// Note that this library should be as fast as possible, reentrant, and
// threadsafe!!
//
-// TODO: Make error message outputs be configurable depending on an option?
+// TODO: Return error messages to caller instead of printing them out directly.
// TODO: Allow passing in an option to ignore the symbol table
//
//===----------------------------------------------------------------------===//
#include "ReaderInternals.h"
+#include "Config/sys/mman.h"
#include "llvm/Bytecode/Reader.h"
#include "llvm/Bytecode/Format.h"
-#include "llvm/GlobalVariable.h"
#include "llvm/Module.h"
-#include "llvm/BasicBlock.h"
#include "llvm/Constants.h"
#include "llvm/iPHINode.h"
#include "llvm/iOther.h"
-#include "llvm/Argument.h"
-#include <sys/types.h>
-#include <sys/stat.h>
-#include <sys/mman.h>
-#include <fcntl.h>
-#include <unistd.h>
+#include "Config/sys/types.h"
+#include "Config/sys/stat.h"
+#include "Config/fcntl.h"
+#include "Config/unistd.h"
#include <algorithm>
-#include <iostream>
-using std::cerr;
-using std::make_pair;
bool BytecodeParser::getTypeSlot(const Type *Ty, unsigned &Slot) {
if (Ty->isPrimitiveType()) {
Slot = Ty->getPrimitiveID();
} else {
- // Check the method level types first...
- TypeValuesListTy::iterator I = find(MethodTypeValues.begin(),
- MethodTypeValues.end(), Ty);
- if (I != MethodTypeValues.end()) {
+ // Check the function level types first...
+ TypeValuesListTy::iterator I = find(FunctionTypeValues.begin(),
+ FunctionTypeValues.end(), Ty);
+ if (I != FunctionTypeValues.end()) {
Slot = FirstDerivedTyID+ModuleTypeValues.size()+
- (&*I - &MethodTypeValues[0]);
+ (&*I - &FunctionTypeValues[0]);
} else {
I = find(ModuleTypeValues.begin(), ModuleTypeValues.end(), Ty);
if (I == ModuleTypeValues.end()) return true; // Didn't find type!
}
const Type *BytecodeParser::getType(unsigned ID) {
- const Type *T = Type::getPrimitiveType((Type::PrimitiveID)ID);
- if (T) return T;
+ if (ID < Type::NumPrimitiveIDs) {
+ const Type *T = Type::getPrimitiveType((Type::PrimitiveID)ID);
+ if (T) return T;
+ }
//cerr << "Looking up Type ID: " << ID << "\n";
-
- const Value *D = getValue(Type::TypeTy, ID, false);
- if (D == 0) return failure<const Type*>(0);
-
- return cast<Type>(D);
+ const Value *V = getValue(Type::TypeTy, ID, false);
+ return cast_or_null<Type>(V);
}
-int BytecodeParser::insertValue(Value *Val, std::vector<ValueList> &ValueTab) {
+int 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 failure<int>(-1);
+ if (getTypeSlot(Val->getType(), type)) return -1;
assert(type != Type::TypeTyID && "Types should never be insertValue'd!");
- if (ValueTab.size() <= type)
- ValueTab.resize(type+1, ValueList());
+ if (ValueTab.size() <= type) {
+ unsigned OldSize = ValueTab.size();
+ ValueTab.resize(type+1);
+ while (OldSize != type+1)
+ ValueTab[OldSize++] = new ValueList();
+ }
//cerr << "insertValue Values[" << type << "][" << ValueTab[type].size()
// << "] = " << Val << "\n";
- ValueTab[type].push_back(Val);
+ ValueTab[type]->push_back(Val);
+
+ bool HasOffset = HasImplicitZeroInitializer &&
+ !Val->getType()->isPrimitiveType();
+
+ return ValueTab[type]->size()-1 + HasOffset;
+}
+
- return ValueTab[type].size()-1;
+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) {
unsigned Num = oNum;
unsigned type; // The type plane it lives in...
- if (getTypeSlot(Ty, type)) return failure<Value*>(0); // TODO: true
+ if (getTypeSlot(Ty, type)) return 0;
if (type == Type::TypeTyID) { // The 'type' plane has implicit values
assert(Create == false);
- const Type *T = Type::getPrimitiveType((Type::PrimitiveID)Num);
- if (T) return (Value*)T; // Asked for a primitive type...
+ 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;
if (Num < ModuleTypeValues.size())
return (Value*)ModuleTypeValues[Num].get();
- // Nope, is it a method level type?
+ // Nope, is it a function level type?
