X-Git-Url: http://demsky.eecs.uci.edu/git/?a=blobdiff_plain;f=lib%2FTarget%2FPowerPC%2FPPCISelLowering.h;h=c93de430fd0576585424bbbd16131e7bc7291bde;hb=e97d93757641299b2be9d10e7e7caf5fd6855331;hp=9461e46198cdf82298b96b473532cd4da43ec02d;hpb=ee8e48d4c9e834a05cea36bb83b2df6b718aaaa4;p=oota-llvm.git diff --git a/lib/Target/PowerPC/PPCISelLowering.h b/lib/Target/PowerPC/PPCISelLowering.h index 9461e46198c..c93de430fd0 100644 --- a/lib/Target/PowerPC/PPCISelLowering.h +++ b/lib/Target/PowerPC/PPCISelLowering.h @@ -12,20 +12,19 @@ // //===----------------------------------------------------------------------===// -#ifndef LLVM_TARGET_POWERPC_PPC32ISELLOWERING_H -#define LLVM_TARGET_POWERPC_PPC32ISELLOWERING_H +#ifndef LLVM_LIB_TARGET_POWERPC_PPCISELLOWERING_H +#define LLVM_LIB_TARGET_POWERPC_PPCISELLOWERING_H #include "PPC.h" #include "PPCInstrInfo.h" #include "PPCRegisterInfo.h" -#include "PPCSubtarget.h" #include "llvm/CodeGen/CallingConvLower.h" #include "llvm/CodeGen/SelectionDAG.h" #include "llvm/Target/TargetLowering.h" namespace llvm { namespace PPCISD { - enum NodeType { + enum NodeType : unsigned { // Start the numbering where the builtin ops and target ops leave off. FIRST_NUMBER = ISD::BUILTIN_OP_END, @@ -62,6 +61,9 @@ namespace llvm { /// VPERM, + /// The CMPB instruction (takes two operands of i32 or i64). + CMPB, + /// Hi/Lo - These represent the high and low 16-bit parts of a global /// address respectively. These nodes have two operands, the first of /// which must be a TargetGlobalAddress, and the second of which must be a @@ -69,23 +71,9 @@ namespace llvm { /// though these are usually folded into other nodes. Hi, Lo, - TOC_ENTRY, - - /// The following three target-specific nodes are used for calls through + /// The following two target-specific nodes are used for calls through /// function pointers in the 64-bit SVR4 ABI. - /// Restore the TOC from the TOC save area of the current stack frame. - /// This is basically a hard coded load instruction which additionally - /// takes/produces a flag. - TOC_RESTORE, - - /// Like a regular LOAD but additionally taking/producing a flag. - LOAD, - - /// LOAD into r2 (also taking/producing a flag). Like TOC_RESTORE, this is - /// a hard coded load instruction. - LOAD_TOC, - /// OPRC, CHAIN = DYNALLOC(CHAIN, NEGSIZE, FRAME_INDEX) /// This instruction is lowered in PPCRegisterInfo::eliminateFrameIndex to /// compute an allocation on the stack. @@ -100,6 +88,12 @@ namespace llvm { /// code. SRL, SRA, SHL, + /// The combination of sra[wd]i and addze used to implemented signed + /// integer division by a power of 2. The first operand is the dividend, + /// and the second is the constant shift amount (representing the + /// divisor). + SRA_ADDZE, + /// CALL - A direct function call. /// CALL_NOP is a call with the special NOP which follows 64-bit /// SVR4 calls. @@ -113,6 +107,10 @@ namespace llvm { /// BCTRL instruction. BCTRL, + /// CHAIN,FLAG = BCTRL(CHAIN, ADDR, INFLAG) - The combination of a bctrl + /// instruction and the TOC reload required on SVR4 PPC64. + BCTRL_LOAD_TOC, + /// Return with a flag operand, matched by 'blr' RET_FLAG, @@ -121,12 +119,25 @@ namespace llvm { /// resultant GPR. Bits corresponding to other CR regs are undefined. MFOCRF, + /// Direct move from a VSX register to a GPR + MFVSR, + + /// Direct move from a GPR to a VSX register (algebraic) + MTVSRA, + + /// Direct move from a GPR to a VSX register (zero) + MTVSRZ, + // FIXME: Remove these once the ANDI glue bug is fixed: /// i1 = ANDIo_1_[EQ|GT]_BIT(i32 or i64 x) - Represents the result of the /// eq or gt