//
//===----------------------------------------------------------------------===//
-#define DEBUG_TYPE "inline-cost"
#include "llvm/Analysis/InlineCost.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SetVector.h"
#include "llvm/ADT/SmallPtrSet.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/ADT/Statistic.h"
+#include "llvm/Analysis/AssumptionCache.h"
+#include "llvm/Analysis/CodeMetrics.h"
#include "llvm/Analysis/ConstantFolding.h"
#include "llvm/Analysis/InstructionSimplify.h"
#include "llvm/Analysis/TargetTransformInfo.h"
+#include "llvm/IR/CallSite.h"
#include "llvm/IR/CallingConv.h"
#include "llvm/IR/DataLayout.h"
+#include "llvm/IR/GetElementPtrTypeIterator.h"
#include "llvm/IR/GlobalAlias.h"
+#include "llvm/IR/InstVisitor.h"
#include "llvm/IR/IntrinsicInst.h"
#include "llvm/IR/Operator.h"
-#include "llvm/InstVisitor.h"
-#include "llvm/Support/CallSite.h"
#include "llvm/Support/Debug.h"
-#include "llvm/Support/GetElementPtrTypeIterator.h"
#include "llvm/Support/raw_ostream.h"
using namespace llvm;
+#define DEBUG_TYPE "inline-cost"
+
STATISTIC(NumCallsAnalyzed, "Number of call sites analyzed");
namespace {
typedef InstVisitor<CallAnalyzer, bool> Base;
friend class InstVisitor<CallAnalyzer, bool>;
- // DataLayout if available, or null.
- const DataLayout *const TD;
-
/// The TargetTransformInfo available for this compilation.
const TargetTransformInfo &TTI;
+ /// The cache of @llvm.assume intrinsics.
+ AssumptionCacheTracker *ACT;
+
// The called function.
Function &F;
+ // The candidate callsite being analyzed. Please do not use this to do
+ // analysis in the caller function; we want the inline cost query to be
+ // easily cacheable. Instead, use the cover function paramHasAttr.
+ CallSite CandidateCS;
+
int Threshold;
int Cost;
bool ExposesReturnsTwice;
bool HasDynamicAlloca;
bool ContainsNoDuplicateCall;
+ bool HasReturn;
+ bool HasIndirectBr;
+ bool HasFrameEscape;
/// Number of bytes allocated statically by the callee.
uint64_t AllocatedSize;
void disableSROA(Value *V);
void accumulateSROACost(DenseMap<Value *, int>::iterator CostIt,
int InstructionCost);
- bool handleSROACandidate(bool IsSROAValid,
- DenseMap<Value *, int>::iterator CostIt,
- int InstructionCost);
bool isGEPOffsetConstant(GetElementPtrInst &GEP);
bool accumulateGEPOffset(GEPOperator &GEP, APInt &Offset);
bool simplifyCallSite(Function *F, CallSite CS);
ConstantInt *stripAndComputeInBoundsConstantOffsets(Value *&V);
+ /// Return true if the given argument to the function being considered for
+ /// inlining has the given attribute set either at the call site or the
+ /// function declaration. Primarily used to inspect call site specific
+ /// attributes since these can be more precise than the ones on the callee
+ /// itself.
+ bool paramHasAttr(Argument *A, Attribute::AttrKind Attr);
+
+ /// Return true if the given value is known non null within the callee if
+ /// inlined through this particular callsite.
+ bool isKnownNonNullInCallee(Value *V);
+
// Custom analysis routines.
- bool analyzeBlock(BasicBlock *BB);
+ bool analyzeBlock(BasicBlock *BB, SmallPtrSetImpl<const Value *> &EphValues);
// Disable several entry points to the visitor so we don't accidentally use
// them by declaring but not defining them here.
bool visitIntToPtr(IntToPtrInst &I);
bool visitCastInst(CastInst &I);
bool visitUnaryInstruction(UnaryInstruction &I);
- bool visitICmp(ICmpInst &I);
+ bool visitCmpInst(CmpInst &I);
bool visitSub(BinaryOperator &I);
bool visitBinaryOperator(BinaryOperator &I);
bool visitLoad(LoadInst &I);
bool visitExtractValue(ExtractValueInst &I);
bool visitInsertValue(InsertValueInst &I);
bool visitCallSite(CallSite CS);
+ bool visitReturnInst(ReturnInst &RI);
+ bool visitBranchInst(BranchInst &BI);
+ bool visitSwitchInst(SwitchInst &SI);
+ bool visitIndirectBrInst(IndirectBrInst &IBI);
+ bool visitResumeInst(ResumeInst &RI);
+ bool visitCleanupReturnInst(CleanupReturnInst &RI);
+ bool visitCatchReturnInst(CatchReturnInst &RI);
+ bool visitUnreachableInst(UnreachableInst &I);
public:
- CallAnalyzer(const DataLayout *TD, const TargetTransformInfo &TTI,
- Function &Callee, int Threshold)
- : TD(TD), TTI(TTI), F(Callee), Threshold(Threshold), Cost(0),
- IsCallerRecursive(false), IsRecursiveCall(false),
+ CallAnalyzer(const TargetTransformInfo &TTI, AssumptionCacheTracker *ACT,
+ Function &Callee, int Threshold, CallSite CSArg)
+ : TTI(TTI), ACT(ACT), F(Callee), CandidateCS(CSArg), Threshold(Threshold),
+ Cost(0), IsCallerRecursive(false), IsRecursiveCall(false),
ExposesReturnsTwice(false), HasDynamicAlloca(false),
- ContainsNoDuplicateCall(false), AllocatedSize(0), NumInstructions(0),
+ ContainsNoDuplicateCall(false), HasReturn(false), HasIndirectBr(false),
+ HasFrameEscape(false), AllocatedSize(0), NumInstructions(0),
NumVectorInstructions(0), FiftyPercentVectorBonus(0),
TenPercentVectorBonus(0), VectorBonus(0), NumConstantArgs(0),
NumConstantOffsetPtrArgs(0), NumAllocaArgs(0), NumConstantPtrCmps(0),
SROACostSavings += InstructionCost;
}
-/// \brief Helper for the common pattern of handling a SROA candidate.
