X-Git-Url: http://demsky.eecs.uci.edu/git/?a=blobdiff_plain;f=lib%2FTarget%2FPowerPC%2FPPCISelLowering.h;h=5e5019b0e557cc08c974ae799a775bff92b7c091;hb=4ecff11794c6d57cbb240fa5ee730dd1cd0e1b65;hp=3d22d0427dbfdce383f93f1f22c8cd48a86c503a;hpb=fb2e752e4175920d0531f2afc93a23d0cdf4db14;p=oota-llvm.git diff --git a/lib/Target/PowerPC/PPCISelLowering.h b/lib/Target/PowerPC/PPCISelLowering.h index 3d22d0427db..5e5019b0e55 100644 --- a/lib/Target/PowerPC/PPCISelLowering.h +++ b/lib/Target/PowerPC/PPCISelLowering.h @@ -15,10 +15,13 @@ #ifndef LLVM_TARGET_POWERPC_PPC32ISELLOWERING_H #define LLVM_TARGET_POWERPC_PPC32ISELLOWERING_H -#include "llvm/Target/TargetLowering.h" -#include "llvm/CodeGen/SelectionDAG.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 { @@ -29,133 +32,137 @@ namespace llvm { /// FSEL - Traditional three-operand fsel node. /// FSEL, - + /// FCFID - The FCFID instruction, taking an f64 operand and producing /// and f64 value containing the FP representation of the integer that /// was temporarily in the f64 operand. FCFID, - - /// FCTI[D,W]Z - The FCTIDZ and FCTIWZ instructions, taking an f32 or f64 + + /// Newer FCFID[US] integer-to-floating-point conversion instructions for + /// unsigned integers and single-precision outputs. + FCFIDU, FCFIDS, FCFIDUS, + + /// FCTI[D,W]Z - The FCTIDZ and FCTIWZ instructions, taking an f32 or f64 /// operand, producing an f64 value containing the integer representation /// of that FP value. FCTIDZ, FCTIWZ, - - /// STFIWX - The STFIWX instruction. The first operand is an input token - /// chain, then an f64 value to store, then an address to store it to, - /// then a SRCVALUE for the address. - STFIWX, - + + /// Newer FCTI[D,W]UZ floating-point-to-integer conversion instructions for + /// unsigned integers. + FCTIDUZ, FCTIWUZ, + + /// Reciprocal estimate instructions (unary FP ops). + FRE, FRSQRTE, + // VMADDFP, VNMSUBFP - The VMADDFP and VNMSUBFP instructions, taking // three v4f32 operands and producing a v4f32 result. VMADDFP, VNMSUBFP, - + /// VPERM - The PPC VPERM Instruction. /// VPERM, - + /// 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 /// Constant. Selected naively, these turn into 'lis G+C' and 'li G+C', /// though these are usually folded into other nodes. Hi, Lo, - + TOC_ENTRY, + /// The following three 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. DYNALLOC, - + /// GlobalBaseReg - On Darwin, this node represents the result of the mflr /// at function entry, used for PIC code. GlobalBaseReg, - + /// These nodes represent the 32-bit PPC shifts that operate on 6-bit /// shift amounts. These nodes are generated by the multi-precision shift /// code. SRL, SRA, SHL, - - /// EXTSW_32 - This is the EXTSW instruction for use with "32-bit" - /// registers. - EXTSW_32, - - /// STD_32 - This is the STD instruction for use with "32-bit" registers. - STD_32, - + /// CALL - A direct function call. - CALL_Darwin, CALL_SVR4, - - /// NOP - Special NOP which follows 64-bit SVR4 calls. - NOP, + /// CALL_NOP is a call with the special NOP which follows 64-bit + /// SVR4 calls. + CALL, CALL_NOP, /// CHAIN,FLAG = MTCTR(VAL, CHAIN[, INFLAG]) - Directly corresponds to a /// MTCTR instruction. MTCTR, - + /// CHAIN,FLAG = BCTRL(CHAIN, INFLAG) - Directly corresponds to a /// BCTRL instruction. - BCTRL_Darwin, BCTRL_SVR4, - + BCTRL, + /// Return with a flag operand, matched by 'blr' RET_FLAG, - - /// R32 = MFCR(CRREG, INFLAG) - Represents the MFCR/MFOCRF instructions. + + /// R32 = MFOCRF(CRREG, INFLAG) - Represents the MFOCRF instruction. /// This copies the bits corresponding to the specified CRREG into the /// resultant