bit of CR0 after executing andi. x, 1. This is used to /// implement truncation of i32 or i64 to i1. ANDIo_1_EQ_BIT, ANDIo_1_GT_BIT, + // READ_TIME_BASE - A read of the 64-bit time-base register on a 32-bit + // target (returns (Lo, Hi)). It takes a chain operand. + READ_TIME_BASE, + // EH_SJLJ_SETJMP - SjLj exception handling setjmp. EH_SJLJ_SETJMP, @@ -164,14 +175,6 @@ namespace llvm { /// F8RC = MFFS - This moves the FPSCR (not modeled) into the register. MFFS, - /// LARX = This corresponds to PPC l{w|d}arx instrcution: load and - /// reserve indexed. This is used to implement atomic operations. - LARX, - - /// STCX = This corresponds to PPC stcx. instrcution: store conditional - /// indexed. This is used to implement atomic operations. - STCX, - /// TC_RETURN - A tail call return. /// operand #0 chain /// operand #1 callee (register or absolute) @@ -187,6 +190,10 @@ namespace llvm { /// on PPC32. PPC32_GOT, + /// GPRC = address of _GLOBAL_OFFSET_TABLE_. Used by general dynamic and + /// local dynamic TLS on PPC32. + PPC32_PICGOT, + /// G8RC = ADDIS_GOT_TPREL_HA %X2, Symbol - Used by the initial-exec /// TLS model, produces an ADDIS8 instruction that adds the GOT /// base to sym\@got\@tprel\@ha. @@ -211,34 +218,46 @@ namespace llvm { /// register to sym\@got\@tlsgd\@ha. ADDIS_TLSGD_HA, - /// G8RC = ADDI_TLSGD_L G8RReg, Symbol - For the general-dynamic TLS + /// %X3 = ADDI_TLSGD_L G8RReg, Symbol - For the general-dynamic TLS /// model, produces an ADDI8 instruction that adds G8RReg to - /// sym\@got\@tlsgd\@l. + /// sym\@got\@tlsgd\@l and stores the result in X3. Hidden by + /// ADDIS_TLSGD_L_ADDR until after register assignment. ADDI_TLSGD_L, - /// G8RC = GET_TLS_ADDR %X3, Symbol - For the general-dynamic TLS - /// model, produces a call to __tls_get_addr(sym\@tlsgd). + /// %X3 = GET_TLS_ADDR %X3, Symbol - For the general-dynamic TLS + /// model, produces a call to __tls_get_addr(sym\@tlsgd). Hidden by + /// ADDIS_TLSGD_L_ADDR until after register assignment. GET_TLS_ADDR, + /// G8RC = ADDI_TLSGD_L_ADDR G8RReg, Symbol, Symbol - Op that + /// combines ADDI_TLSGD_L and GET_TLS_ADDR until expansion following + /// register assignment. + ADDI_TLSGD_L_ADDR, + /// G8RC = ADDIS_TLSLD_HA %X2, Symbol - For the local-dynamic TLS /// model, produces an ADDIS8 instruction that adds the GOT base /// register to sym\@got\@tlsld\@ha. ADDIS_TLSLD_HA, - /// G8RC = ADDI_TLSLD_L G8RReg, Symbol - For the local-dynamic TLS + /// %X3 = ADDI_TLSLD_L G8RReg, Symbol - For the local-dynamic TLS /// model, produces an ADDI8 instruction that adds G8RReg to - /// sym\@got\@tlsld\@l. + /// sym\@got\@tlsld\@l and stores the result in X3. Hidden by + /// ADDIS_TLSLD_L_ADDR until after register assignment. ADDI_TLSLD_L, - /// G8RC = GET_TLSLD_ADDR %X3, Symbol - For the local-dynamic TLS - /// model, produces a call to __tls_get_addr(sym\@tlsld). + /// %X3 = GET_TLSLD_ADDR %X3, Symbol - For the local-dynamic TLS + /// model, produces a call to __tls_get_addr(sym\@tlsld). Hidden by + /// ADDIS_TLSLD_L_ADDR until after register assignment. GET_TLSLD_ADDR, - /// G8RC = ADDIS_DTPREL_HA %X3, Symbol, Chain - For the - /// local-dynamic TLS model, produces an ADDIS8 instruction - /// that adds X3 to sym\@dtprel\@ha. The Chain operand is needed - /// to tie this in place following a copy to %X3 from the result - /// of a GET_TLSLD_ADDR. + /// G8RC = ADDI_TLSLD_L_ADDR G8RReg, Symbol, Symbol - Op that + /// combines ADDI_TLSLD_L and GET_TLSLD_ADDR until expansion + /// following register assignment. + ADDI_TLSLD_L_ADDR, + + /// G8RC = ADDIS_DTPREL_HA %X3, Symbol - For