#include "ReaderInternals.h"
#include "llvm/Bytecode/Reader.h"
#include "llvm/Bytecode/Format.h"
-#include "llvm/Module.h"
#include "llvm/Constants.h"
#include "llvm/iPHINode.h"
#include "llvm/iOther.h"
-#include <sys/types.h>
-#include <sys/stat.h>
-#include <sys/mman.h>
-#include <fcntl.h>
-#include <unistd.h>
+#include "llvm/Module.h"
+#include "Support/StringExtras.h"
+#include "Config/unistd.h"
+#include "Config/sys/mman.h"
+#include "Config/sys/stat.h"
+#include "Config/sys/types.h"
#include <algorithm>
+#include <memory>
-bool BytecodeParser::getTypeSlot(const Type *Ty, unsigned &Slot) {
- if (Ty->isPrimitiveType()) {
- Slot = Ty->getPrimitiveID();
- } else {
- // Check the function level types first...
- TypeValuesListTy::iterator I = find(FunctionTypeValues.begin(),
- FunctionTypeValues.end(), Ty);
- if (I != FunctionTypeValues.end()) {
- Slot = FirstDerivedTyID+ModuleTypeValues.size()+
+static inline void ALIGN32(const unsigned char *&begin,
+ const unsigned char *end) {
+ if (align32(begin, end))
+ throw std::string("Alignment error in buffer: read past end of block.");
+}
+
+unsigned BytecodeParser::getTypeSlot(const Type *Ty) {
+ if (Ty->isPrimitiveType())
+ return Ty->getPrimitiveID();
+
+ // Check the function level types first...
+ TypeValuesListTy::iterator I = find(FunctionTypeValues.begin(),
+ FunctionTypeValues.end(), Ty);
+ if (I != FunctionTypeValues.end())
+ return FirstDerivedTyID + ModuleTypeValues.size() +
(&*I - &FunctionTypeValues[0]);
- } else {
- I = find(ModuleTypeValues.begin(), ModuleTypeValues.end(), Ty);
- if (I == ModuleTypeValues.end()) return true; // Didn't find type!
- Slot = FirstDerivedTyID + (&*I - &ModuleTypeValues[0]);
- }
- }
- //cerr << "getTypeSlot '" << Ty->getName() << "' = " << Slot << "\n";
- return false;
+
+ I = find(ModuleTypeValues.begin(), ModuleTypeValues.end(), Ty);
+ if (I == ModuleTypeValues.end())
+ throw std::string("Didn't find type in ModuleTypeValues.");
+ return FirstDerivedTyID + (&*I - &ModuleTypeValues[0]);
}
const Type *BytecodeParser::getType(unsigned ID) {
- if (ID < Type::NumPrimitiveIDs) {
- const Type *T = Type::getPrimitiveType((Type::PrimitiveID)ID);
- if (T) return T;
- }
+ if (ID < Type::NumPrimitiveIDs)
+ if (const Type *T = Type::getPrimitiveType((Type::PrimitiveID)ID))
+ return T;
//cerr << "Looking up Type ID: " << ID << "\n";
- const Value *V = getValue(Type::TypeTy, ID, false);
- return cast_or_null<Type>(V);
+
+ if (ID < Type::NumPrimitiveIDs)
+ if (const Type *T = Type::getPrimitiveType((Type::PrimitiveID)ID))
+ return T; // Asked for a primitive type...
+
+ // Otherwise, derived types need offset...
+ ID -= FirstDerivedTyID;
+
+ // Is it a module-level type?
+ if (ID < ModuleTypeValues.size())
+ return ModuleTypeValues[ID].get();
+
+ // Nope, is it a function-level type?
