X-Git-Url: http://plrg.eecs.uci.edu/git/?a=blobdiff_plain;f=include%2Fllvm%2FTarget%2FTargetLowering.h;h=774e7d8ac8f4cba8065749482d43a295c1390339;hb=807111a8f434293f84bbd1171a953197e40819d1;hp=01b579b2233440f61f042a1edd57213642609ee9;hpb=f7c6da6fe8ba43205d2a1fb1152720bd72e7ea23;p=oota-llvm.git diff --git a/include/llvm/Target/TargetLowering.h b/include/llvm/Target/TargetLowering.h index 01b579b2233..774e7d8ac8f 100644 --- a/include/llvm/Target/TargetLowering.h +++ b/include/llvm/Target/TargetLowering.h @@ -28,9 +28,12 @@ #include "llvm/CodeGen/RuntimeLibcalls.h" #include "llvm/CodeGen/SelectionDAGNodes.h" #include "llvm/IR/Attributes.h" +#include "llvm/IR/CallSite.h" #include "llvm/IR/CallingConv.h" #include "llvm/IR/InlineAsm.h" -#include "llvm/Support/CallSite.h" +#include "llvm/IR/Instructions.h" +#include "llvm/IR/IRBuilder.h" +#include "llvm/MC/MCRegisterInfo.h" #include "llvm/Target/TargetCallingConv.h" #include "llvm/Target/TargetMachine.h" #include @@ -48,8 +51,10 @@ namespace llvm { class MachineFunction; class MachineInstr; class MachineJumpTableInfo; + class Mangler; class MCContext; class MCExpr; + class MCSymbol; template class SmallVectorImpl; class DataLayout; class TargetRegisterClass; @@ -143,14 +148,15 @@ protected: public: const TargetMachine &getTargetMachine() const { return TM; } - const DataLayout *getDataLayout() const { return TD; } + const DataLayout *getDataLayout() const { return DL; } const TargetLoweringObjectFile &getObjFileLowering() const { return TLOF; } bool isBigEndian() const { return !IsLittleEndian; } bool isLittleEndian() const { return IsLittleEndian; } - // Return the pointer type for the given address space, defaults to - // the pointer type from the data layout. - // FIXME: The default needs to be removed once all the code is updated. + + /// Return the pointer type for the given address space, defaults to + /// the pointer type from the data layout. + /// FIXME: The default needs to be removed once all the code is updated. virtual MVT getPointerTy(uint32_t /*AS*/ = 0) const; unsigned getPointerSizeInBits(uint32_t AS = 0) const; unsigned getPointerTypeSizeInBits(Type *Ty) const; @@ -172,10 +178,38 @@ public: return true; } - /// Return true if a vector of the given type should be split - /// (TypeSplitVector) instead of promoted (TypePromoteInteger) during type - /// legalization. - virtual bool shouldSplitVectorElementType(EVT /*VT*/) const { return false; } + /// Return true if multiple condition registers are available. + bool hasMultipleConditionRegisters() const { + return HasMultipleConditionRegisters; + } + + /// Return true if the target has BitExtract instructions. + bool hasExtractBitsInsn() const { return HasExtractBitsInsn; } + + /// Return the preferred vector type legalization action. + virtual TargetLoweringBase::LegalizeTypeAction + getPreferredVectorAction(EVT VT) const { + // The default action for one element vectors is to scalarize + if (VT.getVectorNumElements() == 1) + return TypeScalarizeVector; + // The default action for other vectors is to promote + return TypePromoteInteger; + } + + // There are two general methods for expanding a BUILD_VECTOR node: + // 1. Use SCALAR_TO_VECTOR on the defined scalar values and then shuffle + // them together. + // 2. Build the vector on the stack and then load it. + // If this function returns true, then method (1) will be used, subject to + // the constraint that all of the necessary shuffles are legal (as determined + // by isShuffleMaskLegal). If this function returns false, then method (2) is + // always used. The vector type, and the number of defined values, are + // provided. + virtual bool + shouldExpandBuildVectorWithShuffles(EVT /* VT */, + unsigned DefinedValues) const { + return DefinedValues < 3; + } /// Return true if integer divide is usually cheaper than a sequence of /// several shifts, adds, and multiplies for this target. @@ -190,8 +224,8 @@ public: return BypassSlowDivWidths; } - /// Return true if pow2 div is cheaper than a chain of srl/add/sra. - bool isPow2DivCheap() const { return Pow2DivIsCheap; } + /// Return true if pow2 sdiv is cheaper than a chain of sra/srl/add/sra. + bool isPow2SDivCheap() const { return Pow2SDivIsCheap; } /// Return true if Flow Control is an expensive operation that should be /// avoided. @@ -203,10 +237,45 @@ public: return PredictableSelectIsExpensive; } - /// Return the ValueType of the result of SETCC operations. Also used to - /// obtain the target's preferred type for the condition operand of SELECT and - /// BRCOND nodes. In the case of BRCOND the argument passed is MVT::Other - /// since there are no other operands to get a type hint from. + /// isLoadBitCastBeneficial() - Return true if the following transform + /// is beneficial. + /// fold (conv (load x)) -> (load (conv*)x) + /// On architectures that don't natively support some vector loads efficiently, + /// casting the load to a smaller vector of larger types and loading + /// is more efficient, however, this can be undone by optimizations in + /// dag combiner. + virtual bool isLoadBitCastBeneficial(EVT /* Load */, EVT /* Bitcast */) const { + return true; + } + + /// \brief Return if the target supports combining a + /// chain like: + /// \code + /// %andResult = and %val1, #imm-with-one-bit-set; + /// %icmpResult = icmp %andResult, 0 + /// br i1 %icmpResult, label %dest1, label %dest2 + /// \endcode + /// into a single machine instruction of a form like: + /// \code + /// brOnBitSet %register, #bitNumber, dest + /// \endcode + bool isMaskAndBranchFoldingLegal() const { + return MaskAndBranchFoldingIsLegal; + } + + /// Return true if target supports floating point exceptions. + bool hasFloatingPointExceptions() const { + return HasFloatingPointExceptions; + } + + /// Return true if target always beneficiates from combining into FMA for a + /// given value type. This must typically return false on targets where FMA + /// takes more cycles to execute than FADD. + virtual bool enableAggressiveFMAFusion(EVT VT) const { + return false; + } + + /// Return the ValueType of the result of SETCC operations. virtual EVT getSetCCResultType(LLVMContext &Context, EVT VT) const; /// Return the ValueType for comparison libcalls. Comparions libcalls include @@ -225,8 +294,17 @@ public: /// selects between the two kinds. For example on X86 a scalar boolean should /// be zero extended from i1, while the elements of a vector of booleans /// should be sign extended from i1. - BooleanContent getBooleanContents(bool isVec) const { - return isVec ? BooleanVectorContents : BooleanContents; + /// + /// Some cpus also treat floating point types the same way as they treat + /// vectors instead of the way they treat scalars. + BooleanContent getBooleanContents(bool isVec, bool isFloat) const { + if (isVec) + return BooleanVectorContents; + return isFloat ? BooleanFloatContents : BooleanContents; + } + + BooleanContent getBooleanContents(EVT Type) const { + return getBooleanContents(Type.isVector(), Type.isFloatingPoint()); } /// Return target scheduling preference. @@ -273,7 +351,7 @@ public: bool isTypeLegal(EVT VT) const { assert(!VT.isSimple() || (unsigned)VT.getSimpleVT().SimpleTy < array_lengthof(RegClassForVT)); - return VT.isSimple() && RegClassForVT[VT.getSimpleVT().SimpleTy] != 0; + return VT.isSimple() && RegClassForVT[VT.getSimpleVT().SimpleTy] != nullptr; } class ValueTypeActionImpl { @@ -283,7 +361,7 @@ public: public: ValueTypeActionImpl() { - std::fill(ValueTypeActions, array_endof(ValueTypeActions), 0); + std::fill(std::begin(ValueTypeActions), std::end(ValueTypeActions), 0); } LegalizeTypeAction getTypeAction(MVT VT) const { @@ -358,10 +436,15 @@ public: EVT memVT; // memory VT const Value* ptrVal; // value representing memory location int offset; // offset off of ptrVal + unsigned size; // the size of the memory location + // (taken from memVT if zero) unsigned align; // alignment bool vol; // is volatile? bool readMem; // reads memory? bool writeMem; // writes memory? + + IntrinsicInfo() : opc(0), ptrVal(nullptr), offset(0), size(0), align(1), + vol(false), readMem(false), writeMem(false) {} }; /// Given an intrinsic, checks if on the target the intrinsic will need to map @@ -449,10 +532,12 @@ public: /// Return how this load with extension should be treated: either it is legal, /// needs to be promoted to a larger size, needs to be expanded to some other /// code sequence, or the target has a custom expander for it. - LegalizeAction getLoadExtAction(unsigned ExtType, MVT VT) const { - assert(ExtType < ISD::LAST_LOADEXT_TYPE && VT < MVT::LAST_VALUETYPE && + LegalizeAction getLoadExtAction(unsigned ExtType, EVT VT) const { + if (VT.isExtended()) return Expand; + unsigned I = (unsigned) VT.getSimpleVT().SimpleTy; + assert(ExtType < ISD::LAST_LOADEXT_TYPE && I < MVT::LAST_VALUETYPE && "Table isn't big enough!"); - return (LegalizeAction)LoadExtActions[VT.SimpleTy][ExtType]; + return (LegalizeAction)LoadExtActions[I][ExtType]; } /// Return true if the specified load with extension is legal on this target. @@ -464,11 +549,13 @@ public: /// Return how this store with truncation should be treated: either it is /// legal, needs to be promoted to a larger size, needs to be expanded to some /// other code sequence, or the target has a custom expander for it. - LegalizeAction getTruncStoreAction(MVT ValVT, MVT MemVT) const { - assert(ValVT < MVT::LAST_VALUETYPE && MemVT < MVT::LAST_VALUETYPE && + LegalizeAction getTruncStoreAction(EVT ValVT, EVT MemVT) const { + if (ValVT.isExtended() || MemVT.isExtended()) return Expand; + unsigned ValI = (unsigned) ValVT.getSimpleVT().SimpleTy; + unsigned MemI = (unsigned) MemVT.getSimpleVT().SimpleTy; + assert(ValI < MVT::LAST_VALUETYPE && MemI < MVT::LAST_VALUETYPE && "Table isn't big enough!"); - return (LegalizeAction)TruncStoreActions[ValVT.SimpleTy] - [MemVT.SimpleTy]; + return (LegalizeAction)TruncStoreActions[ValI][MemI]; } /// Return true if the specified store with truncation is legal on this @@ -592,8 +679,9 @@ public: return getValueType(Ty, AllowUnknown).getSimpleVT(); } - /// Return the desired alignment for ByVal aggregate function arguments in the - /// caller parameter area. This is the actual alignment, not its logarithm. + /// Return the desired alignment for ByVal or InAlloca aggregate function + /// arguments in the caller parameter area. This is the actual alignment, not + /// its logarithm. virtual unsigned getByValTypeAlignment(Type *Ty) const; /// Return the type of registers that this ValueType will eventually require. @@ -656,6 +744,13 @@ public: /// reduce runtime. virtual bool ShouldShrinkFPConstant(EVT) const { return true; } + /// When splitting