diff options
Diffstat (limited to 'lib')
| -rw-r--r-- | lib/CodeGen/IfConversion.cpp | 16 | ||||
| -rw-r--r-- | lib/IR/Instructions.cpp | 11 | ||||
| -rw-r--r-- | lib/Target/X86/X86ISelLowering.cpp | 29 | ||||
| -rw-r--r-- | lib/Transforms/Utils/SimplifyLibCalls.cpp | 4 |
4 files changed, 44 insertions, 16 deletions
diff --git a/lib/CodeGen/IfConversion.cpp b/lib/CodeGen/IfConversion.cpp index b17a253fe23f..af556e2c856e 100644 --- a/lib/CodeGen/IfConversion.cpp +++ b/lib/CodeGen/IfConversion.cpp @@ -912,6 +912,12 @@ void IfConverter::AnalyzeBranches(BBInfo &BBI) { BBI.BrCond.clear(); BBI.IsBrAnalyzable = !TII->analyzeBranch(*BBI.BB, BBI.TrueBB, BBI.FalseBB, BBI.BrCond); + if (!BBI.IsBrAnalyzable) { + BBI.TrueBB = nullptr; + BBI.FalseBB = nullptr; + BBI.BrCond.clear(); + } + SmallVector<MachineOperand, 4> RevCond(BBI.BrCond.begin(), BBI.BrCond.end()); BBI.IsBrReversible = (RevCond.size() == 0) || !TII->reverseBranchCondition(RevCond); @@ -1758,9 +1764,15 @@ bool IfConverter::IfConvertDiamondCommon( if (!BBI1->IsBrAnalyzable) verifySameBranchInstructions(&MBB1, &MBB2); #endif - BBI1->NonPredSize -= TII->removeBranch(*BBI1->BB); - // Remove duplicated instructions. + // Remove duplicated instructions from the tail of MBB1: any branch + // instructions, and the common instructions counted by NumDups2. DI1 = MBB1.end(); + while (DI1 != MBB1.begin()) { + MachineBasicBlock::iterator Prev = std::prev(DI1); + if (!Prev->isBranch() && !Prev->isDebugInstr()) + break; + DI1 = Prev; + } for (unsigned i = 0; i != NumDups2; ) { // NumDups2 only counted non-dbg_value instructions, so this won't // run off the head of the list. diff --git a/lib/IR/Instructions.cpp b/lib/IR/Instructions.cpp index 2e7cad103c12..90cb2c26e082 100644 --- a/lib/IR/Instructions.cpp +++ b/lib/IR/Instructions.cpp @@ -822,6 +822,17 @@ void CallBrInst::init(FunctionType *FTy, Value *Fn, BasicBlock *Fallthrough, setName(NameStr); } +void CallBrInst::updateArgBlockAddresses(unsigned i, BasicBlock *B) { + assert(getNumIndirectDests() > i && "IndirectDest # out of range for callbr"); + if (BasicBlock *OldBB = getIndirectDest(i)) { + BlockAddress *Old = BlockAddress::get(OldBB); + BlockAddress *New = BlockAddress::get(B); + for (unsigned ArgNo = 0, e = getNumArgOperands(); ArgNo != e; ++ArgNo) + if (dyn_cast<BlockAddress>(getArgOperand(ArgNo)) == Old) + setArgOperand(ArgNo, New); + } +} + CallBrInst::CallBrInst(const CallBrInst &CBI) : CallBase(CBI.Attrs, CBI.FTy, CBI.getType(), Instruction::CallBr, OperandTraits<CallBase>::op_end(this) - CBI.getNumOperands(), diff --git a/lib/Target/X86/X86ISelLowering.cpp b/lib/Target/X86/X86ISelLowering.cpp index 87ff5a719fca..0c5b8a79dd62 100644 --- a/lib/Target/X86/X86ISelLowering.cpp +++ b/lib/Target/X86/X86ISelLowering.cpp @@ -7650,17 +7650,22 @@ static SDValue EltsFromConsecutiveLoads(EVT VT, ArrayRef<SDValue> Elts, // IsConsecutiveLoadWithZeros - we need to create a shuffle of the loaded // vector and a zero vector to clear out the zero elements. if (!isAfterLegalize && VT.isVector()) { - SmallVector<int, 4> ClearMask(NumElems, -1); - for (unsigned i = 0; i < NumElems; ++i) { - if (ZeroMask[i]) - ClearMask[i] = i + NumElems; - else if (LoadMask[i]) - ClearMask[i] = i; + unsigned NumMaskElts = VT.getVectorNumElements(); + if ((NumMaskElts % NumElems) == 0) { + unsigned Scale = NumMaskElts / NumElems; + SmallVector<int, 4> ClearMask(NumMaskElts, -1); + for (unsigned i = 0; i < NumElems; ++i) { + if (UndefMask[i]) + continue; + int Offset = ZeroMask[i] ? NumMaskElts : 0; + for (unsigned j = 0; j != Scale; ++j) + ClearMask[(i * Scale) + j] = (i * Scale) + j + Offset; + } + SDValue V = CreateLoad(VT, LDBase); + SDValue Z = VT.isInteger() ? DAG.getConstant(0, DL, VT) + : DAG.getConstantFP(0.0, DL, VT); + return DAG.getVectorShuffle(VT, DL, V, Z, ClearMask); } - SDValue V = CreateLoad(VT, LDBase); - SDValue Z = VT.isInteger() ? DAG.getConstant(0, DL, VT) - : DAG.getConstantFP(0.0, DL, VT); - return DAG.getVectorShuffle(VT, DL, V, Z, ClearMask); } } @@ -31664,8 +31669,8 @@ static bool matchUnaryPermuteShuffle(MVT MaskVT, ArrayRef<int> Mask, if (!ContainsZeros && AllowIntDomain && MaskScalarSizeInBits == 16) { SmallVector<int, 4> RepeatedMask; if (is128BitLaneRepeatedShuffleMask(MaskEltVT, Mask, RepeatedMask)) { - ArrayRef<int> LoMask(Mask.data() + 0, 4); - ArrayRef<int> HiMask(Mask.data() + 4, 4); + ArrayRef<int> LoMask(RepeatedMask.data() + 0, 4); + ArrayRef<int> HiMask(RepeatedMask.data() + 4, 4); // PSHUFLW: permute lower 4 elements only. if (isUndefOrInRange(LoMask, 0, 4) && diff --git a/lib/Transforms/Utils/SimplifyLibCalls.cpp b/lib/Transforms/Utils/SimplifyLibCalls.cpp index e0def81d5eee..e938ae6cb42f 100644 --- a/lib/Transforms/Utils/SimplifyLibCalls.cpp +++ b/lib/Transforms/Utils/SimplifyLibCalls.cpp @@ -1480,8 +1480,8 @@ Value *LibCallSimplifier::optimizePow(CallInst *Pow, IRBuilder<> &B) { if (match(Expo, m_SpecificFP(-1.0))) return B.CreateFDiv(ConstantFP::get(Ty, 1.0), Base, "reciprocal"); - // pow(x, 0.0) -> 1.0 - if (match(Expo, m_SpecificFP(0.0))) + // pow(x, +/-0.0) -> 1.0 + if (match(Expo, m_AnyZeroFP())) return ConstantFP::get(Ty, 1.0); // pow(x, 1.0) -> x |