+ ID -= ModuleTypeValues.size();
+ if (ID < FunctionTypeValues.size())
+ return FunctionTypeValues[ID].get();
+
+ throw std::string("Illegal type reference!");
}
-int BytecodeParser::insertValue(Value *Val, ValueTable &ValueTab) {
+unsigned BytecodeParser::insertValue(Value *Val, ValueTable &ValueTab) {
assert((!HasImplicitZeroInitializer || !isa<Constant>(Val) ||
Val->getType()->isPrimitiveType() ||
!cast<Constant>(Val)->isNullValue()) &&
"Cannot read null values from bytecode!");
- unsigned type;
- if (getTypeSlot(Val->getType(), type)) return -1;
+ unsigned type = getTypeSlot(Val->getType());
assert(type != Type::TypeTyID && "Types should never be insertValue'd!");
if (ValueTab.size() <= type) {
}
//cerr << "insertValue Values[" << type << "][" << ValueTab[type].size()
- // << "] = " << Val << "\n";
+ // << "] = " << Val << "\n";
ValueTab[type]->push_back(Val);
bool HasOffset = HasImplicitZeroInitializer &&
- !Val->getType()->isPrimitiveType();
+ !Val->getType()->isPrimitiveType();
return ValueTab[type]->size()-1 + HasOffset;
}
-void BytecodeParser::setValueTo(ValueTable &ValueTab, unsigned Slot,
- Value *Val) {
- assert(&ValueTab == &ModuleValues && "Can only setValueTo on Module values!");
- unsigned type;
- if (getTypeSlot(Val->getType(), type))
- assert(0 && "getTypeSlot failed!");
-
- assert((!HasImplicitZeroInitializer || Slot != 0) &&
- "Cannot change zero init");
- assert(type < ValueTab.size() && Slot <= ValueTab[type]->size());
- ValueTab[type]->setOperand(Slot-HasImplicitZeroInitializer, Val);
+Value *BytecodeParser::getValue(const Type *Ty, unsigned oNum, bool Create) {
+ return getValue(getTypeSlot(Ty), oNum, Create);
}
-Value *BytecodeParser::getValue(const Type *Ty, unsigned oNum, bool Create) {
+Value *BytecodeParser::getValue(unsigned type, unsigned oNum, bool Create) {
+ assert(type != Type::TypeTyID && "getValue() cannot get types!");
+ assert(type != Type::LabelTyID && "getValue() cannot get blocks!");
unsigned Num = oNum;
- unsigned type; // The type plane it lives in...
-
- if (getTypeSlot(Ty, type)) return 0;
-
- if (type == Type::TypeTyID) { // The 'type' plane has implicit values
- assert(Create == false);
- if (Num < Type::NumPrimitiveIDs) {
- const Type *T = Type::getPrimitiveType((Type::PrimitiveID)Num);
- if (T) return (Value*)T; // Asked for a primitive type...
- }
-
- // Otherwise, derived types need offset...
- Num -= FirstDerivedTyID;
-
- // Is it a module level type?
- if (Num < ModuleTypeValues.size())
- return (Value*)ModuleTypeValues[Num].get();
-
- // Nope, is it a function level type?
- Num -= ModuleTypeValues.size();
- if (Num < FunctionTypeValues.size())
- return (Value*)FunctionTypeValues[Num].get();
-
- return 0;
- }
if (HasImplicitZeroInitializer && type >= FirstDerivedTyID) {
if (Num == 0)
- return Constant::getNullValue(Ty);
+ return Constant::getNullValue(getType(type));
--Num;
}
if (!Create) return 0; // Do not create a placeholder?
- Value *d = 0;
- switch (Ty->getPrimitiveID()) {
- case Type::LabelTyID:
- d = new BBPHolder(Ty, oNum);
- break;
- default:
- d = new ValPHolder(Ty, oNum);
- break;
- }
+ std::pair<unsigned,unsigned> KeyValue(type, oNum);
+ std::map<std::pair<unsigned,unsigned>, Value*>::iterator I =
+ ForwardReferences.lower_bound(KeyValue);
+ if (I != ForwardReferences.end() && I->first == KeyValue)
+ return I->second; // We have already created this placeholder
- assert(d != 0 && "How did we not make something?");
- if (insertValue(d, LateResolveValues) == -1) return 0;
- return d;
+ Value *Val = new Argument(getType(type));
+ ForwardReferences.insert(I, std::make_pair(KeyValue, Val));
+ return Val;
+}
+
+/// getBasicBlock - Get a particular numbered basic block, which might be a
+/// forward reference. This works together with ParseBasicBlock to handle these
+/// forward references in a clean manner.
+///
+BasicBlock *BytecodeParser::getBasicBlock(unsigned ID) {
+ // Make sure there is room in the table...
+ if (ParsedBasicBlocks.size() <= ID) ParsedBasicBlocks.resize(ID+1);
+
+ // First check to see if this is a backwards reference, i.e., ParseBasicBlock
+ // has already created this block, or if the forward reference has already
+ // been created.
+ if (ParsedBasicBlocks[ID])
+ return ParsedBasicBlocks[ID];
+
+ // Otherwise, the basic block has not yet been created. Do so and add it to
+ // the ParsedBasicBlocks list.
+ return ParsedBasicBlocks[ID] = new BasicBlock();
}
/// getConstantValue - Just like getValue, except that it returns a null pointer
}
-bool BytecodeParser::postResolveValues(ValueTable &ValTab) {
- bool Error = false;
- while (!ValTab.empty()) {
- ValueList &DL = *ValTab.back();
- ValTab.pop_back();
-
- while (!DL.empty()) {
- Value *D = DL.back();
- unsigned IDNumber = getValueIDNumberFromPlaceHolder(D);
- DL.pop_back();
-
- Value *NewDef = getValue(D->getType(), IDNumber, false);
- if (NewDef == 0) {
- Error = true; // Unresolved thinger
- std::cerr << "Unresolvable reference found: <"
- << *D->getType() << ">:" << IDNumber <<"!\n";
- } else {
- // Fixup all of the uses of this placeholder def...