-/// Either accumulates the cost savings if the SROA remains valid, or disables
-/// SROA for the candidate.
-bool CallAnalyzer::handleSROACandidate(bool IsSROAValid,
- DenseMap<Value *, int>::iterator CostIt,
- int InstructionCost) {
- if (IsSROAValid) {
- accumulateSROACost(CostIt, InstructionCost);
- return true;
- }
-
- disableSROA(CostIt);
- return false;
-}
-
/// \brief Check whether a GEP's indices are all constant.
///
/// Respects any simplified values known during the analysis of this callsite.
/// Returns false if unable to compute the offset for any reason. Respects any
/// simplified values known during the analysis of this callsite.
bool CallAnalyzer::accumulateGEPOffset(GEPOperator &GEP, APInt &Offset) {
- if (!TD)
- return false;
-
- unsigned IntPtrWidth = TD->getPointerSizeInBits();
+ const DataLayout &DL = F.getParent()->getDataLayout();
+ unsigned IntPtrWidth = DL.getPointerSizeInBits();
assert(IntPtrWidth == Offset.getBitWidth());
for (gep_type_iterator GTI = gep_type_begin(GEP), GTE = gep_type_end(GEP);
// Handle a struct index, which adds its field offset to the pointer.
if (StructType *STy = dyn_cast<StructType>(*GTI)) {
unsigned ElementIdx = OpC->getZExtValue();
- const StructLayout *SL = TD->getStructLayout(STy);
+ const StructLayout *SL = DL.getStructLayout(STy);
Offset += APInt(IntPtrWidth, SL->getElementOffset(ElementIdx));
continue;
}
- APInt TypeSize(IntPtrWidth, TD->getTypeAllocSize(GTI.getIndexedType()));
+ APInt TypeSize(IntPtrWidth, DL.getTypeAllocSize(GTI.getIndexedType()));
Offset += OpC->getValue().sextOrTrunc(IntPtrWidth) * TypeSize;
}
return true;
}
bool CallAnalyzer::visitAlloca(AllocaInst &I) {
- // FIXME: Check whether inlining will turn a dynamic alloca into a static
+ // Check whether inlining will turn a dynamic alloca into a static
// alloca, and handle that case.
+ if (I.isArrayAllocation()) {
+ if (Constant *Size = SimplifiedValues.lookup(I.getArraySize())) {
+ ConstantInt *AllocSize = dyn_cast<ConstantInt>(Size);
+ assert(AllocSize && "Allocation size not a constant int?");
+ Type *Ty = I.getAllocatedType();
+ AllocatedSize += Ty->getPrimitiveSizeInBits() * AllocSize->getZExtValue();
+ return Base::visitAlloca(I);
+ }
+ }
// Accumulate the allocated size.
if (I.isStaticAlloca()) {
+ const DataLayout &DL = F.getParent()->getDataLayout();
Type *Ty = I.getAllocatedType();
- AllocatedSize += (TD ? TD->getTypeAllocSize(Ty) :
- Ty->getPrimitiveSizeInBits());
+ AllocatedSize += DL.getTypeAllocSize(Ty);
}
// We will happily inline static alloca instructions.
// Try to fold GEPs of constant-offset call site argument pointers. This
// requires target data and inbounds GEPs.
- if (TD && I.isInBounds()) {
+ if (I.isInBounds()) {
// Check if we have a base + offset for the pointer.
Value *Ptr = I.getPointerOperand();
std::pair<Value *, APInt> BaseAndOffset = ConstantOffsetPtrs.lookup(Ptr);
// Track base/offset pairs when converted to a plain integer provided the
// integer is large enough to represent the pointer.
unsigned IntegerSize = I.getType()->getScalarSizeInBits();
- if (TD && IntegerSize >= TD->getPointerSizeInBits()) {
+ const DataLayout &DL = F.getParent()->getDataLayout();
+ if (IntegerSize >= DL.getPointerSizeInBits()) {
std::pair<Value *, APInt> BaseAndOffset
= ConstantOffsetPtrs.lookup(I.getOperand(0));
if (BaseAndOffset.first)
if (lookupSROAArgAndCost(I.getOperand(0), SROAArg, CostIt))
SROAArgValues[&I] = SROAArg;
- return isInstructionFree(&I, TD);
+ return TargetTransformInfo::TCC_Free == TTI.getUserCost(&I);
}
bool CallAnalyzer::visitIntToPtr(IntToPtrInst &I) {
// modifications provided the integer is not too large.