a value of the specified type into parts, does the Lo + /// or Hi part come first? This usually follows the endianness, except + /// for ppcf128, where the Hi part always comes first. + bool hasBigEndianPartOrdering(EVT VT) const { + return isBigEndian() || VT == MVT::ppcf128; + } + /// If true, the target has custom DAG combine transformations that it can /// perform for the specified node. bool hasTargetDAGCombine(ISD::NodeType NT) const { @@ -695,14 +790,17 @@ public: /// \brief Determine if the target supports unaligned memory accesses. /// - /// This function returns true if the target allows unaligned memory accesses. - /// of the specified type. If true, it also returns whether the unaligned - /// memory access is "fast" in the second argument by reference. This is used, - /// for example, in situations where an array copy/move/set is converted to a - /// sequence of store operations. It's use helps to ensure that such - /// replacements don't generate code that causes an alignment error (trap) on - /// the target machine. - virtual bool allowsUnalignedMemoryAccesses(EVT, bool * /*Fast*/ = 0) const { + /// This function returns true if the target allows unaligned memory accesses + /// of the specified type in the given address space. If true, it also returns + /// whether the unaligned memory access is "fast" in the last argument by + /// reference. This is used, for example, in situations where an array + /// copy/move/set is converted to a sequence of store operations. Its use + /// helps to ensure that such replacements don't generate code that causes an + /// alignment error (trap) on the target machine. + virtual bool allowsMisalignedMemoryAccesses(EVT, + unsigned AddrSpace = 0, + unsigned Align = 1, + bool * /*Fast*/ = nullptr) const { return false; } @@ -745,11 +843,6 @@ public: return UseUnderscoreLongJmp; } - /// Return whether the target can generate code for jump tables. - bool supportJumpTables() const { - return SupportJumpTables; - } - /// Return integer threshold on number of blocks to use jump tables rather /// than if sequence. int getMinimumJumpTableEntries() const { @@ -826,6 +919,11 @@ public: return 0; } + /// Returns true if a cast between SrcAS and DestAS is a noop. + virtual bool isNoopAddrSpaceCast(unsigned SrcAS, unsigned DestAS) const { + return false; + } + //===--------------------------------------------------------------------===// /// \name Helpers for TargetTransformInfo implementations /// @{ @@ -838,6 +936,89 @@ public: /// @} + //===--------------------------------------------------------------------===// + /// \name Helpers for atomic expansion. + /// @{ + + /// True if AtomicExpandPass should use emitLoadLinked/emitStoreConditional + /// and expand AtomicCmpXchgInst. + virtual bool hasLoadLinkedStoreConditional() const { return false; } + + /// Perform a load-linked operation on Addr, returning a "Value *" with the + /// corresponding pointee type. This may entail some non-trivial operations to + /// truncate or reconstruct types that will be illegal in the backend. See + /// ARMISelLowering for an example implementation. + virtual Value *emitLoadLinked(IRBuilder<> &Builder, Value *Addr, + AtomicOrdering Ord) const { + llvm_unreachable("Load linked unimplemented on this target"); + } + + /// Perform a store-conditional operation to Addr. Return the status of the + /// store. This should be 0 if the store succeeded, non-zero otherwise. + virtual Value *emitStoreConditional(IRBuilder<> &Builder, Value *Val, + Value *Addr, AtomicOrdering Ord) const { + llvm_unreachable("Store conditional unimplemented on this target"); + } + + /// Inserts in the IR a target-specific intrinsic specifying a fence. + /// It is called by AtomicExpandPass before expanding an + /// AtomicRMW/AtomicCmpXchg/AtomicStore/AtomicLoad. + /// RMW and CmpXchg set both IsStore and IsLoad to true. + /// Backends with !getInsertFencesForAtomic() should keep a no-op here. + /// This function should either return a nullptr, or a pointer to an IR-level + /// Instruction*. Even complex fence sequences can be represented by a + /// single Instruction* through an intrinsic to be lowered later. + virtual Instruction* emitLeadingFence(IRBuilder<> &Builder, AtomicOrdering Ord, + bool IsStore, bool IsLoad) const { + assert(!getInsertFencesForAtomic()); + return nullptr; + } + + /// Inserts in the IR a target-specific intrinsic specifying a fence. + /// It is called by AtomicExpandPass after expanding an + /// AtomicRMW/AtomicCmpXchg/AtomicStore/AtomicLoad. + /// RMW and CmpXchg set both IsStore and IsLoad to true. + /// Backends with !getInsertFencesForAtomic() should keep a no-op here. + /// This function should either return a nullptr, or a pointer to an IR-level + /// Instruction*. Even complex fence sequences can be represented by a + /// single Instruction* through an intrinsic to be lowered later. + virtual Instruction* emitTrailingFence(IRBuilder<> &Builder, AtomicOrdering Ord, + bool IsStore, bool IsLoad) const { + assert(!getInsertFencesForAtomic()); + return nullptr; + } + + /// Returns true if the given (atomic) store should be expanded by the + /// IR-level AtomicExpand pass into an "atomic xchg" which ignores its input. + virtual bool shouldExpandAtomicStoreInIR(StoreInst *SI) const { + return false; + } + + /// Returns true if the given (atomic) load