- D->replaceAllUsesWith(NewDef);
-
- // Now that all the uses are gone, delete the placeholder...
- // If we couldn't find a def (error case), then leak a little
- delete D; // memory, 'cause otherwise we can't remove all uses!
- }
- }
- delete &DL;
- }
-
- return Error;
-}
-
-bool BytecodeParser::ParseBasicBlock(const uchar *&Buf, const uchar *EndBuf,
- BasicBlock *&BB) {
- BB = new BasicBlock();
+BasicBlock *BytecodeParser::ParseBasicBlock(const unsigned char *&Buf,
+ const unsigned char *EndBuf,
+ unsigned BlockNo) {
+ BasicBlock *BB;
+ if (ParsedBasicBlocks.size() == BlockNo)
+ ParsedBasicBlocks.push_back(BB = new BasicBlock());
+ else if (ParsedBasicBlocks[BlockNo] == 0)
+ BB = ParsedBasicBlocks[BlockNo] = new BasicBlock();
+ else
+ BB = ParsedBasicBlocks[BlockNo];
while (Buf < EndBuf) {
- Instruction *Inst;
- if (ParseInstruction(Buf, EndBuf, Inst, /*HACK*/BB)) {
- delete BB;
- return true;
- }
-
- if (Inst == 0) { delete BB; return true; }
- if (insertValue(Inst, Values) == -1) { delete BB; return true; }
-
+ Instruction *Inst = ParseInstruction(Buf, EndBuf);
+ insertValue(Inst, Values);
BB->getInstList().push_back(Inst);
-
BCR_TRACE(4, Inst);
}
- return false;
+ return BB;
}
-bool BytecodeParser::ParseSymbolTable(const uchar *&Buf, const uchar *EndBuf,
- SymbolTable *ST) {
+void BytecodeParser::ParseSymbolTable(const unsigned char *&Buf,
+ const unsigned char *EndBuf,
+ SymbolTable *ST,
+ Function *CurrentFunction) {
while (Buf < EndBuf) {
// Symtab block header: [num entries][type id number]
unsigned NumEntries, Typ;
if (read_vbr(Buf, EndBuf, NumEntries) ||
- read_vbr(Buf, EndBuf, Typ)) return true;
+ read_vbr(Buf, EndBuf, Typ)) throw Error_readvbr;
const Type *Ty = getType(Typ);
- if (Ty == 0) return true;
-
- BCR_TRACE(3, "Plane Type: '" << Ty << "' with " << NumEntries <<
- " entries\n");
+ BCR_TRACE(3, "Plane Type: '" << *Ty << "' with " << NumEntries <<
+ " entries\n");
for (unsigned i = 0; i < NumEntries; ++i) {
// Symtab entry: [def slot #][name]
unsigned slot;
- if (read_vbr(Buf, EndBuf, slot)) return true;
+ if (read_vbr(Buf, EndBuf, slot)) throw Error_readvbr;
std::string Name;
if (read(Buf, EndBuf, Name, false)) // Not aligned...
- return true;
-
- Value *V = getValue(Ty, slot, false); // Find mapping...
- if (V == 0) {
- BCR_TRACE(3, "FAILED LOOKUP: Slot #" << slot << "\n");
- return true;
- }
+ throw std::string("Buffer not aligned.");
+
+ Value *V = 0;
+ if (Typ == Type::TypeTyID)
+ V = (Value*)getType(slot);
+ else if (Typ == Type::LabelTyID) {
+ if (CurrentFunction) {
+ // FIXME: THIS IS N^2!!!
+ Function::iterator BlockIterator = CurrentFunction->begin();
+ std::advance(BlockIterator, slot);
+ V = BlockIterator;
+ }
+ } else
+ V = getValue(Typ, slot, false); // Find mapping...
+ if (V == 0) throw std::string("Failed value look-up.");
BCR_TRACE(4, "Map: '" << Name << "' to #" << slot << ":" << *V;
- if (!isa<Instruction>(V)) std::cerr << "\n");
+ if (!isa<Instruction>(V)) std::cerr << "\n");
V->setName(Name, ST);
}
}
- if (Buf > EndBuf) return true;
- return false;
+ if (Buf > EndBuf) throw std::string("Tried to read past end of buffer.");
}
void BytecodeParser::ResolveReferencesToValue(Value *NewV, unsigned Slot) {
GlobalRefs.erase(I); // Remove the map entry for it
}
-
-bool BytecodeParser::ParseFunction(const uchar *&Buf, const uchar *EndBuf) {
- // Clear out the local values table...
- if (FunctionSignatureList.empty()) {
- Error = "Function found, but FunctionSignatureList empty!";
- return true; // Unexpected function!