Value *Op = I.getOperand(0);
unsigned IntegerSize = Op->getType()->getScalarSizeInBits();
- if (TD && IntegerSize <= TD->getPointerSizeInBits()) {
+ const DataLayout &DL = F.getParent()->getDataLayout();
+ if (IntegerSize <= DL.getPointerSizeInBits()) {
std::pair<Value *, APInt> BaseAndOffset = ConstantOffsetPtrs.lookup(Op);
if (BaseAndOffset.first)
ConstantOffsetPtrs[&I] = BaseAndOffset;
if (lookupSROAArgAndCost(Op, SROAArg, CostIt))
SROAArgValues[&I] = SROAArg;
- return isInstructionFree(&I, TD);
+ return TargetTransformInfo::TCC_Free == TTI.getUserCost(&I);
}
bool CallAnalyzer::visitCastInst(CastInst &I) {
// Disable SROA in the face of arbitrary casts we don't whitelist elsewhere.
disableSROA(I.getOperand(0));
- return isInstructionFree(&I, TD);
+ return TargetTransformInfo::TCC_Free == TTI.getUserCost(&I);
}
bool CallAnalyzer::visitUnaryInstruction(UnaryInstruction &I) {
Value *Operand = I.getOperand(0);
- Constant *Ops[1] = { dyn_cast<Constant>(Operand) };
- if (Ops[0] || (Ops[0] = SimplifiedValues.lookup(Operand)))
+ Constant *COp = dyn_cast<Constant>(Operand);
+ if (!COp)
+ COp = SimplifiedValues.lookup(Operand);
+ if (COp) {
+ const DataLayout &DL = F.getParent()->getDataLayout();
if (Constant *C = ConstantFoldInstOperands(I.getOpcode(), I.getType(),
- Ops, TD)) {
+ COp, DL)) {
SimplifiedValues[&I] = C;
return true;
}
+ }
// Disable any SROA on the argument to arbitrary unary operators.
disableSROA(Operand);
return false;
}
-bool CallAnalyzer::visitICmp(ICmpInst &I) {
+bool CallAnalyzer::paramHasAttr(Argument *A, Attribute::AttrKind Attr) {
+ unsigned ArgNo = A->getArgNo();
+ return CandidateCS.paramHasAttr(ArgNo+1, Attr);
+}
+
+bool CallAnalyzer::isKnownNonNullInCallee(Value *V) {
+ // Does the *call site* have the NonNull attribute set on an argument? We
+ // use the attribute on the call site to memoize any analysis done in the
+ // caller. This will also trip if the callee function has a non-null
+ // parameter attribute, but that's a less interesting case because hopefully
+ // the callee would already have been simplified based on that.
+ if (Argument *A = dyn_cast<Argument>(V))
+ if (paramHasAttr(A, Attribute::NonNull))
+ return true;
+
+ // Is this an alloca in the caller? This is distinct from the attribute case
+ // above because attributes aren't updated within the inliner itself and we
+ // always want to catch the alloca derived case.
+ if (isAllocaDerivedArg(V))
+ // We can actually predict the result of comparisons between an
+ // alloca-derived value and null. Note that this fires regardless of
+ // SROA firing.
+ return true;
+
+ return false;
+}
+
+bool CallAnalyzer::visitCmpInst(CmpInst &I) {
Value *LHS = I.getOperand(0), *RHS = I.getOperand(1);
// First try to handle simplified comparisons.
if (!isa<Constant>(LHS))
if (!isa<Constant>(RHS))
if (Constant *SimpleRHS = SimplifiedValues.lookup(RHS))
RHS = SimpleRHS;
- if (Constant *CLHS = dyn_cast<Constant>(LHS))
+ if (Constant *CLHS = dyn_cast<Constant>(LHS)) {
if (Constant *CRHS = dyn_cast<Constant>(RHS))
- if (Constant *C = ConstantExpr::getICmp(I.getPredicate(), CLHS, CRHS)) {
+ if (Constant *C = ConstantExpr::getCompare(I.getPredicate(), CLHS, CRHS)) {
SimplifiedValues[&I] = C;
return true;
}
+ }
+
+ if (I.getOpcode() == Instruction::FCmp)
+ return false;
// Otherwise look for a comparison between constant offset pointers with
// a common base.
Value *LHSBase, *RHSBase;
APInt LHSOffset, RHSOffset;
- llvm::tie(LHSBase, LHSOffset) = ConstantOffsetPtrs.lookup(LHS);
+ std::tie(LHSBase, LHSOffset) = ConstantOffsetPtrs.lookup(LHS);
if (LHSBase) {
- llvm::tie(RHSBase, RHSOffset) = ConstantOffsetPtrs.lookup(RHS);
+ std::tie(RHSBase, RHSOffset) = ConstantOffsetPtrs.lookup(RHS);
if (RHSBase && LHSBase == RHSBase) {
// We have common bases, fold the icmp to a constant based on the
// offsets.
}
// If the comparison is an equality comparison with null, we can simplify it
- // for any alloca-derived argument.
- if (I.isEquality() && isa<ConstantPointerNull>(I.getOperand(1)))
- if (isAllocaDerivedArg(I.getOperand(0))) {
- // We can actually predict the result of comparisons between an
- // alloca-derived value and null. Note that this fires regardless of
- // SROA firing.
- bool IsNotEqual = I.getPredicate() == CmpInst::ICMP_NE;
- SimplifiedValues[&I] = IsNotEqual ? ConstantInt::getTrue(I.getType())
- : ConstantInt::getFalse(I.getType());
- return true;
- }
-
+ // if we know the value (argument) can't be null
+ if (I.isEquality() && isa<ConstantPointerNull>(I.getOperand(1)) &&
+ isKnownNonNullInCallee(I.getOperand(0))) {
+ bool IsNotEqual = I.getPredicate() == CmpInst::ICMP_NE;
+ SimplifiedValues[&I] = IsNotEqual ? ConstantInt::getTrue(I.getType())
+ : ConstantInt::getFalse(I.getType());
+ return true;
+ }
// Finally check for SROA candidates in comparisons.
