#include "llvm/ADT/DepthFirstIterator.h"
#include "llvm/ADT/Hashing.h"
#include "llvm/ADT/MapVector.h"
+#include "llvm/ADT/PostOrderIterator.h"
#include "llvm/ADT/SetVector.h"
#include "llvm/ADT/SmallPtrSet.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/Analysis/AliasAnalysis.h"
+#include "llvm/Analysis/AssumptionCache.h"
#include "llvm/Analysis/CFG.h"
#include "llvm/Analysis/ConstantFolding.h"
+#include "llvm/Analysis/GlobalsModRef.h"
#include "llvm/Analysis/InstructionSimplify.h"
#include "llvm/Analysis/Loads.h"
#include "llvm/Analysis/MemoryBuiltins.h"
#include "llvm/Analysis/MemoryDependenceAnalysis.h"
#include "llvm/Analysis/PHITransAddr.h"
+#include "llvm/Analysis/TargetLibraryInfo.h"
#include "llvm/Analysis/ValueTracking.h"
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/Dominators.h"
#include "llvm/Support/Allocator.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Debug.h"
-#include "llvm/Target/TargetLibraryInfo.h"
+#include "llvm/Support/raw_ostream.h"
#include "llvm/Transforms/Utils/BasicBlockUtils.h"
+#include "llvm/Transforms/Utils/Local.h"
#include "llvm/Transforms/Utils/SSAUpdater.h"
#include <vector>
using namespace llvm;
return e;
}
-/// lookup - Returns the value number of the specified value. Fails if
+/// Returns the value number of the specified value. Fails if
/// the value has not yet been numbered.
uint32_t ValueTable::lookup(Value *V) const {
DenseMap<Value*, uint32_t>::const_iterator VI = valueNumbering.find(V);
return VI->second;
}
-/// lookup_or_add_cmp - Returns the value number of the given comparison,
+/// Returns the value number of the given comparison,
/// assigning it a new number if it did not have one before. Useful when
/// we deduced the result of a comparison, but don't immediately have an
/// instruction realizing that comparison to hand.
return e;
}
-/// clear - Remove all entries from the ValueTable.
+/// Remove all entries from the ValueTable.
void ValueTable::clear() {
valueNumbering.clear();
expressionNumbering.clear();
nextValueNumber = 1;
}
-/// erase - Remove a value from the value numbering.
+/// Remove a value from the value numbering.
void ValueTable::erase(Value *V) {
valueNumbering.erase(V);
}
return cast<MemIntrinsic>(Val.getPointer());
}
- /// MaterializeAdjustedValue - Emit code into this block to adjust the value
- /// defined here to the specified type. This handles various coercion cases.
- Value *MaterializeAdjustedValue(Type *LoadTy, GVN &gvn) const;
+ /// Emit code into this block to adjust the value defined here to the
+ /// specified type. This handles various coercion cases.
+ Value *MaterializeAdjustedValue(LoadInst *LI, GVN &gvn) const;
};
class GVN : public FunctionPass {
bool NoLoads;
MemoryDependenceAnalysis *MD;
DominatorTree *DT;
- const DataLayout *DL;
const TargetLibraryInfo *TLI;
+ AssumptionCache *AC;
SetVector<BasicBlock *> DeadBlocks;
ValueTable VN;
- /// LeaderTable - A mapping from value numbers to lists of Value*'s that
+ /// A mapping from value numbers to lists of Value*'s that
/// have that value number. Use findLeader to query it.
struct LeaderTableEntry {
Value *Val;
DenseMap<uint32_t, LeaderTableEntry> LeaderTable;
BumpPtrAllocator TableAllocator;
+ // Block-local map of equivalent values to their leader, does not
+ // propagate to any successors. Entries added mid-block are applied
+ // to the remaining instructions in the block.
+ SmallMapVector<llvm::Value *, llvm::Constant *, 4> ReplaceWithConstMap;
SmallVector<Instruction*, 8> InstrsToErase;
typedef SmallVector<NonLocalDepResult, 64> LoadDepVect;
bool runOnFunction(Function &F) override;
- /// markInstructionForDeletion - This removes the specified instruction from
+ /// This removes the specified instruction from
/// our various maps and marks it for deletion.
void markInstructionForDeletion(Instruction *I) {
VN.erase(I);
InstrsToErase.push_back(I);
}
- const DataLayout *getDataLayout() const { return DL; }
DominatorTree &getDominatorTree() const { return *DT; }
AliasAnalysis *getAliasAnalysis() const { return VN.getAliasAnalysis(); }
MemoryDependenceAnalysis &getMemDep() const { return *MD; }
private:
- /// addToLeaderTable - Push a new Value to the LeaderTable onto the list for
- /// its value number.
+ /// Push a new Value to the LeaderTable onto the list for its value number.
void addToLeaderTable(uint32_t N, Value *V, const BasicBlock *BB) {
LeaderTableEntry &Curr = LeaderTable[N];
if (!Curr.Val) {
Curr.Next = Node;
}
- /// removeFromLeaderTable - Scan the list of values corresponding to a given
+ /// Scan the list of values corresponding to a given
/// value number, and remove the given instruction if encountered.
void removeFromLeaderTable(uint32_t N, Instruction *I, BasicBlock *BB) {
LeaderTableEntry* Prev = nullptr;
LeaderTableEntry* Curr = &LeaderTable[N];
- while (Curr->Val != I || Curr->BB != BB) {
+ while (Curr && (Curr->Val != I || Curr->BB != BB)) {
Prev = Curr;
Curr = Curr->Next;
}
+ if (!Curr)
+ return;
+
if (Prev) {
Prev->Next = Curr->Next;
} else {
// This transformation requires dominator postdominator info
void getAnalysisUsage(AnalysisUsage &AU) const override {
+ AU.addRequired<AssumptionCacheTracker>();
AU.addRequired<DominatorTreeWrapperPass>();
- AU.addRequired<TargetLibraryInfo>();
+ AU.addRequired<TargetLibraryInfoWrapperPass>();
if (!NoLoads)
AU.addRequired<MemoryDependenceAnalysis>();
- AU.addRequired<AliasAnalysis>();
+ AU.addRequired<AAResultsWrapperPass>();
AU.addPreserved<DominatorTreeWrapperPass>();
- AU.addPreserved<AliasAnalysis>();
+ AU.addPreserved<GlobalsAAWrapperPass>();
}
- // Helper fuctions of redundant load elimination
+ // Helper functions of redundant load elimination
bool processLoad(LoadInst *L);
bool processNonLocalLoad(LoadInst *L);
+ bool processAssumeIntrinsic(IntrinsicInst *II);
void AnalyzeLoadAvailability(LoadInst *LI, LoadDepVect &Deps,
AvailValInBlkVect &ValuesPerBlock,
UnavailBlkVect &UnavailableBlocks);
void dump(DenseMap<uint32_t, Value*> &d);
bool iterateOnFunction(Function &F);
bool performPRE(Function &F);
+ bool performScalarPRE(Instruction *I);
+ bool performScalarPREInsertion(Instruction *Instr, BasicBlock *Pred,
+ unsigned int ValNo);
Value *findLeader(const BasicBlock *BB, uint32_t num);
void cleanupGlobalSets();
void verifyRemoved(const Instruction *I) const;
bool splitCriticalEdges();
BasicBlock *splitCriticalEdges(BasicBlock *Pred, BasicBlock *Succ);
- unsigned replaceAllDominatedUsesWith(Value *From, Value *To,
- const BasicBlockEdge &Root);
- bool propagateEquality(Value *LHS, Value *RHS, const BasicBlockEdge &Root);
+ bool replaceOperandsWithConsts(Instruction *I) const;
+ bool propagateEquality(Value *LHS, Value *RHS, const BasicBlockEdge &Root,
+ bool DominatesByEdge);
bool processFoldableCondBr(BranchInst *BI);
void addDeadBlock(BasicBlock *BB);
void assignValNumForDeadCode();
char GVN::ID = 0;
}
-// createGVNPass - The public interface to this file...
+// The public interface to this file...
FunctionPass *llvm::createGVNPass(bool NoLoads) {
return new GVN(NoLoads);
}
INITIALIZE_PASS_BEGIN(GVN, "gvn", "Global Value Numbering", false, false)
+INITIALIZE_PASS_DEPENDENCY(AssumptionCacheTracker)
INITIALIZE_PASS_DEPENDENCY(MemoryDependenceAnalysis)
INITIALIZE_PASS_DEPENDENCY(DominatorTreeWrapperPass)
-INITIALIZE_PASS_DEPENDENCY(TargetLibraryInfo)
-INITIALIZE_AG_DEPENDENCY(AliasAnalysis)
+INITIALIZE_PASS_DEPENDENCY(TargetLibraryInfoWrapperPass)
+INITIALIZE_PASS_DEPENDENCY(AAResultsWrapperPass)
+INITIALIZE_PASS_DEPENDENCY(GlobalsAAWrapperPass)
INITIALIZE_PASS_END(GVN, "gvn", "Global Value Numbering", false, false)
#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
}
#endif
-/// IsValueFullyAvailableInBlock - Return true if we can prove that the value
+/// Return true if we can prove that the value
/// we're analyzing is fully available in the specified block. As we go, keep
/// track of which blocks we know are fully alive in FullyAvailableBlocks. This
/// map is actually a tri-state map with the following values:
return true;
-// SpeculationFailure - If we get here, we found out that this is not, after
+// If we get here, we found out that this is not, after
// all, a fully-available block. We have a problem if we speculated on this and
// used the speculation to mark other blocks as available.