- }
-
- unsigned isInternal;
- if (read_vbr(Buf, EndBuf, isInternal)) return true;
+void BytecodeParser::ParseFunction(const unsigned char *&Buf,
+ const unsigned char *EndBuf) {
+ if (FunctionSignatureList.empty())
+ throw std::string("FunctionSignatureList empty!");
Function *F = FunctionSignatureList.back().first;
unsigned FunctionSlot = FunctionSignatureList.back().second;
FunctionSignatureList.pop_back();
- F->setInternalLinkage(isInternal != 0);
+
+ // Save the information for future reading of the function
+ LazyFunctionInfo *LFI = new LazyFunctionInfo();
+ LFI->Buf = Buf; LFI->EndBuf = EndBuf; LFI->FunctionSlot = FunctionSlot;
+ LazyFunctionLoadMap[F] = LFI;
+ // Pretend we've `parsed' this function
+ Buf = EndBuf;
+}
+
+void BytecodeParser::materializeFunction(Function* F) {
+ // Find {start, end} pointers and slot in the map. If not there, we're done.
+ std::map<Function*, LazyFunctionInfo*>::iterator Fi =
+ LazyFunctionLoadMap.find(F);
+ if (Fi == LazyFunctionLoadMap.end()) return;
+
+ LazyFunctionInfo *LFI = Fi->second;
+ const unsigned char *Buf = LFI->Buf;
+ const unsigned char *EndBuf = LFI->EndBuf;
+ unsigned FunctionSlot = LFI->FunctionSlot;
+ LazyFunctionLoadMap.erase(Fi);
+ delete LFI;
+
+ GlobalValue::LinkageTypes Linkage = GlobalValue::ExternalLinkage;
+
+ if (!hasInternalMarkerOnly) {
+ unsigned LinkageType;
+ if (read_vbr(Buf, EndBuf, LinkageType))
+ throw std::string("ParseFunction: Error reading from buffer.");
+ if (LinkageType & ~0x3)
+ throw std::string("Invalid linkage type for Function.");
+ Linkage = (GlobalValue::LinkageTypes)LinkageType;
+ } else {
+ // We used to only support two linkage models: internal and external
+ unsigned isInternal;
+ if (read_vbr(Buf, EndBuf, isInternal))
+ throw std::string("ParseFunction: Error reading from buffer.");
+ if (isInternal) Linkage = GlobalValue::InternalLinkage;
+ }
+
+ F->setLinkage(Linkage);
const FunctionType::ParamTypes &Params =F->getFunctionType()->getParamTypes();
Function::aiterator AI = F->abegin();
for (FunctionType::ParamTypes::const_iterator It = Params.begin();
- It != Params.end(); ++It, ++AI) {
- if (insertValue(AI, Values) == -1) {
- Error = "Error reading function arguments!\n";
- return true;
- }
- }
+ It != Params.end(); ++It, ++AI)
+ insertValue(AI, Values);
+
+ // Keep track of how many basic blocks we have read in...
+ unsigned BlockNum = 0;
while (Buf < EndBuf) {
unsigned Type, Size;
const unsigned char *OldBuf = Buf;
- if (readBlock(Buf, EndBuf, Type, Size)) {
- Error = "Error reading Function level block!";
- return true;
- }
+ readBlock(Buf, EndBuf, Type, Size);
switch (Type) {
- case BytecodeFormat::ConstantPool:
+ case BytecodeFormat::ConstantPool: {
BCR_TRACE(2, "BLOCK BytecodeFormat::ConstantPool: {\n");
- if (ParseConstantPool(Buf, Buf+Size, Values, FunctionTypeValues))
- return true;
+ ParseConstantPool(Buf, Buf+Size, Values, FunctionTypeValues);
break;
+ }
case BytecodeFormat::BasicBlock: {
BCR_TRACE(2, "BLOCK BytecodeFormat::BasicBlock: {\n");
- BasicBlock *BB;
- if (ParseBasicBlock(Buf, Buf+Size, BB) ||
- insertValue(BB, Values) == -1)
- return true; // Parse error... :(
-
+ BasicBlock *BB = ParseBasicBlock(Buf, Buf+Size, BlockNum++);
F->getBasicBlockList().push_back(BB);
break;
}
- case BytecodeFormat::SymbolTable:
+ case BytecodeFormat::SymbolTable: {
BCR_TRACE(2, "BLOCK BytecodeFormat::SymbolTable: {\n");
- if (ParseSymbolTable(Buf, Buf+Size, &F->getSymbolTable()))
- return true;
+ ParseSymbolTable(Buf, Buf+Size, &F->getSymbolTable(), F);
break;
+ }
default:
BCR_TRACE(2, "BLOCK <unknown>:ignored! {\n");
Buf += Size;
- if (OldBuf > Buf) return true; // Wrap around!