should be expanded by the + /// IR-level AtomicExpand pass into a load-linked instruction + /// (through emitLoadLinked()). + virtual bool shouldExpandAtomicLoadInIR(LoadInst *LI) const { return false; } + + /// Returns true if the given AtomicRMW should be expanded by the + /// IR-level AtomicExpand pass into a loop using LoadLinked/StoreConditional. + virtual bool shouldExpandAtomicRMWInIR(AtomicRMWInst *RMWI) const { + return false; + } + + /// On some platforms, an AtomicRMW that never actually modifies the value + /// (such as fetch_add of 0) can be turned into a fence followed by an + /// atomic load. This may sound useless, but it makes it possible for the + /// processor to keep the cacheline shared, dramatically improving + /// performance. And such idempotent RMWs are useful for implementing some + /// kinds of locks, see for example (justification + benchmarks): + /// http://www.hpl.hp.com/techreports/2012/HPL-2012-68.pdf + /// This method tries doing that transformation, returning the atomic load if + /// it succeeds, and nullptr otherwise. + /// If shouldExpandAtomicLoadInIR returns true on that load, it will undergo + /// another round of expansion. + virtual LoadInst *lowerIdempotentRMWIntoFencedLoad(AtomicRMWInst *RMWI) const { + return nullptr; + } //===--------------------------------------------------------------------===// // TargetLowering Configuration Methods - These methods should be invoked by // the derived class constructor to configure this object for the target. @@ -847,9 +1028,19 @@ public: virtual void resetOperationActions() {} protected: - /// Specify how the target extends the result of a boolean value from i1 to a - /// wider type. See getBooleanContents. - void setBooleanContents(BooleanContent Ty) { BooleanContents = Ty; } + /// Specify how the target extends the result of integer and floating point + /// boolean values from i1 to a wider type. See getBooleanContents. + void setBooleanContents(BooleanContent Ty) { + BooleanContents = Ty; + BooleanFloatContents = Ty; + } + + /// Specify how the target extends the result of integer and floating point + /// boolean values from i1 to a wider type. See getBooleanContents. + void setBooleanContents(BooleanContent IntTy, BooleanContent FloatTy) { + BooleanContents = IntTy; + BooleanFloatContents = FloatTy; + } /// Specify how the target extends the result of a vector boolean value from a /// vector of i1 to a wider type. See getBooleanContents. @@ -863,22 +1054,17 @@ protected: } /// Indicate whether this target prefers to use _setjmp to implement - /// llvm.setjmp or the non _ version. Defaults to false. + /// llvm.setjmp or the version without _. Defaults to false. void setUseUnderscoreSetJmp(bool Val) { UseUnderscoreSetJmp = Val; } /// Indicate whether this target prefers to use _longjmp to implement - /// llvm.longjmp or the non _ version. Defaults to false. + /// llvm.longjmp or the version without _. Defaults to false. void setUseUnderscoreLongJmp(bool Val) { UseUnderscoreLongJmp = Val; } - /// Indicate whether the target can generate code for jump tables. - void setSupportJumpTables(bool Val) { - SupportJumpTables = Val; - } - /// Indicate the number of blocks to generate jump tables rather than if /// sequence. void setMinimumJumpTableEntries(int Val) { @@ -909,6 +1095,23 @@ protected: SelectIsExpensive = isExpensive; } + /// Tells the code generator that the target has multiple (allocatable) + /// condition registers that can be used to store the results of comparisons + /// for use by selects and conditional branches. With multiple condition + /// registers, the code generator will not aggressively sink comparisons into + /// the blocks of their users. + void setHasMultipleConditionRegisters(bool hasManyRegs = true) { + HasMultipleConditionRegisters = hasManyRegs; + } + + /// Tells the code generator that the target has BitExtract instructions. + /// The code generator will aggressively sink "shift"s into the blocks of + /// their users if the users will generate "and" instructions which can be + /// combined with "shift" to BitExtract instructions. + void setHasExtractBitsInsn(bool hasExtractInsn = true) { + HasExtractBitsInsn = hasExtractInsn; + } + /// Tells the code generator not to expand sequence of operations into a /// separate sequences that increases the amount of flow control. void setJumpIsExpensive(bool isExpensive = true) { @@ -919,15 +1122,21 @@ protected: /// possible, should be replaced by an alternate sequence of instructions not /// containing an integer divide. void setIntDivIsCheap(bool isCheap = true) { IntDivIsCheap = isCheap; } + + /// Tells the code generator that this target supports floating point + /// exceptions and cares about preserving floating point exception behavior. + void setHasFloatingPointExceptions(bool FPExceptions = true) { + HasFloatingPointExceptions = FPExceptions; + } /// Tells the code generator which bitwidths to bypass. void addBypassSlowDiv(unsigned int SlowBitWidth, unsigned int FastBitWidth) { BypassSlowDivWidths[SlowBitWidth] = FastBitWidth; } - /// Tells the code generator that it shouldn't generate srl/add/sra for a - /// signed divide by power of two, and let the target handle it. - void setPow2DivIsCheap(bool isCheap = true) { Pow2DivIsCheap = isCheap; } + /// Tells the code generator that it shouldn't