Num -= ModuleTypeValues.size();
- if (Num < MethodTypeValues.size())
- return (Value*)MethodTypeValues[Num].get();
+ if (Num < FunctionTypeValues.size())
+ return (Value*)FunctionTypeValues[Num].get();
return 0;
}
+ if (HasImplicitZeroInitializer && type >= FirstDerivedTyID) {
+ if (Num == 0)
+ return Constant::getNullValue(Ty);
+ --Num;
+ }
+
if (type < ModuleValues.size()) {
- if (Num < ModuleValues[type].size())
- return ModuleValues[type][Num];
- Num -= ModuleValues[type].size();
+ if (Num < ModuleValues[type]->size())
+ return ModuleValues[type]->getOperand(Num);
+ Num -= ModuleValues[type]->size();
}
- if (Values.size() > type && Values[type].size() > Num)
- return Values[type][Num];
+ if (Values.size() > type && Values[type]->size() > Num)
+ return Values[type]->getOperand(Num);
- if (!Create) return failure<Value*>(0); // Do not create a placeholder?
+ if (!Create) return 0; // Do not create a placeholder?
Value *d = 0;
switch (Ty->getPrimitiveID()) {
- case Type::LabelTyID: d = new BBPHolder(Ty, oNum); break;
- case Type::FunctionTyID:
- cerr << "Creating method pholder! : " << type << ":" << oNum << " "
- << Ty->getName() << "\n";
- d = new MethPHolder(Ty, oNum);
- if (insertValue(d, LateResolveModuleValues) ==-1) return failure<Value*>(0);
- return d;
- default: d = new DefPHolder(Ty, oNum); break;
+ case Type::LabelTyID:
+ d = new BBPHolder(Ty, oNum);
+ break;
+ default:
+ d = new ValPHolder(Ty, oNum);
+ break;
}
assert(d != 0 && "How did we not make something?");
- if (insertValue(d, LateResolveValues) == -1) return failure<Value*>(0);
+ if (insertValue(d, LateResolveValues) == -1) return 0;
return d;
}
+/// getConstantValue - Just like getValue, except that it returns a null pointer
+/// only on error. It always returns a constant (meaning that if the value is
+/// defined, but is not a constant, that is an error). If the specified
+/// constant hasn't been parsed yet, a placeholder is defined and used. Later,
+/// after the real value is parsed, the placeholder is eliminated.
+///
+Constant *BytecodeParser::getConstantValue(const Type *Ty, unsigned Slot) {
+ if (Value *V = getValue(Ty, Slot, false))
+ return dyn_cast<Constant>(V); // If we already have the value parsed...
+
+ std::pair<const Type*, unsigned> Key(Ty, Slot);
+ GlobalRefsType::iterator I = GlobalRefs.lower_bound(Key);
+
+ if (I != GlobalRefs.end() && I->first == Key) {
+ BCR_TRACE(5, "Previous forward ref found!\n");
+ return cast<Constant>(I->second);
+ } else {
+ // Create a placeholder for the constant reference and
+ // keep track of the fact that we have a forward ref to recycle it
+ BCR_TRACE(5, "Creating new forward ref to a constant!\n");
+ Constant *C = new ConstPHolder(Ty, Slot);
+
+ // Keep track of the fact that we have a forward ref to recycle it
+ GlobalRefs.insert(I, std::make_pair(Key, C));
+ return C;
+ }
+}
+
+
bool BytecodeParser::postResolveValues(ValueTable &ValTab) {
bool Error = false;
- for (unsigned ty = 0; ty < ValTab.size(); ++ty) {
- ValueList &DL = ValTab[ty];
- unsigned Size;
- while ((Size = DL.size())) {
- unsigned IDNumber = getValueIDNumberFromPlaceHolder(DL[Size-1]);
+ while (!ValTab.empty()) {
+ ValueList &DL = *ValTab.back();
+ ValTab.pop_back();
- Value *D = DL[Size-1];
+ 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
- cerr << "Unresolvable reference found: <"
- << D->getType()->getDescription() << ">:" << IDNumber << "!\n";
+ std::cerr << "Unresolvable reference found: <"
+ << *D->getType() << ">:" << IDNumber <<"!\n";
} else {
// Fixup all of the uses of this placeholder def...
D->replaceAllUsesWith(NewDef);
delete D; // memory, 'cause otherwise we can't remove all uses!