SpeculationFailure:
}
-/// CanCoerceMustAliasedValueToLoad - Return true if
-/// CoerceAvailableValueToLoadType will succeed.
+/// Return true if CoerceAvailableValueToLoadType will succeed.
static bool CanCoerceMustAliasedValueToLoad(Value *StoredVal,
Type *LoadTy,
const DataLayout &DL) {
return true;
}
-/// CoerceAvailableValueToLoadType - If we saw a store of a value to memory, and
+/// If we saw a store of a value to memory, and
/// then a load from a must-aliased pointer of a different type, try to coerce
-/// the stored value. LoadedTy is the type of the load we want to replace and
-/// InsertPt is the place to insert new instructions.
+/// the stored value. LoadedTy is the type of the load we want to replace.
+/// IRB is IRBuilder used to insert new instructions.
///
/// If we can't do it, return null.
-static Value *CoerceAvailableValueToLoadType(Value *StoredVal,
- Type *LoadedTy,
- Instruction *InsertPt,
+static Value *CoerceAvailableValueToLoadType(Value *StoredVal, Type *LoadedTy,
+ IRBuilder<> &IRB,
const DataLayout &DL) {
if (!CanCoerceMustAliasedValueToLoad(StoredVal, LoadedTy, DL))
return nullptr;
// Pointer to Pointer -> use bitcast.
if (StoredValTy->getScalarType()->isPointerTy() &&
LoadedTy->getScalarType()->isPointerTy())
- return new BitCastInst(StoredVal, LoadedTy, "", InsertPt);
+ return IRB.CreateBitCast(StoredVal, LoadedTy);
// Convert source pointers to integers, which can be bitcast.
if (StoredValTy->getScalarType()->isPointerTy()) {
StoredValTy = DL.getIntPtrType(StoredValTy);
- StoredVal = new PtrToIntInst(StoredVal, StoredValTy, "", InsertPt);
+ StoredVal = IRB.CreatePtrToInt(StoredVal, StoredValTy);
}
Type *TypeToCastTo = LoadedTy;
TypeToCastTo = DL.getIntPtrType(TypeToCastTo);
if (StoredValTy != TypeToCastTo)
- StoredVal = new BitCastInst(StoredVal, TypeToCastTo, "", InsertPt);
+ StoredVal = IRB.CreateBitCast(StoredVal, TypeToCastTo);
// Cast to pointer if the load needs a pointer type.
if (LoadedTy->getScalarType()->isPointerTy())
- StoredVal = new IntToPtrInst(StoredVal, LoadedTy, "", InsertPt);
+ StoredVal = IRB.CreateIntToPtr(StoredVal, LoadedTy);
return StoredVal;
}
// Convert source pointers to integers, which can be manipulated.
if (StoredValTy->getScalarType()->isPointerTy()) {
StoredValTy = DL.getIntPtrType(StoredValTy);
- StoredVal = new PtrToIntInst(StoredVal, StoredValTy, "", InsertPt);
+ StoredVal = IRB.CreatePtrToInt(StoredVal, StoredValTy);
}
// Convert vectors and fp to integer, which can be manipulated.
if (!StoredValTy->isIntegerTy()) {
StoredValTy = IntegerType::get(StoredValTy->getContext(), StoreSize);
- StoredVal = new BitCastInst(StoredVal, StoredValTy, "", InsertPt);
+ StoredVal = IRB.CreateBitCast(StoredVal, StoredValTy);
}
// If this is a big-endian system, we need to shift the value down to the low
// bits so that a truncate will work.
if (DL.isBigEndian()) {
- Constant *Val = ConstantInt::get(StoredVal->getType(), StoreSize-LoadSize);
- StoredVal = BinaryOperator::CreateLShr(StoredVal, Val, "tmp", InsertPt);
+ StoredVal = IRB.CreateLShr(StoredVal, StoreSize - LoadSize, "tmp");
}
// Truncate the integer to the right size now.
Type *NewIntTy = IntegerType::get(StoredValTy->getContext(), LoadSize);
- StoredVal = new TruncInst(StoredVal, NewIntTy, "trunc", InsertPt);
+ StoredVal = IRB.CreateTrunc(StoredVal, NewIntTy, "trunc");
if (LoadedTy == NewIntTy)
return StoredVal;
// If the result is a pointer, inttoptr.
if (LoadedTy->getScalarType()->isPointerTy())
- return new IntToPtrInst(StoredVal, LoadedTy, "inttoptr", InsertPt);
+ return IRB.CreateIntToPtr(StoredVal, LoadedTy, "inttoptr");
// Otherwise, bitcast.
- return new BitCastInst(StoredVal, LoadedTy, "bitcast", InsertPt);
+ return IRB.CreateBitCast(StoredVal, LoadedTy, "bitcast");
}
-/// AnalyzeLoadFromClobberingWrite - This function is called when we have a
+/// This function is called when we have a
/// memdep query of a load that ends up being a clobbering memory write (store,
/// memset, memcpy, memmove). This means that the write *may* provide bits used
/// by the load but we can't be sure because the pointers don't mustalias.
return -1;
int64_t StoreOffset = 0, LoadOffset = 0;
- Value *StoreBase = GetPointerBaseWithConstantOffset(WritePtr,StoreOffset,&DL);
- Value *LoadBase = GetPointerBaseWithConstantOffset(LoadPtr, LoadOffset, &DL);
+ Value *StoreBase =
+ GetPointerBaseWithConstantOffset(WritePtr, StoreOffset, DL);
+ Value *LoadBase = GetPointerBaseWithConstantOffset(LoadPtr, LoadOffset, DL);
if (StoreBase != LoadBase)
return -1;
return LoadOffset-StoreOffset;
}
-/// AnalyzeLoadFromClobberingStore - This function is called when we have a
+/// This function is called when we have a
/// memdep query of a load that ends up being a clobbering store.
static int AnalyzeLoadFromClobberingStore(Type *LoadTy, Value *LoadPtr,
- StoreInst *DepSI,
- const DataLayout &DL) {
+ StoreInst *DepSI) {
// Cannot handle reading from store of first-class aggregate yet.
if (DepSI->getValueOperand()->getType()->isStructTy() ||
DepSI->getValueOperand()->getType()->isArrayTy())
return -1;
+ const DataLayout &DL = DepSI->getModule()->getDataLayout();
Value *StorePtr = DepSI->getPointerOperand();
uint64_t StoreSize =DL.getTypeSizeInBits(DepSI->getValueOperand()->getType());
return AnalyzeLoadFromClobberingWrite(LoadTy, LoadPtr,
StorePtr, StoreSize, DL);
}
-/// AnalyzeLoadFromClobberingLoad - This function is called when we have a
+/// This function is called when we have a
/// memdep query of a load that ends up being clobbered by another load. See if
/// the other load can feed into the second load.
static int AnalyzeLoadFromClobberingLoad(Type *LoadTy, Value *LoadPtr,
// then we should widen it!
int64_t LoadOffs = 0;
const Value *LoadBase =
- GetPointerBaseWithConstantOffset(LoadPtr, LoadOffs, &DL);
+ GetPointerBaseWithConstantOffset(LoadPtr, LoadOffs, DL);
unsigned LoadSize = DL.getTypeStoreSize(LoadTy);
- unsigned Size = MemoryDependenceAnalysis::
- getLoadLoadClobberFullWidthSize(LoadBase, LoadOffs, LoadSize, DepLI, DL);
+ unsigned Size = MemoryDependenceAnalysis::getLoadLoadClobberFullWidthSize(
+ LoadBase, LoadOffs, LoadSize, DepLI);
if (Size == 0) return -1;
return AnalyzeLoadFromClobberingWrite(LoadTy, LoadPtr, DepPtr, Size*8, DL);
Constant *Src = dyn_cast<Constant>(MTI->getSource());
if (!Src) return -1;
- GlobalVariable *GV = dyn_cast<GlobalVariable>(GetUnderlyingObject(Src, &DL));
+ GlobalVariable *GV = dyn_cast<GlobalVariable>(GetUnderlyingObject(Src, DL));
if (!GV || !GV->isConstant()) return -1;
// See if the access is within the bounds of the transfer.