GPR. Bits corresponding to other CR regs are undefined. - MFCR, + MFOCRF, + + // 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, + + // EH_SJLJ_SETJMP - SjLj exception handling setjmp. + EH_SJLJ_SETJMP, + + // EH_SJLJ_LONGJMP - SjLj exception handling longjmp. + EH_SJLJ_LONGJMP, /// RESVEC = VCMP(LHS, RHS, OPC) - Represents one of the altivec VCMP* /// instructions. For lack of better number, we use the opcode number /// encoding for the OPC field to identify the compare. For example, 838 /// is VCMPGTSH. VCMP, - + /// RESVEC, OUTFLAG = VCMPo(LHS, RHS, OPC) - Represents one of the - /// altivec VCMP*o instructions. For lack of better number, we use the + /// altivec VCMP*o instructions. For lack of better number, we use the /// opcode number encoding for the OPC field to identify the compare. For /// example, 838 is VCMPGTSH. VCMPo, - + /// CHAIN = COND_BRANCH CHAIN, CRRC, OPC, DESTBB [, INFLAG] - This /// corresponds to the COND_BRANCH pseudo instruction. CRRC is the /// condition register to branch on, OPC is the branch opcode to use (e.g. /// PPC::BLE), DESTBB is the destination block to branch to, and INFLAG is /// an optional input flag argument. COND_BRANCH, - - /// CHAIN = STBRX CHAIN, GPRC, Ptr, SRCVALUE, 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 - /// i32. - STBRX, - - /// GPRC, CHAIN = LBRX CHAIN, Ptr, SRCVALUE, Type - This is a - /// byte-swapping load instruction. It loads "Type" bits, byte swaps it, - /// then puts it in the bottom bits of the GPRC. TYPE can be either i16 - /// or i32. - LBRX, - - // The following 5 instructions are used only as part of the - // long double-to-int conversion sequence. - - /// OUTFLAG = MFFS F8RC - This moves the FPSCR (not modelled) into the - /// register. - MFFS, - /// OUTFLAG = MTFSB0 INFLAG - This clears a bit in the FPSCR. - MTFSB0, + /// CHAIN = BDNZ CHAIN, DESTBB - These are used to create counter-based + /// loops. + BDNZ, BDZ, - /// OUTFLAG = MTFSB1 INFLAG - This sets a bit in the FPSCR. - MTFSB1, - - /// F8RC, OUTFLAG = FADDRTZ F8RC, F8RC, INFLAG - This is an FADD done with - /// rounding towards zero. It has flags added so it won't move past the - /// FPSCR-setting instructions. + /// F8RC = FADDRTZ F8RC, F8RC - This is an FADD done with rounding + /// towards zero. Used only as part of the long double-to-int + /// conversion sequence. FADDRTZ, - /// MTFSF = F8RC, INFLAG - This moves the register into the FPSCR. - MTFSF, + /// 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. @@ -170,7 +177,125 @@ namespace llvm { /// operand #1 callee (register or absolute) /// operand #2 stack adjustment /// operand #3 optional in flag - TC_RETURN + TC_RETURN, + + /// ch, gl = CR6[UN]SET ch, inglue - Toggle CR bit 6 for SVR4 vararg calls + CR6SET, + CR6UNSET, + + /// GPRC = address of _GLOBAL_OFFSET_TABLE_. Used by initial-exec TLS + /// on PPC32. + PPC32_GOT, + + /// 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. + ADDIS_GOT_TPREL_HA, + + /// G8RC = LD_GOT_TPREL_L Symbol, G8RReg - Used by the initial-exec + /// TLS model, produces a LD instruction with base register G8RReg + /// and offset sym\@got\@tprel\@l. This completes the addition that + /// finds the offset of "sym" relative to the thread pointer. + LD_GOT_TPREL_L, + + /// G8RC = ADD_TLS G8RReg, Symbol - Used by the initial-exec TLS + /// model, produces an ADD instruction that adds the contents of + /// G8RReg to the thread pointer. Symbol contains a relocation + /// sym\@tls which is to be replaced by the thread pointer and + /// identifies to the linker that the instruction is part of a + /// TLS sequence. + ADD_TLS, + + /// G8RC = ADDIS_TLSGD_HA %X2, Symbol - For the general-dynamic TLS + /// model, produces an ADDIS8 instruction that adds the GOT base + /// register to sym\@got\@tlsgd\@ha. + ADDIS_TLSGD_HA, + + /// G8RC = ADDI_TLSGD_L G8RReg, Symbol - For the general-dynamic TLS + /// model, produces an ADDI8 instruction that adds G8RReg to + /// sym\@got\@tlsgd\@l. + ADDI_TLSGD_L, + + /// G8RC = GET_TLS_ADDR %X3, Symbol - For the general-dynamic TLS + /// model, produces a call to __tls_get_addr(sym\@tlsgd). + GET_TLS_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 + /// model, produces an ADDI8 instruction that adds G8RReg to + /// sym\@got\@tlsld\@l. + ADDI_TLSLD_L, + + /// G8RC = GET_TLSLD_ADDR %X3, Symbol - For the local-dynamic TLS + /// model, produces a call to __tls_get_addr(sym\@tlsld). + 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. + ADDIS_DTPREL_HA, + + /// G8RC = ADDI_DTPREL_L G8RReg, Symbol - For the local-dynamic TLS + /// model, produces an ADDI8 instruction that adds G8RReg to + /// sym\@got\@dtprel\@l. + ADDI_DTPREL_L, + + /// VRRC = VADD_SPLAT Elt, EltSize - Temporary node to be expanded + /// during instruction selection to optimize a BUILD_VECTOR into + /// operations on splats. This is necessary to avoid losing these + /// optimizations due to constant folding. + VADD_SPLAT, + + /// CHAIN = SC CHAIN, Imm128 - System call. The 7-bit unsigned + /// operand identifies the operating system entry point. + SC, + + /// 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 + /// i32. + STBRX = ISD::FIRST_TARGET_MEMORY_OPCODE, + + /// GPRC, CHAIN = LBRX CHAIN, Ptr, Type - This is a + /// byte-swapping load instruction. It loads "Type" bits, byte swaps it, + /// then puts it in the bottom bits of the GPRC. TYPE can be either i16 + /// or i32. + LBRX, + + /// STFIWX - The STFIWX instruction. The first operand is an input token + /// chain, then an f64 value to store, then an address to store it to. + STFIWX, + + /// GPRC, CHAIN = LFIWAX CHAIN, Ptr - This is a floating-point + /// load which sign-extends from a 32-bit integer value into the + /// destination 64-bit register. + LFIWAX, + + /// GPRC, CHAIN = LFIWZX CHAIN, Ptr - This is a floating-point + /// load which zero-extends from a 32-bit integer value into the + /// 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, + + /// 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, + + /// 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 }; } @@ -178,225 +303,280 @@ 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, bool isUnary, + 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, bool isUnary, + 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); + bool isUnary, 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); - + bool isUnary, SelectionDAG &DAG); + /// isVSLDOIShuffleMask - If this is a vsldoi shuffle mask, return the shift /// amount, otherwise return -1. - int isVSLDOIShuffleMask(SDNode *N, bool isUnary); - + int isVSLDOIShuffleMask(SDNode *N, bool isUnary, 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); } - + class PPCTargetLowering : public TargetLowering { - int VarArgsFrameIndex; // FrameIndex for start of varargs area. - int VarArgsStackOffset; // StackOffset for start of stack - // arguments. - unsigned VarArgsNumGPR; // Index of the first unused integer - // register for parameter passing. - unsigned VarArgsNumFPR; // Index of the first unused double - // register for parameter passing. const PPCSubtarget &PPCSubTarget; + public: explicit PPCTargetLowering(PPCTargetMachine &TM); - + /// 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; + + MVT getScalarShiftAmountTy(EVT