}
}
+ delete &DL;
}
return Error;
}
-bool BytecodeParser::ParseBasicBlock(const uchar *&Buf, const uchar *EndBuf,
+bool BytecodeParser::ParseBasicBlock(const unsigned char *&Buf,
+ const unsigned char *EndBuf,
BasicBlock *&BB) {
BB = new BasicBlock();
while (Buf < EndBuf) {
Instruction *Inst;
- if (ParseInstruction(Buf, EndBuf, Inst)) {
+ if (ParseInstruction(Buf, EndBuf, Inst, /*HACK*/BB)) {
delete BB;
- return failure(true);
+ return true;
}
- if (Inst == 0) { delete BB; return failure(true); }
- if (insertValue(Inst, Values) == -1) { delete BB; return failure(true); }
+ if (Inst == 0) { delete BB; return true; }
+ if (insertValue(Inst, Values) == -1) { delete BB; return true; }
BB->getInstList().push_back(Inst);
return false;
}
-bool BytecodeParser::ParseSymbolTable(const uchar *&Buf, const uchar *EndBuf,
+bool BytecodeParser::ParseSymbolTable(const unsigned char *&Buf,
+ const unsigned char *EndBuf,
SymbolTable *ST) {
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 failure(true);
+ read_vbr(Buf, EndBuf, Typ)) return true;
const Type *Ty = getType(Typ);
- if (Ty == 0) return failure(true);
+ if (Ty == 0) return true;
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 failure(true);
+ if (read_vbr(Buf, EndBuf, slot)) return true;
std::string Name;
if (read(Buf, EndBuf, Name, false)) // Not aligned...
- return failure(true);
+ return true;
- Value *D = getValue(Ty, slot, false); // Find mapping...
- if (D == 0) {
+ Value *V = getValue(Ty, slot, false); // Find mapping...
+ if (V == 0) {
BCR_TRACE(3, "FAILED LOOKUP: Slot #" << slot << "\n");
- return failure(true);
+ return true;
}
- BCR_TRACE(4, "Map: '" << Name << "' to #" << slot << ":" << D;
- if (!isa<Instruction>(D)) cerr << "\n");
+ BCR_TRACE(4, "Map: '" << Name << "' to #" << slot << ":" << *V;
+ if (!isa<Instruction>(V)) std::cerr << "\n");
- D->setName(Name, ST);
+ V->setName(Name, ST);
}
}
- if (Buf > EndBuf) return failure(true);
+ if (Buf > EndBuf) return true;
return false;
}
-// DeclareNewGlobalValue - Patch up forward references to global values in the
-// form of ConstantPointerRef.
-//
-void BytecodeParser::DeclareNewGlobalValue(GlobalValue *GV, unsigned Slot) {
- // Check to see if there is a forward reference to this global variable...
- // if there is, eliminate it and patch the reference to use the new def'n.
- GlobalRefsType::iterator I = GlobalRefs.find(make_pair(GV->getType(), Slot));
-
- if (I != GlobalRefs.end()) {
- GlobalVariable *OldGV = I->second; // Get the placeholder...
- BCR_TRACE(3, "Mutating CPPR Forward Ref!\n");
-
- // Loop over all of the uses of the GlobalValue. The only thing they are
- // allowed to be at this point is ConstantPointerRef's.
- assert(OldGV->use_size() == 1 && "Only one reference should exist!");
- while (!OldGV->use_empty()) {
- User *U = OldGV->use_back(); // Must be a ConstantPointerRef...
- ConstantPointerRef *CPPR = cast<ConstantPointerRef>(U);
- assert(CPPR->getValue() == OldGV && "Something isn't happy");
-
- BCR_TRACE(4, "Mutating Forward Ref!\n");
-
- // Change the const pool reference to point to the real global variable
- // now. This should drop a use from the OldGV.
- CPPR->mutateReference(GV);
- }
-
- // Remove GV from the module...
- GV->getParent()->getGlobalList().remove(OldGV);
- delete OldGV; // Delete the old placeholder
-
- // Remove the map entry for the global now that it has been created...
- GlobalRefs.erase(I);
- }
+void BytecodeParser::ResolveReferencesToValue(Value *NewV, unsigned Slot) {
+ GlobalRefsType::iterator I = GlobalRefs.find(std::make_pair(NewV->getType(),
+ Slot));
+ if (I == GlobalRefs.end()) return; // Never forward referenced?
+
+ BCR_TRACE(3, "Mutating forward refs!\n");
+ Value *VPH = I->second; // Get the placeholder...
+
+ VPH->replaceAllUsesWith(NewV);
+
+ // If this is a global variable being resolved, remove the placeholder from
+ // the module...
+ if (GlobalValue* GVal = dyn_cast<GlobalValue>(NewV))
+ GVal->getParent()->getGlobalList().remove(cast<GlobalVariable>(VPH));
+
+ delete VPH; // Delete the old placeholder
+ GlobalRefs.erase(I); // Remove the map entry for it
}
-bool BytecodeParser::ParseMethod(const uchar *&Buf, const uchar *EndBuf,
- Module *C) {
+
+bool BytecodeParser::ParseFunction(const unsigned char *&Buf,
+ const unsigned char *EndBuf) {
// Clear out the local values table...
- Values.clear();
- if (MethodSignatureList.empty()) {
+ if (FunctionSignatureList.empty()) {
Error = "Function found, but FunctionSignatureList empty!";
- return failure(true); // Unexpected method!
+ return true; // Unexpected function!