Type::getInt8PtrTy(Src->getContext(), AS));
Constant *OffsetCst =
ConstantInt::get(Type::getInt64Ty(Src->getContext()), (unsigned)Offset);
- Src = ConstantExpr::getGetElementPtr(Src, OffsetCst);
+ Src = ConstantExpr::getGetElementPtr(Type::getInt8Ty(Src->getContext()), Src,
+ OffsetCst);
Src = ConstantExpr::getBitCast(Src, PointerType::get(LoadTy, AS));
- if (ConstantFoldLoadFromConstPtr(Src, &DL))
+ if (ConstantFoldLoadFromConstPtr(Src, DL))
return Offset;
return -1;
}
-/// GetStoreValueForLoad - This function is called when we have a
+/// This function is called when we have a
/// memdep query of a load that ends up being a clobbering store. This means
/// that the store provides bits used by the load but we the pointers don't
/// mustalias. Check this case to see if there is anything more we can do
uint64_t StoreSize = (DL.getTypeSizeInBits(SrcVal->getType()) + 7) / 8;
uint64_t LoadSize = (DL.getTypeSizeInBits(LoadTy) + 7) / 8;
- IRBuilder<> Builder(InsertPt->getParent(), InsertPt);
+ IRBuilder<> Builder(InsertPt);
// Compute which bits of the stored value are being used by the load. Convert
// to an integer type to start with.
if (LoadSize != StoreSize)
SrcVal = Builder.CreateTrunc(SrcVal, IntegerType::get(Ctx, LoadSize*8));
- return CoerceAvailableValueToLoadType(SrcVal, LoadTy, InsertPt, DL);
+ return CoerceAvailableValueToLoadType(SrcVal, LoadTy, Builder, DL);
}
-/// GetLoadValueForLoad - This function is called when we have a
+/// This function is called when we have a
/// memdep query of a load that ends up being a clobbering load. This means
/// that the load *may* provide bits used by the load but we can't be sure
/// because the pointers don't mustalias. Check this case to see if there is
static Value *GetLoadValueForLoad(LoadInst *SrcVal, unsigned Offset,
Type *LoadTy, Instruction *InsertPt,
GVN &gvn) {
- const DataLayout &DL = *gvn.getDataLayout();
+ const DataLayout &DL = SrcVal->getModule()->getDataLayout();
// If Offset+LoadTy exceeds the size of SrcVal, then we must be wanting to
// widen SrcVal out to a larger load.
unsigned SrcValSize = DL.getTypeStoreSize(SrcVal->getType());
}
-/// GetMemInstValueForLoad - This function is called when we have a
+/// This function is called when we have a
/// memdep query of a load that ends up being a clobbering mem intrinsic.
static Value *GetMemInstValueForLoad(MemIntrinsic *SrcInst, unsigned Offset,
Type *LoadTy, Instruction *InsertPt,
LLVMContext &Ctx = LoadTy->getContext();
uint64_t LoadSize = DL.getTypeSizeInBits(LoadTy)/8;
- IRBuilder<> Builder(InsertPt->getParent(), InsertPt);
+ IRBuilder<> Builder(InsertPt);
// We know that this method is only called when the mem transfer fully
// provides the bits for the load.
++NumBytesSet;
}
- return CoerceAvailableValueToLoadType(Val, LoadTy, InsertPt, DL);
+ return CoerceAvailableValueToLoadType(Val, LoadTy, Builder, DL);
}
// Otherwise, this is a memcpy/memmove from a constant global.
Type::getInt8PtrTy(Src->getContext(), AS));
Constant *OffsetCst =
ConstantInt::get(Type::getInt64Ty(Src->getContext()), (unsigned)Offset);
- Src = ConstantExpr::getGetElementPtr(Src, OffsetCst);
+ Src = ConstantExpr::getGetElementPtr(Type::getInt8Ty(Src->getContext()), Src,
+ OffsetCst);
Src = ConstantExpr::getBitCast(Src, PointerType::get(LoadTy, AS));
- return ConstantFoldLoadFromConstPtr(Src, &DL);
+ return ConstantFoldLoadFromConstPtr(Src, DL);
}
-/// ConstructSSAForLoadSet - Given a set of loads specified by ValuesPerBlock,
+/// Given a set of loads specified by ValuesPerBlock,
/// construct SSA form, allowing us to eliminate LI. This returns the value
/// that should be used at LI's definition site.
static Value *ConstructSSAForLoadSet(LoadInst *LI,
gvn.getDominatorTree().properlyDominates(ValuesPerBlock[0].BB,
LI->getParent())) {
assert(!ValuesPerBlock[0].isUndefValue() && "Dead BB dominate this block");
- return ValuesPerBlock[0].MaterializeAdjustedValue(LI->getType(), gvn);
+ return ValuesPerBlock[0].MaterializeAdjustedValue(LI, gvn);
}
// Otherwise, we have to construct SSA form.
SSAUpdater SSAUpdate(&NewPHIs);
SSAUpdate.Initialize(LI->getType(), LI->getName());
- Type *LoadTy = LI->getType();
-
for (unsigned i = 0, e = ValuesPerBlock.size(); i != e; ++i) {
const AvailableValueInBlock &AV = ValuesPerBlock[i];
BasicBlock *BB = AV.BB;
if (SSAUpdate.HasValueForBlock(BB))
continue;
- SSAUpdate.AddAvailableValue(BB, AV.MaterializeAdjustedValue(LoadTy, gvn));
+ SSAUpdate.AddAvailableValue(BB, AV.MaterializeAdjustedValue(LI, gvn));
}
// Perform PHI construction.
- Value *V = SSAUpdate.GetValueInMiddleOfBlock(LI->getParent());
-
- // If new PHI nodes were created, notify alias analysis.
- if (V->getType()->getScalarType()->isPointerTy()) {
- AliasAnalysis *AA = gvn.getAliasAnalysis();
-
- for (unsigned i = 0, e = NewPHIs.size(); i != e; ++i)
- AA->copyValue(LI, NewPHIs[i]);
-
- // Now that we've copied information to the new PHIs, scan through
- // them again and inform alias analysis that we've added potentially
- // escaping uses to any values that are operands to these PHIs.
- for (unsigned i = 0, e = NewPHIs.size(); i != e; ++i) {
- PHINode *P = NewPHIs[i];
- for (unsigned ii = 0, ee = P->getNumIncomingValues(); ii != ee; ++ii) {
- unsigned jj = PHINode::getOperandNumForIncomingValue(ii);
- AA->addEscapingUse(P->getOperandUse(jj));
- }
- }
- }
-
- return V;
+ return SSAUpdate.GetValueInMiddleOfBlock(LI->getParent());
}
-Value *AvailableValueInBlock::MaterializeAdjustedValue(Type *LoadTy, GVN &gvn) const {
+Value *AvailableValueInBlock::MaterializeAdjustedValue(LoadInst *LI,
+ GVN &gvn) const {
Value *Res;
+ Type *LoadTy = LI->getType();
+ const DataLayout &DL = LI->getModule()->getDataLayout();
if (isSimpleValue()) {
Res = getSimpleValue();
if (Res->getType() != LoadTy) {
- const DataLayout *DL = gvn.getDataLayout();
- assert(DL && "Need target data to handle type mismatch case");
- Res = GetStoreValueForLoad(Res, Offset, LoadTy, BB->getTerminator(),
- *DL);
-
+ Res = GetStoreValueForLoad(Res, Offset, LoadTy, BB->getTerminator(), DL);
+
DEBUG(dbgs() << "GVN COERCED NONLOCAL VAL:\nOffset: " << Offset << " "
<< *getSimpleValue() << '\n'
<< *Res << '\n' << "\n\n\n");
<< *Res << '\n' << "\n\n\n");
}
} else if (isMemIntrinValue()) {
- const DataLayout *DL = gvn.getDataLayout();
- assert(DL && "Need target data to handle type mismatch case");
- Res = GetMemInstValueForLoad(getMemIntrinValue(), Offset,
- LoadTy, BB->getTerminator(), *DL);
+ Res = GetMemInstValueForLoad(getMemIntrinValue(), Offset, LoadTy,
+ BB->getTerminator(), DL);
DEBUG(dbgs() << "GVN COERCED NONLOCAL MEM INTRIN:\nOffset: " << Offset
<< " " << *getMemIntrinValue() << '\n'
<< *Res << '\n' << "\n\n\n");
// dependencies that produce an unknown value for the load (such as a call
// that could potentially clobber the load).
unsigned NumDeps = Deps.size();
+ const DataLayout &DL = LI->getModule()->getDataLayout();
for (unsigned i = 0, e = NumDeps; i != e; ++i) {
BasicBlock *DepBB = Deps[i].getBB();
MemDepResult DepInfo = Deps[i].getResult();
// read by the load, we can extract the bits we need for the load from the
// stored value.
if (StoreInst *DepSI = dyn_cast<StoreInst>(DepInfo.getInst())) {
- if (DL && Address) {
- int Offset = AnalyzeLoadFromClobberingStore(LI->getType(), Address,
- DepSI, *DL);
+ if (Address) {
+ int Offset =
+ AnalyzeLoadFromClobberingStore(LI->getType(), Address, DepSI);
if (Offset != -1) {
ValuesPerBlock.push_back(AvailableValueInBlock::get(DepBB,
DepSI->getValueOperand(),
if (LoadInst *DepLI = dyn_cast<LoadInst>(DepInfo.getInst())) {
// If this is a clobber and L is the first instruction in its block, then
// we have the first instruction in the entry block.
- if (DepLI != LI && Address && DL) {
- int Offset = AnalyzeLoadFromClobberingLoad(LI->getType(), Address,
- DepLI, *DL);
+ if (DepLI != LI && Address) {
+ int Offset =
+ AnalyzeLoadFromClobberingLoad(LI->getType(), Address, DepLI, DL);
if (Offset != -1) {
ValuesPerBlock.push_back(AvailableValueInBlock::getLoad(DepBB,DepLI,
// If the clobbering value is a memset/memcpy/memmove, see if we can
// forward a value on from it.
if (MemIntrinsic *DepMI = dyn_cast<MemIntrinsic>(DepInfo.getInst())) {
- if (DL && Address) {
+ if (Address) {
int Offset = AnalyzeLoadFromClobberingMemInst(LI->getType(), Address,
- DepMI, *DL);
+ DepMI, DL);
if (Offset != -1) {
ValuesPerBlock.push_back(AvailableValueInBlock::getMI(DepBB, DepMI,
Offset));
if (S->getValueOperand()->getType() != LI->getType()) {
// If the stored value is larger or equal to the loaded value, we can
// reuse it.
