned getAttributeID(const AttrListPtr &PAL) const { + if (PAL.isEmpty()) return 0; // Null maps to zero. + AttributeMapType::const_iterator I = AttributeMap.find(PAL.getRawPointer()); + assert(I != AttributeMap.end() && "Attribute not in ValueEnumerator!"); + return I->second; + } + + /// getFunctionConstantRange - Return the range of values that corresponds to + /// function-local constants. + void getFunctionConstantRange(unsigned &Start, unsigned &End) const { + Start = FirstFuncConstantID; + End = FirstInstID; + } + + const ValueList &getValues() const { return Values; } + const TypeList &getTypes() const { return Types; } + const std::vector<const BasicBlock*> &getBasicBlocks() const { + return BasicBlocks; + } + const std::vector<AttrListPtr> &getAttributes() const { + return Attributes; + } + + /// PurgeAggregateValues - If there are any aggregate values at the end of the + /// value list, remove them and return the count of the remaining values. If + /// there are none, return -1. + int PurgeAggregateValues(); + + /// incorporateFunction/purgeFunction - If you'd like to deal with a function, + /// use these two methods to get its data into the ValueEnumerator! + /// + void incorporateFunction(const Function &F); + void purgeFunction(); + +private: + void OptimizeConstants(unsigned CstStart, unsigned CstEnd); + + void EnumerateValue(const Value *V); + void EnumerateType(const Type *T); + void EnumerateOperandType(const Value *V); + void EnumerateAttributes(const AttrListPtr &PAL); + + void EnumerateTypeSymbolTable(const TypeSymbolTable &ST); + void EnumerateValueSymbolTable(const ValueSymbolTable &ST); +}; + +} // End llvm namespace + +#endif |