}
- const PointerType *PMTy = MethodSignatureList.front().first; // PtrMeth
- const FunctionType *MTy = dyn_cast<FunctionType>(PMTy->getElementType());
- if (MTy == 0) return failure(true); // Not ptr to method!
-
- unsigned isInternal;
- if (read_vbr(Buf, EndBuf, isInternal)) return failure(true);
+ GlobalValue::LinkageTypes Linkage = GlobalValue::ExternalLinkage;
- unsigned MethSlot = MethodSignatureList.front().second;
- MethodSignatureList.pop_front();
- Function *M = new Function(MTy, isInternal != 0);
+ if (!hasInternalMarkerOnly) {
+ unsigned LinkageType;
+ if (read_vbr(Buf, EndBuf, LinkageType)) return true;
+ if (LinkageType & ~0x3) return true;
+ Linkage = (GlobalValue::LinkageTypes)LinkageType;
+ } else {
+ // We used to only support two linkage models: internal and external
+ unsigned isInternal;
+ if (read_vbr(Buf, EndBuf, isInternal)) return true;
+ if (isInternal) Linkage = GlobalValue::InternalLinkage;
+ }
- BCR_TRACE(2, "METHOD TYPE: " << MTy << "\n");
+ Function *F = FunctionSignatureList.back().first;
+ unsigned FunctionSlot = FunctionSignatureList.back().second;
+ FunctionSignatureList.pop_back();
+ F->setLinkage(Linkage);
- const FunctionType::ParamTypes &Params = MTy->getParamTypes();
+ const FunctionType::ParamTypes &Params =F->getFunctionType()->getParamTypes();
+ Function::aiterator AI = F->abegin();
for (FunctionType::ParamTypes::const_iterator It = Params.begin();
- It != Params.end(); ++It) {
- Argument *FA = new Argument(*It);
- if (insertValue(FA, Values) == -1) {
- Error = "Error reading method arguments!\n";
- delete M; return failure(true);
+ It != Params.end(); ++It, ++AI) {
+ if (insertValue(AI, Values) == -1) {
+ Error = "Error reading function arguments!\n";
+ return true;
}
- M->getArgumentList().push_back(FA);
}
while (Buf < EndBuf) {
unsigned Type, Size;
- const uchar *OldBuf = Buf;
+ const unsigned char *OldBuf = Buf;
if (readBlock(Buf, EndBuf, Type, Size)) {
Error = "Error reading Function level block!";
- delete M; return failure(true);
+ return true;
}
switch (Type) {
case BytecodeFormat::ConstantPool:
BCR_TRACE(2, "BLOCK BytecodeFormat::ConstantPool: {\n");
- if (ParseConstantPool(Buf, Buf+Size, Values, MethodTypeValues)) {
- delete M; return failure(true);
- }
+ if (ParseConstantPool(Buf, Buf+Size, Values, FunctionTypeValues))
+ return true;
break;
case BytecodeFormat::BasicBlock: {
BCR_TRACE(2, "BLOCK BytecodeFormat::BasicBlock: {\n");
BasicBlock *BB;
if (ParseBasicBlock(Buf, Buf+Size, BB) ||
- insertValue(BB, Values) == -1) {
- delete M; return failure(true); // Parse error... :(
- }
+ insertValue(BB, Values) == -1)
+ return true; // Parse error... :(
- M->getBasicBlocks().push_back(BB);
+ F->getBasicBlockList().push_back(BB);
break;
}
case BytecodeFormat::SymbolTable:
BCR_TRACE(2, "BLOCK BytecodeFormat::SymbolTable: {\n");
- if (ParseSymbolTable(Buf, Buf+Size, M->getSymbolTableSure())) {
- delete M; return failure(true);
- }
+ if (ParseSymbolTable(Buf, Buf+Size, &F->getSymbolTable()))
+ return true;
break;
default:
BCR_TRACE(2, "BLOCK <unknown>:ignored! {\n");
Buf += Size;
- if (OldBuf > Buf) return failure(true); // Wrap around!
+ if (OldBuf > Buf) return true; // Wrap around!
break;
}
BCR_TRACE(2, "} end block\n");
if (align32(Buf, EndBuf)) {
Error = "Error aligning Function level block!";
- delete M; // Malformed bc file, read past end of block.
- return failure(true);
+ return true; // Malformed bc file, read past end of block.
}
}
- if (postResolveValues(LateResolveValues) ||
- postResolveValues(LateResolveModuleValues)) {
- Error = "Error resolving method values!";
- delete M; return failure(true); // Unresolvable references!
+ if (postResolveValues(LateResolveValues)) {
+ Error = "Error resolving function values!";
+ return true; // Unresolvable references!
}
- Value *MethPHolder = getValue(PMTy, MethSlot, false);
- assert(MethPHolder && "Something is broken no placeholder found!");
- assert(isa<Function>(MethPHolder) && "Not a function?");
-
- unsigned type; // Type slot
- assert(!getTypeSlot(MTy, type) && "How can meth type not exist?");
- getTypeSlot(PMTy, type);
+ ResolveReferencesToValue(F, FunctionSlot);
- C->getFunctionList().push_back(M);
-
- // Replace placeholder with the real method pointer...
