X-Git-Url: http://plrg.eecs.uci.edu/git/?a=blobdiff_plain;f=lib%2FTarget%2FMips%2FMipsISelLowering.cpp;h=5b87111d5add9495122fe40c2ae2e65207b4a60c;hb=61d2d3d2457554575ca00638c46db8dce5b0e158;hp=04f90833d82790066443956a3c67c1522e5a5c0c;hpb=7e0df9aa2966d0462e34511524a4958e226b74ee;p=oota-llvm.git diff --git a/lib/Target/Mips/MipsISelLowering.cpp b/lib/Target/Mips/MipsISelLowering.cpp index 04f90833d82..5b87111d5ad 100644 --- a/lib/Target/Mips/MipsISelLowering.cpp +++ b/lib/Target/Mips/MipsISelLowering.cpp @@ -11,7 +11,6 @@ // selection DAG. // //===----------------------------------------------------------------------===// -#define DEBUG_TYPE "mips-lower" #include "MipsISelLowering.h" #include "InstPrinter/MipsInstPrinter.h" #include "MCTargetDesc/MipsBaseInfo.h" @@ -20,10 +19,12 @@ #include "MipsTargetMachine.h" #include "MipsTargetObjectFile.h" #include "llvm/ADT/Statistic.h" +#include "llvm/ADT/StringSwitch.h" #include "llvm/CodeGen/CallingConvLower.h" #include "llvm/CodeGen/MachineFrameInfo.h" #include "llvm/CodeGen/MachineFunction.h" #include "llvm/CodeGen/MachineInstrBuilder.h" +#include "llvm/CodeGen/MachineJumpTableInfo.h" #include "llvm/CodeGen/MachineRegisterInfo.h" #include "llvm/CodeGen/SelectionDAGISel.h" #include "llvm/CodeGen/ValueTypes.h" @@ -38,6 +39,8 @@ using namespace llvm; +#define DEBUG_TYPE "mips-lower" + STATISTIC(NumTailCalls, "Number of tail calls"); static cl::opt @@ -49,26 +52,132 @@ NoZeroDivCheck("mno-check-zero-division", cl::Hidden, cl::desc("MIPS: Don't trap on integer division by zero."), cl::init(false)); -static const uint16_t O32IntRegs[4] = { +cl::opt +EnableMipsFastISel("mips-fast-isel", cl::Hidden, + cl::desc("Allow mips-fast-isel to be used"), + cl::init(false)); + +static const MCPhysReg O32IntRegs[4] = { Mips::A0, Mips::A1, Mips::A2, Mips::A3 }; -static const uint16_t Mips64IntRegs[8] = { +static const MCPhysReg Mips64IntRegs[8] = { Mips::A0_64, Mips::A1_64, Mips::A2_64, Mips::A3_64, Mips::T0_64, Mips::T1_64, Mips::T2_64, Mips::T3_64 }; -static const uint16_t Mips64DPRegs[8] = { +static const MCPhysReg Mips64DPRegs[8] = { Mips::D12_64, Mips::D13_64, Mips::D14_64, Mips::D15_64, Mips::D16_64, Mips::D17_64, Mips::D18_64, Mips::D19_64 }; +static bool originalTypeIsF128(const Type *Ty, const SDNode *CallNode); + +namespace { +class MipsCCState : public CCState { +private: + /// Identify lowered values that originated from f128 arguments and record + /// this for use by RetCC_MipsN. + void + PreAnalyzeCallResultForF128(const SmallVectorImpl &Ins, + const TargetLowering::CallLoweringInfo &CLI) { + for (unsigned i = 0; i < Ins.size(); ++i) + OriginalArgWasF128.push_back( + originalTypeIsF128(CLI.RetTy, CLI.Callee.getNode())); + } + + /// Identify lowered values that originated from f128 arguments and record + /// this for use by RetCC_MipsN. + void PreAnalyzeReturnForF128(const SmallVectorImpl &Outs) { + const MachineFunction &MF = getMachineFunction(); + for (unsigned i = 0; i < Outs.size(); ++i) + OriginalArgWasF128.push_back( + originalTypeIsF128(MF.getFunction()->getReturnType(), nullptr)); + } + + /// Identify lowered values that originated from f128 arguments and record + /// this. + void PreAnalyzeCallOperandsForF128( + const SmallVectorImpl &Outs, + std::vector &FuncArgs, SDNode *CallNode) { + const MachineFunction &MF = getMachineFunction(); + for (unsigned i = 0; i < Outs.size(); ++i) + OriginalArgWasF128.push_back( + originalTypeIsF128(FuncArgs[Outs[i].OrigArgIndex].Ty, CallNode)); + } + + /// Identify lowered values that originated from f128 arguments and record + /// this. + void + PreAnalyzeFormalArgumentsForF128(const SmallVectorImpl &Ins) { + const MachineFunction &MF = getMachineFunction(); + for (unsigned i = 0; i < Ins.size(); ++i) { + Function::const_arg_iterator FuncArg = MF.getFunction()->arg_begin(); + std::advance(FuncArg, Ins[i].OrigArgIndex); + + OriginalArgWasF128.push_back( + originalTypeIsF128(FuncArg->getType(), nullptr)); + } + } + + /// Records whether the value has been lowered from an f128. + SmallVector OriginalArgWasF128; + +public: + // FIXME: Remove this from a public inteface ASAP. It's a temporary trap door + // to allow analyzeCallOperands to be removed incrementally. + void PreAnalyzeCallOperandsForF128_( + const SmallVectorImpl &Outs, + std::vector &FuncArgs, SDNode *CallNode) { + PreAnalyzeCallOperandsForF128(Outs, FuncArgs, CallNode); + } + // FIXME: Remove this from a public inteface ASAP. It's a temporary trap door + // to allow analyzeFormalArguments to be removed incrementally. + void + PreAnalyzeFormalArgumentsForF128_(const SmallVectorImpl &Ins) { + PreAnalyzeFormalArgumentsForF128(Ins); + } + // FIXME: Remove this from a public inteface ASAP. It's a temporary trap door + // to clean up after the above functions. + void ClearOriginalArgWasF128() { OriginalArgWasF128.clear(); } + + MipsCCState(CallingConv::ID CC, bool isVarArg, MachineFunction &MF, + SmallVectorImpl &locs, LLVMContext &C) + : CCState(CC, isVarArg, MF, locs, C) {} + + void AnalyzeCallResult(const SmallVectorImpl &Ins, + CCAssignFn Fn, + const TargetLowering::CallLoweringInfo &CLI) { + PreAnalyzeCallResultForF128(Ins, CLI); + CCState::AnalyzeCallResult(Ins, Fn); + OriginalArgWasF128.clear(); + } + + void AnalyzeReturn(const SmallVectorImpl &Outs, + CCAssignFn Fn) { + PreAnalyzeReturnForF128(Outs); + CCState::AnalyzeReturn(Outs, Fn); + OriginalArgWasF128.clear(); + } + + bool CheckReturn(const SmallVectorImpl &ArgsFlags, + CCAssignFn Fn) { + PreAnalyzeReturnForF128(ArgsFlags); + bool Return = CCState::CheckReturn(ArgsFlags, Fn); + OriginalArgWasF128.clear(); + return Return; + } + + bool WasOriginalArgF128(unsigned ValNo) { return OriginalArgWasF128[ValNo]; } +}; +} + // If I is a shifted mask, set the size (Size) and the first bit of the // mask (Pos), and return true. // For example, if I is 0x003ff800, (Pos, Size) = (11, 11). static bool isShiftedMask(uint64_t I, uint64_t &Pos, uint64_t &Size) { if (!isShiftedMask_64(I)) - return false; + return false; Size = CountPopulation_64(I); Pos = countTrailingZeros(I); @@ -80,72 +189,35 @@ SDValue MipsTargetLowering::getGlobalReg(SelectionDAG &DAG, EVT Ty) const { return DAG.getRegister(FI->getGlobalBaseReg(), Ty); } -static SDValue getTargetNode(SDValue Op, SelectionDAG &DAG, unsigned Flag) { - EVT Ty = Op.getValueType(); +SDValue MipsTargetLowering::getTargetNode(GlobalAddressSDNode *N, EVT Ty, + SelectionDAG &DAG, + unsigned Flag) const { + return DAG.getTargetGlobalAddress(N->getGlobal(), SDLoc(N), Ty, 0, Flag); +} - if (GlobalAddressSDNode *N = dyn_cast(Op)) - return DAG.getTargetGlobalAddress(N->getGlobal(), SDLoc(Op), Ty, 0, - Flag); - if (ExternalSymbolSDNode *N = dyn_cast(Op)) - return DAG.getTargetExternalSymbol(N->getSymbol(), Ty, Flag); - if (BlockAddressSDNode *N = dyn_cast(Op)) - return DAG.getTargetBlockAddress(N->getBlockAddress(), Ty, 0, Flag); - if (JumpTableSDNode *N = dyn_cast(Op)) - return DAG.getTargetJumpTable(N->getIndex(), Ty, Flag); - if (ConstantPoolSDNode *N = dyn_cast(Op)) - return DAG.getTargetConstantPool(N->getConstVal(), Ty, N->getAlignment(), - N->getOffset(), Flag); - - llvm_unreachable("Unexpected node type."); - return SDValue(); +SDValue MipsTargetLowering::getTargetNode(ExternalSymbolSDNode *N, EVT Ty, + SelectionDAG &DAG, + unsigned Flag) const { + return DAG.getTargetExternalSymbol(N->getSymbol(), Ty, Flag); } -static