+ if (OldBuf > Buf)
+ throw std::string("Wrapped around reading bytecode.");
break;
}
BCR_TRACE(2, "} end block\n");
- if (align32(Buf, EndBuf)) {
- Error = "Error aligning Function level block!";
- return true; // Malformed bc file, read past end of block.
- }
+ // Malformed bc file if read past end of block.
+ ALIGN32(Buf, EndBuf);
}
- if (postResolveValues(LateResolveValues)) {
- Error = "Error resolving function values!";
- return true; // Unresolvable references!
+ // Make sure there were no references to non-existant basic blocks.
+ if (BlockNum != ParsedBasicBlocks.size())
+ throw std::string("Illegal basic block operand reference");
+ ParsedBasicBlocks.clear();
+
+
+ // Resolve forward references
+ while (!ForwardReferences.empty()) {
+ std::map<std::pair<unsigned,unsigned>, Value*>::iterator I = ForwardReferences.begin();
+ unsigned type = I->first.first;
+ unsigned Slot = I->first.second;
+ Value *PlaceHolder = I->second;
+ ForwardReferences.erase(I);
+
+ Value *NewVal = getValue(type, Slot, false);
+ if (NewVal == 0)
+ throw std::string("Unresolvable reference found: <" +
+ PlaceHolder->getType()->getDescription() + ">:" +
+ utostr(Slot) + ".");
+
+ // Fixup all of the uses of this placeholder def...
+ PlaceHolder->replaceAllUsesWith(NewVal);
+
+ // Now that all the uses are gone, delete the placeholder...
+ // If we couldn't find a def (error case), then leak a little
+ // memory, because otherwise we can't remove all uses!
+ delete PlaceHolder;
}
- ResolveReferencesToValue(F, FunctionSlot);
-
- // Clear out function level types...
+ // Clear out function-level types...
FunctionTypeValues.clear();
freeTable(Values);
- return false;
}
-bool BytecodeParser::ParseModuleGlobalInfo(const uchar *&Buf, const uchar *End){
- if (!FunctionSignatureList.empty()) {
- Error = "Two ModuleGlobalInfo packets found!";
- return true; // Two ModuleGlobal blocks?
- }
+void BytecodeParser::ParseModuleGlobalInfo(const unsigned char *&Buf,
+ const unsigned char *End) {
+ if (!FunctionSignatureList.empty())
+ throw std::string("Two ModuleGlobalInfo packets found!");
// Read global variables...
unsigned VarType;
- if (read_vbr(Buf, End, VarType)) return true;
+ if (read_vbr(Buf, End, VarType)) throw Error_readvbr;
while (VarType != Type::VoidTyID) { // List is terminated by Void
- // VarType Fields: bit0 = isConstant, bit1 = hasInitializer,
- // bit2 = isInternal, bit3+ = slot#
- const Type *Ty = getType(VarType >> 3);
- if (!Ty || !isa<PointerType>(Ty)) {
- Error = "Global not pointer type! Ty = " + Ty->getDescription();
- return true;
+ unsigned SlotNo;
+ GlobalValue::LinkageTypes Linkage;
+
+ if (!hasInternalMarkerOnly) {
+ // VarType Fields: bit0 = isConstant, bit1 = hasInitializer,
+ // bit2,3 = Linkage, bit4+ = slot#
+ SlotNo = VarType >> 4;
+ Linkage = (GlobalValue::LinkageTypes)((VarType >> 2) & 3);
+ } else {
+ // VarType Fields: bit0 = isConstant, bit1 = hasInitializer,
+ // bit2 = isInternal, bit3+ = slot#
+ SlotNo = VarType >> 3;
+ Linkage = (VarType & 4) ? GlobalValue::InternalLinkage :
+ GlobalValue::ExternalLinkage;
}
+ const Type *Ty = getType(SlotNo);
+ if (!isa<PointerType>(Ty))
+ throw std::string("Global not pointer type! Ty = " +
+ Ty->getDescription());
+
const Type *ElTy = cast<PointerType>(Ty)->getElementType();
// Create the global variable...
- GlobalVariable *GV = new GlobalVariable(ElTy, VarType & 1, VarType & 4,
+ GlobalVariable *GV = new GlobalVariable(ElTy, VarType & 1, Linkage,
0, "", TheModule);
- int DestSlot = insertValue(GV, ModuleValues);
- if (DestSlot == -1) return true;
BCR_TRACE(2, "Global Variable of type: " << *Ty << "\n");
- ResolveReferencesToValue(GV, (unsigned)DestSlot);
+ ResolveReferencesToValue(GV, insertValue(GV, ModuleValues));
- if (VarType & 2) { // Does it have an initalizer?