LHSTy) const override { return MVT::i32; } /// getSetCCResultType - Return the ISD::SETCC ValueType - virtual MVT::SimpleValueType getSetCCResultType(EVT VT) const; + EVT getSetCCResultType(LLVMContext &Context, 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 /// can be more efficiently represented with [r+imm]. bool SelectAddressRegReg(SDValue N, SDValue &Base, SDValue &Index, SelectionDAG &DAG) const; - + /// SelectAddressRegImm - Returns true if the address N can be represented /// by a base register plus a signed 16-bit displacement [r+imm], and if it - /// is not better represented as reg+reg. + /// is not better represented as reg+reg. If Aligned is true, only accept + /// displacements suitable for STD and friends, i.e. multiples of 4. bool SelectAddressRegImm(SDValue N, SDValue &Disp, SDValue &Base, - SelectionDAG &DAG) const; - + SelectionDAG &DAG, bool Aligned) const; + /// SelectAddressRegRegOnly - Given the specified addressed, force it to be /// represented as an indexed [r+r] operation. bool SelectAddressRegRegOnly(SDValue N, SDValue &Base, SDValue &Index, SelectionDAG &DAG) const; - /// SelectAddressRegImmShift - Returns true if the address N can be - /// represented by a base register plus a signed 14-bit displacement - /// [r+imm*4]. Suitable for use by STD and friends. - bool SelectAddressRegImmShift(SDValue N, SDValue &Disp, SDValue &Base, - SelectionDAG &DAG) const; + Sched::Preference getSchedulingPreference(SDNode *N) const override; - /// LowerOperation - Provide custom lowering hooks for some operations. /// - virtual SDValue LowerOperation(SDValue Op, SelectionDAG &DAG); + 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); - - virtual SDValue PerformDAGCombine(SDNode *N, DAGCombinerInfo &DCI) const; - - virtual void computeMaskedBitsForTargetNode(const SDValue Op, - const APInt &Mask, - APInt &KnownZero, - APInt &KnownOne, - const SelectionDAG &DAG, - unsigned Depth = 0) const; - - virtual MachineBasicBlock *EmitInstrWithCustomInserter(MachineInstr *MI, - MachineBasicBlock *MBB, - DenseMap *EM) const; - MachineBasicBlock *EmitAtomicBinary(MachineInstr *MI, + void ReplaceNodeResults(SDNode *N, SmallVectorImpl&Results, + SelectionDAG &DAG) const override; + + SDValue PerformDAGCombine(SDNode *N, DAGCombinerInfo &DCI) 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; + + MachineBasicBlock * + EmitInstrWithCustomInserter(MachineInstr *MI, + MachineBasicBlock *MBB) const override; + MachineBasicBlock *EmitAtomicBinary(MachineInstr *MI, MachineBasicBlock *MBB, bool is64Bit, unsigned BinOpcode) const; - MachineBasicBlock *EmitPartwordAtomicBinary(MachineInstr *MI, - MachineBasicBlock *MBB, + MachineBasicBlock *EmitPartwordAtomicBinary(MachineInstr *MI, + MachineBasicBlock *MBB, bool is8bit, unsigned Opcode) const; - - ConstraintType getConstraintType(const std::string &Constraint) const; - std::pair + + MachineBasicBlock *emitEHSjLjSetJmp(MachineInstr *MI, + MachineBasicBlock *MBB) const; + + MachineBasicBlock *emitEHSjLjLongJmp(MachineInstr *MI, + MachineBasicBlock *MBB) 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 override; + + std::pair getRegForInlineAsmConstraint(const std::string &Constraint, - EVT VT) const; + 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(const 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. If hasMemory is - /// true it means one of the asm constraint of the inline asm instruction - /// being processed is 'm'. - virtual void LowerAsmOperandForConstraint(SDValue Op, - char ConstraintLetter, - bool hasMemory, - std::vector &Ops, - SelectionDAG &DAG) const; - + /// vector. If it is