SDValue getAddrNonPIC(SDValue Op, SelectionDAG &DAG) { - SDLoc DL(Op); - EVT Ty = Op.getValueType(); - SDValue Hi = getTargetNode(Op, DAG, MipsII::MO_ABS_HI); - SDValue Lo = getTargetNode(Op, DAG, MipsII::MO_ABS_LO); - return DAG.getNode(ISD::ADD, DL, Ty, - DAG.getNode(MipsISD::Hi, DL, Ty, Hi), - DAG.getNode(MipsISD::Lo, DL, Ty, Lo)); +SDValue MipsTargetLowering::getTargetNode(BlockAddressSDNode *N, EVT Ty, + SelectionDAG &DAG, + unsigned Flag) const { + return DAG.getTargetBlockAddress(N->getBlockAddress(), Ty, 0, Flag); } -SDValue MipsTargetLowering::getAddrLocal(SDValue Op, SelectionDAG &DAG, - bool HasMips64) const { - SDLoc DL(Op); - EVT Ty = Op.getValueType(); - unsigned GOTFlag = HasMips64 ? MipsII::MO_GOT_PAGE : MipsII::MO_GOT; - SDValue GOT = DAG.getNode(MipsISD::Wrapper, DL, Ty, getGlobalReg(DAG, Ty), - getTargetNode(Op, DAG, GOTFlag)); - SDValue Load = DAG.getLoad(Ty, DL, DAG.getEntryNode(), GOT, - MachinePointerInfo::getGOT(), false, false, false, - 0); - unsigned LoFlag = HasMips64 ? MipsII::MO_GOT_OFST : MipsII::MO_ABS_LO; - SDValue Lo = DAG.getNode(MipsISD::Lo, DL, Ty, getTargetNode(Op, DAG, LoFlag)); - return DAG.getNode(ISD::ADD, DL, Ty, Load, Lo); -} - -SDValue MipsTargetLowering::getAddrGlobal(SDValue Op, SelectionDAG &DAG, +SDValue MipsTargetLowering::getTargetNode(JumpTableSDNode *N, EVT Ty, + SelectionDAG &DAG, unsigned Flag) const { - SDLoc DL(Op); - EVT Ty = Op.getValueType(); - SDValue Tgt = DAG.getNode(MipsISD::Wrapper, DL, Ty, getGlobalReg(DAG, Ty), - getTargetNode(Op, DAG, Flag)); - return DAG.getLoad(Ty, DL, DAG.getEntryNode(), Tgt, - MachinePointerInfo::getGOT(), false, false, false, 0); + return DAG.getTargetJumpTable(N->getIndex(), Ty, Flag); } -SDValue MipsTargetLowering::getAddrGlobalLargeGOT(SDValue Op, SelectionDAG &DAG, - unsigned HiFlag, - unsigned LoFlag) const { - SDLoc DL(Op); - EVT Ty = Op.getValueType(); - SDValue Hi = DAG.getNode(MipsISD::Hi, DL, Ty, getTargetNode(Op, DAG, HiFlag)); - Hi = DAG.getNode(ISD::ADD, DL, Ty, Hi, getGlobalReg(DAG, Ty)); - SDValue Wrapper = DAG.getNode(MipsISD::Wrapper, DL, Ty, Hi, - getTargetNode(Op, DAG, LoFlag)); - return DAG.getLoad(Ty, DL, DAG.getEntryNode(), Wrapper, - MachinePointerInfo::getGOT(), false, false, false, 0); +SDValue MipsTargetLowering::getTargetNode(ConstantPoolSDNode *N, EVT Ty, + SelectionDAG &DAG, + unsigned Flag) const { + return DAG.getTargetConstantPool(N->getConstVal(), Ty, N->getAlignment(), + N->getOffset(), Flag); } const char *MipsTargetLowering::getTargetNodeName(unsigned Opcode) const { @@ -163,8 +235,9 @@ const char *MipsTargetLowering::getTargetNodeName(unsigned Opcode) const { case MipsISD::CMovFP_T: return "MipsISD::CMovFP_T"; case MipsISD::CMovFP_F: return "MipsISD::CMovFP_F"; case MipsISD::TruncIntFP: return "MipsISD::TruncIntFP"; - case MipsISD::ExtractLOHI: return "MipsISD::ExtractLOHI"; - case MipsISD::InsertLOHI: return "MipsISD::InsertLOHI"; + case MipsISD::MFHI: return "MipsISD::MFHI"; + case MipsISD::MFLO: return "MipsISD::MFLO"; + case MipsISD::MTLOHI: return "MipsISD::MTLOHI"; case MipsISD::Mult: return "MipsISD::Mult"; case MipsISD::Multu: return "MipsISD::Multu"; case MipsISD::MAdd: return "MipsISD::MAdd"; @@ -225,20 +298,30 @@ const char *MipsTargetLowering::getTargetNodeName(unsigned Opcode) const { case MipsISD::VEXTRACT_ZEXT_ELT: return "MipsISD::VEXTRACT_ZEXT_ELT"; case MipsISD::VNOR: return "MipsISD::VNOR"; case MipsISD::VSHF: return "MipsISD::VSHF"; - default: return NULL; + case MipsISD::SHF: return "MipsISD::SHF"; + case MipsISD::ILVEV: return "MipsISD::ILVEV"; + case MipsISD::ILVOD: return "MipsISD::ILVOD"; + case MipsISD::ILVL: return "MipsISD::ILVL"; + case MipsISD::ILVR: return "MipsISD::ILVR"; + case MipsISD::PCKEV: return "MipsISD::PCKEV"; + case MipsISD::PCKOD: return "MipsISD::PCKOD"; + case MipsISD::INSVE: return "MipsISD::INSVE"; + default: return nullptr; } } -MipsTargetLowering:: -MipsTargetLowering(MipsTargetMachine &TM) - : TargetLowering(TM, new MipsTargetObjectFile()), - Subtarget(&TM.getSubtarget()), - HasMips64(Subtarget->hasMips64()), IsN64(Subtarget->isABI_N64()), - IsO32(Subtarget->isABI_O32()) { +MipsTargetLowering::MipsTargetLowering(const MipsTargetMachine &TM, + const MipsSubtarget &STI) + : TargetLowering(TM, new MipsTargetObjectFile()), Subtarget(STI) { // Mips does not have i1 type, so use i32 for // setcc operations results (slt, sgt, ...). setBooleanContents(ZeroOrOneBooleanContent); setBooleanVectorContents(ZeroOrNegativeOneBooleanContent); + // The cmp.cond.fmt instruction in MIPS32r6/MIPS64r6 uses 0 and -1 like MSA + // does. Integer booleans still use 0 and 1. + if (Subtarget.hasMips32r6()) + setBooleanContents(ZeroOrOneBooleanContent, + ZeroOrNegativeOneBooleanContent); // Load extented operations for i1 types must be promoted setLoadExtAction(ISD::EXTLOAD, MVT::i1, Promote); @@ -270,17 +353,11 @@ MipsTargetLowering(MipsTargetMachine &TM) setOperationAction(ISD::SETCC, MVT::f32, Custom); setOperationAction(ISD::SETCC, MVT::f64, Custom); setOperationAction(ISD::BRCOND, MVT::Other, Custom); - setOperationAction(ISD::VASTART, MVT::Other, Custom); setOperationAction(ISD::FCOPYSIGN, MVT::f32, Custom); setOperationAction(ISD::FCOPYSIGN, MVT::f64, Custom); setOperationAction(ISD::FP_TO_SINT, MVT::i32, Custom); - if (!TM.Options.NoNaNsFPMath) { - setOperationAction(ISD::FABS, MVT::f32, Custom); - setOperationAction(ISD::FABS, MVT::f64, Custom); - } - - if (HasMips64) { + if (Subtarget.isGP64bit()) { setOperationAction(ISD::GlobalAddress, MVT::i64, Custom); setOperationAction(ISD::BlockAddress, MVT::i64, Custom); setOperationAction(ISD::GlobalTLSAddress, MVT::i64, Custom); @@ -292,14 +369,14 @@ MipsTargetLowering(MipsTargetMachine &TM) setOperationAction(ISD::FP_TO_SINT, MVT::i64, Custom); } - if (!HasMips64) { + if (!Subtarget.isGP64bit()) { setOperationAction(ISD::SHL_PARTS, MVT::i32, Custom); setOperationAction(ISD::SRA_PARTS, MVT::i32, Custom); setOperationAction(ISD::SRL_PARTS, MVT::i32, Custom); } setOperationAction(ISD::ADD, MVT::i32, Custom); - if (HasMips64) + if (Subtarget.isGP64bit()) setOperationAction(ISD::ADD, MVT::i64, Custom); setOperationAction(ISD::SDIV, MVT::i32, Expand); @@ -316,14 +393,20 @@ MipsTargetLowering(MipsTargetMachine &TM) setOperationAction(ISD::BR_CC, MVT::f64, Expand); setOperationAction(ISD::BR_CC, MVT::i32, Expand); setOperationAction(ISD::BR_CC, MVT::i64, Expand); - setOperationAction(ISD::SELECT_CC, MVT::Other, Expand); + setOperationAction(ISD::SELECT_CC, MVT::i32, Expand); + setOperationAction(ISD::SELECT_CC, MVT::i64, Expand); setOperationAction(ISD::UINT_TO_FP, MVT::i32, Expand); setOperationAction(ISD::UINT_TO_FP, MVT::i64, Expand); setOperationAction(ISD::FP_TO_UINT, MVT::i32, Expand); setOperationAction(ISD::FP_TO_UINT, MVT::i64, Expand); setOperationAction(ISD::SIGN_EXTEND_INREG, MVT::i1, Expand); - setOperationAction(ISD::CTPOP, MVT::i32, Expand); - setOperationAction(ISD::CTPOP, MVT::i64, Expand); + if (Subtarget.hasCnMips()) { + setOperationAction(ISD::CTPOP, MVT::i32, Legal); + setOperationAction(ISD::CTPOP, MVT::i64, Legal); + } else { + setOperationAction(ISD::CTPOP, MVT::i32, Expand); + setOperationAction(ISD::CTPOP, MVT::i64, Expand); + } setOperationAction(ISD::CTTZ, MVT::i32, Expand); setOperationAction(ISD::CTTZ, MVT::i64, Expand); setOperationAction(ISD::CTTZ_ZERO_UNDEF, MVT::i32, Expand); @@ -335,10 +418,10 @@ MipsTargetLowering(MipsTargetMachine &TM) setOperationAction(ISD::DYNAMIC_STACKALLOC, MVT::i32, Expand); setOperationAction(ISD::DYNAMIC_STACKALLOC, MVT::i64, Expand); - if (!Subtarget->hasMips32r2()) + if (!Subtarget.hasMips32r2()) setOperationAction(ISD::ROTR, MVT::i32, Expand); - if (!Subtarget->hasMips64r2()) + if (!Subtarget.hasMips64r2()) setOperationAction(ISD::ROTR, MVT::i64, Expand); setOperationAction(ISD::FSIN, MVT::f32, Expand); @@ -359,14 +442,10 @@ MipsTargetLowering(MipsTargetMachine &TM) setOperationAction(ISD::FREM, MVT::f32, Expand); setOperationAction(ISD::FREM, MVT::f64, Expand); - if (!TM.Options.NoNaNsFPMath) { - setOperationAction(ISD::FNEG, MVT::f32, Expand); - setOperationAction(ISD::FNEG, MVT::f64, Expand); - } - setOperationAction(ISD::EH_RETURN, MVT::Other, Custom); - setOperationAction(ISD::VAARG, MVT::Other, Expand); + setOperationAction(ISD::VASTART, MVT::Other, Custom); + setOperationAction(ISD::VAARG, MVT::Other, Custom); setOperationAction(ISD::VACOPY, MVT::Other, Expand); setOperationAction(ISD::VAEND, MVT::Other, Expand); @@ -381,22 +460,23 @@ MipsTargetLowering(MipsTargetMachine &TM) setInsertFencesForAtomic(true); - if (!Subtarget->hasSEInReg()) { + if (!Subtarget.hasMips32r2()) { setOperationAction(ISD::SIGN_EXTEND_INREG, MVT::i8, Expand); setOperationAction(ISD::SIGN_EXTEND_INREG, MVT::i16, Expand); } - if (!Subtarget->hasBitCount()) { + // MIPS16 lacks MIPS32's clz and clo instructions. + if (!Subtarget.hasMips32() || Subtarget.inMips16Mode()) setOperationAction(ISD::CTLZ, MVT::i32, Expand); + if (!Subtarget.hasMips64()) setOperationAction(ISD::CTLZ, MVT::i64, Expand); - } - if (!Subtarget->hasSwap()) { + if (!Subtarget.hasMips32r2()) setOperationAction(ISD::BSWAP, MVT::i32, Expand); + if (!Subtarget.hasMips64r2()) setOperationAction(ISD::BSWAP, MVT::i64, Expand); - } - if (HasMips64) { + if (Subtarget.isGP64bit()) { setLoadExtAction(ISD::SEXTLOAD, MVT::i32, Custom); setLoadExtAction(ISD::ZEXTLOAD, MVT::i32, Custom); setLoadExtAction(ISD::EXTLOAD, MVT::i32, Custom); @@ -412,21 +492,39 @@ MipsTargetLowering(MipsTargetMachine &TM) setTargetDAGCombine(ISD::OR); setTargetDAGCombine(ISD::ADD); - setMinFunctionAlignment(HasMips64 ? 3 : 2); + setMinFunctionAlignment(Subtarget.isGP64bit() ? 3 : 2); + + // The arguments on the stack are defined in terms of 4-byte slots on O32 + // and 8-byte slots on N32/N64. + setMinStackArgumentAlignment( + (Subtarget.isABI_N32() || Subtarget.isABI_N64()) ? 8 : 4); - setStackPointerRegisterToSaveRestore(IsN64 ? Mips::SP_64 : Mips::SP); + setStackPointerRegisterToSaveRestore(Subtarget.isABI_N64() ? Mips::SP_64 + : Mips::SP); - setExceptionPointerRegister(IsN64 ? Mips::A0_64 : Mips::A0); - setExceptionSelectorRegister(IsN64 ? Mips::A1_64 : Mips::A1); + setExceptionPointerRegister(Subtarget.isABI_N64() ? Mips::A0_64 : Mips::A0); + setExceptionSelectorRegister(Subtarget.isABI_N64() ? Mips::A1_64 : Mips::A1); MaxStoresPerMemcpy = 16; + + isMicroMips = Subtarget.inMicroMipsMode(); } -const MipsTargetLowering *MipsTargetLowering::create(MipsTargetMachine &TM) { - if (TM.getSubtargetImpl()->inMips16Mode()) - return llvm::createMips16TargetLowering(TM); +const MipsTargetLowering *MipsTargetLowering::create(const MipsTargetMachine &TM, + const MipsSubtarget &STI) { + if (STI.inMips16Mode()) + return llvm::createMips16TargetLowering(TM, STI); + + return llvm::createMipsSETargetLowering(TM, STI); +} - return llvm::createMipsSETargetLowering(TM); +// Create a fast isel object. +FastISel * +MipsTargetLowering::createFastISel(FunctionLoweringInfo &funcInfo, + const TargetLibraryInfo *libInfo) const { + if (!EnableMipsFastISel) + return TargetLowering::createFastISel(funcInfo, libInfo); + return Mips::createFastISel(funcInfo, libInfo); } EVT MipsTargetLowering::getSetCCResultType(LLVMContext &, EVT VT) const { @@ -437,7 +535,7 @@ EVT MipsTargetLowering::getSetCCResultType(LLVMContext &, EVT VT) const { static SDValue performDivRemCombine(SDNode *N, SelectionDAG &DAG, TargetLowering::DAGCombinerInfo &DCI, - const MipsSubtarget *Subtarget) { + const MipsSubtarget &Subtarget) { if (DCI.isBeforeLegalizeOps()) return SDValue(); @@ -547,7 +645,7 @@ static SDValue createCMovFP(SelectionDAG &DAG, SDValue Cond, SDValue True, static SDValue performSELECTCombine(SDNode *N, SelectionDAG &DAG, TargetLowering::DAGCombinerInfo &DCI, - const MipsSubtarget *Subtarget) { + const MipsSubtarget &Subtarget) { if (DCI.isBeforeLegalizeOps()) return SDValue(); @@ -563,28 +661,74 @@ static SDValue performSELECTCombine(SDNode *N, SelectionDAG &DAG, if (!FalseTy.isInteger()) return SDValue(); - ConstantSDNode *CN = dyn_cast(False); + ConstantSDNode *FalseC = dyn_cast(False); - if (!CN || CN->getZExtValue()) + // If the RHS (False) is 0, we swap the order of the operands + // of ISD::SELECT (obviously also inverting the condition) so that we can + // take advantage of conditional moves using the $0 register. + // Example: + // return (a != 0) ? x : 0; + // load $reg, x + // movz $reg, $0, a + if (!FalseC) return SDValue(); const SDLoc DL(N); - ISD::CondCode CC = cast(SetCC.getOperand(2))->get(); + + if (!FalseC->getZExtValue()) { + ISD::CondCode CC = cast(SetCC.getOperand(2))->get(); + SDValue True = N->getOperand(1); + + SetCC = DAG.getSetCC(DL, SetCC.getValueType(), SetCC.getOperand(0), + SetCC.getOperand(1), ISD::getSetCCInverse(CC, true)); + + return DAG.getNode(ISD::SELECT, DL, FalseTy, SetCC, False, True); + } + + // If both operands are integer constants there's a possibility that we + // can do some interesting optimizations. SDValue True = N->getOperand(1); + ConstantSDNode *TrueC = dyn_cast(True); - SetCC = DAG.getSetCC(DL, SetCC.getValueType(), SetCC.getOperand(0), - SetCC.getOperand(1), ISD::getSetCCInverse(CC, true)); + if (!TrueC || !True.getValueType().isInteger()) + return SDValue(); + + // We'll also ignore MVT::i64 operands as this optimizations proves + // to be ineffective because of the required sign extensions as the result + // of a SETCC operator is always MVT::i32 for non-vector types. + if (True.getValueType() == MVT::i64) + return SDValue(); + + int64_t Diff = TrueC->getSExtValue() - FalseC->getSExtValue(); + + // 1) (a < x) ? y : y-1 + // slti $reg1, a, x + // addiu $reg2, $reg1, y-1 + if (Diff == 1) + return DAG.getNode(ISD::ADD, DL, SetCC.getValueType(), SetCC, False); + + // 2) (a < x) ? y-1 : y + // slti $reg1, a, x + // xor $reg1, $reg1, 1 + // addiu $reg2, $reg1, y-1 + if (Diff == -1) { + ISD::CondCode CC = cast(SetCC.getOperand(2))->get(); + SetCC = DAG.getSetCC(DL, SetCC.getValueType(), SetCC.getOperand(0), + SetCC.getOperand(1), ISD::getSetCCInverse(CC, true)); + return DAG.getNode(ISD::ADD, DL, SetCC.getValueType(), SetCC, True); + } - return DAG.getNode(ISD::SELECT, DL, FalseTy, SetCC, False, True); + // Couldn't optimize. + return SDValue(); } static SDValue performANDCombine(SDNode *N, SelectionDAG &DAG, TargetLowering::DAGCombinerInfo &DCI, - const MipsSubtarget *Subtarget) { + const MipsSubtarget &Subtarget) { // Pattern match EXT. // $dst = and ((sra or srl) $src , pos), (2**size - 1) // => ext $dst, $src, size, pos - if (DCI.isBeforeLegalizeOps() || !Subtarget->hasMips32r2()) + if (DCI.isBeforeLegalizeOps() || !Subtarget.hasExtractInsert()) return SDValue(); SDValue ShiftRight = N->getOperand(0), Mask = N->getOperand(1); @@ -620,12 +764,12 @@ static SDValue performANDCombine(SDNode *N, SelectionDAG &DAG, static SDValue performORCombine(SDNode *N, SelectionDAG &DAG, TargetLowering::DAGCombinerInfo &DCI, - const MipsSubtarget *Subtarget) { + const MipsSubtarget &Subtarget) { // Pattern match INS. // $dst = or (and $src1 , mask0), (and (shl $src, pos), mask1), // where mask1 = (2**size - 1) << pos, mask0 = ~mask1 // => ins $dst, $src, size, pos, $src1 - if (DCI.isBeforeLegalizeOps() || !Subtarget->hasMips32r2()) + if (DCI.isBeforeLegalizeOps() || !Subtarget.hasExtractInsert()) return SDValue(); SDValue And0 = N->getOperand(0), And1 = N->getOperand(1); @@ -674,7 +818,7 @@ static SDValue performORCombine(SDNode *N, SelectionDAG &DAG, static SDValue performADDCombine(SDNode *N, SelectionDAG &DAG, TargetLowering::DAGCombinerInfo &DCI, - const MipsSubtarget *Subtarget) { + const MipsSubtarget &Subtarget) { // (add v0, (add v1, abs_lo(tjt))) => (add (add v0, v1), abs_lo(tjt)) if (DCI.isBeforeLegalizeOps()) @@ -755,8 +899,8 @@ LowerOperation(SDValue Op, SelectionDAG &DAG) const case ISD::SELECT_CC: return lowerSELECT_CC(Op, DAG); case ISD::SETCC: return lowerSETCC(Op, DAG); case ISD::VASTART: return lowerVASTART(Op, DAG); + case ISD::VAARG: return lowerVAARG(Op, DAG); case ISD::FCOPYSIGN: return lowerFCOPYSIGN(Op, DAG); - case ISD::FABS: return lowerFABS(Op, DAG); case ISD::FRAMEADDR: return lowerFRAMEADDR(Op, DAG); case ISD::RETURNADDR: return lowerRETURNADDR(Op, DAG); case ISD::EH_RETURN: return lowerEH_RETURN(Op, DAG); @@ -787,23 +931,31 @@ addLiveIn(MachineFunction &MF, unsigned PReg, const TargetRegisterClass *RC) return VReg; } -static MachineBasicBlock *expandPseudoDIV(MachineInstr *MI, - MachineBasicBlock &MBB, - const TargetInstrInfo &TII, - bool Is64Bit) { +static MachineBasicBlock *insertDivByZeroTrap(MachineInstr *MI, + MachineBasicBlock &MBB, + const TargetInstrInfo &TII, + bool Is64Bit) { if (NoZeroDivCheck) return &MBB; // Insert instruction "teq $divisor_reg, $zero, 7". MachineBasicBlock::iterator I(MI); MachineInstrBuilder MIB; - MIB = BuildMI(MBB, llvm::next(I), MI->getDebugLoc(), TII.get(Mips::TEQ)) - .addOperand(MI->getOperand(2)).addReg(Mips::ZERO).addImm(7); + MachineOperand &Divisor = MI->getOperand(2); + MIB = BuildMI(MBB, std::next(I), MI->getDebugLoc(), TII.get(Mips::TEQ)) + .addReg(Divisor.getReg(), getKillRegState(Divisor.isKill())) + .addReg(Mips::ZERO).addImm(7); // Use the 32-bit sub-register if this is a 64-bit division. if (Is64Bit) MIB->getOperand(0).setSubReg(Mips::sub_32); + // Clear Divisor's kill flag. + Divisor.setIsKill(false); + + // We would normally delete the original instruction here but in this case + // we only needed to inject an additional instruction rather than replace it. + return &MBB; } @@ -886,10 +1038,22 @@ MipsTargetLowering::EmitInstrWithCustomInserter(MachineInstr *MI, return emitAtomicCmpSwap(MI, BB, 8); case Mips::PseudoSDIV: case Mips::PseudoUDIV: - return expandPseudoDIV(MI, *BB, *getTargetMachine().getInstrInfo(), false); + case Mips::DIV: + case Mips::DIVU: + case Mips::MOD: + case Mips::MODU: + return insertDivByZeroTrap( + MI, *BB, *getTargetMachine().getSubtargetImpl()->getInstrInfo(), false); case Mips::PseudoDSDIV: case Mips::PseudoDUDIV: - return expandPseudoDIV(MI, *BB, *getTargetMachine().getInstrInfo(), true); + case Mips::DDIV: + case Mips::DDIVU: + case Mips::DMOD: + case Mips::DMODU: + return insertDivByZeroTrap( + MI, *BB, *getTargetMachine().getSubtargetImpl()->getInstrInfo(), true); + case Mips::SEL_D: + return emitSEL_D(MI, BB); } } @@ -904,21 +1068,26 @@ MipsTargetLowering::emitAtomicBinary(MachineInstr *MI, MachineBasicBlock *BB, MachineFunction *MF = BB->getParent(); MachineRegisterInfo &RegInfo = MF->getRegInfo(); const TargetRegisterClass *RC = getRegClassFor(MVT::getIntegerVT(Size * 8)); - const TargetInstrInfo *TII = getTargetMachine().getInstrInfo(); + const TargetInstrInfo *TII = + getTargetMachine().getSubtargetImpl()->getInstrInfo(); DebugLoc DL = MI->getDebugLoc(); unsigned LL, SC, AND, NOR, ZERO, BEQ; if (Size == 4) { - LL = Mips::LL; - SC = Mips::SC; + if (isMicroMips) { + LL = Mips::LL_MM; + SC = Mips::SC_MM; + } else { + LL = Subtarget.hasMips32r6() ? Mips::LL_R6 : Mips::LL; + SC = Subtarget.hasMips32r6() ? Mips::SC_R6 : Mips::SC; + } AND = Mips::AND; NOR = Mips::NOR; ZERO = Mips::ZERO; BEQ = Mips::BEQ; - } - else { - LL = Mips::LLD; - SC = Mips::SCD; + } else { + LL = Subtarget.hasMips64r6() ? Mips::LLD_R6 : Mips::LLD; + SC = Subtarget.hasMips64r6() ? Mips::SCD_R6 : Mips::SCD; AND = Mips::AND64; NOR = Mips::NOR64; ZERO = Mips::ZERO_64; @@ -944,8 +1113,7 @@ MipsTargetLowering::emitAtomicBinary(MachineInstr *MI, MachineBasicBlock *BB, // Transfer the remainder of BB and its successor edges to exitMBB. exitMBB->splice(exitMBB->begin(), BB, - llvm::next(MachineBasicBlock::iterator(MI)), - BB->end()); + std::next(MachineBasicBlock::iterator(MI)), BB->end()); exitMBB->transferSuccessorsAndUpdatePHIs(BB); // thisMBB: @@ -976,23 +1144,53 @@ MipsTargetLowering::emitAtomicBinary(MachineInstr *MI, MachineBasicBlock *BB, BuildMI(BB, DL, TII->get(SC), Success).addReg(StoreVal).addReg(Ptr).addImm(0); BuildMI(BB, DL, TII->get(BEQ)).addReg(Success).addReg(ZERO).addMBB(loopMBB); - MI->eraseFromParent(); // The instruction is gone now. + MI->eraseFromParent(); // The instruction is gone now. return exitMBB; } -MachineBasicBlock * -MipsTargetLowering::emitAtomicBinaryPartword(MachineInstr *MI, - MachineBasicBlock *BB, - unsigned Size, unsigned BinOpcode, - bool Nand) const { +MachineBasicBlock *MipsTargetLowering::emitSignExtendToI32InReg( + MachineInstr *MI, MachineBasicBlock *BB, unsigned Size, unsigned DstReg, + unsigned SrcReg) const { + const TargetInstrInfo *TII = + getTargetMachine().getSubtargetImpl()->getInstrInfo(); + DebugLoc DL = MI->getDebugLoc(); + + if (Subtarget.hasMips32r2() && Size == 1) { + BuildMI(BB, DL, TII->get(Mips::SEB), DstReg).addReg(SrcReg); + return BB; + } + + if (Subtarget.hasMips32r2() && Size == 2) { + BuildMI(BB, DL, TII->get(Mips::SEH), DstReg).addReg(SrcReg); + return BB; + } + + MachineFunction *MF = BB->getParent(); + MachineRegisterInfo &RegInfo = MF->getRegInfo(); + const TargetRegisterClass *RC = getRegClassFor(MVT::i32); + unsigned ScrReg = RegInfo.createVirtualRegister(RC); + + assert(Size < 32); + int64_t ShiftImm = 32 - (Size * 8); + + BuildMI(BB, DL, TII->get(Mips::SLL), ScrReg).addReg(SrcReg).addImm(ShiftImm); + BuildMI(BB, DL, TII->get(Mips::SRA), DstReg).addReg(ScrReg).addImm(ShiftImm); + + return BB; +} + +MachineBasicBlock *MipsTargetLowering::emitAtomicBinaryPartword( + MachineInstr *MI, MachineBasicBlock *BB, unsigned Size, unsigned BinOpcode, + bool Nand) const { assert((Size == 1 || Size == 2) && - "Unsupported size for EmitAtomicBinaryPartial."); + "Unsupported size for EmitAtomicBinaryPartial."); MachineFunction *MF = BB->getParent(); MachineRegisterInfo &RegInfo = MF->getRegInfo(); const TargetRegisterClass *RC = getRegClassFor(MVT::i32); - const TargetInstrInfo *TII = getTargetMachine().getInstrInfo(); + const TargetInstrInfo *TII = + getTargetMachine().getSubtargetImpl()->getInstrInfo(); DebugLoc DL = MI->getDebugLoc(); unsigned Dest = MI->getOperand(0).getReg(); @@ -1015,7 +1213,6 @@ MipsTargetLowering::emitAtomicBinaryPartword(MachineInstr *MI, unsigned StoreVal = RegInfo.createVirtualRegister(RC); unsigned