generate sra/srl/add/sra for a + /// signed divide by power of two; let the target handle it. + void setPow2SDivIsCheap(bool isCheap = true) { Pow2SDivIsCheap = isCheap; } /// Add the specified register class as an available regclass for the /// specified value type. This indicates the selector can handle values of @@ -1112,7 +1321,7 @@ public: int64_t BaseOffs; bool HasBaseReg; int64_t Scale; - AddrMode() : BaseGV(0), BaseOffs(0), HasBaseReg(false), Scale(0) {} + AddrMode() : BaseGV(nullptr), BaseOffs(0), HasBaseReg(false), Scale(0) {} }; /// Return true if the addressing mode represented by AM is legal for this @@ -1149,6 +1358,14 @@ public: return true; } + /// Return true if it's significantly cheaper to shift a vector by a uniform + /// scalar than by an amount which will vary across each lane. On x86, for + /// example, there is a "psllw" instruction for the former case, but no simple + /// instruction for a general "a << b" operation on vectors. + virtual bool isVectorShiftByScalarCheap(Type *Ty) const { + return false; + } + /// Return true if it's free to truncate a value of type Ty1 to type /// Ty2. e.g. On x86 it's free to truncate a i32 value in register EAX to i16 /// by referencing its sub-register AX. @@ -1256,6 +1473,15 @@ public: return false; } + /// \brief Return true if it is beneficial to convert a load of a constant to + /// just the constant itself. + /// On some targets it might be more efficient to use a combination of + /// arithmetic instructions to materialize the constant instead of loading it + /// from a constant pool. + virtual bool shouldConvertConstantLoadToIntImm(const APInt &Imm, + Type *Ty) const { + return false; + } //===--------------------------------------------------------------------===// // Runtime Library hooks // @@ -1294,7 +1520,7 @@ public: private: const TargetMachine &TM; - const DataLayout *TD; + const DataLayout *DL; const TargetLoweringObjectFile &TLOF; /// True if this is a little endian target. @@ -1304,6 +1530,19 @@ private: /// the select operations if possible. bool SelectIsExpensive; + /// Tells the code generator that the target has multiple (allocatable) + /// condition registers that can be used to store the results of comparisons + /// for use by selects and conditional branches. With multiple condition + /// registers, the code generator will not aggressively sink comparisons into + /// the blocks of their users. + bool HasMultipleConditionRegisters; + + /// Tells the code generator that the target has BitExtract instructions. + /// The code generator will aggressively sink "shift"s into the blocks of + /// their users if the users will generate "and" instructions which can be + /// combined with "shift" to BitExtract instructions. + bool HasExtractBitsInsn; + /// Tells the code generator not to expand integer divides by constants into a /// sequence of muls, adds, and shifts. This is a hack until a real cost /// model is in place. If we ever optimize for size, this will be set to true @@ -1316,15 +1555,19 @@ private: /// div/rem when the operands are positive and less than 256. DenseMap BypassSlowDivWidths; - /// Tells the code generator that it shouldn't generate srl/add/sra for a - /// signed divide by power of two, and let the target handle it. - bool Pow2DivIsCheap; + /// Tells the code generator that it shouldn't generate sra/srl/add/sra for a + /// signed divide by power of two; let the target handle it. + bool Pow2SDivIsCheap; /// Tells the code generator that it shouldn't generate extra flow control /// instructions and should attempt to combine flow control instructions via /// predication. bool JumpIsExpensive; + /// Whether the target supports or cares about preserving floating point + /// exception behavior. + bool HasFloatingPointExceptions; + /// This target prefers to use _setjmp to implement llvm.setjmp. /// /// Defaults to false. @@ -1335,10 +1578,6 @@ private: /// Defaults to false. bool UseUnderscoreLongJmp; - /// Whether the target can generate code for jumptables. If it's not true, - /// then each jumptable must be lowered into if-then-else's. - bool SupportJumpTables; - /// Number of blocks threshold to use jump tables. int MinimumJumpTableEntries; @@ -1346,6 +1585,10 @@ private: /// a type wider than i1. See getBooleanContents. BooleanContent BooleanContents; + /// Information about the contents of the high-bits in boolean values held in + /// a type wider than i1. See getBooleanContents. + BooleanContent BooleanFloatContents; + /// Information about the contents of the high-bits in boolean vector values /// when the element type is wider than i1. See getBooleanContents. BooleanContent BooleanVectorContents; @@ -1462,7 +1705,7 @@ public: LegalizeTypeAction LA = ValueTypeActions.getTypeAction(SVT); assert( - (LA == TypeLegal || + (LA == TypeLegal || LA == TypeSoftenFloat || ValueTypeActions.getTypeAction(NVT) != TypePromoteInteger) && "Promote may not follow Expand or Promote"); @@ -1664,10 +1907,18 @@ protected: /// the branch is usually predicted right. bool PredictableSelectIsExpensive; + /// MaskAndBranchFoldingIsLegal - Indicates if the target supports folding + /// a mask of a single bit, a compare, and a branch into a single instruction. + bool MaskAndBranchFoldingIsLegal; + protected: /// Return true if the value types that can be represented by the specified /// register