+ if (VarType & 2) { // Does it have an initializer?
unsigned InitSlot;
- if (read_vbr(Buf, End, InitSlot)) return true;
+ if (read_vbr(Buf, End, InitSlot)) throw Error_readvbr;
GlobalInits.push_back(std::make_pair(GV, InitSlot));
}
- if (read_vbr(Buf, End, VarType)) return true;
+ if (read_vbr(Buf, End, VarType)) throw Error_readvbr;
}
// Read the function objects for all of the functions that are coming
unsigned FnSignature;
- if (read_vbr(Buf, End, FnSignature)) return true;
+ if (read_vbr(Buf, End, FnSignature)) throw Error_readvbr;
while (FnSignature != Type::VoidTyID) { // List is terminated by Void
const Type *Ty = getType(FnSignature);
- if (!Ty || !isa<PointerType>(Ty) ||
- !isa<FunctionType>(cast<PointerType>(Ty)->getElementType())) {
- Error = "Function not ptr to func type! Ty = " + Ty->getDescription();
- return true;
- }
+ if (!isa<PointerType>(Ty) ||
+ !isa<FunctionType>(cast<PointerType>(Ty)->getElementType()))
+ throw std::string("Function not ptr to func type! Ty = " +
+ Ty->getDescription());
// We create functions by passing the underlying FunctionType to create...
Ty = cast<PointerType>(Ty)->getElementType();
// this placeholder is replaced.
// Insert the placeholder...
- Function *Func = new Function(cast<FunctionType>(Ty), false, "", TheModule);
- int DestSlot = insertValue(Func, ModuleValues);
- if (DestSlot == -1) return true;
- ResolveReferencesToValue(Func, (unsigned)DestSlot);
+ Function *Func = new Function(cast<FunctionType>(Ty),
+ GlobalValue::InternalLinkage, "", TheModule);
+ unsigned DestSlot = insertValue(Func, ModuleValues);
+ ResolveReferencesToValue(Func, DestSlot);
// Keep track of this information in a list that is emptied as functions are
// loaded...
//
FunctionSignatureList.push_back(std::make_pair(Func, DestSlot));
- if (read_vbr(Buf, End, FnSignature)) return true;
+ if (read_vbr(Buf, End, FnSignature)) throw Error_readvbr;
BCR_TRACE(2, "Function of type: " << Ty << "\n");
}
- if (align32(Buf, End)) return true;
+ ALIGN32(Buf, End);
// Now that the function signature list is set up, reverse it so that we can
// remove elements efficiently from the back of the vector.
// we don't understand, so we transparently ignore them.
//
Buf = End;
- return false;
}
-bool BytecodeParser::ParseVersionInfo(const uchar *&Buf, const uchar *EndBuf) {
+void BytecodeParser::ParseVersionInfo(const unsigned char *&Buf,
+ const unsigned char *EndBuf) {
unsigned Version;
- if (read_vbr(Buf, EndBuf, Version)) return true;
+ if (read_vbr(Buf, EndBuf, Version)) throw Error_readvbr;
// Unpack version number: low four bits are for flags, top bits = version
- isBigEndian = Version & 1;
- hasLongPointers = Version & 2;
- RevisionNum = Version >> 4;
+ Module::Endianness Endianness;
+ Module::PointerSize PointerSize;
+ Endianness = (Version & 1) ? Module::BigEndian : Module::LittleEndian;
+ PointerSize = (Version & 2) ? Module::Pointer64 : Module::Pointer32;
+
+ bool hasNoEndianness = Version & 4;
+ bool hasNoPointerSize = Version & 8;
+
+ RevisionNum = Version >> 4;
+
+ // Default values for the current bytecode version
HasImplicitZeroInitializer = true;
+ hasInternalMarkerOnly = false;
+ FirstDerivedTyID = 14;
switch (RevisionNum) {
case 0: // Initial revision
// only valid with a 14 in the flags values. Also, it does not support
// encoding zero initializers for arrays compactly.
//
- if (Version != 14) return true; // Unknown revision 0 flags?
- FirstDerivedTyID = 14;
+ if (Version != 14) throw std::string("Unknown revision 0 flags?");
HasImplicitZeroInitializer = false;
- isBigEndian = hasLongPointers = true;
+ Endianness = Module::BigEndian;
+ PointerSize = Module::Pointer64;
+ hasInternalMarkerOnly = true;
+ hasNoEndianness = hasNoPointerSize = false;
break;
case 1:
- // Version #1 has two bit fields: isBigEndian and hasLongPointers
- FirstDerivedTyID = 14;
+ // Version #1 has four bit fields: isBigEndian, hasLongPointers,
+ // hasNoEndianness, and hasNoPointerSize.