- ModuleValues[type][MethSlot] = M;
-
- // Clear out method level types...
- MethodTypeValues.clear();
-
- // If anyone is using the placeholder make them use the real method instead
- MethPHolder->replaceAllUsesWith(M);
-
- // We don't need the placeholder anymore!
- delete MethPHolder;
-
- // If the method is empty, we don't need the method argument entries...
- if (M->isExternal())
- M->getArgumentList().delete_all();
-
- DeclareNewGlobalValue(M, MethSlot);
+ // Clear out function level types...
+ FunctionTypeValues.clear();
+ freeTable(Values);
return false;
}
-bool BytecodeParser::ParseModuleGlobalInfo(const uchar *&Buf, const uchar *End,
- Module *Mod) {
- if (!MethodSignatureList.empty()) {
+bool BytecodeParser::ParseModuleGlobalInfo(const unsigned char *&Buf,
+ const unsigned char *End){
+ if (!FunctionSignatureList.empty()) {
Error = "Two ModuleGlobalInfo packets found!";
- return failure(true); // Two ModuleGlobal blocks?
+ return true; // Two ModuleGlobal blocks?
}
// Read global variables...
unsigned VarType;
- if (read_vbr(Buf, End, VarType)) return failure(true);
+ if (read_vbr(Buf, End, VarType)) return true;
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 || !Ty->isPointerType()) {
- Error = "Global not pointer type! Ty = " + Ty->getDescription();
- return failure(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 PointerType *PTy = cast<const PointerType>(Ty);
- const Type *ElTy = PTy->getElementType();
-
- Constant *Initializer = 0;
- if (VarType & 2) { // Does it have an initalizer?
- // Do not improvise... values must have been stored in the constant pool,
- // which should have been read before now.
- //
- unsigned InitSlot;
- if (read_vbr(Buf, End, InitSlot)) return failure(true);
-
- Value *V = getValue(ElTy, InitSlot, false);
- if (V == 0) return failure(true);
- Initializer = cast<Constant>(V);
+ const Type *Ty = getType(SlotNo);
+ if (!Ty || !isa<PointerType>(Ty)) {
+ Error = "Global not pointer type! Ty = " + Ty->getDescription();
+ return true;
}
+ const Type *ElTy = cast<PointerType>(Ty)->getElementType();
+
// Create the global variable...
- GlobalVariable *GV = new GlobalVariable(ElTy, VarType & 1, VarType & 4,
- Initializer);
+ GlobalVariable *GV = new GlobalVariable(ElTy, VarType & 1, Linkage,
+ 0, "", TheModule);
int DestSlot = insertValue(GV, ModuleValues);
- if (DestSlot == -1) return failure(true);
-
- Mod->getGlobalList().push_back(GV);
-
- DeclareNewGlobalValue(GV, unsigned(DestSlot));
+ if (DestSlot == -1) return true;
+ BCR_TRACE(2, "Global Variable of type: " << *Ty << "\n");
+ ResolveReferencesToValue(GV, (unsigned)DestSlot);
- BCR_TRACE(2, "Global Variable of type: " << PTy->getDescription()
- << " into slot #" << DestSlot << "\n");
-
- if (read_vbr(Buf, End, VarType)) return failure(true);
+ if (VarType & 2) { // Does it have an initalizer?
+ unsigned InitSlot;
+ if (read_vbr(Buf, End, InitSlot)) return true;
+ GlobalInits.push_back(std::make_pair(GV, InitSlot));
+ }
+ if (read_vbr(Buf, End, VarType)) return true;
}
- // Read the method signatures for all of the methods that are coming, and
- // create fillers in the Value tables.
- unsigned MethSignature;
- if (read_vbr(Buf, End, MethSignature)) return failure(true);
- while (MethSignature != Type::VoidTyID) { // List is terminated by Void
- const Type *Ty = getType(MethSignature);
+ // Read the function objects for all of the functions that are coming
+ unsigned FnSignature;
+ if (read_vbr(Buf, End, FnSignature)) return true;
+ 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 failure(true);
+ return true;
}
-
- // We create methods by passing the underlying FunctionType to create...
+
+ // We create functions by passing the underlying FunctionType to create...
Ty = cast<PointerType>(Ty)->getElementType();
- // When the ModuleGlobalInfo section is read, we load the type of each
- // method and the 'ModuleValues' slot that it lands in. We then load a
- // placeholder into its slot to reserve it. When the method is loaded, this
- // placeholder is replaced.
+ // When the ModuleGlobalInfo section is read, we load the type of each
+ // function and the 'ModuleValues' slot that it lands in. We then load a
+ // placeholder into its slot to reserve it. When the function is loaded,
+ // this placeholder is replaced.