- if (!DL || !CanCoerceMustAliasedValueToLoad(S->getValueOperand(),
- LI->getType(), *DL)) {
+ if (!CanCoerceMustAliasedValueToLoad(S->getValueOperand(),
+ LI->getType(), DL)) {
UnavailableBlocks.push_back(DepBB);
continue;
}
if (LD->getType() != LI->getType()) {
// If the stored value is larger or equal to the loaded value, we can
// reuse it.
- if (!DL || !CanCoerceMustAliasedValueToLoad(LD, LI->getType(),*DL)) {
+ if (!CanCoerceMustAliasedValueToLoad(LD, LI->getType(), DL)) {
UnavailableBlocks.push_back(DepBB);
continue;
}
return false;
}
- if (LoadBB->isLandingPad()) {
+ if (LoadBB->isEHPad()) {
DEBUG(dbgs()
- << "COULD NOT PRE LOAD BECAUSE OF LANDING PAD CRITICAL EDGE '"
+ << "COULD NOT PRE LOAD BECAUSE OF AN EH PAD CRITICAL EDGE '"
<< Pred->getName() << "': " << *LI << '\n');
return false;
}
// Check if the load can safely be moved to all the unavailable predecessors.
bool CanDoPRE = true;
+ const DataLayout &DL = LI->getModule()->getDataLayout();
SmallVector<Instruction*, 8> NewInsts;
for (auto &PredLoad : PredLoads) {
BasicBlock *UnavailablePred = PredLoad.first;
// If all preds have a single successor, then we know it is safe to insert
// the load on the pred (?!?), so we can insert code to materialize the
// pointer if it is not available.
- PHITransAddr Address(LI->getPointerOperand(), DL);
+ PHITransAddr Address(LI->getPointerOperand(), DL, AC);
Value *LoadPtr = nullptr;
LoadPtr = Address.PHITranslateWithInsertion(LoadBB, UnavailablePred,
*DT, NewInsts);
LI->getAlignment(),
UnavailablePred->getTerminator());
- // Transfer the old load's TBAA tag to the new load.
- if (MDNode *Tag = LI->getMetadata(LLVMContext::MD_tbaa))
- NewLoad->setMetadata(LLVMContext::MD_tbaa, Tag);
+ // Transfer the old load's AA tags to the new load.
+ AAMDNodes Tags;
+ LI->getAAMetadata(Tags);
+ if (Tags)
+ NewLoad->setAAMetadata(Tags);
+
+ if (auto *InvGroupMD = LI->getMetadata(LLVMContext::MD_invariant_group))
+ NewLoad->setMetadata(LLVMContext::MD_invariant_group, InvGroupMD);
// Transfer DebugLoc.
NewLoad->setDebugLoc(LI->getDebugLoc());
LI->replaceAllUsesWith(V);
if (isa<PHINode>(V))
V->takeName(LI);
+ if (Instruction *I = dyn_cast<Instruction>(V))
+ I->setDebugLoc(LI->getDebugLoc());
if (V->getType()->getScalarType()->isPointerTy())
MD->invalidateCachedPointerInfo(V);
markInstructionForDeletion(LI);
return true;
}
-/// processNonLocalLoad - Attempt to eliminate a load whose dependencies are
+/// Attempt to eliminate a load whose dependencies are
/// non-local by performing PHI construction.
bool GVN::processNonLocalLoad(LoadInst *LI) {
// Step 1: Find the non-local dependencies of the load.
LoadDepVect Deps;
- AliasAnalysis::Location Loc = VN.getAliasAnalysis()->getLocation(LI);
- MD->getNonLocalPointerDependency(Loc, true, LI->getParent(), Deps);
+ MD->getNonLocalPointerDependency(LI, Deps);
// If we had to process more than one hundred blocks to find the
// dependencies, this load isn't worth worrying about. Optimizing
return false;
}
+ // If this load follows a GEP, see if we can PRE the indices before analyzing.
+ if (GetElementPtrInst *GEP = dyn_cast<GetElementPtrInst>(LI->getOperand(0))) {
+ for (GetElementPtrInst::op_iterator OI = GEP->idx_begin(),
+ OE = GEP->idx_end();
+ OI != OE; ++OI)
+ if (Instruction *I = dyn_cast<Instruction>(OI->get()))
+ performScalarPRE(I);
+ }
+
// Step 2: Analyze the availability of the load
AvailValInBlkVect ValuesPerBlock;
UnavailBlkVect UnavailableBlocks;
if (isa<PHINode>(V))
V->takeName(LI);
+ if (Instruction *I = dyn_cast<Instruction>(V))
+ if (LI->getDebugLoc())
+ I->setDebugLoc(LI->getDebugLoc());
if (V->getType()->getScalarType()->isPointerTy())
MD->invalidateCachedPointerInfo(V);
markInstructionForDeletion(LI);
return PerformLoadPRE(LI, ValuesPerBlock, UnavailableBlocks);
}
+bool GVN::processAssumeIntrinsic(IntrinsicInst *IntrinsicI) {
+ assert(IntrinsicI->getIntrinsicID() == Intrinsic::assume &&
+ "This function can only be called with llvm.assume intrinsic");
+ Value *V = IntrinsicI->getArgOperand(0);
+
+ if (ConstantInt *Cond = dyn_cast<ConstantInt>(V)) {
+ if (Cond->isZero()) {
+ Type *Int8Ty = Type::getInt8Ty(V->getContext());
+ // Insert a new store to null instruction before the load to indicate that
+ // this code is not reachable. FIXME: We could insert unreachable
+ // instruction directly because we can modify the CFG.
+ new StoreInst(UndefValue::get(Int8Ty),
+ Constant::getNullValue(Int8Ty->getPointerTo()),
+ IntrinsicI);
+ }
+ markInstructionForDeletion(IntrinsicI);
+ return false;
+ }
+
+ Constant *True = ConstantInt::getTrue(V->getContext());
+ bool Changed = false;
+
+ for (BasicBlock *Successor : successors(IntrinsicI->getParent())) {
+ BasicBlockEdge Edge(IntrinsicI->getParent(), Successor);
+
+ // This property is only true in dominated successors, propagateEquality
+ // will check dominance for us.
+ Changed |= propagateEquality(V, True, Edge, false);
+ }
+
+ // We can replace assume value with true, which covers cases like this:
+ // call void @llvm.assume(i1 %cmp)
+ // br i1 %cmp, label %bb1, label %bb2 ; will change %cmp to true
+ ReplaceWithConstMap[V] = True;
+
+ // If one of *cmp *eq operand is const, adding it to map will cover this:
+ // %cmp = fcmp oeq float 3.000000e+00, %0 ; const on lhs could happen
+ // call void @llvm.assume(i1 %cmp)
+ // ret float %0 ; will change it to ret float 3.000000e+00
+ if (auto *CmpI = dyn_cast<CmpInst>(V)) {
+ if (CmpI->getPredicate() == CmpInst::Predicate::ICMP_EQ ||
+ CmpI->getPredicate() == CmpInst::Predicate::FCMP_OEQ ||
+ (CmpI->getPredicate() == CmpInst::Predicate::FCMP_UEQ &&
+ CmpI->getFastMathFlags().noNaNs())) {
+ Value *CmpLHS = CmpI->getOperand(0);
+ Value *CmpRHS = CmpI->getOperand(1);
+ if (isa<Constant>(CmpLHS))
+ std::swap(CmpLHS, CmpRHS);
+ auto *RHSConst = dyn_cast<Constant>(CmpRHS);
+
+ // If only one operand is constant.
+ if (RHSConst != nullptr && !isa<Constant>(CmpLHS))
+ ReplaceWithConstMap[CmpLHS] = RHSConst;
+ }
+ }
+ return Changed;
+}
static void patchReplacementInstruction(Instruction *I, Value *Repl) {
// Patch the replacement so that it is not more restrictive than the value
// being replaced.
BinaryOperator *Op = dyn_cast<BinaryOperator>(I);
BinaryOperator *ReplOp = dyn_cast<BinaryOperator>(Repl);
- if (Op && ReplOp && isa<OverflowingBinaryOperator>(Op) &&
- isa<OverflowingBinaryOperator>(ReplOp)) {
- if (ReplOp->hasNoSignedWrap() && !Op->hasNoSignedWrap())
- ReplOp->setHasNoSignedWrap(false);
- if (ReplOp->hasNoUnsignedWrap() && !Op->hasNoUnsignedWrap())
- ReplOp->setHasNoUnsignedWrap(false);
- }
+ if (Op && ReplOp)
+ ReplOp->andIRFlags(Op);
+
if (Instruction *ReplInst = dyn_cast<Instruction>(Repl)) {
- SmallVector<std::pair<unsigned, MDNode*>, 4> Metadata;
- ReplInst->getAllMetadataOtherThanDebugLoc(Metadata);
- for (int i = 0, n = Metadata.size(); i < n; ++i) {
- unsigned Kind = Metadata[i].first;
- MDNode *IMD = I->getMetadata(Kind);
- MDNode *ReplMD = Metadata[i].second;
- switch(Kind) {
- default:
- ReplInst->setMetadata(Kind, nullptr); // Remove unknown metadata
- break;
- case LLVMContext::MD_dbg:
- llvm_unreachable("getAllMetadataOtherThanDebugLoc returned a MD_dbg");
- case LLVMContext::MD_tbaa:
- ReplInst->setMetadata(Kind, MDNode::getMostGenericTBAA(IMD, ReplMD));
- break;
- case LLVMContext::MD_range:
- ReplInst->setMetadata(Kind, MDNode::getMostGenericRange(IMD, ReplMD));
- break;
- case LLVMContext::MD_prof:
- llvm_unreachable("MD_prof in a non-terminator instruction");
- break;
- case LLVMContext::MD_fpmath:
- ReplInst->setMetadata(Kind, MDNode::getMostGenericFPMath(IMD, ReplMD));
- break;
- case LLVMContext::MD_invariant_load:
- // Only set the !invariant.load if it is present in both instructions.