Value *SROAArg;
DenseMap<Value *, int>::iterator CostIt;
Value *LHS = I.getOperand(0), *RHS = I.getOperand(1);
Value *LHSBase, *RHSBase;
APInt LHSOffset, RHSOffset;
- llvm::tie(LHSBase, LHSOffset) = ConstantOffsetPtrs.lookup(LHS);
+ std::tie(LHSBase, LHSOffset) = ConstantOffsetPtrs.lookup(LHS);
if (LHSBase) {
- llvm::tie(RHSBase, RHSOffset) = ConstantOffsetPtrs.lookup(RHS);
+ std::tie(RHSBase, RHSOffset) = ConstantOffsetPtrs.lookup(RHS);
if (RHSBase && LHSBase == RHSBase) {
// We have common bases, fold the subtract to a constant based on the
// offsets.
bool CallAnalyzer::visitBinaryOperator(BinaryOperator &I) {
Value *LHS = I.getOperand(0), *RHS = I.getOperand(1);
+ const DataLayout &DL = F.getParent()->getDataLayout();
if (!isa<Constant>(LHS))
if (Constant *SimpleLHS = SimplifiedValues.lookup(LHS))
LHS = SimpleLHS;
if (!isa<Constant>(RHS))
if (Constant *SimpleRHS = SimplifiedValues.lookup(RHS))
RHS = SimpleRHS;
- Value *SimpleV = SimplifyBinOp(I.getOpcode(), LHS, RHS, TD);
+ Value *SimpleV = nullptr;
+ if (auto FI = dyn_cast<FPMathOperator>(&I))
+ SimpleV =
+ SimplifyFPBinOp(I.getOpcode(), LHS, RHS, FI->getFastMathFlags(), DL);
+ else
+ SimpleV = SimplifyBinOp(I.getOpcode(), LHS, RHS, DL);
+
if (Constant *C = dyn_cast_or_null<Constant>(SimpleV)) {
SimplifiedValues[&I] = C;
return true;
bool CallAnalyzer::visitLoad(LoadInst &I) {
Value *SROAArg;
DenseMap<Value *, int>::iterator CostIt;
- if (lookupSROAArgAndCost(I.getOperand(0), SROAArg, CostIt)) {
+ if (lookupSROAArgAndCost(I.getPointerOperand(), SROAArg, CostIt)) {
if (I.isSimple()) {
accumulateSROACost(CostIt, InlineConstants::InstrCost);
return true;
bool CallAnalyzer::visitStore(StoreInst &I) {
Value *SROAArg;
DenseMap<Value *, int>::iterator CostIt;
- if (lookupSROAArgAndCost(I.getOperand(0), SROAArg, CostIt)) {
+ if (lookupSROAArgAndCost(I.getPointerOperand(), SROAArg, CostIt)) {
if (I.isSimple()) {
accumulateSROACost(CostIt, InlineConstants::InstrCost);
return true;
}
bool CallAnalyzer::visitCallSite(CallSite CS) {
- if (CS.isCall() && cast<CallInst>(CS.getInstruction())->canReturnTwice() &&
- !F.getAttributes().hasAttribute(AttributeSet::FunctionIndex,
- Attribute::ReturnsTwice)) {
+ if (CS.hasFnAttr(Attribute::ReturnsTwice) &&
+ !F.hasFnAttribute(Attribute::ReturnsTwice)) {
// This aborts the entire analysis.
ExposesReturnsTwice = true;
return false;
}
if (CS.isCall() &&
- cast<CallInst>(CS.getInstruction())->hasFnAttr(Attribute::NoDuplicate))
+ cast<CallInst>(CS.getInstruction())->cannotDuplicate())
ContainsNoDuplicateCall = true;
if (Function *F = CS.getCalledFunction()) {
case Intrinsic::memmove:
// SROA can usually chew through these intrinsics, but they aren't free.
return false;
+ case Intrinsic::localescape:
+ HasFrameEscape = true;
+ return false;
}
}
return false;
}
- if (!callIsSmall(CS)) {
+ if (TTI.isLoweredToCall(F)) {
// We account for the average 1 instruction per call argument setup
// here.
Cost += CS.arg_size() * InlineConstants::InstrCost;
// during devirtualization and so we want to give it a hefty bonus for
// inlining, but cap that bonus in the event that inlining wouldn't pan
// out. Pretend to inline the function, with a custom threshold.
- CallAnalyzer CA(TD, TTI, *F, InlineConstants::IndirectCallThreshold);
+ CallAnalyzer CA(TTI, ACT, *F, InlineConstants::IndirectCallThreshold, CS);
if (CA.analyzeCall(CS)) {
// We were able to inline the indirect call! Subtract the cost from the
// bonus we want to apply, but don't go below zero.
return Base::visitCallSite(CS);
}
+bool CallAnalyzer::visitReturnInst(ReturnInst &RI) {
+ // At least one return instruction will be free after inlining.
+ bool Free = !HasReturn;
+ HasReturn = true;
+ return Free;
+}
+
+bool CallAnalyzer::visitBranchInst(BranchInst &BI) {
+ // We model unconditional branches as essentially free -- they really
+ // shouldn't exist at all, but handling them makes the behavior of the
+ // inliner more regular and predictable. Interestingly, conditional branches
+ // which will fold away are also free.
+ return BI.isUnconditional() || isa<ConstantInt>(BI.getCondition()) ||
+ dyn_cast_or_null<ConstantInt>(
+ SimplifiedValues.lookup(BI.getCondition()));
+}
+
+bool CallAnalyzer::visitSwitchInst(SwitchInst &SI) {
+ // We model unconditional switches as free, see the comments on handling
+ // branches.