// Insert the placeholder...
- Value *Val = new MethPHolder(Ty, 0);
- if (insertValue(Val, ModuleValues) == -1) return failure(true);
-
- // Figure out which entry of its typeslot it went into...
- unsigned TypeSlot;
- if (getTypeSlot(Val->getType(), TypeSlot)) return failure(true);
-
- unsigned SlotNo = ModuleValues[TypeSlot].size()-1;
-
- // Keep track of this information in a linked list that is emptied as
- // methods are loaded...
+ Function *Func = new Function(cast<FunctionType>(Ty),
+ GlobalValue::InternalLinkage, "", TheModule);
+ int DestSlot = insertValue(Func, ModuleValues);
+ if (DestSlot == -1) return true;
+ ResolveReferencesToValue(Func, (unsigned)DestSlot);
+
+ // Keep track of this information in a list that is emptied as functions are
+ // loaded...
//
- MethodSignatureList.push_back(
- make_pair(cast<const PointerType>(Val->getType()), SlotNo));
- if (read_vbr(Buf, End, MethSignature)) return failure(true);
+ FunctionSignatureList.push_back(std::make_pair(Func, DestSlot));
+
+ if (read_vbr(Buf, End, FnSignature)) return true;
BCR_TRACE(2, "Function of type: " << Ty << "\n");
}
- if (align32(Buf, End)) return failure(true);
+ if (align32(Buf, End)) return true;
+
+ // Now that the function signature list is set up, reverse it so that we can
+ // remove elements efficiently from the back of the vector.
+ std::reverse(FunctionSignatureList.begin(), FunctionSignatureList.end());
// This is for future proofing... in the future extra fields may be added that
// we don't understand, so we transparently ignore them.
return false;
}
-bool BytecodeParser::ParseModule(const uchar *Buf, const uchar *EndBuf,
- Module *&C) {
+bool BytecodeParser::ParseVersionInfo(const unsigned char *&Buf,
+ const unsigned char *EndBuf) {
+ unsigned Version;
+ if (read_vbr(Buf, EndBuf, Version)) return true;
+
+ // Unpack version number: low four bits are for flags, top bits = version
+ isBigEndian = Version & 1;
+ hasLongPointers = Version & 2;
+ RevisionNum = Version >> 4;
+
+ // Default values for the current bytecode version
+ HasImplicitZeroInitializer = true;
+ hasInternalMarkerOnly = false;
+ FirstDerivedTyID = 14;
+
+ switch (RevisionNum) {
+ case 0: // Initial revision
+ // Version #0 didn't have any of the flags stored correctly, and in fact as
+ // 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?
+ HasImplicitZeroInitializer = false;
+ isBigEndian = hasLongPointers = true;
+ hasInternalMarkerOnly = true;
+ break;
+ case 1:
+ // Version #1 has two bit fields: isBigEndian and hasLongPointers
+ 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;
+ }
+
+ TheModule->setEndianness(isBigEndian ? Module::BigEndian :
+ Module::LittleEndian);
+ TheModule->setPointerSize(hasLongPointers ? Module::Pointer64 :
+ Module::Pointer32);
+
+ BCR_TRACE(1, "Bytecode Rev = " << (unsigned)RevisionNum << "\n");
+ BCR_TRACE(1, "BigEndian/LongPointers = " << isBigEndian << ","
+ << hasLongPointers << "\n");
+ BCR_TRACE(1, "HasImplicitZeroInit = " << HasImplicitZeroInitializer << "\n");
+ return false;
+}
+bool BytecodeParser::ParseModule(const unsigned char *Buf,
+ const unsigned char *EndBuf) {
unsigned Type, Size;
- if (readBlock(Buf, EndBuf, Type, Size)) return failure(true);
+ if (readBlock(Buf, EndBuf, Type, Size)) return true;
if (Type != BytecodeFormat::Module || Buf+Size != EndBuf) {
Error = "Expected Module packet!";
- return failure(true); // Hrm, not a class?
+ return true; // Hrm, not a class?
}
BCR_TRACE(0, "BLOCK BytecodeFormat::Module: {\n");
- MethodSignatureList.clear(); // Just in case...
+ FunctionSignatureList.clear(); // Just in case...
// Read into instance variables...