- ReplInst->setMetadata(Kind, IMD);
- break;
- }
- }
+ // FIXME: If both the original and replacement value are part of the
+ // same control-flow region (meaning that the execution of one
+ // guarantees the execution of the other), then we can combine the
+ // noalias scopes here and do better than the general conservative
+ // answer used in combineMetadata().
+
+ // In general, GVN unifies expressions over different control-flow
+ // regions, and so we need a conservative combination of the noalias
+ // scopes.
+ static const unsigned KnownIDs[] = {
+ LLVMContext::MD_tbaa, LLVMContext::MD_alias_scope,
+ LLVMContext::MD_noalias, LLVMContext::MD_range,
+ LLVMContext::MD_fpmath, LLVMContext::MD_invariant_load,
+ LLVMContext::MD_invariant_group};
+ combineMetadata(ReplInst, I, KnownIDs);
}
}
I->replaceAllUsesWith(Repl);
}
-/// processLoad - Attempt to eliminate a load, first by eliminating it
+/// Attempt to eliminate a load, first by eliminating it
/// locally, and then attempting non-local elimination if that fails.
bool GVN::processLoad(LoadInst *L) {
if (!MD)
// ... to a pointer that has been loaded from before...
MemDepResult Dep = MD->getDependency(L);
+ const DataLayout &DL = L->getModule()->getDataLayout();
// If we have a clobber and target data is around, see if this is a clobber
// that we can fix up through code synthesis.
- if (Dep.isClobber() && DL) {
+ if (Dep.isClobber()) {
// Check to see if we have something like this:
// store i32 123, i32* %P
// %A = bitcast i32* %P to i8*
// access code.
Value *AvailVal = nullptr;
if (StoreInst *DepSI = dyn_cast<StoreInst>(Dep.getInst())) {
- int Offset = AnalyzeLoadFromClobberingStore(L->getType(),
- L->getPointerOperand(),
- DepSI, *DL);
+ int Offset = AnalyzeLoadFromClobberingStore(
+ L->getType(), L->getPointerOperand(), DepSI);
if (Offset != -1)
AvailVal = GetStoreValueForLoad(DepSI->getValueOperand(), Offset,
- L->getType(), L, *DL);
+ L->getType(), L, DL);
}
// Check to see if we have something like this:
if (DepLI == L)
return false;
- int Offset = AnalyzeLoadFromClobberingLoad(L->getType(),
- L->getPointerOperand(),
- DepLI, *DL);
+ int Offset = AnalyzeLoadFromClobberingLoad(
+ L->getType(), L->getPointerOperand(), DepLI, DL);
if (Offset != -1)
AvailVal = GetLoadValueForLoad(DepLI, Offset, L->getType(), L, *this);
}
// If the clobbering value is a memset/memcpy/memmove, see if we can forward
// a value on from it.
if (MemIntrinsic *DepMI = dyn_cast<MemIntrinsic>(Dep.getInst())) {
- int Offset = AnalyzeLoadFromClobberingMemInst(L->getType(),
- L->getPointerOperand(),
- DepMI, *DL);
+ int Offset = AnalyzeLoadFromClobberingMemInst(
+ L->getType(), L->getPointerOperand(), DepMI, DL);
if (Offset != -1)
- AvailVal = GetMemInstValueForLoad(DepMI, Offset, L->getType(), L, *DL);
+ AvailVal = GetMemInstValueForLoad(DepMI, Offset, L->getType(), L, DL);
}
if (AvailVal) {
++NumGVNLoad;
return true;
}
- }
- // If the value isn't available, don't do anything!
- if (Dep.isClobber()) {
+ // If the value isn't available, don't do anything!
DEBUG(
// fast print dep, using operator<< on instruction is too slow.
dbgs() << "GVN: load ";
// actually have the same type. See if we know how to reuse the stored
// value (depending on its type).
if (StoredVal->getType() != L->getType()) {
- if (DL) {
- StoredVal = CoerceAvailableValueToLoadType(StoredVal, L->getType(),
- L, *DL);
- if (!StoredVal)
- return false;
-
- DEBUG(dbgs() << "GVN COERCED STORE:\n" << *DepSI << '\n' << *StoredVal
- << '\n' << *L << "\n\n\n");
- }
- else
+ IRBuilder<> Builder(L);
+ StoredVal =
+ CoerceAvailableValueToLoadType(StoredVal, L->getType(), Builder, DL);
+ if (!StoredVal)
return false;
+
+ DEBUG(dbgs() << "GVN COERCED STORE:\n" << *DepSI << '\n' << *StoredVal
+ << '\n' << *L << "\n\n\n");
}
// Remove it!
// the same type. See if we know how to reuse the previously loaded value
// (depending on its type).
if (DepLI->getType() != L->getType()) {
- if (DL) {
- AvailableVal = CoerceAvailableValueToLoadType(DepLI, L->getType(),
- L, *DL);
- if (!AvailableVal)
- return false;
-
- DEBUG(dbgs() << "GVN COERCED LOAD:\n" << *DepLI << "\n" << *AvailableVal
- << "\n" << *L << "\n\n\n");
- }
- else
+ IRBuilder<> Builder(L);
+ AvailableVal =
+ CoerceAvailableValueToLoadType(DepLI, L->getType(), Builder, DL);
+ if (!AvailableVal)
return false;
+
+ DEBUG(dbgs() << "GVN COERCED LOAD:\n" << *DepLI << "\n" << *AvailableVal
+ << "\n" << *L << "\n\n\n");
}
// Remove it!
return false;
}
-// findLeader - In order to find a leader for a given value number at a
+// In order to find a leader for a given value number at a
// specific basic block, we first obtain the list of all Values for that number,
// and then scan the list to find one whose block dominates the block in
// question. This is fast because dominator tree queries consist of only
return Val;
}
-/// replaceAllDominatedUsesWith - Replace all uses of 'From' with 'To' if the
-/// use is dominated by the given basic block. Returns the number of uses that
-/// were replaced.
-unsigned GVN::replaceAllDominatedUsesWith(Value *From, Value *To,
- const BasicBlockEdge &Root) {
- unsigned Count = 0;
- for (Value::use_iterator UI = From->use_begin(), UE = From->use_end();
- UI != UE; ) {
- Use &U = *UI++;
-
- if (DT->dominates(Root, U)) {
- U.set(To);
- ++Count;
- }
- }
- return Count;
-}
-
-/// isOnlyReachableViaThisEdge - There is an edge from 'Src' to 'Dst'. Return
+/// There is an edge from 'Src' to 'Dst'. Return
/// true if every path from the entry block to 'Dst' passes via this edge. In
/// particular 'Dst' must not be reachable via another edge from 'Src'.
static bool isOnlyReachableViaThisEdge(const BasicBlockEdge &E,
return Pred != nullptr;
}
-/// propagateEquality - The given values are known to be equal in every block
+// Tries to replace instruction with const, using information from
+// ReplaceWithConstMap.
+bool GVN::replaceOperandsWithConsts(Instruction *Instr) const {
+ bool Changed = false;
+ for (unsigned OpNum = 0; OpNum < Instr->getNumOperands(); ++OpNum) {
+ Value *Operand = Instr->getOperand(OpNum);
+ auto it = ReplaceWithConstMap.find(Operand);
+ if (it != ReplaceWithConstMap.end()) {
+ assert(!isa<Constant>(Operand) &&
+ "Replacing constants with constants is invalid");
+ DEBUG(dbgs() << "GVN replacing: " << *Operand << " with " << *it->second
+ << " in instruction " << *Instr << '\n');
+ Instr->setOperand(OpNum, it->second);
+ Changed = true;
+ }
+ }
+ return Changed;
+}
+
+/// The given values are known to be equal in every block
/// dominated by 'Root'. Exploit this, for example by replacing 'LHS' with
/// 'RHS' everywhere in the scope. Returns whether a change was made.
-bool GVN::propagateEquality(Value *LHS, Value *RHS,
- const BasicBlockEdge &Root) {
+/// If DominatesByEdge is false, then it means that it is dominated by Root.End.
+bool GVN::propagateEquality(Value *LHS, Value *RHS, const BasicBlockEdge &Root,
+ bool DominatesByEdge) {
SmallVector<std::pair<Value*, Value*>, 4> Worklist;
Worklist.push_back(std::make_pair(LHS, RHS));
bool Changed = false;
std::pair<Value*, Value*> Item = Worklist.pop_back_val();
LHS = Item.first; RHS = Item.second;
- if (LHS == RHS) continue;
+ if (LHS == RHS)
+ continue;
assert(LHS->getType() == RHS->getType() && "Equality but unequal types!");
// Don't try to propagate equalities between constants.