class are all legal. bool isLegalRC(const TargetRegisterClass *RC) const; + + /// Replace/modify any TargetFrameIndex operands with a targte-dependent + /// sequence of memory operands that is recognized by PrologEpilogInserter. + MachineBasicBlock *emitPatchPoint(MachineInstr *MI, MachineBasicBlock *MBB) const; }; /// This class defines information used to lower LLVM code to legal SelectionDAG @@ -1797,15 +2048,16 @@ public: /// Determine which of the bits specified in Mask are known to be either zero /// or one and return them in the KnownZero/KnownOne bitsets. - virtual void computeMaskedBitsForTargetNode(const SDValue Op, - APInt &KnownZero, - APInt &KnownOne, - const SelectionDAG &DAG, - unsigned Depth = 0) const; + virtual void computeKnownBitsForTargetNode(const SDValue Op, + APInt &KnownZero, + APInt &KnownOne, + const SelectionDAG &DAG, + unsigned Depth = 0) const; /// This method can be implemented by targets that want to expose additional /// information about sign bits to the DAG Combiner. virtual unsigned ComputeNumSignBitsForTargetNode(SDValue Op, + const SelectionDAG &DAG, unsigned Depth = 0) const; struct DAGCombinerInfo { @@ -1836,6 +2088,14 @@ public: void CommitTargetLoweringOpt(const TargetLoweringOpt &TLO); }; + /// Return if the N is a constant or constant vector equal to the true value + /// from getBooleanContents(). + bool isConstTrueVal(const SDNode *N) const; + + /// Return if the N is a constant or constant vector equal to the false value + /// from getBooleanContents(). + bool isConstFalseVal(const SDNode *N) const; + /// Try to simplify a setcc built with the specified operands and cc. If it is /// unable to simplify it, return a null SDValue. SDValue SimplifySetCC(EVT VT, SDValue N0, SDValue N1, @@ -1862,6 +2122,15 @@ public: /// virtual SDValue PerformDAGCombine(SDNode *N, DAGCombinerInfo &DCI) const; + /// Return true if it is profitable to move a following shift through this + // node, adjusting any immediate operands as necessary to preserve semantics. + // This transformation may not be desirable if it disrupts a particularly + // auspicious target-specific tree (e.g. bitfield extraction in AArch64). + // By default, it returns true. + virtual bool isDesirableToCommuteWithShift(const SDNode *N /*Op*/) const { + return true; + } + /// Return true if the target has native support for the specified value type /// and it is 'desirable' to use the type for the given node type. e.g. On x86 /// i16 is legal, but undesirable since i16 instruction encodings are longer @@ -1914,12 +2183,15 @@ public: bool isSRet : 1; bool isNest : 1; bool isByVal : 1; + bool isInAlloca : 1; bool isReturned : 1; uint16_t Alignment; ArgListEntry() : isSExt(false), isZExt(false), isInReg(false), - isSRet(false), isNest(false), isByVal(false), isReturned(false), - Alignment(0) { } + isSRet(false), isNest(false), isByVal(false), isInAlloca(false), + isReturned(false), Alignment(0) { } + + void setAttributes(ImmutableCallSite *CS, unsigned AttrIdx); }; typedef std::vector ArgListTy; @@ -1944,7 +2216,7 @@ public: unsigned NumFixedArgs; CallingConv::ID CallConv; SDValue Callee; - ArgListTy &Args; + ArgListTy Args; SelectionDAG &DAG; SDLoc DL; ImmutableCallSite *CS; @@ -1952,33 +2224,95 @@ public: SmallVector OutVals; SmallVector Ins; + CallLoweringInfo(SelectionDAG &DAG) + : RetTy(nullptr), RetSExt(false), RetZExt(false), IsVarArg(false), + IsInReg(false), DoesNotReturn(false), IsReturnValueUsed(true), + IsTailCall(false), NumFixedArgs(-1), CallConv(CallingConv::C), + DAG(DAG), CS(nullptr) {} + + CallLoweringInfo &setDebugLoc(SDLoc dl) { + DL = dl; + return *this; + } + + CallLoweringInfo &setChain(SDValue InChain) { + Chain = InChain; + return *this; + } + + CallLoweringInfo &setCallee(CallingConv::ID CC, Type *ResultType, + SDValue Target, ArgListTy &&ArgsList, + unsigned FixedArgs = -1) { + RetTy = ResultType; + Callee = Target; + CallConv = CC; + NumFixedArgs = + (FixedArgs == static_cast(-1) ? Args.size() : FixedArgs); + Args = std::move(ArgsList); + return *this; + } + + CallLoweringInfo &setCallee(Type *ResultType, FunctionType *FTy, + SDValue Target, ArgListTy &&ArgsList, + ImmutableCallSite &Call) { + RetTy = ResultType; + + IsInReg = Call.paramHasAttr(0, Attribute::InReg); + DoesNotReturn = Call.doesNotReturn(); + IsVarArg = FTy->isVarArg(); + IsReturnValueUsed = !Call.getInstruction()->use_empty(); + RetSExt = Call.paramHasAttr(0, Attribute::SExt); + RetZExt = Call.paramHasAttr(0, Attribute::ZExt); + + Callee = Target; + + CallConv = Call.getCallingConv(); + NumFixedArgs = FTy->getNumParams(); + Args = std::move(ArgsList); + + CS = &Call; + + return *this; + } + + CallLoweringInfo &setInRegister(bool Value = true) { + IsInReg = Value; + return *this; + } + + CallLoweringInfo &setNoReturn(bool Value = true) { + DoesNotReturn = Value; + return *this; + } + + CallLoweringInfo &setVarArg(bool Value = true) { + IsVarArg = Value; + return *this; + } + + CallLoweringInfo &setTailCall(bool Value = true) { + IsTailCall = Value; + return *this; + } + + CallLoweringInfo &setDiscardResult(bool Value = true) { + IsReturnValueUsed = !Value; + return *this; + } + + CallLoweringInfo &setSExtResult(bool Value = true) { + RetSExt = Value; + return *this; + } - /// Constructs a call lowering context based on the ImmutableCallSite \p cs. - CallLoweringInfo(SDValue