MaskedOldVal1 = RegInfo.createVirtualRegister(RC); unsigned SrlRes = RegInfo.createVirtualRegister(RC); - unsigned SllRes = RegInfo.createVirtualRegister(RC); unsigned Success = RegInfo.createVirtualRegister(RC); // insert new blocks after the current block @@ -1031,7 +1228,7 @@ MipsTargetLowering::emitAtomicBinaryPartword(MachineInstr *MI, // Transfer the remainder of BB and its successor edges to exitMBB. exitMBB->splice(exitMBB->begin(), BB, - llvm::next(MachineBasicBlock::iterator(MI)), BB->end()); + std::next(MachineBasicBlock::iterator(MI)), BB->end()); exitMBB->transferSuccessorsAndUpdatePHIs(BB); BB->addSuccessor(loopMBB); @@ -1055,7 +1252,7 @@ MipsTargetLowering::emitAtomicBinaryPartword(MachineInstr *MI, BuildMI(BB, DL, TII->get(Mips::AND), AlignedAddr) .addReg(Ptr).addReg(MaskLSB2); BuildMI(BB, DL, TII->get(Mips::ANDi), PtrLSB2).addReg(Ptr).addImm(3); - if (Subtarget->isLittle()) { + if (Subtarget.isLittle()) { BuildMI(BB, DL, TII->get(Mips::SLL), ShiftAmt).addReg(PtrLSB2).addImm(3); } else { unsigned Off = RegInfo.createVirtualRegister(RC); @@ -1104,7 +1301,7 @@ MipsTargetLowering::emitAtomicBinaryPartword(MachineInstr *MI, // and newval, binopres, mask BuildMI(BB, DL, TII->get(BinOpcode), BinOpRes).addReg(OldVal).addReg(Incr2); BuildMI(BB, DL, TII->get(Mips::AND), NewVal).addReg(BinOpRes).addReg(Mask); - } else {// atomic.swap + } else { // atomic.swap // and newval, incr2, mask BuildMI(BB, DL, TII->get(Mips::AND), NewVal).addReg(Incr2).addReg(Mask); } @@ -1121,46 +1318,40 @@ MipsTargetLowering::emitAtomicBinaryPartword(MachineInstr *MI, // sinkMBB: // and maskedoldval1,oldval,mask // srl srlres,maskedoldval1,shiftamt - // sll sllres,srlres,24 - // sra dest,sllres,24 + // sign_extend dest,srlres BB = sinkMBB; - int64_t ShiftImm = (Size == 1) ? 24 : 16; BuildMI(BB, DL, TII->get(Mips::AND), MaskedOldVal1) .addReg(OldVal).addReg(Mask); BuildMI(BB, DL, TII->get(Mips::SRLV), SrlRes) .addReg(MaskedOldVal1).addReg(ShiftAmt); - BuildMI(BB, DL, TII->get(Mips::SLL), SllRes) - .addReg(SrlRes).addImm(ShiftImm); - BuildMI(BB, DL, TII->get(Mips::SRA), Dest) - .addReg(SllRes).addImm(ShiftImm); + BB = emitSignExtendToI32InReg(MI, BB, Size, Dest, SrlRes); - MI->eraseFromParent(); // The instruction is gone now. + MI->eraseFromParent(); // The instruction is gone now. return exitMBB; } -MachineBasicBlock * -MipsTargetLowering::emitAtomicCmpSwap(MachineInstr *MI, - MachineBasicBlock *BB, - unsigned Size) const { +MachineBasicBlock * MipsTargetLowering::emitAtomicCmpSwap(MachineInstr *MI, + MachineBasicBlock *BB, + unsigned Size) const { assert((Size == 4 || Size == 8) && "Unsupported size for EmitAtomicCmpSwap."); MachineFunction *MF = BB->getParent(); MachineRegisterInfo &RegInfo = MF->getRegInfo(); const TargetRegisterClass *RC = getRegClassFor(MVT::getIntegerVT(Size * 8)); - const TargetInstrInfo *TII = getTargetMachine().getInstrInfo(); + const TargetInstrInfo *TII = + getTargetMachine().getSubtargetImpl()->getInstrInfo(); DebugLoc DL = MI->getDebugLoc(); unsigned LL, SC, ZERO, BNE, BEQ; if (Size == 4) { - LL = Mips::LL; - SC = Mips::SC; + LL = isMicroMips ? Mips::LL_MM : Mips::LL; + SC = isMicroMips ? Mips::SC_MM : Mips::SC; ZERO = Mips::ZERO; BNE = Mips::BNE; BEQ = Mips::BEQ; - } - else { + } else { LL = Mips::LLD; SC = Mips::SCD; ZERO = Mips::ZERO_64; @@ -1188,7 +1379,7 @@ MipsTargetLowering::emitAtomicCmpSwap(MachineInstr *MI, // Transfer the remainder of BB and its successor edges to exitMBB. exitMBB->splice(exitMBB->begin(), BB, - llvm::next(MachineBasicBlock::iterator(MI)), BB->end()); + std::next(MachineBasicBlock::iterator(MI)), BB->end()); exitMBB->transferSuccessorsAndUpdatePHIs(BB); // thisMBB: @@ -1217,7 +1408,7 @@ MipsTargetLowering::emitAtomicCmpSwap(MachineInstr *MI, BuildMI(BB, DL, TII->get(BEQ)) .addReg(Success).addReg(ZERO).addMBB(loop1MBB); - MI->eraseFromParent(); // The instruction is gone now. + MI->eraseFromParent(); // The instruction is gone now. return exitMBB; } @@ -1232,7 +1423,8 @@ MipsTargetLowering::emitAtomicCmpSwapPartword(MachineInstr *MI, MachineFunction *MF = BB->getParent(); MachineRegisterInfo &RegInfo = MF->getRegInfo(); const TargetRegisterClass *RC = getRegClassFor(MVT::i32); - const TargetInstrInfo *TII = getTargetMachine().getInstrInfo(); + const TargetInstrInfo *TII = + getTargetMachine().getSubtargetImpl()->getInstrInfo(); DebugLoc DL = MI->getDebugLoc(); unsigned Dest = MI->getOperand(0).getReg(); @@ -1256,7 +1448,6 @@ MipsTargetLowering::emitAtomicCmpSwapPartword(MachineInstr *MI, unsigned MaskedOldVal1 = RegInfo.createVirtualRegister(RC); unsigned StoreVal = RegInfo.createVirtualRegister(RC); unsigned SrlRes = RegInfo.createVirtualRegister(RC); - unsigned SllRes = RegInfo.createVirtualRegister(RC); unsigned Success = RegInfo.createVirtualRegister(RC); // insert new blocks after the current block @@ -1274,7 +1465,7 @@ MipsTargetLowering::emitAtomicCmpSwapPartword(MachineInstr *MI, // Transfer the remainder of BB and its successor edges to exitMBB. exitMBB->splice(exitMBB->begin(), BB, - llvm::next(MachineBasicBlock::iterator(MI)), BB->end()); + std::next(MachineBasicBlock::iterator(MI)), BB->end()); exitMBB->transferSuccessorsAndUpdatePHIs(BB); BB->addSuccessor(loop1MBB); @@ -1303,7 +1494,7 @@ MipsTargetLowering::emitAtomicCmpSwapPartword(MachineInstr *MI, BuildMI(BB, DL, TII->get(Mips::AND), AlignedAddr) .addReg(Ptr).addReg(MaskLSB2); BuildMI(BB, DL, TII->get(Mips::ANDi), PtrLSB2).addReg(Ptr).addImm(3); - if (Subtarget->isLittle()) { + if (Subtarget.isLittle()) { BuildMI(BB, DL, TII->get(Mips::SLL), ShiftAmt).addReg(PtrLSB2).addImm(3); } else { unsigned Off = RegInfo.createVirtualRegister(RC); @@ -1353,23 +1544,46 @@ MipsTargetLowering::emitAtomicCmpSwapPartword(MachineInstr *MI, // sinkMBB: // srl srlres,maskedoldval0,shiftamt - // sll sllres,srlres,24 - // sra dest,sllres,24 + // sign_extend dest,srlres BB = sinkMBB; - int64_t ShiftImm = (Size == 1) ? 24 : 16; BuildMI(BB, DL, TII->get(Mips::SRLV), SrlRes) .addReg(MaskedOldVal0).addReg(ShiftAmt); - BuildMI(BB, DL, TII->get(Mips::SLL), SllRes) - .addReg(SrlRes).addImm(ShiftImm); - BuildMI(BB, DL, TII->get(Mips::SRA), Dest) - .addReg(SllRes).addImm(ShiftImm); + BB = emitSignExtendToI32InReg(MI, BB, Size, Dest, SrlRes); MI->eraseFromParent(); // The instruction is gone now. return exitMBB; } +MachineBasicBlock *MipsTargetLowering::emitSEL_D(MachineInstr *MI, + MachineBasicBlock *BB) const { + MachineFunction *MF = BB->getParent(); + const TargetRegisterInfo *TRI = + getTargetMachine().getSubtargetImpl()->getRegisterInfo(); + const TargetInstrInfo *TII = + getTargetMachine().getSubtargetImpl()->getInstrInfo(); + MachineRegisterInfo &RegInfo = MF->getRegInfo(); + DebugLoc DL = MI->getDebugLoc(); + MachineBasicBlock::iterator II(MI); + + unsigned Fc = MI->getOperand(1).getReg(); + const auto &FGR64RegClass = TRI->getRegClass(Mips::FGR64RegClassID); + + unsigned Fc2 = RegInfo.createVirtualRegister(FGR64RegClass); + + BuildMI(*BB, II, DL, TII->get(Mips::SUBREG_TO_REG), Fc2) + .addImm(0) + .addReg(Fc) + .addImm(Mips::sub_lo); + + // We don't erase the original instruction, we just replace the condition + // register with the 64-bit super-register. + MI->getOperand(1).setReg(Fc2); + + return BB; +} + //===----------------------------------------------------------------------===// // Misc Lower Operation implementation //===----------------------------------------------------------------------===// @@ -1389,10 +1603,11 @@ SDValue MipsTargetLowering::lowerBR_JT(SDValue Op, SelectionDAG &DAG) const { EVT MemVT = EVT::getIntegerVT(*DAG.getContext(), EntrySize * 8); Addr = DAG.getExtLoad(ISD::SEXTLOAD, DL, PTy, Chain, Addr, MachinePointerInfo::getJumpTable(), MemVT, false, false, - 0); + false, 0); Chain = Addr.getValue(1); - if ((getTargetMachine().getRelocationModel() == Reloc::PIC_) || IsN64) { + if ((getTargetMachine().getRelocationModel() == Reloc::PIC_) || + Subtarget.isABI_N64()) { // For PIC, the sequence is: // BRIND(load(Jumptable + index) + RelocBase) // RelocBase can be JumpTable, GOT or some sort of global base. @@ -1403,15 +1618,14 @@ SDValue MipsTargetLowering::lowerBR_JT(SDValue Op, SelectionDAG &DAG) const { return DAG.getNode(ISD::BRIND, DL, MVT::Other, Chain, Addr); } -SDValue MipsTargetLowering:: -lowerBRCOND(SDValue Op, SelectionDAG &DAG) const -{ +SDValue MipsTargetLowering::lowerBRCOND(SDValue Op, SelectionDAG &DAG) const { // The first operand is the chain, the second is the condition, the third is // the block to branch to if the condition is true. SDValue Chain = Op.getOperand(0); SDValue Dest = Op.getOperand(2); SDLoc DL(Op); + assert(!Subtarget.hasMips32r6() && !Subtarget.hasMips64r6()); SDValue CondRes = createFPCmp(DAG, Op.getOperand(1)); // Return if flag is not set by a floating point comparison. @@ -1431,6 +1645,7 @@ lowerBRCOND(SDValue Op, SelectionDAG &DAG) const SDValue MipsTargetLowering:: lowerSELECT(SDValue Op, SelectionDAG &DAG) const { + assert(!Subtarget.hasMips32r6() && !Subtarget.hasMips64r6()); SDValue Cond = createFPCmp(DAG, Op.getOperand(0)); // Return if flag is not set by a floating point comparison. @@ -1456,6 +1671,7 @@ lowerSELECT_CC(SDValue Op, SelectionDAG &DAG) const } SDValue MipsTargetLowering::lowerSETCC(SDValue Op, SelectionDAG &DAG) const { + assert(!Subtarget.hasMips32r6() && !Subtarget.hasMips64r6()); SDValue Cond = createFPCmp(DAG, Op); assert(Cond.getOpcode() == MipsISD::FPCmp && @@ -1471,9 +1687,12 @@ SDValue MipsTargetLowering::lowerGlobalAddress(SDValue Op, SelectionDAG &DAG) const { // FIXME there isn't actually debug info here SDLoc DL(Op); - const GlobalValue *GV = cast(Op)->getGlobal(); + EVT Ty = Op.getValueType(); + GlobalAddressSDNode *N = cast(Op); + const GlobalValue *GV = N->getGlobal(); - if (getTargetMachine().getRelocationModel() != Reloc::PIC_ && !IsN64) { + if (getTargetMachine().getRelocationModel() != Reloc::PIC_ && + !Subtarget.isABI_N64()) { const MipsTargetObjectFile &TLOF = (const MipsTargetObjectFile&)getObjFileLowering(); @@ -1482,32 +1701,42 @@ SDValue MipsTargetLowering::lowerGlobalAddress(SDValue Op, SDValue GA = DAG.getTargetGlobalAddress(GV, DL, MVT::i32, 0, MipsII::MO_GPREL); SDValue GPRelNode = DAG.getNode(MipsISD::GPRel, DL, - DAG.getVTList(MVT::i32), &GA, 1); + DAG.getVTList(MVT::i32), GA); SDValue GPReg = DAG.getRegister(Mips::GP, MVT::i32); return DAG.getNode(ISD::ADD, DL, MVT::i32, GPReg, GPRelNode); } // %hi/%lo relocation - return getAddrNonPIC(Op, DAG); + return getAddrNonPIC(N, Ty, DAG); } if (GV->hasInternalLinkage() || (GV->hasLocalLinkage() && !isa(GV))) - return getAddrLocal(Op, DAG, HasMips64); + return getAddrLocal(N, Ty, DAG, + Subtarget.isABI_N32() || Subtarget.isABI_N64()); if (LargeGOT) - return getAddrGlobalLargeGOT(Op, DAG, MipsII::MO_GOT_HI16, - MipsII::MO_GOT_LO16); + return getAddrGlobalLargeGOT(N, Ty, DAG, MipsII::MO_GOT_HI16, + MipsII::MO_GOT_LO16, DAG.getEntryNode(), + MachinePointerInfo::getGOT()); - return getAddrGlobal(Op, DAG, - HasMips64 ? MipsII::MO_GOT_DISP : MipsII::MO_GOT16); + return getAddrGlobal(N, Ty, DAG, + (Subtarget.isABI_N32() || Subtarget.isABI_N64()) + ? MipsII::MO_GOT_DISP + : MipsII::MO_GOT16, + DAG.getEntryNode(), MachinePointerInfo::getGOT()); } SDValue MipsTargetLowering::lowerBlockAddress(SDValue Op, SelectionDAG &DAG) const { - if (getTargetMachine().getRelocationModel() != Reloc::PIC_ && !IsN64) - return getAddrNonPIC(Op, DAG); + BlockAddressSDNode *N = cast(Op); + EVT Ty = Op.getValueType(); - return getAddrLocal(Op, DAG, HasMips64); + if (getTargetMachine().getRelocationModel() != Reloc::PIC_ && + !Subtarget.isABI_N64()) + return getAddrNonPIC(N, Ty, DAG); + + return getAddrLocal(N, Ty, DAG, + Subtarget.isABI_N32() || Subtarget.isABI_N64()); } SDValue MipsTargetLowering:: @@ -1543,11 +1772,9 @@ lowerGlobalTLSAddress(SDValue Op, SelectionDAG &DAG) const Entry.Ty = PtrTy; Args.push_back(Entry); - TargetLowering::CallLoweringInfo CLI(DAG.getEntryNode(), PtrTy, - false, false, false, false, 0, CallingConv::C, - /*IsTailCall=*/false, /*doesNotRet=*/false, - /*isReturnValueUsed=*/true, - TlsGetAddr, Args, DAG, DL); + TargetLowering::CallLoweringInfo CLI(DAG); + CLI.setDebugLoc(DL).setChain(DAG.getEntryNode()) + .setCallee(CallingConv::C, PtrTy, TlsGetAddr, std::move(Args), 0); std::pair CallResult = LowerCallTo(CLI); SDValue Ret = CallResult.first; @@ -1594,10 +1821,15 @@ lowerGlobalTLSAddress(SDValue Op, SelectionDAG &DAG) const SDValue MipsTargetLowering:: lowerJumpTable(SDValue Op, SelectionDAG &DAG) const { - if (getTargetMachine().getRelocationModel() != Reloc::PIC_ && !IsN64) - return getAddrNonPIC(Op, DAG); + JumpTableSDNode *N = cast(Op); + EVT Ty = Op.getValueType(); - return getAddrLocal(Op, DAG, HasMips64); + if (getTargetMachine().getRelocationModel() != Reloc::PIC_ && + !Subtarget.isABI_N64()) + return getAddrNonPIC(N, Ty, DAG); + + return getAddrLocal(N, Ty, DAG, + Subtarget.isABI_N32() || Subtarget.isABI_N64()); } SDValue MipsTargetLowering:: @@ -1612,11 +1844,15 @@ lowerConstantPool(SDValue Op, SelectionDAG &DAG) const // SDValue GPRelNode = DAG.getNode(MipsISD::GPRel, MVT::i32, CP); // SDValue GOT = DAG.getGLOBAL_OFFSET_TABLE(MVT::i32); // ResNode = DAG.getNode(ISD::ADD, MVT::i32, GOT, GPRelNode); + ConstantPoolSDNode *N = cast(Op); + EVT Ty = Op.getValueType(); - if (getTargetMachine().getRelocationModel() != Reloc::PIC_ && !IsN64) - return getAddrNonPIC(Op, DAG); + if (getTargetMachine().getRelocationModel() != Reloc::PIC_ && + !Subtarget.isABI_N64()) + return getAddrNonPIC(N, Ty, DAG); - return getAddrLocal(Op, DAG, HasMips64); + return getAddrLocal(N, Ty, DAG, + Subtarget.isABI_N32() || Subtarget.isABI_N64()); } SDValue MipsTargetLowering::lowerVASTART(SDValue Op, SelectionDAG &DAG) const { @@ -1634,7 +1870,67 @@ SDValue MipsTargetLowering::lowerVASTART(SDValue Op, SelectionDAG &DAG) const { MachinePointerInfo(SV), false, false, 0); } -static SDValue lowerFCOPYSIGN32(SDValue Op, SelectionDAG &DAG, bool HasR2) { +SDValue MipsTargetLowering::lowerVAARG(SDValue Op, SelectionDAG &DAG) const { + SDNode *Node = Op.getNode(); + EVT VT = Node->getValueType(0); + SDValue Chain = Node->getOperand(0); + SDValue VAListPtr = Node->getOperand(1); + unsigned Align = Node->getConstantOperandVal(3); + const Value *SV = cast(Node->getOperand(2))->getValue(); + SDLoc DL(Node); + unsigned ArgSlotSizeInBytes = + (Subtarget.isABI_N32() || Subtarget.isABI_N64()) ? 