- if (isa<Constant>(LHS) && isa<Constant>(RHS)) continue;
+ if (isa<Constant>(LHS) && isa<Constant>(RHS))
+ continue;
// Prefer a constant on the right-hand side, or an Argument if no constants.
if (isa<Constant>(LHS) || (isa<Argument>(LHS) && !isa<Constant>(RHS)))
std::swap(LHS, RHS);
assert((isa<Argument>(LHS) || isa<Instruction>(LHS)) && "Unexpected value!");
- // If there is no obvious reason to prefer the left-hand side over the right-
- // hand side, ensure the longest lived term is on the right-hand side, so the
- // shortest lived term will be replaced by the longest lived. This tends to
- // expose more simplifications.
+ // If there is no obvious reason to prefer the left-hand side over the
+ // right-hand side, ensure the longest lived term is on the right-hand side,
+ // so the shortest lived term will be replaced by the longest lived.
+ // This tends to expose more simplifications.
uint32_t LVN = VN.lookup_or_add(LHS);
if ((isa<Argument>(LHS) && isa<Argument>(RHS)) ||
(isa<Instruction>(LHS) && isa<Instruction>(RHS))) {
- // Move the 'oldest' value to the right-hand side, using the value number as
- // a proxy for age.
+ // Move the 'oldest' value to the right-hand side, using the value number
+ // as a proxy for age.
uint32_t RVN = VN.lookup_or_add(RHS);
if (LVN < RVN) {
std::swap(LHS, RHS);
// LHS always has at least one use that is not dominated by Root, this will
// never do anything if LHS has only one use.
if (!LHS->hasOneUse()) {
- unsigned NumReplacements = replaceAllDominatedUsesWith(LHS, RHS, Root);
+ unsigned NumReplacements =
+ DominatesByEdge
+ ? replaceDominatedUsesWith(LHS, RHS, *DT, Root)
+ : replaceDominatedUsesWith(LHS, RHS, *DT, Root.getEnd());
+
Changed |= NumReplacements > 0;
NumGVNEqProp += NumReplacements;
}
- // Now try to deduce additional equalities from this one. For example, if the
- // known equality was "(A != B)" == "false" then it follows that A and B are
- // equal in the scope. Only boolean equalities with an explicit true or false
- // RHS are currently supported.
+ // Now try to deduce additional equalities from this one. For example, if
+ // the known equality was "(A != B)" == "false" then it follows that A and B
+ // are equal in the scope. Only boolean equalities with an explicit true or
+ // false RHS are currently supported.
if (!RHS->getType()->isIntegerTy(1))
// Not a boolean equality - bail out.
continue;
// If we are propagating an equality like "(A == B)" == "true" then also
// propagate the equality A == B. When propagating a comparison such as
// "(A >= B)" == "true", replace all instances of "A < B" with "false".
- if (ICmpInst *Cmp = dyn_cast<ICmpInst>(LHS)) {
+ if (CmpInst *Cmp = dyn_cast<CmpInst>(LHS)) {
Value *Op0 = Cmp->getOperand(0), *Op1 = Cmp->getOperand(1);
// If "A == B" is known true, or "A != B" is known false, then replace
(isKnownFalse && Cmp->getPredicate() == CmpInst::ICMP_NE))
Worklist.push_back(std::make_pair(Op0, Op1));
+ // Handle the floating point versions of equality comparisons too.
+ if ((isKnownTrue && Cmp->getPredicate() == CmpInst::FCMP_OEQ) ||
+ (isKnownFalse && Cmp->getPredicate() == CmpInst::FCMP_UNE)) {
+
+ // Floating point -0.0 and 0.0 compare equal, so we can only
+ // propagate values if we know that we have a constant and that
+ // its value is non-zero.
+
+ // FIXME: We should do this optimization if 'no signed zeros' is
+ // applicable via an instruction-level fast-math-flag or some other
+ // indicator that relaxed FP semantics are being used.
+
+ if (isa<ConstantFP>(Op1) && !cast<ConstantFP>(Op1)->isZero())
+ Worklist.push_back(std::make_pair(Op0, Op1));
+ }
+
// If "A >= B" is known true, replace "A < B" with false everywhere.
CmpInst::Predicate NotPred = Cmp->getInversePredicate();
Constant *NotVal = ConstantInt::get(Cmp->getType(), isKnownFalse);
- // Since we don't have the instruction "A < B" immediately to hand, work out
- // the value number that it would have and use that to find an appropriate
- // instruction (if any).
+ // Since we don't have the instruction "A < B" immediately to hand, work
+ // out the value number that it would have and use that to find an
+ // appropriate instruction (if any).
uint32_t NextNum = VN.getNextUnusedValueNumber();
uint32_t Num = VN.lookup_or_add_cmp(Cmp->getOpcode(), NotPred, Op0, Op1);
// If the number we were assigned was brand new then there is no point in
Value *NotCmp = findLeader(Root.getEnd(), Num);
if (NotCmp && isa<Instruction>(NotCmp)) {
unsigned NumReplacements =
- replaceAllDominatedUsesWith(NotCmp, NotVal, Root);
+ DominatesByEdge
+ ? replaceDominatedUsesWith(NotCmp, NotVal, *DT, Root)
+ : replaceDominatedUsesWith(NotCmp, NotVal, *DT,
+ Root.getEnd());
Changed |= NumReplacements > 0;
NumGVNEqProp += NumReplacements;
}
return Changed;
}
-/// processInstruction - When calculating availability, handle an instruction
+/// When calculating availability, handle an instruction
/// by inserting it into the appropriate sets
bool GVN::processInstruction(Instruction *I) {
// Ignore dbg info intrinsics.
// to value numbering it. Value numbering often exposes redundancies, for
// example if it determines that %y is equal to %x then the instruction
// "%z = and i32 %x, %y" becomes "%z = and i32 %x, %x" which we now simplify.
- if (Value *V = SimplifyInstruction(I, DL, TLI, DT)) {
+ const DataLayout &DL = I->getModule()->getDataLayout();
+ if (Value *V = SimplifyInstruction(I, DL, TLI, DT, AC)) {
I->replaceAllUsesWith(V);
if (MD && V->getType()->getScalarType()->isPointerTy())
MD->invalidateCachedPointerInfo(V);
return true;
}
+ if (IntrinsicInst *IntrinsicI = dyn_cast<IntrinsicInst>(I))
+ if (IntrinsicI->getIntrinsicID() == Intrinsic::assume)
+ return processAssumeIntrinsic(IntrinsicI);
+
if (LoadInst *LI = dyn_cast<LoadInst>(I)) {
if (processLoad(LI))
return true;
Value *TrueVal = ConstantInt::getTrue(TrueSucc->getContext());
BasicBlockEdge TrueE(Parent, TrueSucc);
- Changed |= propagateEquality(BranchCond, TrueVal, TrueE);
+ Changed |= propagateEquality(BranchCond, TrueVal, TrueE, true);
Value *FalseVal = ConstantInt::getFalse(FalseSucc->getContext());
BasicBlockEdge FalseE(Parent, FalseSucc);
- Changed |= propagateEquality(BranchCond, FalseVal, FalseE);
+ Changed |= propagateEquality(BranchCond, FalseVal, FalseE, true);
return Changed;
}
// If there is only a single edge, propagate the case value into it.
if (SwitchEdges.lookup(Dst) == 1) {
BasicBlockEdge E(Parent, Dst);
- Changed |= propagateEquality(SwitchCond, i.getCaseValue(), E);
+ Changed |= propagateEquality(SwitchCond, i.getCaseValue(), E, true);
}
}
return Changed;
// Instructions with void type don't return a value, so there's
// no point in trying to find redundancies in them.
- if (I->getType()->isVoidTy()) return false;
+ if (I->getType()->isVoidTy())
+ return false;
uint32_t NextNum = VN.getNextUnusedValueNumber();
unsigned Num = VN.lookup_or_add(I);
if (!NoLoads)
MD = &getAnalysis<MemoryDependenceAnalysis>();
DT = &getAnalysis<DominatorTreeWrapperPass>().getDomTree();
- DataLayoutPass *DLP = getAnalysisIfAvailable<DataLayoutPass>();
- DL = DLP ? &DLP->getDataLayout() : nullptr;
- TLI = &getAnalysis<TargetLibraryInfo>();
- VN.setAliasAnalysis(&getAnalysis<AliasAnalysis>());
+ AC = &getAnalysis<AssumptionCacheTracker>().getAssumptionCache(F);
+ TLI = &getAnalysis<TargetLibraryInfoWrapperPass>().getTLI();
+ VN.setAliasAnalysis(&getAnalysis<AAResultsWrapperPass>().getAAResults());
VN.setMemDep(MD);
VN.setDomTree(DT);
// Merge unconditional branches, allowing PRE to catch more
// optimization opportunities.
for (Function::iterator FI = F.begin(), FE = F.end(); FI != FE; ) {
- BasicBlock *BB = FI++;
+ BasicBlock *BB = &*FI++;
- bool removedBlock = MergeBlockIntoPredecessor(BB, this);
+ bool removedBlock =
+ MergeBlockIntoPredecessor(BB, DT, /* LoopInfo */ nullptr, MD);
if (removedBlock) ++NumGVNBlocks;
Changed |= removedBlock;
return Changed;
}
-
bool GVN::processBlock(BasicBlock *BB) {
// FIXME: Kill off InstrsToErase by doing erasing eagerly in a helper function
// (and incrementing BI before processing an instruction).
if (DeadBlocks.count(BB))
return false;
+ // Clearing map before every BB because it can be used only for single BB.