chain, Type *retTy, - FunctionType *FTy, bool isTailCall, SDValue callee, - ArgListTy &args, SelectionDAG &dag, SDLoc dl, - ImmutableCallSite &cs) - : Chain(chain), RetTy(retTy), RetSExt(cs.paramHasAttr(0, Attribute::SExt)), - RetZExt(cs.paramHasAttr(0, Attribute::ZExt)), IsVarArg(FTy->isVarArg()), - IsInReg(cs.paramHasAttr(0, Attribute::InReg)), - DoesNotReturn(cs.doesNotReturn()), - IsReturnValueUsed(!cs.getInstruction()->use_empty()), - IsTailCall(isTailCall), NumFixedArgs(FTy->getNumParams()), - CallConv(cs.getCallingConv()), Callee(callee), Args(args), DAG(dag), - DL(dl), CS(&cs) {} - - /// Constructs a call lowering context based on the provided call - /// information. - CallLoweringInfo(SDValue chain, Type *retTy, bool retSExt, bool retZExt, - bool isVarArg, bool isInReg, unsigned numFixedArgs, - CallingConv::ID callConv, bool isTailCall, - bool doesNotReturn, bool isReturnValueUsed, SDValue callee, - ArgListTy &args, SelectionDAG &dag, SDLoc dl) - : Chain(chain), RetTy(retTy), RetSExt(retSExt), RetZExt(retZExt), - IsVarArg(isVarArg), IsInReg(isInReg), DoesNotReturn(doesNotReturn), - IsReturnValueUsed(isReturnValueUsed), IsTailCall(isTailCall), - NumFixedArgs(numFixedArgs), CallConv(callConv), Callee(callee), - Args(args), DAG(dag), DL(dl), CS(NULL) {} + CallLoweringInfo &setZExtResult(bool Value = true) { + RetZExt = Value; + return *this; + } + + ArgListTy &getArgs() { + return Args; + } }; /// This function lowers an abstract call to a function into an actual call. @@ -2041,18 +2375,58 @@ public: return false; } + /// Return the builtin name for the __builtin___clear_cache intrinsic + /// Default is to invoke the clear cache library call + virtual const char * getClearCacheBuiltinName() const { + return "__clear_cache"; + } + + /// Return the register ID of the name passed in. Used by named register + /// global variables extension. There is no target-independent behaviour + /// so the default action is to bail. + virtual unsigned getRegisterByName(const char* RegName, EVT VT) const { + report_fatal_error("Named registers not implemented for this target"); + } + /// Return the type that should be used to zero or sign extend a /// zeroext/signext integer argument or return value. FIXME: Most C calling /// convention requires the return type to be promoted, but this is not true /// all the time, e.g. i1 on x86-64. It is also not necessary for non-C /// calling conventions. The frontend should handle this and include all of /// the necessary information. - virtual MVT getTypeForExtArgOrReturn(MVT VT, + virtual EVT getTypeForExtArgOrReturn(LLVMContext &Context, EVT VT, ISD::NodeType /*ExtendKind*/) const { - MVT MinVT = getRegisterType(MVT::i32); + EVT MinVT = getRegisterType(Context, MVT::i32); return VT.bitsLT(MinVT) ? MinVT : VT; } + /// For some targets, an LLVM struct type must be broken down into multiple + /// simple types, but the calling convention specifies that the entire struct + /// must be passed in a block of consecutive registers. + virtual bool + functionArgumentNeedsConsecutiveRegisters(Type *Ty, CallingConv::ID CallConv, + bool isVarArg) const { + return false; + } + + /// Returns a 0 terminated array of registers that can be safely used as + /// scratch registers. + virtual const MCPhysReg *getScratchRegisters(CallingConv::ID CC) const { + return nullptr; + } + + /// This callback is used to prepare for a volatile or atomic load. + /// It takes a chain node as input and returns the chain for the load itself. + /// + /// Having a callback like this is necessary for targets like SystemZ, + /// which allows a CPU to reuse the result of a previous load indefinitely, + /// even if a cache-coherent store is performed by another CPU. The default + /// implementation does nothing. + virtual SDValue prepareVolatileOrAtomicLoad(SDValue Chain, SDLoc DL, + SelectionDAG &DAG) const { + return Chain; + } + /// This callback is invoked by the type legalizer to legalize nodes with an /// illegal operand type but legal result types. It replaces the /// LowerOperation callback in the type Legalizer. The reason we can not do @@ -2099,9 +2473,13 @@ public: /// target does not support "fast" ISel. virtual FastISel *createFastISel(FunctionLoweringInfo &, const TargetLibraryInfo *) const { - return 0; + return nullptr; } + + bool verifyReturnAddressArgumentIsConstant(SDValue Op, + SelectionDAG &DAG) const; + //===--------------------------------------------------------------------===// // Inline Asm Support hooks // @@ -2165,21 +2543,11 @@ public: /// operand it matches. unsigned getMatchedOperand() const; - /// Copy constructor for copying from an AsmOperandInfo. - AsmOperandInfo(const AsmOperandInfo &info) - : InlineAsm::ConstraintInfo(info), - ConstraintCode(info.ConstraintCode), - ConstraintType(info.ConstraintType), - CallOperandVal(info.CallOperandVal), - ConstraintVT(info.ConstraintVT) { - } - /// Copy constructor for copying from a ConstraintInfo. - AsmOperandInfo(const InlineAsm::ConstraintInfo &info) - : InlineAsm::ConstraintInfo(info), - ConstraintType(TargetLowering::C_Unknown), - CallOperandVal(0), ConstraintVT(MVT::Other) { - } + AsmOperandInfo(InlineAsm::ConstraintInfo Info) + : InlineAsm::ConstraintInfo(std::move(Info)), + ConstraintType(TargetLowering::C_Unknown), CallOperandVal(nullptr), + ConstraintVT(MVT::Other) {} }; typedef