+ if (isa<ConstantInt>(SI.getCondition()))
+ return true;
+ if (Value *V = SimplifiedValues.lookup(SI.getCondition()))
+ if (isa<ConstantInt>(V))
+ return true;
+
+ // Otherwise, we need to accumulate a cost proportional to the number of
+ // distinct successor blocks. This fan-out in the CFG cannot be represented
+ // for free even if we can represent the core switch as a jumptable that
+ // takes a single instruction.
+ //
+ // NB: We convert large switches which are just used to initialize large phi
+ // nodes to lookup tables instead in simplify-cfg, so this shouldn't prevent
+ // inlining those. It will prevent inlining in cases where the optimization
+ // does not (yet) fire.
+ SmallPtrSet<BasicBlock *, 8> SuccessorBlocks;
+ SuccessorBlocks.insert(SI.getDefaultDest());
+ for (auto I = SI.case_begin(), E = SI.case_end(); I != E; ++I)
+ SuccessorBlocks.insert(I.getCaseSuccessor());
+ // Add cost corresponding to the number of distinct destinations. The first
+ // we model as free because of fallthrough.
+ Cost += (SuccessorBlocks.size() - 1) * InlineConstants::InstrCost;
+ return false;
+}
+
+bool CallAnalyzer::visitIndirectBrInst(IndirectBrInst &IBI) {
+ // We never want to inline functions that contain an indirectbr. This is
+ // incorrect because all the blockaddress's (in static global initializers
+ // for example) would be referring to the original function, and this
+ // indirect jump would jump from the inlined copy of the function into the
+ // original function which is extremely undefined behavior.
+ // FIXME: This logic isn't really right; we can safely inline functions with
+ // indirectbr's as long as no other function or global references the
+ // blockaddress of a block within the current function.
+ HasIndirectBr = true;
+ return false;
+}
+
+bool CallAnalyzer::visitResumeInst(ResumeInst &RI) {
+ // FIXME: It's not clear that a single instruction is an accurate model for
+ // the inline cost of a resume instruction.
+ return false;
+}
+
+bool CallAnalyzer::visitCleanupReturnInst(CleanupReturnInst &CRI) {
+ // FIXME: It's not clear that a single instruction is an accurate model for
+ // the inline cost of a cleanupret instruction.
+ return false;
+}
+
+bool CallAnalyzer::visitCatchReturnInst(CatchReturnInst &CRI) {
+ // FIXME: It's not clear that a single instruction is an accurate model for
+ // the inline cost of a cleanupret instruction.
+ return false;
+}
+
+bool CallAnalyzer::visitUnreachableInst(UnreachableInst &I) {
+ // FIXME: It might be reasonably to discount the cost of instructions leading
+ // to unreachable as they have the lowest possible impact on both runtime and
+ // code size.
+ return true; // No actual code is needed for unreachable.
+}
+
bool CallAnalyzer::visitInstruction(Instruction &I) {
// Some instructions are free. All of the free intrinsics can also be
// handled by SROA, etc.
- if (isInstructionFree(&I, TD))
+ if (TargetTransformInfo::TCC_Free == TTI.getUserCost(&I))
return true;
// We found something we don't understand or can't handle. Mark any SROA-able
/// aborts early if the threshold has been exceeded or an impossible to inline
/// construct has been detected. It returns false if inlining is no longer
/// viable, and true if inlining remains viable.
-bool CallAnalyzer::analyzeBlock(BasicBlock *BB) {
- for (BasicBlock::iterator I = BB->begin(), E = llvm::prior(BB->end());
- I != E; ++I) {
+bool CallAnalyzer::analyzeBlock(BasicBlock *BB,
+ SmallPtrSetImpl<const Value *> &EphValues) {
+ for (BasicBlock::iterator I = BB->begin(), E = BB->end(); I != E; ++I) {
+ // FIXME: Currently, the number of instructions in a function regardless of
+ // our ability to simplify them during inline to constants or dead code,
+ // are actually used by the vector bonus heuristic. As long as that's true,
+ // we have to special case debug intrinsics here to prevent differences in
+ // inlining due to debug symbols. Eventually, the number of unsimplified
+ // instructions shouldn't factor into the cost computation, but until then,
+ // hack around it here.
+ if (isa<DbgInfoIntrinsic>(I))
+ continue;
+
+ // Skip ephemeral values.
+ if (EphValues.count(I))
+ continue;
+
++NumInstructions;
if (isa<ExtractElementInst>(I) || I->getType()->isVectorTy())
++NumVectorInstructions;
+ // If the instruction is floating point, and the target says this operation is
+ // expensive or the function has the "use-soft-float" attribute, this may
+ // eventually become a library call. Treat the cost as such.
+ if (I->getType()->isFloatingPointTy()) {
+ bool hasSoftFloatAttr = false;
+
+ // If the function has the "use-soft-float" attribute, mark it as expensive.
+ if (F.hasFnAttribute("use-soft-float")) {
+ Attribute Attr = F.getFnAttribute("use-soft-float");
+ StringRef Val = Attr.getValueAsString();
+ if (Val == "true")
+ hasSoftFloatAttr = true;
+ }
+
+ if (TTI.getFPOpCost(I->getType()) == TargetTransformInfo::TCC_Expensive ||
+ hasSoftFloatAttr)
+ Cost += InlineConstants::CallPenalty;
+ }
+
// If the instruction simplified to a constant, there is no cost to this
// instruction. Visit the instructions using our InstVisitor to account for
// all of the per-instruction logic. The visit tree returns true if we
Cost += InlineConstants::InstrCost;
// If the visit this instruction detected an uninlinable pattern, abort.