- if (read_vbr(Buf, EndBuf, FirstDerivedTyID)) return failure(true);
- if (align32(Buf, EndBuf)) return failure(true);
- BCR_TRACE(1, "FirstDerivedTyID = " << FirstDerivedTyID << "\n");
+ if (ParseVersionInfo(Buf, EndBuf)) return true;
+ if (align32(Buf, EndBuf)) return true;
- TheModule = C = new Module();
while (Buf < EndBuf) {
- const uchar *OldBuf = Buf;
- if (readBlock(Buf, EndBuf, Type, Size)) { delete C; return failure(true); }
+ const unsigned char *OldBuf = Buf;
+ if (readBlock(Buf, EndBuf, Type, Size)) return true;
switch (Type) {
- case BytecodeFormat::ConstantPool:
- BCR_TRACE(1, "BLOCK BytecodeFormat::ConstantPool: {\n");
- if (ParseConstantPool(Buf, Buf+Size, ModuleValues, ModuleTypeValues)) {
- delete C; return failure(true);
- }
+ case BytecodeFormat::GlobalTypePlane:
+ BCR_TRACE(1, "BLOCK BytecodeFormat::GlobalTypePlane: {\n");
+ if (ParseGlobalTypes(Buf, Buf+Size)) return true;
break;
case BytecodeFormat::ModuleGlobalInfo:
BCR_TRACE(1, "BLOCK BytecodeFormat::ModuleGlobalInfo: {\n");
+ if (ParseModuleGlobalInfo(Buf, Buf+Size)) return true;
+ break;
- if (ParseModuleGlobalInfo(Buf, Buf+Size, C)) {
- delete C; return failure(true);
- }
+ case BytecodeFormat::ConstantPool:
+ BCR_TRACE(1, "BLOCK BytecodeFormat::ConstantPool: {\n");
+ if (ParseConstantPool(Buf, Buf+Size, ModuleValues, ModuleTypeValues))
+ return true;
break;
case BytecodeFormat::Function: {
BCR_TRACE(1, "BLOCK BytecodeFormat::Function: {\n");
- if (ParseMethod(Buf, Buf+Size, C)) {
- delete C; return failure(true); // Error parsing method
- }
+ if (ParseFunction(Buf, Buf+Size))
+ return true; // Error parsing function
break;
}
case BytecodeFormat::SymbolTable:
BCR_TRACE(1, "BLOCK BytecodeFormat::SymbolTable: {\n");
- if (ParseSymbolTable(Buf, Buf+Size, C->getSymbolTableSure())) {
- delete C; return failure(true);
- }
+ if (ParseSymbolTable(Buf, Buf+Size, &TheModule->getSymbolTable()))
+ return true;
break;
default:
Error = "Expected Module Block!";
Buf += Size;
- if (OldBuf > Buf) return failure(true); // Wrap around!
+ if (OldBuf > Buf) return true; // Wrap around!
break;
}
BCR_TRACE(1, "} end block\n");
- if (align32(Buf, EndBuf)) { delete C; return failure(true); }
+ if (align32(Buf, EndBuf)) return true;
+ }
+
+ // After the module constant pool has been read, we can safely initialize
+ // global variables...
+ while (!GlobalInits.empty()) {
+ GlobalVariable *GV = GlobalInits.back().first;
+ unsigned Slot = GlobalInits.back().second;
+ GlobalInits.pop_back();
+
+ // Look up the initializer value...
+ if (Value *V = getValue(GV->getType()->getElementType(), Slot, false)) {
+ if (GV->hasInitializer()) return true;
+ GV->setInitializer(cast<Constant>(V));
+ } else
+ return true;
}
- if (!MethodSignatureList.empty()) { // Expected more methods!
+ if (!FunctionSignatureList.empty()) { // Expected more functions!
Error = "Function expected, but bytecode stream at end!";
- return failure(true);
+ return true;
}
BCR_TRACE(0, "} end block\n\n");
return false;
}
-Module *BytecodeParser::ParseBytecode(const uchar *Buf, const uchar *EndBuf) {
- LateResolveValues.clear();
+static inline Module *Error(std::string *ErrorStr, const char *Message) {
+ if (ErrorStr) *ErrorStr = Message;
+ return 0;
+}
+
+Module *BytecodeParser::ParseBytecode(const unsigned char *Buf,
+ const unsigned char *EndBuf,
+ const std::string &ModuleID) {
unsigned Sig;
// Read and check signature...
if (read(Buf, EndBuf, Sig) ||
- Sig != ('l' | ('l' << 8) | ('v' << 16) | 'm' << 24)) {
- Error = "Invalid bytecode signature!";
- return failure<Module*>(0); // Invalid signature!
+ Sig != ('l' | ('l' << 8) | ('v' << 16) | 'm' << 24))
+ return ::Error(&Error, "Invalid bytecode signature!");
+
+ TheModule = new Module(ModuleID);
+ if (ParseModule(Buf, EndBuf)) {
+ freeState(); // Must destroy handles before deleting module!