+ hasInternalMarkerOnly = true;
+ break;
+ case 2:
+ // Version #2 added information about all 4 linkage types instead of just
+ // having internal and external.
break;
default:
- Error = "Unknown bytecode version number!";
- return true;
+ throw std::string("Unknown bytecode version number!");
}
+ if (hasNoEndianness) Endianness = Module::AnyEndianness;
+ if (hasNoPointerSize) PointerSize = Module::AnyPointerSize;
+
+ TheModule->setEndianness(Endianness);
+ TheModule->setPointerSize(PointerSize);
BCR_TRACE(1, "Bytecode Rev = " << (unsigned)RevisionNum << "\n");
- BCR_TRACE(1, "BigEndian/LongPointers = " << isBigEndian << ","
- << hasLongPointers << "\n");
+ BCR_TRACE(1, "Endianness/PointerSize = " << Endianness << ","
+ << PointerSize << "\n");
BCR_TRACE(1, "HasImplicitZeroInit = " << HasImplicitZeroInitializer << "\n");
- return false;
}
-bool BytecodeParser::ParseModule(const uchar *Buf, const uchar *EndBuf) {
+void BytecodeParser::ParseModule(const unsigned char *Buf,
+ const unsigned char *EndBuf) {
unsigned Type, Size;
- if (readBlock(Buf, EndBuf, Type, Size)) return true;
- if (Type != BytecodeFormat::Module || Buf+Size != EndBuf) {
- Error = "Expected Module packet!";
- return true; // Hrm, not a class?
- }
+ readBlock(Buf, EndBuf, Type, Size);
+ if (Type != BytecodeFormat::Module || Buf+Size != EndBuf)
+ throw std::string("Expected Module packet! B: "+
+ utostr((unsigned)(intptr_t)Buf) + ", S: "+utostr(Size)+
+ " E: "+utostr((unsigned)(intptr_t)EndBuf)); // Hrm, not a class?
BCR_TRACE(0, "BLOCK BytecodeFormat::Module: {\n");
FunctionSignatureList.clear(); // Just in case...
// Read into instance variables...
- if (ParseVersionInfo(Buf, EndBuf)) return true;
- if (align32(Buf, EndBuf)) return true;
+ ParseVersionInfo(Buf, EndBuf);
+ ALIGN32(Buf, EndBuf);
while (Buf < EndBuf) {
const unsigned char *OldBuf = Buf;
- if (readBlock(Buf, EndBuf, Type, Size)) return true;
+ readBlock(Buf, EndBuf, Type, Size);
switch (Type) {
case BytecodeFormat::GlobalTypePlane:
BCR_TRACE(1, "BLOCK BytecodeFormat::GlobalTypePlane: {\n");
- if (ParseGlobalTypes(Buf, Buf+Size)) return true;
+ ParseGlobalTypes(Buf, Buf+Size);
break;
case BytecodeFormat::ModuleGlobalInfo:
BCR_TRACE(1, "BLOCK BytecodeFormat::ModuleGlobalInfo: {\n");
- if (ParseModuleGlobalInfo(Buf, Buf+Size)) return true;
+ ParseModuleGlobalInfo(Buf, Buf+Size);
break;
case BytecodeFormat::ConstantPool:
BCR_TRACE(1, "BLOCK BytecodeFormat::ConstantPool: {\n");
- if (ParseConstantPool(Buf, Buf+Size, ModuleValues, ModuleTypeValues))
- return true;
+ ParseConstantPool(Buf, Buf+Size, ModuleValues, ModuleTypeValues);
break;
case BytecodeFormat::Function: {
BCR_TRACE(1, "BLOCK BytecodeFormat::Function: {\n");
- if (ParseFunction(Buf, Buf+Size))
- return true; // Error parsing function
+ ParseFunction(Buf, Buf+Size);
break;
}
case BytecodeFormat::SymbolTable:
BCR_TRACE(1, "BLOCK BytecodeFormat::SymbolTable: {\n");
- if (ParseSymbolTable(Buf, Buf+Size, &TheModule->getSymbolTable()))
- return true;
+ ParseSymbolTable(Buf, Buf+Size, &TheModule->getSymbolTable(), 0);
break;
default:
- Error = "Expected Module Block!";
Buf += Size;
- if (OldBuf > Buf) return true; // Wrap around!
+ if (OldBuf > Buf) throw std::string("Expected Module Block!");
break;
}
BCR_TRACE(1, "} end block\n");
- if (align32(Buf, EndBuf)) return true;
+ ALIGN32(Buf, EndBuf);
}
// After the module constant pool has been read, we can safely initialize
// Look up the initializer value...
if (Value *V = getValue(GV->getType()->getElementType(), Slot, false)) {
- if (GV->hasInitializer()) return true;
+ if (GV->hasInitializer())
+ throw std::string("Global *already* has an initializer?!");
GV->setInitializer(cast<Constant>(V));
} else
- return true;
+ throw std::string("Cannot find initializer value.");
}
- if (!FunctionSignatureList.empty()) { // Expected more functions!