8 : 4; + + SDValue VAListLoad = DAG.getLoad(getPointerTy(), DL, Chain, VAListPtr, + MachinePointerInfo(SV), false, false, false, + 0); + SDValue VAList = VAListLoad; + + // Re-align the pointer if necessary. + // It should only ever be necessary for 64-bit types on O32 since the minimum + // argument alignment is the same as the maximum type alignment for N32/N64. + // + // FIXME: We currently align too often. The code generator doesn't notice + // when the pointer is still aligned from the last va_arg (or pair of + // va_args for the i64 on O32 case). + if (Align > getMinStackArgumentAlignment()) { + assert(((Align & (Align-1)) == 0) && "Expected Align to be a power of 2"); + + VAList = DAG.getNode(ISD::ADD, DL, VAList.getValueType(), VAList, + DAG.getConstant(Align - 1, + VAList.getValueType())); + + VAList = DAG.getNode(ISD::AND, DL, VAList.getValueType(), VAList, + DAG.getConstant(-(int64_t)Align, + VAList.getValueType())); + } + + // Increment the pointer, VAList, to the next vaarg. + unsigned ArgSizeInBytes = getDataLayout()->getTypeAllocSize(VT.getTypeForEVT(*DAG.getContext())); + SDValue Tmp3 = DAG.getNode(ISD::ADD, DL, VAList.getValueType(), VAList, + DAG.getConstant(RoundUpToAlignment(ArgSizeInBytes, ArgSlotSizeInBytes), + VAList.getValueType())); + // Store the incremented VAList to the legalized pointer + Chain = DAG.getStore(VAListLoad.getValue(1), DL, Tmp3, VAListPtr, + MachinePointerInfo(SV), false, false, 0); + + // In big-endian mode we must adjust the pointer when the load size is smaller + // than the argument slot size. We must also reduce the known alignment to + // match. For example in the N64 ABI, we must add 4 bytes to the offset to get + // the correct half of the slot, and reduce the alignment from 8 (slot + // alignment) down to 4 (type alignment). + if (!Subtarget.isLittle() && ArgSizeInBytes < ArgSlotSizeInBytes) { + unsigned Adjustment = ArgSlotSizeInBytes - ArgSizeInBytes; + VAList = DAG.getNode(ISD::ADD, DL, VAListPtr.getValueType(), VAList, + DAG.getIntPtrConstant(Adjustment)); + } + // Load the actual argument out of the pointer VAList + return DAG.getLoad(VT, DL, Chain, VAList, MachinePointerInfo(), false, false, + false, 0); +} + +static SDValue lowerFCOPYSIGN32(SDValue Op, SelectionDAG &DAG, + bool HasExtractInsert) { EVT TyX = Op.getOperand(0).getValueType(); EVT TyY = Op.getOperand(1).getValueType(); SDValue Const1 = DAG.getConstant(1, MVT::i32); @@ -1653,7 +1949,7 @@ static SDValue lowerFCOPYSIGN32(SDValue Op, SelectionDAG &DAG, bool HasR2) { DAG.getNode(MipsISD::ExtractElementF64, DL, MVT::i32, Op.getOperand(1), Const1); - if (HasR2) { + if (HasExtractInsert) { // ext E, Y, 31, 1 ; extract bit31 of Y // ins X, E, 31, 1 ; insert extracted bit at bit31 of X SDValue E = DAG.getNode(MipsISD::Ext, DL, MVT::i32, Y, Const31, Const1); @@ -1679,7 +1975,8 @@ static SDValue lowerFCOPYSIGN32(SDValue Op, SelectionDAG &DAG, bool HasR2) { return DAG.getNode(MipsISD::BuildPairF64, DL, MVT::f64, LowX, Res); } -static SDValue lowerFCOPYSIGN64(SDValue Op, SelectionDAG &DAG, bool HasR2) { +static SDValue lowerFCOPYSIGN64(SDValue Op, SelectionDAG &DAG, + bool HasExtractInsert) { unsigned WidthX = Op.getOperand(0).getValueSizeInBits(); unsigned WidthY = Op.getOperand(1).getValueSizeInBits(); EVT TyX = MVT::getIntegerVT(WidthX), TyY = MVT::getIntegerVT(WidthY); @@ -1690,7 +1987,7 @@ static SDValue lowerFCOPYSIGN64(SDValue Op, SelectionDAG &DAG, bool HasR2) { SDValue X = DAG.getNode(ISD::BITCAST, DL, TyX, Op.getOperand(0)); SDValue Y = DAG.getNode(ISD::BITCAST, DL, TyY, Op.getOperand(1)); - if (HasR2) { + if (HasExtractInsert) { // ext E, Y, width(Y) - 1, 1 ; extract bit width(Y)-1 of Y // ins X, E, width(X) - 1, 1 ; insert extracted bit at bit width(X)-1 of X SDValue E = DAG.getNode(MipsISD::Ext, DL, TyY, Y, @@ -1729,67 +2026,10 @@ static SDValue lowerFCOPYSIGN64(SDValue Op, SelectionDAG &DAG, bool HasR2) { SDValue MipsTargetLowering::lowerFCOPYSIGN(SDValue Op, SelectionDAG &DAG) const { - if (Subtarget->hasMips64()) - return lowerFCOPYSIGN64(Op, DAG, Subtarget->hasMips32r2()); + if (Subtarget.isGP64bit()) + return lowerFCOPYSIGN64(Op, DAG, Subtarget.hasExtractInsert()); - return lowerFCOPYSIGN32(Op, DAG, Subtarget->hasMips32r2()); -} - -static SDValue lowerFABS32(SDValue Op, SelectionDAG &DAG, bool HasR2) { - SDValue Res, Const1 = DAG.getConstant(1, MVT::i32); - SDLoc DL(Op); - - // If operand is of type f64, extract the upper 32-bit. Otherwise, bitcast it - // to i32. - SDValue X = (Op.getValueType() == MVT::f32) ? - DAG.getNode(ISD::BITCAST, DL, MVT::i32, Op.getOperand(0)) : - DAG.getNode(MipsISD::ExtractElementF64, DL, MVT::i32, Op.getOperand(0), - Const1); - - // Clear MSB. - if (HasR2) - Res = DAG.getNode(MipsISD::Ins, DL, MVT::i32, - DAG.getRegister(Mips::ZERO, MVT::i32), - DAG.getConstant(31, MVT::i32), Const1, X); - else { - SDValue SllX = DAG.getNode(ISD::SHL, DL, MVT::i32, X, Const1); - Res = DAG.getNode(ISD::SRL, DL, MVT::i32, SllX, Const1); - } - - if (Op.getValueType() == MVT::f32) - return DAG.getNode(ISD::BITCAST, DL, MVT::f32, Res); - - SDValue LowX = DAG.getNode(MipsISD::ExtractElementF64, DL, MVT::i32, - Op.getOperand(0), DAG.getConstant(0, MVT::i32)); - return DAG.getNode(MipsISD::BuildPairF64, DL, MVT::f64, LowX, Res); -} - -static SDValue lowerFABS64(SDValue Op, SelectionDAG &DAG, bool HasR2) { - SDValue Res, Const1 = DAG.getConstant(1, MVT::i32); - SDLoc DL(Op); - - // Bitcast to integer node. - SDValue X = DAG.getNode(ISD::BITCAST, DL, MVT::i64, Op.getOperand(0)); - - // Clear MSB. - if (HasR2) - Res = DAG.getNode(MipsISD::Ins, DL, MVT::i64, - DAG.getRegister(Mips::ZERO_64, MVT::i64), - DAG.getConstant(63, MVT::i32), Const1, X); - else { - SDValue SllX = DAG.getNode(ISD::SHL, DL, MVT::i64, X, Const1); - Res = DAG.getNode(ISD::SRL, DL, MVT::i64, SllX, Const1); - } - - return DAG.getNode(ISD::BITCAST, DL, MVT::f64, Res); -} - -SDValue -MipsTargetLowering::lowerFABS(SDValue Op, SelectionDAG &DAG) const { - if (Subtarget->hasMips64() && (Op.getValueType() == MVT::f64)) - return lowerFABS64(Op, DAG, Subtarget->hasMips32r2()); - - return lowerFABS32(Op, DAG, Subtarget->hasMips32r2()); + return lowerFCOPYSIGN32(Op, DAG, Subtarget.hasExtractInsert()); } SDValue MipsTargetLowering:: @@ -1802,13 +2042,17 @@ lowerFRAMEADDR(SDValue Op, SelectionDAG &DAG) const { MFI->setFrameAddressIsTaken(true); EVT VT = Op.getValueType(); SDLoc DL(Op); - SDValue FrameAddr = DAG.getCopyFromReg(DAG.getEntryNode(), DL, - IsN64 ? Mips::FP_64 : Mips::FP, VT); + SDValue FrameAddr = + DAG.getCopyFromReg(DAG.getEntryNode(), DL, + Subtarget.isABI_N64() ? Mips::FP_64 : Mips::FP, VT); return FrameAddr; } SDValue MipsTargetLowering::lowerRETURNADDR(SDValue Op, SelectionDAG &DAG) const { + if (verifyReturnAddressArgumentIsConstant(Op, DAG)) + return SDValue(); + // check the depth assert((cast(Op.getOperand(0))->getZExtValue() == 0) && "Return address can be determined only for current frame."); @@ -1816,7 +2060,7 @@ SDValue MipsTargetLowering::lowerRETURNADDR(SDValue Op, MachineFunction &MF = DAG.getMachineFunction(); MachineFrameInfo *MFI = MF.getFrameInfo(); MVT VT = Op.getSimpleValueType(); - unsigned RA = IsN64 ? Mips::RA_64 : Mips::RA; + unsigned RA = Subtarget.isABI_N64() ? Mips::RA_64 : Mips::RA; MFI->setReturnAddressIsTaken(true); // Return RA, which contains the return address. Mark it an implicit live-in. @@ -1838,12 +2082,12 @@ SDValue MipsTargetLowering::lowerEH_RETURN(SDValue Op, SelectionDAG &DAG) SDValue Offset = Op.getOperand(1); SDValue Handler = Op.getOperand(2); SDLoc DL(Op); - EVT Ty = IsN64 ? MVT::i64 : MVT::i32; + EVT Ty = Subtarget.isABI_N64() ? MVT::i64 : MVT::i32; // Store stack offset in V1, store jump target in V0. Glue CopyToReg