- if (IsRecursiveCall || ExposesReturnsTwice || HasDynamicAlloca)
+ if (IsRecursiveCall || ExposesReturnsTwice || HasDynamicAlloca ||
+ HasIndirectBr || HasFrameEscape)
return false;
// If the caller is a recursive function then we don't want to inline
AllocatedSize > InlineConstants::TotalAllocaSizeRecursiveCaller)
return false;
- if (NumVectorInstructions > NumInstructions/2)
- VectorBonus = FiftyPercentVectorBonus;
- else if (NumVectorInstructions > NumInstructions/10)
- VectorBonus = TenPercentVectorBonus;
- else
- VectorBonus = 0;
-
- // Check if we've past the threshold so we don't spin in huge basic
- // blocks that will never inline.
- if (Cost > (Threshold + VectorBonus))
+ // Check if we've past the maximum possible threshold so we don't spin in
+ // huge basic blocks that will never inline.
+ if (Cost > Threshold)
return false;
}
/// returns 0 if V is not a pointer, and returns the constant '0' if there are
/// no constant offsets applied.
ConstantInt *CallAnalyzer::stripAndComputeInBoundsConstantOffsets(Value *&V) {
- if (!TD || !V->getType()->isPointerTy())
- return 0;
+ if (!V->getType()->isPointerTy())
+ return nullptr;
- unsigned IntPtrWidth = TD->getPointerSizeInBits();
+ const DataLayout &DL = F.getParent()->getDataLayout();
+ unsigned IntPtrWidth = DL.getPointerSizeInBits();
APInt Offset = APInt::getNullValue(IntPtrWidth);
// Even though we don't look through PHI nodes, we could be called on an
do {
if (GEPOperator *GEP = dyn_cast<GEPOperator>(V)) {
if (!GEP->isInBounds() || !accumulateGEPOffset(*GEP, Offset))
- return 0;
+ return nullptr;
V = GEP->getPointerOperand();
} else if (Operator::getOpcode(V) == Instruction::BitCast) {
V = cast<Operator>(V)->getOperand(0);
break;
}
assert(V->getType()->isPointerTy() && "Unexpected operand type!");
- } while (Visited.insert(V));
+ } while (Visited.insert(V).second);
- Type *IntPtrTy = TD->getIntPtrType(V->getContext());
+ Type *IntPtrTy = DL.getIntPtrType(V->getContext());
return cast<ConstantInt>(ConstantInt::get(IntPtrTy, Offset));
}
bool CallAnalyzer::analyzeCall(CallSite CS) {
++NumCallsAnalyzed;
- // Track whether the post-inlining function would have more than one basic
- // block. A single basic block is often intended for inlining. Balloon the
- // threshold by 50% until we pass the single-BB phase.
- bool SingleBB = true;
- int SingleBBBonus = Threshold / 2;
- Threshold += SingleBBBonus;
-
// Perform some tweaks to the cost and threshold based on the direct
// callsite information.
// We want to more aggressively inline vector-dense kernels, so up the
// threshold, and we'll lower it if the % of vector instructions gets too
- // low.
+ // low. Note that these bonuses are some what arbitrary and evolved over time
+ // by accident as much as because they are principled bonuses.
+ //
+ // FIXME: It would be nice to remove all such bonuses. At least it would be
+ // nice to base the bonus values on something more scientific.
assert(NumInstructions == 0);
assert(NumVectorInstructions == 0);
- FiftyPercentVectorBonus = Threshold;
- TenPercentVectorBonus = Threshold / 2;
+ FiftyPercentVectorBonus = 3 * Threshold / 2;
+ TenPercentVectorBonus = 3 * Threshold / 4;
+ const DataLayout &DL = F.getParent()->getDataLayout();
+
+ // Track whether the post-inlining function would have more than one basic
+ // block. A single basic block is often intended for inlining. Balloon the
+ // threshold by 50% until we pass the single-BB phase.
+ bool SingleBB = true;
+ int SingleBBBonus = Threshold / 2;
+
+ // Speculatively apply all possible bonuses to Threshold. If cost exceeds
+ // this Threshold any time, and cost cannot decrease, we can stop processing
+ // the rest of the function body.
+ Threshold += (SingleBBBonus + FiftyPercentVectorBonus);
// Give out bonuses per argument, as the instructions setting them up will
// be gone after inlining.
for (unsigned I = 0, E = CS.arg_size(); I != E; ++I) {
- if (TD && CS.isByValArgument(I)) {
+ if (CS.isByValArgument(I)) {
// We approximate the number of loads and stores needed by dividing the
// size of the byval type by the target's pointer size.
PointerType *PTy = cast<PointerType>(CS.getArgument(I)->getType());
- unsigned TypeSize = TD->getTypeSizeInBits(PTy->getElementType());
- unsigned PointerSize = TD->getPointerSizeInBits();
+ unsigned TypeSize = DL.getTypeSizeInBits(PTy->getElementType());
+ unsigned PointerSize = DL.getPointerSizeInBits();
// Ceiling division.
unsigned NumStores = (TypeSize + PointerSize - 1) / PointerSize;
Instruction *Instr = CS.getInstruction();
if (InvokeInst *II = dyn_cast<InvokeInst>(Instr)) {
if (isa<UnreachableInst>(II->getNormalDest()->begin()))
- Threshold = 1;
+ Threshold = 0;
} else if (isa<UnreachableInst>(++BasicBlock::iterator(Instr)))
- Threshold = 1;
+ Threshold = 0;
// If this function uses the coldcc calling convention, prefer not to inline
// it.
Function *Caller = CS.getInstruction()->getParent()->getParent();
// Check if the caller function is recursive itself.