invalid, don't add anything to Ops. + void LowerAsmOperandForConstraint(SDValue Op, + std::string &Constraint, + std::vector &Ops, + SelectionDAG &DAG) const override; + /// 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, const Type *Ty)const; - - /// isLegalAddressImmediate - Return true if the integer value can be used - /// as the offset of the target addressing mode for load / store of the - /// given type. - virtual bool isLegalAddressImmediate(int64_t V, const Type *Ty) const; - - /// isLegalAddressImmediate - Return true if the GlobalValue can be used as - /// the offset of the target addressing mode. - virtual bool isLegalAddressImmediate(GlobalValue *GV) const; - - virtual bool + 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; + + /// \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; + + /// getOptimalMemOpType - Returns the target specific optimal type for load + /// and store operations as a result of memset, memcpy, and memmove + /// lowering. If DstAlign is zero that means it's safe to destination + /// alignment can satisfy any constraint. Similarly if SrcAlign is zero it + /// means there isn't a need to check it against alignment requirement, + /// probably because the source does not need to be loaded. If 'IsMemset' is + /// true, that means it's expanding a memset. If 'ZeroMemset' is true, that + /// means it's a memset of zero. 'MemcpyStrSrc' indicates whether the memcpy + /// 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. + EVT + getOptimalMemOpType(uint64_t Size, unsigned DstAlign, unsigned SrcAlign, + bool IsMemset, bool ZeroMemset, bool MemcpyStrSrc, + MachineFunction &MF) const override; + + /// Is unaligned memory access allowed for the given type, and is it fast + /// relative to software emulation. + bool allowsUnalignedMemoryAccesses(EVT VT, + unsigned AddrSpace, + 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. + bool isFMAFasterThanFMulAndFAdd(EVT VT) 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. + FastISel *createFastISel(FunctionLoweringInfo &FuncInfo, + const TargetLibraryInfo *LibInfo) const override; + + private: + SDValue getFramePointerFrameIndex(SelectionDAG & DAG) const; + SDValue getReturnAddrFrameIndex(SelectionDAG & DAG) const; + + bool IsEligibleForTailCallOptimization(SDValue Callee, CallingConv::ID CalleeCC, bool isVarArg, const SmallVectorImpl &Ins, SelectionDAG& DAG) const; - virtual bool isOffsetFoldingLegal(const GlobalAddressSDNode *GA) const; - - virtual EVT getOptimalMemOpType(uint64_t Size, unsigned Align, - bool isSrcConst, bool isSrcStr, - SelectionDAG &DAG) const; - - /// getFunctionAlignment - Return the Log2 alignment of this function. - virtual unsigned getFunctionAlignment(const Function *F) const; - - private: - SDValue getFramePointerFrameIndex(SelectionDAG & DAG) const; - SDValue getReturnAddrFrameIndex(SelectionDAG & DAG) const; - SDValue EmitTailCallLoadFPAndRetAddr(SelectionDAG & DAG, int SPDiff, SDValue Chain, SDValue &LROpOut, SDValue &FPOpOut, bool isDarwinABI, - DebugLoc dl); - - SDValue LowerRETURNADDR(SDValue Op, SelectionDAG &DAG); - SDValue LowerFRAMEADDR(SDValue Op, SelectionDAG &DAG); - SDValue LowerConstantPool(SDValue Op, SelectionDAG &DAG); - SDValue LowerGlobalAddress(SDValue Op, SelectionDAG &DAG); - SDValue LowerGlobalTLSAddress(SDValue Op, SelectionDAG &DAG); - SDValue LowerJumpTable(SDValue Op, SelectionDAG &DAG); - SDValue LowerSETCC(SDValue Op, SelectionDAG &DAG); - SDValue LowerTRAMPOLINE(SDValue Op, SelectionDAG &DAG); + SDLoc dl) const; + + SDValue