std::vector AsmOperandInfoVector; @@ -2206,7 +2574,7 @@ public: /// Op, otherwise an empty SDValue can be passed. virtual void ComputeConstraintToUse(AsmOperandInfo &OpInfo, SDValue Op, - SelectionDAG *DAG = 0) const; + SelectionDAG *DAG = nullptr) const; /// Given a constraint, return the type of constraint it is for this target. virtual ConstraintType getConstraintType(const std::string &Constraint) const; @@ -2240,21 +2608,83 @@ public: // SDValue BuildExactSDIV(SDValue Op1, SDValue Op2, SDLoc dl, SelectionDAG &DAG) const; - SDValue BuildSDIV(SDNode *N, SelectionDAG &DAG, bool IsAfterLegalization, - std::vector *Created) const; - SDValue BuildUDIV(SDNode *N, SelectionDAG &DAG, bool IsAfterLegalization, - std::vector *Created) const; + SDValue BuildSDIV(SDNode *N, const APInt &Divisor, SelectionDAG &DAG, + bool IsAfterLegalization, + std::vector *Created) const; + SDValue BuildUDIV(SDNode *N, const APInt &Divisor, SelectionDAG &DAG, + bool IsAfterLegalization, + std::vector *Created) const; + virtual SDValue BuildSDIVPow2(SDNode *N, const APInt &Divisor, + SelectionDAG &DAG, + std::vector *Created) const { + return SDValue(); + } + + /// Hooks for building estimates in place of slower divisions and square + /// roots. + + /// Return a reciprocal square root estimate value for the input operand. + /// The RefinementSteps output is the number of Newton-Raphson refinement + /// iterations required to generate a sufficient (though not necessarily + /// IEEE-754 compliant) estimate for the value type. + /// A target may choose to implement its own refinement within this function. + /// If that's true, then return '0' as the number of RefinementSteps to avoid + /// any further refinement of the estimate. + /// An empty SDValue return means no estimate sequence can be created. + virtual SDValue getRsqrtEstimate(SDValue Operand, + DAGCombinerInfo &DCI, + unsigned &RefinementSteps) const { + return SDValue(); + } + + /// Return a reciprocal estimate value for the input operand. + /// The RefinementSteps output is the number of Newton-Raphson refinement + /// iterations required to generate a sufficient (though not necessarily + /// IEEE-754 compliant) estimate for the value type. + /// A target may choose to implement its own refinement within this function. + /// If that's true, then return '0' as the number of RefinementSteps to avoid + /// any further refinement of the estimate. + /// An empty SDValue return means no estimate sequence can be created. + virtual SDValue getRecipEstimate(SDValue Operand, + DAGCombinerInfo &DCI, + unsigned &RefinementSteps) const { + return SDValue(); + } + + //===--------------------------------------------------------------------===// + // Legalization utility functions + // + + /// Expand a MUL into two nodes. One that computes the high bits of + /// the result and one that computes the low bits. + /// \param HiLoVT The value type to use for the Lo and Hi nodes. + /// \param LL Low bits of the LHS of the MUL. You can use this parameter + /// if you want to control how low bits are extracted from the LHS. + /// \param LH High bits of the LHS of the MUL. See LL for meaning. + /// \param RL Low bits of the RHS of the MUL. See LL for meaning + /// \param RH High bits of the RHS of the MUL. See LL for meaning. + /// \returns true if the node has been expanded. false if it has not + bool expandMUL(SDNode *N, SDValue &Lo, SDValue &Hi, EVT HiLoVT, + SelectionDAG &DAG, SDValue LL = SDValue(), + SDValue LH = SDValue(), SDValue RL = SDValue(), + SDValue RH = SDValue()) const; + + /// Expand float(f32) to SINT(i64) conversion + /// \param N Node to expand + /// \param Result output after conversion + /// \returns True, if the expansion was successful, false otherwise + bool expandFP_TO_SINT(SDNode *N, SDValue &Result, SelectionDAG &DAG) const; //===--------------------------------------------------------------------===// // Instruction Emitting Hooks // - // This method should be implemented by targets that mark instructions with - // the 'usesCustomInserter' flag. These instructions are special in various - // ways, which require special support to insert. The specified MachineInstr - // is created but not inserted into any basic blocks, and this method is - // called to expand it into a sequence of instructions, potentially also - // creating new basic blocks and control flow. + /// This method should be implemented by targets that mark instructions with + /// the 'usesCustomInserter' flag. These instructions are special in various + /// ways, which require special support to insert. The specified MachineInstr + /// is created but not inserted into any basic blocks, and this method is + /// called to expand it into a sequence of instructions, potentially also + /// creating new basic blocks and control flow. virtual MachineBasicBlock * EmitInstrWithCustomInserter(MachineInstr *MI, MachineBasicBlock *MBB) const; @@ -2264,6 +2694,12 @@ public: /// ARM 's' setting instructions. virtual void AdjustInstrPostInstrSelection(MachineInstr *MI, SDNode *Node) const; + + /// If this function returns true, SelectionDAGBuilder emits a + /// LOAD_STACK_GUARD node when it is lowering Intrinsic::stackprotector. + virtual bool useLoadStackGuardNode() const { + return false; + } }; /// Given an LLVM IR type and return type attributes, compute the return value