the local-dynamic TLS + /// model, produces an ADDIS8 instruction that adds X3 to + /// sym\@dtprel\@ha. ADDIS_DTPREL_HA, /// G8RC = ADDI_DTPREL_L G8RReg, Symbol - For the local-dynamic TLS @@ -256,6 +275,39 @@ namespace llvm { /// operand identifies the operating system entry point. SC, + /// CHAIN = CLRBHRB CHAIN - Clear branch history rolling buffer. + CLRBHRB, + + /// GPRC, CHAIN = MFBHRBE CHAIN, Entry, Dummy - Move from branch + /// history rolling buffer entry. + MFBHRBE, + + /// CHAIN = RFEBB CHAIN, State - Return from event-based branch. + RFEBB, + + /// VSRC, CHAIN = XXSWAPD CHAIN, VSRC - Occurs only for little + /// endian. Maps to an xxswapd instruction that corrects an lxvd2x + /// or stxvd2x instruction. The chain is necessary because the + /// sequence replaces a load and needs to provide the same number + /// of outputs. + XXSWAPD, + + /// QVFPERM = This corresponds to the QPX qvfperm instruction. + QVFPERM, + + /// QVGPCI = This corresponds to the QPX qvgpci instruction. + QVGPCI, + + /// QVALIGNI = This corresponds to the QPX qvaligni instruction. + QVALIGNI, + + /// QVESPLATI = This corresponds to the QPX qvesplati instruction. + QVESPLATI, + + /// QBFLT = Access the underlying QPX floating-point boolean + /// representation. + QBFLT, + /// CHAIN = STBRX CHAIN, GPRC, Ptr, Type - This is a /// byte-swapping store instruction. It byte-swaps the low "Type" bits of /// the GPRC input, then stores it through Ptr. Type can be either i16 or @@ -282,20 +334,24 @@ namespace llvm { /// destination 64-bit register. LFIWZX, - /// G8RC = ADDIS_TOC_HA %X2, Symbol - For medium and large code model, - /// produces an ADDIS8 instruction that adds the TOC base register to - /// sym\@toc\@ha. - ADDIS_TOC_HA, + /// VSRC, CHAIN = LXVD2X_LE CHAIN, Ptr - Occurs only for little endian. + /// Maps directly to an lxvd2x instruction that will be followed by + /// an xxswapd. + LXVD2X, + + /// CHAIN = STXVD2X CHAIN, VSRC, Ptr - Occurs only for little endian. + /// Maps directly to an stxvd2x instruction that will be preceded by + /// an xxswapd. + STXVD2X, - /// G8RC = LD_TOC_L Symbol, G8RReg - For medium and large code model, - /// produces a LD instruction with base register G8RReg and offset - /// sym\@toc\@l. Preceded by an ADDIS_TOC_HA to form a full 32-bit offset. - LD_TOC_L, + /// QBRC, CHAIN = QVLFSb CHAIN, Ptr + /// The 4xf32 load used for v4i1 constants. + QVLFSb, - /// G8RC = ADDI_TOC_L G8RReg, Symbol - For medium code model, produces - /// an ADDI8 instruction that adds G8RReg to sym\@toc\@l. - /// Preceded by an ADDIS_TOC_HA to form a full 32-bit offset. - ADDI_TOC_L + /// GPRC = TOC_ENTRY GA, TOC + /// Loads the entry for GA from the TOC, where the TOC base is given by + /// the last operand. + TOC_ENTRY }; } @@ -303,68 +359,90 @@ namespace llvm { namespace PPC { /// isVPKUHUMShuffleMask - Return true if this is the shuffle mask for a /// VPKUHUM instruction. - bool isVPKUHUMShuffleMask(ShuffleVectorSDNode *N, bool isUnary); + bool isVPKUHUMShuffleMask(ShuffleVectorSDNode *N, unsigned ShuffleKind, + SelectionDAG &DAG); /// isVPKUWUMShuffleMask - Return true if this is the shuffle mask for a /// VPKUWUM instruction. - bool isVPKUWUMShuffleMask(ShuffleVectorSDNode *N, bool isUnary); + bool isVPKUWUMShuffleMask(ShuffleVectorSDNode *N, unsigned ShuffleKind, + SelectionDAG &DAG); + + /// isVPKUDUMShuffleMask - Return true if this is the shuffle mask for a + /// VPKUDUM instruction. + bool isVPKUDUMShuffleMask(ShuffleVectorSDNode *N, unsigned ShuffleKind, + SelectionDAG &DAG); /// isVMRGLShuffleMask - Return true if this is a shuffle