+ ReplaceWithConstMap.clear();
bool ChangedFunction = false;
for (BasicBlock::iterator BI = BB->begin(), BE = BB->end();
BI != BE;) {
- ChangedFunction |= processInstruction(BI);
+ if (!ReplaceWithConstMap.empty())
+ ChangedFunction |= replaceOperandsWithConsts(&*BI);
+ ChangedFunction |= processInstruction(&*BI);
+
if (InstrsToErase.empty()) {
++BI;
continue;
return ChangedFunction;
}
-/// performPRE - Perform a purely local form of PRE that looks for diamond
-/// control flow patterns and attempts to perform simple PRE at the join point.
-bool GVN::performPRE(Function &F) {
- bool Changed = false;
+// Instantiate an expression in a predecessor that lacked it.
+bool GVN::performScalarPREInsertion(Instruction *Instr, BasicBlock *Pred,
+ unsigned int ValNo) {
+ // Because we are going top-down through the block, all value numbers
+ // will be available in the predecessor by the time we need them. Any
+ // that weren't originally present will have been instantiated earlier
+ // in this loop.
+ bool success = true;
+ for (unsigned i = 0, e = Instr->getNumOperands(); i != e; ++i) {
+ Value *Op = Instr->getOperand(i);
+ if (isa<Argument>(Op) || isa<Constant>(Op) || isa<GlobalValue>(Op))
+ continue;
+
+ if (Value *V = findLeader(Pred, VN.lookup(Op))) {
+ Instr->setOperand(i, V);
+ } else {
+ success = false;
+ break;
+ }
+ }
+
+ // Fail out if we encounter an operand that is not available in
+ // the PRE predecessor. This is typically because of loads which
+ // are not value numbered precisely.
+ if (!success)
+ return false;
+
+ Instr->insertBefore(Pred->getTerminator());
+ Instr->setName(Instr->getName() + ".pre");
+ Instr->setDebugLoc(Instr->getDebugLoc());
+ VN.add(Instr, ValNo);
+
+ // Update the availability map to include the new instruction.
+ addToLeaderTable(ValNo, Instr, Pred);
+ return true;
+}
+
+bool GVN::performScalarPRE(Instruction *CurInst) {
SmallVector<std::pair<Value*, BasicBlock*>, 8> predMap;
- for (BasicBlock *CurrentBlock : depth_first(&F.getEntryBlock())) {
- // Nothing to PRE in the entry block.
- if (CurrentBlock == &F.getEntryBlock()) continue;
- // Don't perform PRE on a landing pad.
- if (CurrentBlock->isLandingPad()) continue;
+ if (isa<AllocaInst>(CurInst) || isa<TerminatorInst>(CurInst) ||
+ isa<PHINode>(CurInst) || CurInst->getType()->isVoidTy() ||
+ CurInst->mayReadFromMemory() || CurInst->mayHaveSideEffects() ||
+ isa<DbgInfoIntrinsic>(CurInst))
+ return false;
- for (BasicBlock::iterator BI = CurrentBlock->begin(),
- BE = CurrentBlock->end(); BI != BE; ) {
- Instruction *CurInst = BI++;
-
- if (isa<AllocaInst>(CurInst) ||
- isa<TerminatorInst>(CurInst) || isa<PHINode>(CurInst) ||
- CurInst->getType()->isVoidTy() ||
- CurInst->mayReadFromMemory() || CurInst->mayHaveSideEffects() ||
- isa<DbgInfoIntrinsic>(CurInst))
- continue;
+ // Don't do PRE on compares. The PHI would prevent CodeGenPrepare from
+ // sinking the compare again, and it would force the code generator to
+ // move the i1 from processor flags or predicate registers into a general
+ // purpose register.
+ if (isa<CmpInst>(CurInst))
+ return false;
- // Don't do PRE on compares. The PHI would prevent CodeGenPrepare from
- // sinking the compare again, and it would force the code generator to
- // move the i1 from processor flags or predicate registers into a general
- // purpose register.
- if (isa<CmpInst>(CurInst))
- continue;
+ // We don't currently value number ANY inline asm calls.
+ if (CallInst *CallI = dyn_cast<CallInst>(CurInst))
+ if (CallI->isInlineAsm())
+ return false;
- // We don't currently value number ANY inline asm calls.
- if (CallInst *CallI = dyn_cast<CallInst>(CurInst))
- if (CallI->isInlineAsm())
- continue;
+ uint32_t ValNo = VN.lookup(CurInst);
+
+ // Look for the predecessors for PRE opportunities. We're
+ // only trying to solve the basic diamond case, where
+ // a value is computed in the successor and one predecessor,
+ // but not the other. We also explicitly disallow cases
+ // where the successor is its own predecessor, because they're
+ // more complicated to get right.
+ unsigned NumWith = 0;
+ unsigned NumWithout = 0;
+ BasicBlock *PREPred = nullptr;
+ BasicBlock *CurrentBlock = CurInst->getParent();
+ predMap.clear();
+
+ for (pred_iterator PI = pred_begin(CurrentBlock), PE = pred_end(CurrentBlock);
+ PI != PE; ++PI) {
+ BasicBlock *P = *PI;
+ // We're not interested in PRE where the block is its
+ // own predecessor, or in blocks with predecessors
+ // that are not reachable.
+ if (P == CurrentBlock) {
+ NumWithout = 2;
+ break;
+ } else if (!DT->isReachableFromEntry(P)) {
+ NumWithout = 2;
+ break;
+ }
- uint32_t ValNo = VN.lookup(CurInst);
-
- // Look for the predecessors for PRE opportunities. We're
- // only trying to solve the basic diamond case, where
- // a value is computed in the successor and one predecessor,
- // but not the other. We also explicitly disallow cases
- // where the successor is its own predecessor, because they're
- // more complicated to get right.
- unsigned NumWith = 0;
- unsigned NumWithout = 0;
- BasicBlock *PREPred = nullptr;
- predMap.clear();
-
- for (pred_iterator PI = pred_begin(CurrentBlock),
- PE = pred_end(CurrentBlock); PI != PE; ++PI) {
- BasicBlock *P = *PI;
- // We're not interested in PRE where the block is its
- // own predecessor, or in blocks with predecessors
- // that are not reachable.
- if (P == CurrentBlock) {
- NumWithout = 2;
- break;
- } else if (!DT->isReachableFromEntry(P)) {
- NumWithout = 2;
- break;
- }
+ Value *predV = findLeader(P, ValNo);
+ if (!predV) {
+ predMap.push_back(std::make_pair(static_cast<Value *>(nullptr), P));
+ PREPred = P;
+ ++NumWithout;
+ } else if (predV == CurInst) {
+ /* CurInst dominates this predecessor. */
+ NumWithout = 2;
+ break;
+ } else {
+ predMap.push_back(std::make_pair(predV, P));
+ ++NumWith;
+ }
+ }
- Value* predV = findLeader(P, ValNo);
- if (!predV) {
- predMap.push_back(std::make_pair(static_cast<Value *>(nullptr), P));
- PREPred = P;
- ++NumWithout;
- } else if (predV == CurInst) {
- /* CurInst dominates this predecessor. */
- NumWithout = 2;
- break;
- } else {
- predMap.push_back(std::make_pair(predV, P));
- ++NumWith;
- }
- }
+ // Don't do PRE when it might increase code size, i.e. when
+ // we would need to insert instructions in more than one pred.
+ if (NumWithout > 1 || NumWith == 0)
+ return false;
- // Don't do PRE when it might increase code size, i.e. when
- // we would need to insert instructions in more than one pred.
- if (NumWithout != 1 || NumWith == 0)
- continue;
+ // We may have a case where all predecessors have the instruction,
+ // and we just need to insert a phi node. Otherwise, perform
+ // insertion.
+ Instruction *PREInstr = nullptr;
- // Don't do PRE across indirect branch.
- if (isa<IndirectBrInst>(PREPred->getTerminator()))
- continue;
+ if (NumWithout != 0) {
+ // Don't do PRE across indirect branch.
+ if (isa<IndirectBrInst>(PREPred->getTerminator()))
+ return false;
- // We can't do PRE safely on a critical edge, so instead we schedule
- // the edge to be split and perform the PRE the next time we iterate
- // on the function.
- unsigned SuccNum = GetSuccessorNumber(PREPred, CurrentBlock);
- if (isCriticalEdge(PREPred->getTerminator(), SuccNum)) {
- toSplit.push_back(std::make_pair(PREPred->getTerminator(), SuccNum));
- continue;
- }
+ // We can't do PRE safely on a critical edge, so instead we schedule
+ // the edge to be split and perform the PRE the next time we iterate
+ // on the function.
+ unsigned SuccNum = GetSuccessorNumber(PREPred, CurrentBlock);
+ if (isCriticalEdge(PREPred->getTerminator(), SuccNum)) {
+ toSplit.push_back(std::make_pair(PREPred->getTerminator(), SuccNum));
+ return false;
+ }
+ // We need to insert somewhere, so let's give it a shot
+ PREInstr = CurInst->clone();
+ if (!performScalarPREInsertion(PREInstr, PREPred, ValNo)) {
+ // If we failed insertion, make sure we remove the instruction.
+ DEBUG(verifyRemoved(PREInstr));
+ delete PREInstr;
+ return false;
+ }
+ }
- // Instantiate the expression in the predecessor that lacked it.
- // Because we are going top-down through the block, all value numbers
- // will be available in the predecessor by the time we need them. Any
- // that weren't originally present will have been instantiated earlier
- // in this loop.