LowerRETURNADDR(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerFRAMEADDR(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerConstantPool(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerBlockAddress(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerGlobalTLSAddress(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerGlobalAddress(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerJumpTable(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerSETCC(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerINIT_TRAMPOLINE(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerADJUST_TRAMPOLINE(SDValue Op, SelectionDAG &DAG) const; SDValue LowerVASTART(SDValue Op, SelectionDAG &DAG, - int VarArgsFrameIndex, int VarArgsStackOffset, - unsigned VarArgsNumGPR, unsigned VarArgsNumFPR, - const PPCSubtarget &Subtarget); - SDValue LowerVAARG(SDValue Op, SelectionDAG &DAG, int VarArgsFrameIndex, - int VarArgsStackOffset, unsigned VarArgsNumGPR, - unsigned VarArgsNumFPR, const PPCSubtarget &Subtarget); + const PPCSubtarget &Subtarget) const; + SDValue LowerVAARG(SDValue Op, SelectionDAG &DAG, + const PPCSubtarget &Subtarget) const; + SDValue LowerVACOPY(SDValue Op, SelectionDAG &DAG, + const PPCSubtarget &Subtarget) const; SDValue LowerSTACKRESTORE(SDValue Op, SelectionDAG &DAG, - const PPCSubtarget &Subtarget); + const PPCSubtarget &Subtarget) const; SDValue LowerDYNAMIC_STACKALLOC(SDValue Op, SelectionDAG &DAG, - const PPCSubtarget &Subtarget); - SDValue LowerSELECT_CC(SDValue Op, SelectionDAG &DAG); - SDValue LowerFP_TO_INT(SDValue Op, SelectionDAG &DAG, DebugLoc dl); - SDValue LowerSINT_TO_FP(SDValue Op, SelectionDAG &DAG); - SDValue LowerFLT_ROUNDS_(SDValue Op, SelectionDAG &DAG); - SDValue LowerSHL_PARTS(SDValue Op, SelectionDAG &DAG); - SDValue LowerSRL_PARTS(SDValue Op, SelectionDAG &DAG); - SDValue LowerSRA_PARTS(SDValue Op, SelectionDAG &DAG); - SDValue LowerBUILD_VECTOR(SDValue Op, SelectionDAG &DAG); - SDValue LowerVECTOR_SHUFFLE(SDValue Op, SelectionDAG &DAG); - SDValue LowerINTRINSIC_WO_CHAIN(SDValue Op, SelectionDAG &DAG); - SDValue LowerSCALAR_TO_VECTOR(SDValue Op, SelectionDAG &DAG); - SDValue LowerMUL(SDValue Op, SelectionDAG &DAG); + const PPCSubtarget &Subtarget) const; + SDValue LowerLOAD(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerSTORE(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerTRUNCATE(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerSELECT_CC(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerFP_TO_INT(SDValue Op, SelectionDAG &DAG, SDLoc dl) const; + SDValue LowerINT_TO_FP(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerFLT_ROUNDS_(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerSHL_PARTS(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerSRL_PARTS(SDValue Op, SelectionDAG &DAG) const; + 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 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 LowerCallResult(SDValue Chain, SDValue InFlag, CallingConv::ID CallConv, bool isVarArg, const SmallVectorImpl &Ins, - DebugLoc dl, SelectionDAG &DAG, - SmallVectorImpl &InVals); - SDValue FinishCall(CallingConv::ID CallConv, DebugLoc dl, bool isTailCall, + SDLoc dl, SelectionDAG &DAG, + SmallVectorImpl &InVals) const; + SDValue FinishCall(CallingConv::ID CallConv, SDLoc dl, bool isTailCall, bool isVarArg, SelectionDAG &DAG, SmallVector, 8> @@ -405,57 +585,124 @@ namespace llvm { SDValue &Callee, int SPDiff, unsigned NumBytes, const SmallVectorImpl &Ins, - SmallVectorImpl &InVals); + SmallVectorImpl &InVals) const; - virtual SDValue + SDValue LowerFormalArguments(SDValue Chain, CallingConv::ID CallConv, bool isVarArg, const SmallVectorImpl &Ins, - DebugLoc dl, SelectionDAG &DAG, - SmallVectorImpl &InVals); - - virtual SDValue - LowerCall(SDValue Chain, SDValue Callee, - CallingConv::ID CallConv, bool isVarArg, bool isTailCall, - const SmallVectorImpl &Outs, - const SmallVectorImpl &Ins, - DebugLoc dl, SelectionDAG &DAG, - SmallVectorImpl &InVals); - - virtual SDValue + SDLoc dl, SelectionDAG &DAG, + SmallVectorImpl &InVals) const override; + + SDValue + LowerCall(TargetLowering::CallLoweringInfo &CLI, + SmallVectorImpl &InVals) const override; + + bool + CanLowerReturn(CallingConv::ID CallConv, MachineFunction &MF, + bool isVarArg, + const SmallVectorImpl &Outs, + LLVMContext &Context) const override; + + SDValue LowerReturn(SDValue Chain, CallingConv::ID CallConv, bool isVarArg, const SmallVectorImpl &Outs, - DebugLoc dl, SelectionDAG &DAG); + const SmallVectorImpl &OutVals, + 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, const SmallVectorImpl &Ins, - DebugLoc dl, SelectionDAG &DAG, - SmallVectorImpl &InVals); + SDLoc dl, SelectionDAG &DAG, + SmallVectorImpl &InVals) const; + SDValue + LowerFormalArguments_64SVR4(SDValue Chain, + CallingConv::ID CallConv, bool isVarArg, + const SmallVectorImpl &Ins, + SDLoc dl, SelectionDAG &DAG, + SmallVectorImpl &InVals) const; SDValue - LowerFormalArguments_SVR4(SDValue Chain, - CallingConv::ID CallConv, bool isVarArg, - const SmallVectorImpl &Ins, - DebugLoc dl, SelectionDAG &DAG, - SmallVectorImpl &InVals); + LowerFormalArguments_32SVR4(SDValue Chain, + CallingConv::ID CallConv, bool isVarArg, + const SmallVectorImpl &Ins, + SDLoc dl, SelectionDAG &DAG, + SmallVectorImpl &InVals) const; + + SDValue + createMemcpyOutsideCallSeq(SDValue Arg, SDValue PtrOff, + SDValue CallSeqStart, ISD::ArgFlagsTy Flags, + SelectionDAG &DAG, SDLoc dl) const; SDValue LowerCall_Darwin(SDValue Chain, SDValue Callee, - CallingConv::ID CallConv, bool isVarArg, bool isTailCall, + CallingConv::ID CallConv, + bool isVarArg, bool isTailCall, + const SmallVectorImpl &Outs, + const SmallVectorImpl &OutVals, + const SmallVectorImpl &Ins, + SDLoc dl, SelectionDAG &DAG, + SmallVectorImpl &InVals) const; + SDValue + LowerCall_64SVR4(SDValue Chain, SDValue Callee, + CallingConv::ID CallConv, + bool isVarArg, bool isTailCall, const SmallVectorImpl &Outs, + const SmallVectorImpl &OutVals, const SmallVectorImpl &Ins, - DebugLoc dl, SelectionDAG &DAG, - SmallVectorImpl &InVals); + SDLoc dl, SelectionDAG &DAG, + SmallVectorImpl &InVals) const; SDValue - LowerCall_SVR4(SDValue Chain, SDValue Callee, - CallingConv::ID CallConv, bool isVarArg, bool isTailCall, + LowerCall_32SVR4(SDValue Chain, SDValue Callee, CallingConv::ID CallConv, + bool isVarArg, bool isTailCall, const SmallVectorImpl &Outs, + const SmallVectorImpl &OutVals, const SmallVectorImpl &Ins, - DebugLoc dl, SelectionDAG &DAG, - SmallVectorImpl &InVals); + SDLoc dl, SelectionDAG &DAG, + SmallVectorImpl &InVals) 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; + + CCAssignFn *useFastISelCCs(unsigned Flag) const; }; + + namespace PPC { + FastISel *createFastISel(FunctionLoweringInfo &FuncInfo, + const TargetLibraryInfo *LibInfo); + } + + bool CC_PPC32_SVR4_Custom_Dummy(unsigned &ValNo, MVT &ValVT, MVT &LocVT, + CCValAssign::LocInfo &LocInfo, + ISD::ArgFlagsTy &ArgFlags, + CCState &State); + + bool CC_PPC32_SVR4_Custom_AlignArgRegs(unsigned &ValNo, MVT &ValVT, + MVT &LocVT, + CCValAssign::LocInfo &LocInfo, + ISD::ArgFlagsTy &ArgFlags, + CCState &State); + + bool CC_PPC32_SVR4_Custom_AlignFPArgRegs(unsigned &ValNo, MVT &ValVT, + MVT &LocVT, + CCValAssign::LocInfo &LocInfo, + ISD::ArgFlagsTy &ArgFlags, + CCState &State); } #endif // LLVM_TARGET_POWERPC_PPC32ISELLOWERING_H