mask suitable for /// a VRGL* instruction with the specified unit size (1,2 or 4 bytes). bool isVMRGLShuffleMask(ShuffleVectorSDNode *N, unsigned UnitSize, - bool isUnary); + unsigned ShuffleKind, SelectionDAG &DAG); /// isVMRGHShuffleMask - Return true if this is a shuffle mask suitable for /// a VRGH* instruction with the specified unit size (1,2 or 4 bytes). bool isVMRGHShuffleMask(ShuffleVectorSDNode *N, unsigned UnitSize, - bool isUnary); + unsigned ShuffleKind, SelectionDAG &DAG); - /// isVSLDOIShuffleMask - If this is a vsldoi shuffle mask, return the shift - /// amount, otherwise return -1. - int isVSLDOIShuffleMask(SDNode *N, bool isUnary); + /// isVSLDOIShuffleMask - If this is a vsldoi shuffle mask, return the + /// shift amount, otherwise return -1. + int isVSLDOIShuffleMask(SDNode *N, unsigned ShuffleKind, + SelectionDAG &DAG); /// isSplatShuffleMask - Return true if the specified VECTOR_SHUFFLE operand /// specifies a splat of a single element that is suitable for input to /// VSPLTB/VSPLTH/VSPLTW. bool isSplatShuffleMask(ShuffleVectorSDNode *N, unsigned EltSize); - /// isAllNegativeZeroVector - Returns true if all elements of build_vector - /// are -0.0. - bool isAllNegativeZeroVector(SDNode *N); - /// getVSPLTImmediate - Return the appropriate VSPLT* immediate to splat the /// specified isSplatShuffleMask VECTOR_SHUFFLE mask. - unsigned getVSPLTImmediate(SDNode *N, unsigned EltSize); + unsigned getVSPLTImmediate(SDNode *N, unsigned EltSize, SelectionDAG &DAG); /// get_VSPLTI_elt - If this is a build_vector of constants which can be /// formed by using a vspltis[bhw] instruction of the specified element /// size, return the constant being splatted. The ByteSize field indicates /// the number of bytes of each element [124] -> [bhw]. SDValue get_VSPLTI_elt(SDNode *N, unsigned ByteSize, SelectionDAG &DAG); + + /// If this is a qvaligni shuffle mask, return the shift + /// amount, otherwise return -1. + int isQVALIGNIShuffleMask(SDNode *N); } class PPCTargetLowering : public TargetLowering { - const PPCSubtarget &PPCSubTarget; + const PPCSubtarget &Subtarget; public: - explicit PPCTargetLowering(PPCTargetMachine &TM); + explicit PPCTargetLowering(const PPCTargetMachine &TM, + const PPCSubtarget &STI); /// getTargetNodeName() - This method returns the name of a target specific /// DAG node. - virtual const char *getTargetNodeName(unsigned Opcode) const; + const char *getTargetNodeName(unsigned Opcode) const override; - virtual MVT getScalarShiftAmountTy(EVT LHSTy) const { return MVT::i32; } + MVT getScalarShiftAmountTy(EVT LHSTy) const override { return MVT::i32; } + + bool isCheapToSpeculateCttz() const override { + return true; + } + + bool isCheapToSpeculateCtlz() const override { + return true; + } /// getSetCCResultType - Return the ISD::SETCC ValueType - virtual EVT getSetCCResultType(LLVMContext &Context, EVT VT) const; + EVT getSetCCResultType(LLVMContext &Context, EVT VT) const override; + + /// 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. + bool enableAggressiveFMAFusion(EVT VT) const override; /// getPreIndexedAddressParts - returns true by value, base pointer and /// offset pointer and addressing mode by reference if the node's address /// can be legally represented as pre-indexed load / store address. - virtual bool getPreIndexedAddressParts(SDNode *N, SDValue &Base, - SDValue &Offset, - ISD::MemIndexedMode &AM, - SelectionDAG &DAG) const; + bool getPreIndexedAddressParts(SDNode *N, SDValue &Base, + SDValue &Offset, + ISD::MemIndexedMode &AM, + SelectionDAG &DAG) const override; /// SelectAddressRegReg - Given the specified addressed, check to see if it /// can be represented as an indexed [r+r] operation. Returns false if it @@ -384,31 +462,47 @@ namespace llvm { bool SelectAddressRegRegOnly(SDValue N, SDValue &Base, SDValue &Index, SelectionDAG &DAG) const; - Sched::Preference getSchedulingPreference(SDNode *N) const; + Sched::Preference getSchedulingPreference(SDNode *N) const override; /// LowerOperation - Provide custom lowering hooks for some operations. /// - virtual SDValue LowerOperation(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerOperation(SDValue Op, SelectionDAG &DAG) const override; /// ReplaceNodeResults - Replace the results of node with an illegal result /// type with new values built out of custom code. /// - virtual void ReplaceNodeResults(SDNode *N, SmallVectorImpl&Results, - SelectionDAG &DAG) const; + void ReplaceNodeResults(SDNode *N, SmallVectorImpl&Results, + SelectionDAG &DAG) const override; + + SDValue expandVSXLoadForLE(SDNode *N, DAGCombinerInfo &DCI) const; + SDValue expandVSXStoreForLE(SDNode *N, DAGCombinerInfo &DCI) const; + + SDValue PerformDAGCombine(SDNode *N, DAGCombinerInfo &DCI) const override; + + SDValue BuildSDIVPow2(SDNode *N, const APInt &Divisor, SelectionDAG &DAG, + std::vector *Created) const override; + + unsigned getRegisterByName(const char* RegName, EVT VT) const override; + + void computeKnownBitsForTargetNode(const SDValue Op, + APInt &KnownZero, + APInt &KnownOne, + const SelectionDAG &DAG, + unsigned Depth = 0) const override; - virtual SDValue PerformDAGCombine(SDNode *N, DAGCombinerInfo &DCI) const; + unsigned getPrefLoopAlignment(MachineLoop *ML) const override; - virtual void computeMaskedBitsForTargetNode(const SDValue Op, - APInt &KnownZero, - APInt &KnownOne, - const SelectionDAG &DAG, - unsigned Depth = 0) const; + Instruction* emitLeadingFence(IRBuilder<> &Builder, AtomicOrdering Ord, + bool IsStore, bool IsLoad) const override; + Instruction* emitTrailingFence(IRBuilder<> &Builder, AtomicOrdering Ord, + bool IsStore, bool IsLoad) const override; - virtual MachineBasicBlock * + MachineBasicBlock * EmitInstrWithCustomInserter(MachineInstr *MI, - MachineBasicBlock *MBB) const; + MachineBasicBlock *MBB) const override; MachineBasicBlock *EmitAtomicBinary(MachineInstr *MI, - MachineBasicBlock *MBB, bool is64Bit, + MachineBasicBlock *MBB, + unsigned AtomicSize, unsigned BinOpcode) const; MachineBasicBlock *EmitPartwordAtomicBinary(MachineInstr *MI, MachineBasicBlock *MBB, @@ -420,34 +514,82 @@ namespace llvm { MachineBasicBlock *emitEHSjLjLongJmp(MachineInstr *MI, MachineBasicBlock *MBB) const; - ConstraintType getConstraintType(const std::string &Constraint) const; + ConstraintType + getConstraintType(const std::string &Constraint) const override; /// Examine constraint string and operand type and determine a weight value. /// The operand object must already have been set up with the operand type. ConstraintWeight getSingleConstraintMatchWeight( - AsmOperandInfo &info, const char *constraint) const; + AsmOperandInfo &info, const char *constraint) const override; - std::pair - getRegForInlineAsmConstraint(const std::string &Constraint, - MVT VT) const; + std::pair + getRegForInlineAsmConstraint(const TargetRegisterInfo *TRI, + const std::string &Constraint, + MVT VT) const override; /// getByValTypeAlignment - Return the desired alignment for ByVal aggregate /// function arguments in the caller parameter area. This is the actual /// alignment, not its logarithm. - unsigned getByValTypeAlignment(Type *Ty) const; + unsigned getByValTypeAlignment(Type *Ty) const override; /// LowerAsmOperandForConstraint - Lower the specified operand into the Ops /// vector. If it is invalid, don't add anything to Ops. - virtual void LowerAsmOperandForConstraint(SDValue Op, - std::string &Constraint, - std::vector &Ops, - SelectionDAG &DAG) const; + void LowerAsmOperandForConstraint(SDValue Op, + std::string &Constraint, + std::vector &Ops, + SelectionDAG &DAG) const override; + + unsigned getInlineAsmMemConstraint( + const std::string &ConstraintCode) const override { + if (ConstraintCode == "es") + return InlineAsm::Constraint_es; + else if (ConstraintCode == "o") + return InlineAsm::Constraint_o; + else if (ConstraintCode == "Q") + return InlineAsm::Constraint_Q; + else if (ConstraintCode == "Z") + return InlineAsm::Constraint_Z; + else if (ConstraintCode == "Zy") + return InlineAsm::Constraint_Zy; + return TargetLowering::getInlineAsmMemConstraint(ConstraintCode); + } /// isLegalAddressingMode - Return true if the addressing mode represented /// by AM is legal for this target, for a load/store of the specified type. - virtual bool isLegalAddressingMode(const AddrMode &AM, Type *Ty)const; + bool isLegalAddressingMode(const AddrMode &AM, Type *Ty) const override; + + /// isLegalICmpImmediate - Return true if the specified immediate is legal + /// icmp immediate, that is the target has icmp instructions which can + /// compare a register against the immediate without having to materialize + /// the immediate into a register. + bool isLegalICmpImmediate(int64_t Imm) const override; + + /// isLegalAddImmediate - Return true if the specified immediate is legal + /// add immediate, that is the target has add instructions which can + /// add a register and the immediate without having to materialize + /// the immediate into a register. + bool isLegalAddImmediate(int64_t Imm) const override; + + /// isTruncateFree - Return true if it's free to truncate a value of + /// type Ty1 to type Ty2. e.g. On PPC it's free to truncate a i64 value in + /// register X1 to i32 by referencing its sub-register R1. + bool isTruncateFree(Type *Ty1, Type *Ty2) const override; + bool isTruncateFree(EVT VT1, EVT VT2) const override; + + bool isZExtFree(SDValue Val, EVT VT2) const override; + + bool isFPExtFree(EVT VT) const override; + + /// \brief Returns true if it is beneficial to convert a load of a constant + /// to just the constant itself. + bool shouldConvertConstantLoadToIntImm(const APInt &Imm, + Type *Ty) const override; + + bool isOffsetFoldingLegal(const GlobalAddressSDNode *GA) const override; - virtual bool isOffsetFoldingLegal(const GlobalAddressSDNode *GA) const; + bool getTgtMemIntrinsic(IntrinsicInfo &Info, + const CallInst &I, + unsigned Intrinsic) const override; /// getOptimalMemOpType - Returns the target specific optimal type for load /// and store operations as a result of memset, memcpy, and memmove @@ -460,29 +602,78 @@ namespace llvm { /// source is constant so it does not need to be loaded. /// It returns EVT::Other if the type should be determined using generic /// target-independent logic. - virtual EVT + EVT getOptimalMemOpType(uint64_t Size, unsigned DstAlign, unsigned SrcAlign, bool IsMemset, bool ZeroMemset, bool MemcpyStrSrc, - MachineFunction &MF) const; + MachineFunction &MF) const override; /// Is unaligned memory access allowed for the given type, and is it fast /// relative to software emulation. - virtual bool allowsUnalignedMemoryAccesses(EVT VT, - unsigned AddrSpace, - bool *Fast = 0) const; + bool allowsMisalignedMemoryAccesses(EVT VT, + unsigned AddrSpace, + unsigned Align = 1, + bool *Fast = nullptr) const override; /// isFMAFasterThanFMulAndFAdd - Return true if an FMA operation is faster /// than a pair of fmul and fadd instructions. fmuladd intrinsics will be /// expanded to FMAs when this method returns true, otherwise fmuladd is /// expanded to fmul + fadd. - virtual bool isFMAFasterThanFMulAndFAdd(EVT VT) const; + bool isFMAFasterThanFMulAndFAdd(EVT VT) const override; + + const MCPhysReg *getScratchRegisters(CallingConv::ID CC) const override; + + // Should we expand the build vector with shuffles? + bool + shouldExpandBuildVectorWithShuffles(EVT VT, + unsigned DefinedValues) const override; /// createFastISel - This method returns a target-specific FastISel object, /// or null if the target does not support "fast" instruction selection. - virtual FastISel *createFastISel(FunctionLoweringInfo &FuncInfo, - const TargetLibraryInfo *LibInfo) const; + FastISel *createFastISel(FunctionLoweringInfo &FuncInfo, + const TargetLibraryInfo *LibInfo) const override; + + /// \brief Returns true if an argument of type Ty needs to be passed in a + /// contiguous block of registers in calling convention CallConv. + bool functionArgumentNeedsConsecutiveRegisters( + Type *Ty, CallingConv::ID CallConv, bool isVarArg) const override { + // We support any array type as "consecutive" block in the parameter + // save area. The element type defines the alignment requirement and + // whether the argument should go in GPRs, FPRs, or VRs if available. + // + // Note that clang uses this capability both to implement the ELFv2 + // homogeneous float/vector aggregate ABI, and to avoid having to use + // "byval" when passing aggregates that might fully fit in registers. + return Ty->isArrayTy(); + } private: + + struct ReuseLoadInfo { + SDValue Ptr; + SDValue Chain; + SDValue ResChain; + MachinePointerInfo MPI; + bool IsInvariant; + unsigned Alignment; + AAMDNodes AAInfo; + const MDNode *Ranges; + + ReuseLoadInfo() : IsInvariant(false), Alignment(0), Ranges(nullptr) {} + }; + + bool canReuseLoadAddress(SDValue Op, EVT MemVT, ReuseLoadInfo &RLI, + SelectionDAG &DAG, + ISD::LoadExtType ET = ISD::NON_EXTLOAD) const; + void spliceIntoChain(SDValue ResChain, SDValue NewResChain, + SelectionDAG &DAG) const; + + void LowerFP_TO_INTForReuse(SDValue Op, ReuseLoadInfo &RLI, + SelectionDAG &DAG, SDLoc dl) const; + SDValue LowerFP_TO_INTDirectMove(SDValue Op, SelectionDAG &DAG, + SDLoc dl) const; + SDValue LowerINT_TO_FPDirectMove(SDValue Op, SelectionDAG &DAG, + SDLoc dl) const; + SDValue getFramePointerFrameIndex(SelectionDAG & DAG) const; SDValue getReturnAddrFrameIndex(SelectionDAG & DAG) const; @@ -533,60 +724,60 @@ namespace llvm { SDValue LowerSRA_PARTS(SDValue Op, SelectionDAG &DAG) const; SDValue LowerBUILD_VECTOR(SDValue Op, SelectionDAG &DAG) const; SDValue LowerVECTOR_SHUFFLE(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerEXTRACT_VECTOR_ELT(SDValue Op, SelectionDAG &DAG) const; SDValue LowerINTRINSIC_WO_CHAIN(SDValue Op, SelectionDAG &DAG) const; SDValue LowerSCALAR_TO_VECTOR(SDValue Op, SelectionDAG &DAG) const; SDValue LowerSIGN_EXTEND_INREG(SDValue Op, SelectionDAG &DAG) const; SDValue LowerMUL(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerVectorLoad(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerVectorStore(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerCallResult(SDValue Chain, SDValue InFlag, CallingConv::ID CallConv, bool isVarArg, const SmallVectorImpl &Ins, SDLoc dl, SelectionDAG &DAG, SmallVectorImpl &InVals) const; SDValue FinishCall(CallingConv::ID CallConv, SDLoc dl, bool