bool runOnSCC(CallGraphSCC &SCC) override;
static char ID; // Pass identification, replacement for typeid
explicit ArgPromotion(unsigned maxElements = 3)
- : CallGraphSCCPass(ID), DL(nullptr), maxElements(maxElements) {
+ : CallGraphSCCPass(ID), maxElements(maxElements) {
initializeArgPromotionPass(*PassRegistry::getPassRegistry());
}
/// A vector used to hold the indices of a single GEP instruction
typedef std::vector<uint64_t> IndicesVector;
- const DataLayout *DL;
private:
- bool isDenselyPacked(Type *type);
+ bool isDenselyPacked(Type *type, const DataLayout &DL);
bool canPaddingBeAccessed(Argument *Arg);
CallGraphNode *PromoteArguments(CallGraphNode *CGN);
bool isSafeToPromoteArgument(Argument *Arg, bool isByVal) const;
}
/// \brief Checks if a type could have padding bytes.
-bool ArgPromotion::isDenselyPacked(Type *type) {
+bool ArgPromotion::isDenselyPacked(Type *type, const DataLayout &DL) {
// There is no size information, so be conservative.
if (!type->isSized())
// If the alloc size is not equal to the storage size, then there are padding
// bytes. For x86_fp80 on x86-64, size: 80 alloc size: 128.
- if (!DL || DL->getTypeSizeInBits(type) != DL->getTypeAllocSizeInBits(type))
+ if (DL.getTypeSizeInBits(type) != DL.getTypeAllocSizeInBits(type))
return false;
if (!isa<CompositeType>(type))
// For homogenous sequential types, check for padding within members.
if (SequentialType *seqTy = dyn_cast<SequentialType>(type))
- return isa<PointerType>(seqTy) || isDenselyPacked(seqTy->getElementType());
+ return isa<PointerType>(seqTy) ||
+ isDenselyPacked(seqTy->getElementType(), DL);
// Check for padding within and between elements of a struct.
StructType *StructTy = cast<StructType>(type);
- const StructLayout *Layout = DL->getStructLayout(StructTy);
+ const StructLayout *Layout = DL.getStructLayout(StructTy);
uint64_t StartPos = 0;
for (unsigned i = 0, E = StructTy->getNumElements(); i < E; ++i) {
Type *ElTy = StructTy->getElementType(i);
- if (!isDenselyPacked(ElTy))
+ if (!isDenselyPacked(ElTy, DL))
return false;
if (StartPos != Layout->getElementOffsetInBits(i))
return false;
- StartPos += DL->getTypeAllocSizeInBits(ElTy);
+ StartPos += DL.getTypeAllocSizeInBits(ElTy);
}
return true;
// Make sure that it is local to this module.
if (!F || !F->hasLocalLinkage()) return nullptr;
- DL = &F->getParent()->getDataLayout();
-
// First check: see if there are any pointer arguments! If not, quick exit.
SmallVector<Argument*, 16> PointerArgs;
for (Function::arg_iterator I = F->arg_begin(), E = F->arg_end(); I != E; ++I)
// IR, while in the callee the classification is determined dynamically based
// on the number of registers consumed so far.
if (F->isVarArg()) return nullptr;
+ const DataLayout &DL = F->getParent()->getDataLayout();
// Check to see which arguments are promotable. If an argument is promotable,
// add it to ArgsToPromote.
// packed or if we can prove the padding bytes are never accessed. This does
// not apply to inalloca.
bool isSafeToPromote =
- PtrArg->hasByValAttr() &&
- (isDenselyPacked(AgTy) || !canPaddingBeAccessed(PtrArg));
+ PtrArg->hasByValAttr() &&
+ (isDenselyPacked(AgTy, DL) || !canPaddingBeAccessed(PtrArg));
if (isSafeToPromote) {
if (StructType *STy = dyn_cast<StructType>(AgTy)) {
if (maxElements > 0 && STy->getNumElements() > maxElements) {
/// AllCallersPassInValidPointerForArgument - Return true if we can prove that
/// all callees pass in a valid pointer for the specified function argument.
-static bool AllCallersPassInValidPointerForArgument(Argument *Arg,
- const DataLayout *DL) {
+static bool AllCallersPassInValidPointerForArgument(Argument *Arg) {
Function *Callee = Arg->getParent();
+ const DataLayout &DL = Callee->getParent()->getDataLayout();
unsigned ArgNo = Arg->getArgNo();
GEPIndicesSet ToPromote;
// If the pointer is always valid, any load with first index 0 is valid.
- if (isByValOrInAlloca || AllCallersPassInValidPointerForArgument(Arg, DL))
+ if (isByValOrInAlloca || AllCallersPassInValidPointerForArgument(Arg))
SafeToUnconditionallyLoad.insert(IndicesVector(1, 0));
// First, iterate the entry block and mark loads of (geps of) arguments as