- for (Value::use_iterator U = Caller->use_begin(), E = Caller->use_end();
- U != E; ++U) {
- CallSite Site(cast<Value>(*U));
+ for (User *U : Caller->users()) {
+ CallSite Site(U);
if (!Site)
continue;
Instruction *I = Site.getInstruction();
}
}
- // Track whether we've seen a return instruction. The first return
- // instruction is free, as at least one will usually disappear in inlining.
- bool HasReturn = false;
-
// Populate our simplified values by mapping from function arguments to call
// arguments with known important simplifications.
CallSite::arg_iterator CAI = CS.arg_begin();
NumConstantOffsetPtrArgs = ConstantOffsetPtrs.size();
NumAllocaArgs = SROAArgValues.size();
+ // FIXME: If a caller has multiple calls to a callee, we end up recomputing
+ // the ephemeral values multiple times (and they're completely determined by
+ // the callee, so this is purely duplicate work).
+ SmallPtrSet<const Value *, 32> EphValues;
+ CodeMetrics::collectEphemeralValues(&F, &ACT->getAssumptionCache(F), EphValues);
+
// The worklist of live basic blocks in the callee *after* inlining. We avoid
// adding basic blocks of the callee which can be proven to be dead for this
// particular call site in order to get more accurate cost estimates. This
for (unsigned Idx = 0; Idx != BBWorklist.size(); ++Idx) {
// Bail out the moment we cross the threshold. This means we'll under-count
// the cost, but only when undercounting doesn't matter.
- if (Cost > (Threshold + VectorBonus))
+ if (Cost > Threshold)
break;
BasicBlock *BB = BBWorklist[Idx];
if (BB->empty())
continue;
- // Handle the terminator cost here where we can track returns and other
- // function-wide constructs.
- TerminatorInst *TI = BB->getTerminator();
-
- // We never want to inline functions that contain an indirectbr. This is
- // incorrect because all the blockaddress's (in static global initializers
- // for example) would be referring to the original function, and this
- // indirect jump would jump from the inlined copy of the function into the
- // original function which is extremely undefined behavior.
- // FIXME: This logic isn't really right; we can safely inline functions
- // with indirectbr's as long as no other function or global references the
- // blockaddress of a block within the current function. And as a QOI issue,
- // if someone is using a blockaddress without an indirectbr, and that
- // reference somehow ends up in another function or global, we probably
- // don't want to inline this function.
- if (isa<IndirectBrInst>(TI))
+ // Disallow inlining a blockaddress. A blockaddress only has defined
+ // behavior for an indirect branch in the same function, and we do not
+ // currently support inlining indirect branches. But, the inliner may not
+ // see an indirect branch that ends up being dead code at a particular call
+ // site. If the blockaddress escapes the function, e.g., via a global
+ // variable, inlining may lead to an invalid cross-function reference.
+ if (BB->hasAddressTaken())
return false;
- if (!HasReturn && isa<ReturnInst>(TI))
- HasReturn = true;
- else
- Cost += InlineConstants::InstrCost;
-
// Analyze the cost of this block. If we blow through the threshold, this
// returns false, and we can bail on out.
- if (!analyzeBlock(BB)) {
- if (IsRecursiveCall || ExposesReturnsTwice || HasDynamicAlloca)
+ if (!analyzeBlock(BB, EphValues)) {
+ if (IsRecursiveCall || ExposesReturnsTwice || HasDynamicAlloca ||
+ HasIndirectBr || HasFrameEscape)
return false;
// If the caller is a recursive function then we don't want to inline
break;
}
+ TerminatorInst *TI = BB->getTerminator();
+
// Add in the live successors by first checking whether we have terminator
// that may be simplified based on the values simplified by this call.
if (BranchInst *BI = dyn_cast<BranchInst>(TI)) {
}
}
- // If this is a noduplicate call, we can still inline as long as
+ // If this is a noduplicate call, we can still inline as long as
// inlining this would cause the removal of the caller (so the instruction
// is not actually duplicated, just moved).
if (!OnlyOneCallAndLocalLinkage && ContainsNoDuplicateCall)
return false;
- Threshold += VectorBonus;
+ // We applied the maximum possible vector bonus at the beginning. Now,
+ // subtract the excess bonus, if any, from the Threshold before
+ // comparing against Cost.
+ if (NumVectorInstructions <= NumInstructions / 10)
+ Threshold -= FiftyPercentVectorBonus;
+ else if (NumVectorInstructions <= NumInstructions / 2)
+ Threshold -= (FiftyPercentVectorBonus - TenPercentVectorBonus);
return Cost < Threshold;
}
#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
/// \brief Dump stats about this call's analysis.