isTailCall, - bool isVarArg, + bool isVarArg, bool IsPatchPoint, SelectionDAG &DAG, SmallVector, 8> &RegsToPass, - SDValue InFlag, SDValue Chain, + SDValue InFlag, SDValue Chain, SDValue CallSeqStart, SDValue &Callee, int SPDiff, unsigned NumBytes, const SmallVectorImpl &Ins, - SmallVectorImpl &InVals) const; + SmallVectorImpl &InVals, + ImmutableCallSite *CS) const; - virtual SDValue + SDValue LowerFormalArguments(SDValue Chain, CallingConv::ID CallConv, bool isVarArg, const SmallVectorImpl &Ins, SDLoc dl, SelectionDAG &DAG, - SmallVectorImpl &InVals) const; + SmallVectorImpl &InVals) const override; - virtual SDValue + SDValue LowerCall(TargetLowering::CallLoweringInfo &CLI, - SmallVectorImpl &InVals) const; + SmallVectorImpl &InVals) const override; - virtual bool + bool CanLowerReturn(CallingConv::ID CallConv, MachineFunction &MF, bool isVarArg, const SmallVectorImpl &Outs, - LLVMContext &Context) const; + LLVMContext &Context) const override; - virtual SDValue + SDValue LowerReturn(SDValue Chain, CallingConv::ID CallConv, bool isVarArg, const SmallVectorImpl &Outs, const SmallVectorImpl &OutVals, - SDLoc dl, SelectionDAG &DAG) const; + SDLoc dl, SelectionDAG &DAG) const override; SDValue extendArgForPPC64(ISD::ArgFlagsTy Flags, EVT ObjectVT, SelectionDAG &DAG, SDValue ArgVal, SDLoc dl) const; - void - setMinReservedArea(MachineFunction &MF, SelectionDAG &DAG, - unsigned nAltivecParamsAtEnd, - unsigned MinReservedArea, bool isPPC64) const; - SDValue LowerFormalArguments_Darwin(SDValue Chain, CallingConv::ID CallConv, bool isVarArg, @@ -614,37 +805,46 @@ namespace llvm { SDValue LowerCall_Darwin(SDValue Chain, SDValue Callee, CallingConv::ID CallConv, - bool isVarArg, bool isTailCall, + bool isVarArg, bool isTailCall, bool IsPatchPoint, const SmallVectorImpl &Outs, const SmallVectorImpl &OutVals, const SmallVectorImpl &Ins, SDLoc dl, SelectionDAG &DAG, - SmallVectorImpl &InVals) const; + SmallVectorImpl &InVals, + ImmutableCallSite *CS) const; SDValue LowerCall_64SVR4(SDValue Chain, SDValue Callee, CallingConv::ID CallConv, - bool isVarArg, bool isTailCall, + bool isVarArg, bool isTailCall, bool IsPatchPoint, const SmallVectorImpl &Outs, const SmallVectorImpl &OutVals, const SmallVectorImpl &Ins, SDLoc dl, SelectionDAG &DAG, - SmallVectorImpl &InVals) const; + SmallVectorImpl &InVals, + ImmutableCallSite *CS) const; SDValue LowerCall_32SVR4(SDValue Chain, SDValue Callee, CallingConv::ID CallConv, - bool isVarArg, bool isTailCall, + bool isVarArg, bool isTailCall, bool IsPatchPoint, const SmallVectorImpl &Outs, const SmallVectorImpl &OutVals, const SmallVectorImpl &Ins, SDLoc dl, SelectionDAG &DAG, - SmallVectorImpl &InVals) const; + SmallVectorImpl &InVals, + ImmutableCallSite *CS) const; SDValue lowerEH_SJLJ_SETJMP(SDValue Op, SelectionDAG &DAG) const; SDValue lowerEH_SJLJ_LONGJMP(SDValue Op, SelectionDAG &DAG) const; SDValue DAGCombineExtBoolTrunc(SDNode *N, DAGCombinerInfo &DCI) const; SDValue DAGCombineTruncBoolExt(SDNode *N, DAGCombinerInfo &DCI) const; - SDValue DAGCombineFastRecip(SDValue Op, DAGCombinerInfo &DCI) const; - SDValue DAGCombineFastRecipFSQRT(SDValue Op, DAGCombinerInfo &DCI) const; + SDValue combineFPToIntToFP(SDNode *N, DAGCombinerInfo &DCI) const; + + SDValue getRsqrtEstimate(SDValue Operand, DAGCombinerInfo &DCI, + unsigned &RefinementSteps, + bool &UseOneConstNR) const override; + SDValue getRecipEstimate(SDValue Operand, DAGCombinerInfo &DCI, + unsigned &RefinementSteps) const override; + bool combineRepeatedFPDivisors(unsigned NumUsers) const override; CCAssignFn *useFastISelCCs(unsigned Flag) const; };