- Instruction *PREInstr = CurInst->clone();
- bool success = true;
- for (unsigned i = 0, e = CurInst->getNumOperands(); i != e; ++i) {
- Value *Op = PREInstr->getOperand(i);
- if (isa<Argument>(Op) || isa<Constant>(Op) || isa<GlobalValue>(Op))
- continue;
+ // Either we should have filled in the PRE instruction, or we should
+ // not have needed insertions.
+ assert (PREInstr != nullptr || NumWithout == 0);
- if (Value *V = findLeader(PREPred, VN.lookup(Op))) {
- PREInstr->setOperand(i, V);
- } else {
- success = false;
- break;
- }
- }
+ ++NumGVNPRE;
- // Fail out if we encounter an operand that is not available in
- // the PRE predecessor. This is typically because of loads which
- // are not value numbered precisely.
- if (!success) {
- DEBUG(verifyRemoved(PREInstr));
- delete PREInstr;
- continue;
- }
+ // Create a PHI to make the value available in this block.
+ PHINode *Phi =
+ PHINode::Create(CurInst->getType(), predMap.size(),
+ CurInst->getName() + ".pre-phi", &CurrentBlock->front());
+ for (unsigned i = 0, e = predMap.size(); i != e; ++i) {
+ if (Value *V = predMap[i].first)
+ Phi->addIncoming(V, predMap[i].second);
+ else
+ Phi->addIncoming(PREInstr, PREPred);
+ }
- PREInstr->insertBefore(PREPred->getTerminator());
- PREInstr->setName(CurInst->getName() + ".pre");
- PREInstr->setDebugLoc(CurInst->getDebugLoc());
- VN.add(PREInstr, ValNo);
- ++NumGVNPRE;
-
- // Update the availability map to include the new instruction.
- addToLeaderTable(ValNo, PREInstr, PREPred);
-
- // Create a PHI to make the value available in this block.
- PHINode* Phi = PHINode::Create(CurInst->getType(), predMap.size(),
- CurInst->getName() + ".pre-phi",
- CurrentBlock->begin());
- for (unsigned i = 0, e = predMap.size(); i != e; ++i) {
- if (Value *V = predMap[i].first)
- Phi->addIncoming(V, predMap[i].second);
- else
- Phi->addIncoming(PREInstr, PREPred);
- }
+ VN.add(Phi, ValNo);
+ addToLeaderTable(ValNo, Phi, CurrentBlock);
+ Phi->setDebugLoc(CurInst->getDebugLoc());
+ CurInst->replaceAllUsesWith(Phi);
+ if (MD && Phi->getType()->getScalarType()->isPointerTy())
+ MD->invalidateCachedPointerInfo(Phi);
+ VN.erase(CurInst);
+ removeFromLeaderTable(ValNo, CurInst, CurrentBlock);
- VN.add(Phi, ValNo);
- addToLeaderTable(ValNo, Phi, CurrentBlock);
- Phi->setDebugLoc(CurInst->getDebugLoc());
- CurInst->replaceAllUsesWith(Phi);
- if (Phi->getType()->getScalarType()->isPointerTy()) {
- // Because we have added a PHI-use of the pointer value, it has now
- // "escaped" from alias analysis' perspective. We need to inform
- // AA of this.
- for (unsigned ii = 0, ee = Phi->getNumIncomingValues(); ii != ee;
- ++ii) {
- unsigned jj = PHINode::getOperandNumForIncomingValue(ii);
- VN.getAliasAnalysis()->addEscapingUse(Phi->getOperandUse(jj));
- }
+ DEBUG(dbgs() << "GVN PRE removed: " << *CurInst << '\n');
+ if (MD)
+ MD->removeInstruction(CurInst);
+ DEBUG(verifyRemoved(CurInst));
+ CurInst->eraseFromParent();
+ ++NumGVNInstr;
+
+ return true;
+}
- if (MD)
- MD->invalidateCachedPointerInfo(Phi);
- }
- VN.erase(CurInst);
- removeFromLeaderTable(ValNo, CurInst, CurrentBlock);
+/// Perform a purely local form of PRE that looks for diamond
+/// control flow patterns and attempts to perform simple PRE at the join point.
+bool GVN::performPRE(Function &F) {
+ bool Changed = false;
+ for (BasicBlock *CurrentBlock : depth_first(&F.getEntryBlock())) {
+ // Nothing to PRE in the entry block.
+ if (CurrentBlock == &F.getEntryBlock())
+ continue;
- DEBUG(dbgs() << "GVN PRE removed: " << *CurInst << '\n');
- if (MD) MD->removeInstruction(CurInst);
- DEBUG(verifyRemoved(CurInst));
- CurInst->eraseFromParent();
- Changed = true;
+ // Don't perform PRE on an EH pad.
+ if (CurrentBlock->isEHPad())
+ continue;
+
+ for (BasicBlock::iterator BI = CurrentBlock->begin(),
+ BE = CurrentBlock->end();
+ BI != BE;) {
+ Instruction *CurInst = &*BI++;
+ Changed = performScalarPRE(CurInst);
}
}
/// Split the critical edge connecting the given two blocks, and return
/// the block inserted to the critical edge.
BasicBlock *GVN::splitCriticalEdges(BasicBlock *Pred, BasicBlock *Succ) {
- BasicBlock *BB = SplitCriticalEdge(Pred, Succ, this);
+ BasicBlock *BB =
+ SplitCriticalEdge(Pred, Succ, CriticalEdgeSplittingOptions(DT));
if (MD)
MD->invalidateCachedPredecessors();
return BB;
}
-/// splitCriticalEdges - Split critical edges found during the previous
+/// Split critical edges found during the previous
/// iteration that may enable further optimization.
bool GVN::splitCriticalEdges() {
if (toSplit.empty())
return false;
do {
std::pair<TerminatorInst*, unsigned> Edge = toSplit.pop_back_val();
- SplitCriticalEdge(Edge.first, Edge.second, this);
+ SplitCriticalEdge(Edge.first, Edge.second,
+ CriticalEdgeSplittingOptions(DT));
} while (!toSplit.empty());
if (MD) MD->invalidateCachedPredecessors();
return true;
}
-/// iterateOnFunction - Executes one iteration of GVN
+/// Executes one iteration of GVN
bool GVN::iterateOnFunction(Function &F) {
cleanupGlobalSets();
// Top-down walk of the dominator tree
bool Changed = false;
-#if 0
- // Needed for value numbering with phi construction to work.
- ReversePostOrderTraversal<Function*> RPOT(&F);
- for (ReversePostOrderTraversal<Function*>::rpo_iterator RI = RPOT.begin(),
- RE = RPOT.end(); RI != RE; ++RI)
- Changed |= processBlock(*RI);
-#else
// Save the blocks this function have before transformation begins. GVN may
// split critical edge, and hence may invalidate the RPO/DT iterator.
//
std::vector<BasicBlock *> BBVect;
BBVect.reserve(256);
- for (DomTreeNode *x : depth_first(DT->getRootNode()))
- BBVect.push_back(x->getBlock());
+ // Needed for value numbering with phi construction to work.
+ ReversePostOrderTraversal<Function *> RPOT(&F);
+ for (ReversePostOrderTraversal<Function *>::rpo_iterator RI = RPOT.begin(),
+ RE = RPOT.end();
+ RI != RE; ++RI)
+ BBVect.push_back(*RI);
for (std::vector<BasicBlock *>::iterator I = BBVect.begin(), E = BBVect.end();
I != E; I++)
Changed |= processBlock(*I);
-#endif
return Changed;
}
TableAllocator.Reset();
}
-/// verifyRemoved - Verify that the specified instruction does not occur in our
+/// Verify that the specified instruction does not occur in our
/// internal data structures.
void GVN::verifyRemoved(const Instruction *Inst) const {
VN.verifyRemoved(Inst);
}
}
-// BB is declared dead, which implied other blocks become dead as well. This
-// function is to add all these blocks to "DeadBlocks". For the dead blocks'
-// live successors, update their phi nodes by replacing the operands
-// corresponding to dead blocks with UndefVal.
-//
+/// BB is declared dead, which implied other blocks become dead as well. This
+/// function is to add all these blocks to "DeadBlocks". For the dead blocks'
+/// live successors, update their phi nodes by replacing the operands
+/// corresponding to dead blocks with UndefVal.
void GVN::addDeadBlock(BasicBlock *BB) {
SmallVector<BasicBlock *, 4> NewDead;
SmallSetVector<BasicBlock *, 4> DF;
// R be the target of the dead out-coming edge.
// 1) Identify the set of dead blocks implied by the branch's dead outcoming
// edge. The result of this step will be {X| X is dominated by R}
-// 2) Identify those blocks which haves at least one dead prodecessor. The
+// 2) Identify those blocks which haves at least one dead predecessor. The
// result of this step will be dominance-frontier(R).
// 3) Update the PHIs in DF(R) by replacing the operands corresponding to
// dead blocks with "UndefVal" in an hope these PHIs will optimized away.
if (!BI || BI->isUnconditional())
return false;
+ // If a branch has two identical successors, we cannot declare either dead.
+ if (BI->getSuccessor(0) == BI->getSuccessor(1))
+ return false;
+
ConstantInt *Cond = dyn_cast<ConstantInt>(BI->getCondition());
if (!Cond)
return false;
return true;
}
-// performPRE() will trigger assert if it come across an instruciton without
+// performPRE() will trigger assert if it comes across an instruction without
// associated val-num. As it normally has far more live instructions than dead
// instructions, it makes more sense just to "fabricate" a val-number for the
// dead code than checking if instruction involved is dead or not.