- Error = "Function expected, but bytecode stream at end!";
- return true;
- }
+ if (!FunctionSignatureList.empty())
+ throw std::string("Function expected, but bytecode stream ended!");
BCR_TRACE(0, "} end block\n\n");
- return false;
}
-static inline Module *Error(std::string *ErrorStr, const char *Message) {
- if (ErrorStr) *ErrorStr = Message;
- return 0;
-}
+void
+BytecodeParser::ParseBytecode(const unsigned char *Buf, unsigned Length,
+ const std::string &ModuleID) {
+
+ unsigned char *EndBuf = (unsigned char*)(Buf + Length);
-Module *BytecodeParser::ParseBytecode(const uchar *Buf, const uchar *EndBuf) {
- unsigned Sig;
// Read and check signature...
+ unsigned Sig;
if (read(Buf, EndBuf, Sig) ||
- Sig != ('l' | ('l' << 8) | ('v' << 16) | 'm' << 24))
- return ::Error(&Error, "Invalid bytecode signature!");
-
- TheModule = new Module();
- if (ParseModule(Buf, EndBuf)) {
+ Sig != ('l' | ('l' << 8) | ('v' << 16) | ('m' << 24)))
+ throw std::string("Invalid bytecode signature!");
+
+ TheModule = new Module(ModuleID);
+ try {
+ ParseModule(Buf, EndBuf);
+ } catch (std::string &Error) {
+ freeState(); // Must destroy handles before deleting module!
delete TheModule;
TheModule = 0;
+ throw;
}
- return TheModule;
-}
-
-
-Module *ParseBytecodeBuffer(const unsigned char *Buffer, unsigned Length,
- std::string *ErrorStr) {
- BytecodeParser Parser;
- Module *R = Parser.ParseBytecode(Buffer, Buffer+Length);
- if (ErrorStr) *ErrorStr = Parser.getError();
- return R;
-}
-
-
-/// FDHandle - Simple handle class to make sure a file descriptor gets closed
-/// when the object is destroyed.
-class FDHandle {
- int FD;
-public:
- FDHandle(int fd) : FD(fd) {}
- operator int() const { return FD; }
- ~FDHandle() {
- if (FD != -1) close(FD);
- }
-};
-
-// Parse and return a class file...
-//
-Module *ParseBytecodeFile(const std::string &Filename, std::string *ErrorStr) {
- Module *Result = 0;
-
- if (Filename != std::string("-")) { // Read from a file...
- FDHandle FD = open(Filename.c_str(), O_RDONLY);
- if (FD == -1)
- return Error(ErrorStr, "Error opening file!");
-
- // Stat the file to get its length...
- struct stat StatBuf;
- if (fstat(FD, &StatBuf) == -1 || StatBuf.st_size == 0)
- return Error(ErrorStr, "Error stat'ing file!");
-
- // mmap in the file all at once...
- int Length = StatBuf.st_size;
- unsigned char *Buffer = (unsigned char*)mmap(0, Length, PROT_READ,
- MAP_PRIVATE, FD, 0);
- if (Buffer == (unsigned char*)MAP_FAILED)
- return Error(ErrorStr, "Error mmapping file!");
-
- // Parse the bytecode we mmapped in
- Result = ParseBytecodeBuffer(Buffer, Length, ErrorStr);
-
- // Unmmap the bytecode...
- munmap((char*)Buffer, Length);
- } else { // Read from stdin
- int BlockSize;
- uchar Buffer[4096*4];
- std::vector<unsigned char> FileData;
-
- // Read in all of the data from stdin, we cannot mmap stdin...
- while ((BlockSize = read(0 /*stdin*/, Buffer, 4096*4))) {
- if (BlockSize == -1)
- return Error(ErrorStr, "Error reading from stdin!");
-
- FileData.insert(FileData.end(), Buffer, Buffer+BlockSize);
- }
-
- if (FileData.empty())
- return Error(ErrorStr, "Standard Input empty!");
-
-#define ALIGN_PTRS 0
-#if ALIGN_PTRS
- uchar *Buf = (uchar*)mmap(0, FileData.size(), PROT_READ|PROT_WRITE,
- MAP_PRIVATE|MAP_ANONYMOUS, -1, 0);
- assert((Buf != (uchar*)-1) && "mmap returned error!");
- memcpy(Buf, &FileData[0], FileData.size());
-#else
- unsigned char *Buf = &FileData[0];
-#endif
-
- Result = ParseBytecodeBuffer(Buf, FileData.size(), ErrorStr);
-
-#if ALIGN_PTRS
- munmap((char*)Buf, FileData.size()); // Free mmap'd data area
-#endif
- }
-
- return Result;
}