+ delete TheModule;
+ TheModule = 0;
}
-
- Module *Result;
- if (ParseModule(Buf, EndBuf, Result)) return 0;
- return Result;
+ return TheModule;
}
-Module *ParseBytecodeBuffer(const uchar *Buffer, unsigned Length) {
+Module *ParseBytecodeBuffer(const unsigned char *Buffer, unsigned Length,
+ const std::string &ModuleID, std::string *ErrorStr){
BytecodeParser Parser;
- return Parser.ParseBytecode(Buffer, Buffer+Length);
+ unsigned char *PtrToDelete = 0;
+ if ((intptr_t)Buffer & 3) { // If the buffer is not 4 byte aligned...
+ // Allocate a new buffer to hold the bytecode...
+ PtrToDelete = new unsigned char[Length+4];
+ unsigned Offset = 4-((intptr_t)PtrToDelete & 3); // Make sure it's aligned
+ memcpy(PtrToDelete+Offset, Buffer, Length); // Copy it over
+ Buffer = PtrToDelete+Offset;
+ }
+
+ Module *R = Parser.ParseBytecode(Buffer, Buffer+Length, ModuleID);
+ if (ErrorStr) *ErrorStr = Parser.getError();
+
+ delete [] PtrToDelete; // Delete alignment buffer if neccesary
+ 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) {
- struct stat StatBuf;
Module *Result = 0;
if (Filename != std::string("-")) { // Read from a file...
- int FD = open(Filename.c_str(), O_RDONLY);
- if (FD == -1) {
- if (ErrorStr) *ErrorStr = "Error opening file!";
- return failure<Module*>(0);
- }
+ FDHandle FD = open(Filename.c_str(), O_RDONLY);
+ if (FD == -1)
+ return Error(ErrorStr, "Error opening file!");
- if (fstat(FD, &StatBuf) == -1) { close(FD); return failure<Module*>(0); }
+ // 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;
- if (Length == 0) {
- if (ErrorStr) *ErrorStr = "Error stat'ing file!";
- close(FD); return failure<Module*>(0);
- }
- uchar *Buffer = (uchar*)mmap(0, Length, PROT_READ,
- MAP_PRIVATE, FD, 0);
- if (Buffer == (uchar*)-1) {
- if (ErrorStr) *ErrorStr = "Error mmapping file!";
- close(FD); return failure<Module*>(0);
- }
+ 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!");
- BytecodeParser Parser;
- Result = Parser.ParseBytecode(Buffer, Buffer+Length);
+ // Parse the bytecode we mmapped in
+ Result = ParseBytecodeBuffer(Buffer, Length, Filename, ErrorStr);
+ // Unmmap the bytecode...
munmap((char*)Buffer, Length);
- close(FD);
- if (ErrorStr) *ErrorStr = Parser.getError();
} else { // Read from stdin
- size_t FileSize = 0;
int BlockSize;
- uchar Buffer[4096], *FileData = 0;
- while ((BlockSize = read(0, Buffer, 4))) {
- if (BlockSize == -1) { free(FileData); return failure<Module*>(0); }
+ unsigned char Buffer[4096*4];
+ std::vector<unsigned char> FileData;
- FileData = (uchar*)realloc(FileData, FileSize+BlockSize);
- memcpy(FileData+FileSize, Buffer, BlockSize);
- FileSize += BlockSize;
- }
+ // 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!");
- if (FileSize == 0) {
- if (ErrorStr) *ErrorStr = "Standard Input empty!";
- free(FileData); return failure<Module*>(0);
+ FileData.insert(FileData.end(), Buffer, Buffer+BlockSize);
}
-#define ALIGN_PTRS 1
+ if (FileData.empty())
+ return Error(ErrorStr, "Standard Input empty!");
+
+#define ALIGN_PTRS 0
#if ALIGN_PTRS
- uchar *Buf = (uchar*)mmap(0, FileSize, PROT_READ|PROT_WRITE,
- MAP_PRIVATE|MAP_ANONYMOUS, -1, 0);
- assert((Buf != (uchar*)-1) && "mmap returned error!");
- memcpy(Buf, FileData, FileSize);
- free(FileData);
+ unsigned char *Buf =
+ (unsigned char*)mmap(0, FileData.size(), PROT_READ|PROT_WRITE,
+ MAP_PRIVATE|MAP_ANONYMOUS, -1, 0);
+ assert((Buf != (unsigned char*)-1) && "mmap returned error!");
+ memcpy(Buf, &FileData[0], FileData.size());
#else
- uchar *Buf = FileData;
+ unsigned char *Buf = &FileData[0];
#endif
- BytecodeParser Parser;
- Result = Parser.ParseBytecode(Buf, Buf+FileSize);
+ Result = ParseBytecodeBuffer(Buf, FileData.size(), "<stdin>", ErrorStr);
#if ALIGN_PTRS
- munmap((char*)Buf, FileSize); // Free mmap'd data area
-#else
- free(FileData); // Free realloc'd block of memory
+ munmap((char*)Buf, FileData.size()); // Free mmap'd data area
#endif
-
- if (ErrorStr) *ErrorStr = Parser.getError();
}
return Result;