void CallAnalyzer::dump() {
-#define DEBUG_PRINT_STAT(x) llvm::dbgs() << " " #x ": " << x << "\n"
+#define DEBUG_PRINT_STAT(x) dbgs() << " " #x ": " << x << "\n"
DEBUG_PRINT_STAT(NumConstantArgs);
DEBUG_PRINT_STAT(NumConstantOffsetPtrArgs);
DEBUG_PRINT_STAT(NumAllocaArgs);
DEBUG_PRINT_STAT(NumConstantPtrCmps);
DEBUG_PRINT_STAT(NumConstantPtrDiffs);
DEBUG_PRINT_STAT(NumInstructionsSimplified);
+ DEBUG_PRINT_STAT(NumInstructions);
DEBUG_PRINT_STAT(SROACostSavings);
DEBUG_PRINT_STAT(SROACostSavingsLost);
DEBUG_PRINT_STAT(ContainsNoDuplicateCall);
+ DEBUG_PRINT_STAT(Cost);
+ DEBUG_PRINT_STAT(Threshold);
#undef DEBUG_PRINT_STAT
}
#endif
INITIALIZE_PASS_BEGIN(InlineCostAnalysis, "inline-cost", "Inline Cost Analysis",
true, true)
-INITIALIZE_AG_DEPENDENCY(TargetTransformInfo)
+INITIALIZE_PASS_DEPENDENCY(TargetTransformInfoWrapperPass)
+INITIALIZE_PASS_DEPENDENCY(AssumptionCacheTracker)
INITIALIZE_PASS_END(InlineCostAnalysis, "inline-cost", "Inline Cost Analysis",
true, true)
char InlineCostAnalysis::ID = 0;
-InlineCostAnalysis::InlineCostAnalysis() : CallGraphSCCPass(ID), TD(0) {}
+InlineCostAnalysis::InlineCostAnalysis() : CallGraphSCCPass(ID) {}
InlineCostAnalysis::~InlineCostAnalysis() {}
void InlineCostAnalysis::getAnalysisUsage(AnalysisUsage &AU) const {
AU.setPreservesAll();
- AU.addRequired<TargetTransformInfo>();
+ AU.addRequired<AssumptionCacheTracker>();
+ AU.addRequired<TargetTransformInfoWrapperPass>();
CallGraphSCCPass::getAnalysisUsage(AU);
}
bool InlineCostAnalysis::runOnSCC(CallGraphSCC &SCC) {
- TD = getAnalysisIfAvailable<DataLayout>();
- TTI = &getAnalysis<TargetTransformInfo>();
+ TTIWP = &getAnalysis<TargetTransformInfoWrapperPass>();
+ ACT = &getAnalysis<AssumptionCacheTracker>();
return false;
}
return getInlineCost(CS, CS.getCalledFunction(), Threshold);
}
+/// \brief Test that two functions either have or have not the given attribute
+/// at the same time.
+template<typename AttrKind>
+static bool attributeMatches(Function *F1, Function *F2, AttrKind Attr) {
+ return F1->getFnAttribute(Attr) == F2->getFnAttribute(Attr);
+}
+
+/// \brief Test that there are no attribute conflicts between Caller and Callee
+/// that prevent inlining.
+static bool functionsHaveCompatibleAttributes(Function *Caller,
+ Function *Callee,
+ TargetTransformInfo &TTI) {
+ return TTI.areInlineCompatible(Caller, Callee) &&
+ attributeMatches(Caller, Callee, Attribute::SanitizeAddress) &&
+ attributeMatches(Caller, Callee, Attribute::SanitizeMemory) &&
+ attributeMatches(Caller, Callee, Attribute::SanitizeThread);
+}
+
InlineCost InlineCostAnalysis::getInlineCost(CallSite CS, Function *Callee,
int Threshold) {
// Cannot inline indirect calls.
// Calls to functions with always-inline attributes should be inlined
// whenever possible.
- if (Callee->getAttributes().hasAttribute(AttributeSet::FunctionIndex,
- Attribute::AlwaysInline)) {
+ if (CS.hasFnAttr(Attribute::AlwaysInline)) {
if (isInlineViable(*Callee))
return llvm::InlineCost::getAlways();
return llvm::InlineCost::getNever();
}
+ // Never inline functions with conflicting attributes (unless callee has
+ // always-inline attribute).
+ if (!functionsHaveCompatibleAttributes(CS.getCaller(), Callee,
+ TTIWP->getTTI(*Callee)))
+ return llvm::InlineCost::getNever();
+
+ // Don't inline this call if the caller has the optnone attribute.
+ if (CS.getCaller()->hasFnAttribute(Attribute::OptimizeNone))
+ return llvm::InlineCost::getNever();
+
// Don't inline functions which can be redefined at link-time to mean
// something else. Don't inline functions marked noinline or call sites
// marked noinline.
if (Callee->mayBeOverridden() ||
- Callee->getAttributes().hasAttribute(AttributeSet::FunctionIndex,
- Attribute::NoInline) ||
- CS.isNoInline())
+ Callee->hasFnAttribute(Attribute::NoInline) || CS.isNoInline())
return llvm::InlineCost::getNever();
DEBUG(llvm::dbgs() << " Analyzing call of " << Callee->getName()
<< "...\n");
- CallAnalyzer CA(TD, *TTI, *Callee, Threshold);
+ CallAnalyzer CA(TTIWP->getTTI(*Callee), ACT, *Callee, Threshold, CS);
bool ShouldInline = CA.analyzeCall(CS);
DEBUG(CA.dump());
}
bool InlineCostAnalysis::isInlineViable(Function &F) {
- bool ReturnsTwice =
- F.getAttributes().hasAttribute(AttributeSet::FunctionIndex,
- Attribute::ReturnsTwice);
+ bool ReturnsTwice = F.hasFnAttribute(Attribute::ReturnsTwice);
for (Function::iterator BI = F.begin(), BE = F.end(); BI != BE; ++BI) {
- // Disallow inlining of functions which contain an indirect branch.
- if (isa<IndirectBrInst>(BI->getTerminator()))
+ // Disallow inlining of functions which contain indirect branches or
+ // blockaddresses.
+ if (isa<IndirectBrInst>(BI->getTerminator()) || BI->hasAddressTaken())
return false;
for (BasicBlock::iterator II = BI->begin(), IE = BI->end(); II != IE;
if (!ReturnsTwice && CS.isCall() &&
cast<CallInst>(CS.getInstruction())->canReturnTwice())
return false;
+
+ // Disallow inlining functions that call @llvm.localescape. Doing this
+ // correctly would require major changes to the inliner.
+ if (CS.getCalledFunction() &&
+ CS.getCalledFunction()->getIntrinsicID() ==
+ llvm::Intrinsic::localescape)
+ return false;
}
}