and // EH_RETURN nodes, so that instructions are emitted back-to-back. - unsigned OffsetReg = IsN64 ? Mips::V1_64 : Mips::V1; - unsigned AddrReg = IsN64 ? Mips::V0_64 : Mips::V0; + unsigned OffsetReg = Subtarget.isABI_N64() ? Mips::V1_64 : Mips::V1; + unsigned AddrReg = Subtarget.isABI_N64() ? Mips::V0_64 : Mips::V0; Chain = DAG.getCopyToReg(Chain, DL, OffsetReg, Offset, SDValue()); Chain = DAG.getCopyToReg(Chain, DL, AddrReg, Handler, Chain.getValue(1)); return DAG.getNode(MipsISD::EH_RETURN, DL, MVT::Other, Chain, @@ -1890,7 +2134,7 @@ SDValue MipsTargetLowering::lowerShiftLeftParts(SDValue Op, Hi = DAG.getNode(ISD::SELECT, DL, MVT::i32, Cond, ShiftLeftLo, Or); SDValue Ops[2] = {Lo, Hi}; - return DAG.getMergeValues(Ops, 2, DL); + return DAG.getMergeValues(Ops, DL); } SDValue MipsTargetLowering::lowerShiftRightParts(SDValue Op, SelectionDAG &DAG, @@ -1931,7 +2175,7 @@ SDValue MipsTargetLowering::lowerShiftRightParts(SDValue Op, SelectionDAG &DAG, ShiftRightHi); SDValue Ops[2] = {Lo, Hi}; - return DAG.getMergeValues(Ops, 2, DL); + return DAG.getMergeValues(Ops, DL); } static SDValue createLoadLR(unsigned Opc, SelectionDAG &DAG, LoadSDNode *LD, @@ -1947,7 +2191,7 @@ static SDValue createLoadLR(unsigned Opc, SelectionDAG &DAG, LoadSDNode *LD, DAG.getConstant(Offset, BasePtrVT)); SDValue Ops[] = { Chain, Ptr, Src }; - return DAG.getMemIntrinsicNode(Opc, DL, VTList, Ops, 3, MemVT, + return DAG.getMemIntrinsicNode(Opc, DL, VTList, Ops, MemVT, LD->getMemOperand()); } @@ -1956,12 +2200,15 @@ SDValue MipsTargetLowering::lowerLOAD(SDValue Op, SelectionDAG &DAG) const { LoadSDNode *LD = cast(Op); EVT MemVT = LD->getMemoryVT(); + if (Subtarget.systemSupportsUnalignedAccess()) + return Op; + // Return if load is aligned or if MemVT is neither i32 nor i64. if ((LD->getAlignment() >= MemVT.getSizeInBits() / 8) || ((MemVT != MVT::i32) && (MemVT != MVT::i64))) return SDValue(); - bool IsLittle = Subtarget->isLittle(); + bool IsLittle = Subtarget.isLittle(); EVT VT = Op.getValueType(); ISD::LoadExtType ExtType = LD->getExtensionType(); SDValue Chain = LD->getChain(), Undef = DAG.getUNDEF(VT); @@ -2010,7 +2257,7 @@ SDValue MipsTargetLowering::lowerLOAD(SDValue Op, SelectionDAG &DAG) const { SDValue SLL = DAG.getNode(ISD::SHL, DL, MVT::i64, LWR, Const32); SDValue SRL = DAG.getNode(ISD::SRL, DL, MVT::i64, SLL, Const32); SDValue Ops[] = { SRL, LWR.getValue(1) }; - return DAG.getMergeValues(Ops, 2, DL); + return DAG.getMergeValues(Ops, DL); } static SDValue createStoreLR(unsigned Opc, SelectionDAG &DAG, StoreSDNode *SD, @@ -2025,7 +2272,7 @@ static SDValue createStoreLR(unsigned Opc, SelectionDAG &DAG, StoreSDNode *SD, DAG.getConstant(Offset, BasePtrVT)); SDValue Ops[] = { Chain, Value, Ptr }; - return DAG.getMemIntrinsicNode(Opc, DL, VTList, Ops, 3, MemVT, + return DAG.getMemIntrinsicNode(Opc, DL, VTList, Ops, MemVT, SD->getMemOperand()); } @@ -2079,9 +2326,10 @@ SDValue MipsTargetLowering::lowerSTORE(SDValue Op, SelectionDAG &DAG) const { EVT MemVT = SD->getMemoryVT(); // Lower unaligned integer stores. - if ((SD->getAlignment() < MemVT.getSizeInBits() / 8) && + if (!Subtarget.systemSupportsUnalignedAccess() && + (SD->getAlignment() < MemVT.getSizeInBits() / 8) && ((MemVT == MVT::i32) || (MemVT == MVT::i64))) - return lowerUnalignedIntStore(SD, DAG, Subtarget->isLittle()); + return lowerUnalignedIntStore(SD, DAG, Subtarget.isLittle()); return lowerFP_TO_SINT_STORE(SD, DAG); } @@ -2128,25 +2376,20 @@ SDValue MipsTargetLowering::lowerFP_TO_SINT(SDValue Op, // an argument. Otherwise, passed in A1, A2, A3 and stack. // f64 - Only passed in two aliased f32 registers if no int reg has been used // yet to hold an argument. Otherwise, use A2, A3 and stack. If A1 is -// not used, it must be shadowed. If only A3 is avaiable, shadow it and +// not used, it must be shadowed. If only A3 is available, shadow it and // go to stack. // // For vararg functions, all arguments are passed in A0, A1, A2, A3 and stack. //===----------------------------------------------------------------------===// -static bool CC_MipsO32(unsigned ValNo, MVT ValVT, - MVT LocVT, CCValAssign::LocInfo LocInfo, - ISD::ArgFlagsTy ArgFlags, CCState &State, - const uint16_t *F64Regs) { +static bool CC_MipsO32(unsigned ValNo, MVT ValVT, MVT LocVT, + CCValAssign::LocInfo LocInfo, ISD::ArgFlagsTy ArgFlags, + CCState &State, const MCPhysReg *F64Regs) { - static const unsigned IntRegsSize=4, FloatRegsSize=2; + static const unsigned IntRegsSize = 4, FloatRegsSize = 2; - static const uint16_t IntRegs[] = { - Mips::A0, Mips::A1, Mips::A2, Mips::A3 - }; - static const uint16_t F32Regs[] = { - Mips::F12, Mips::F14 - }; + static const MCPhysReg IntRegs[] = { Mips::A0, Mips::A1, Mips::A2, Mips::A3 }; + static const MCPhysReg F32Regs[] = { Mips::F12, Mips::F14 }; // Do not process byval args here. if (ArgFlags.isByVal()) @@ -2218,7 +2461,7 @@ static bool CC_MipsO32(unsigned ValNo, MVT ValVT, static bool CC_MipsO32_FP32(unsigned ValNo, MVT ValVT, MVT LocVT, CCValAssign::LocInfo LocInfo, ISD::ArgFlagsTy ArgFlags, CCState &State) { - static const uint16_t F64Regs[] = { Mips::D6, Mips::D7 }; + static const MCPhysReg F64Regs[] = { Mips::D6, Mips::D7 }; return CC_MipsO32(ValNo, ValVT, LocVT, LocInfo, ArgFlags, State, F64Regs); } @@ -2226,7 +2469,7 @@ static bool CC_MipsO32_FP32(unsigned ValNo, MVT ValVT, static bool CC_MipsO32_FP64(unsigned ValNo, MVT ValVT, MVT LocVT, CCValAssign::LocInfo LocInfo, ISD::ArgFlagsTy ArgFlags, CCState &State) { - static const uint16_t F64Regs[] = { Mips::D12_64, Mips::D12_64 }; + static const MCPhysReg F64Regs[] = { Mips::D12_64, Mips::D14_64 }; return CC_MipsO32(ValNo, ValVT, LocVT, LocInfo, ArgFlags, State, F64Regs); } @@ -2265,15 +2508,21 @@ void MipsTargetLowering:: getOpndList(SmallVectorImpl &Ops, std::deque< std::pair > &RegsToPass, bool IsPICCall, bool GlobalOrExternal, bool InternalLinkage, - CallLoweringInfo &CLI, SDValue Callee, SDValue Chain) const { + bool IsCallReloc, CallLoweringInfo &CLI, SDValue Callee, + SDValue Chain) const { // Insert node "GP copy globalreg" before call to function. // // R_MIPS_CALL* operators (emitted when non-internal functions are called // in PIC mode) allow symbols to be resolved via lazy binding. // The lazy binding stub requires GP to point to the GOT. - if (IsPICCall && !InternalLinkage) { - unsigned GPReg = IsN64 ? Mips::GP_64 : Mips::GP; - EVT Ty = IsN64 ? MVT::i64 : MVT::i32; + // Note that we don't need GP to point to the GOT for indirect calls + // (when R_MIPS_CALL* is not used for the call) because Mips linker generates + // lazy binding stub for a function only when R_MIPS_CALL* are the only relocs + // used for the function (that is, Mips linker doesn't generate lazy binding + // stub for a function whose address is taken in the program). + if (IsPICCall && !InternalLinkage && IsCallReloc) { + unsigned GPReg = Subtarget.isABI_N64() ? Mips::GP_64 : Mips::GP; + EVT Ty = Subtarget.isABI_N64() ? MVT::i64 : MVT::i32; RegsToPass.push_back(std::make_pair(GPReg, getGlobalReg(CLI.DAG, Ty))); } @@ -2296,14 +2545,15 @@ getOpndList(SmallVectorImpl &Ops, RegsToPass[i].second.getValueType())); // Add a register mask operand representing the call-preserved registers. - const TargetRegisterInfo *TRI = getTargetMachine().getRegisterInfo(); + const TargetRegisterInfo *TRI = + getTargetMachine().getSubtargetImpl()->getRegisterInfo(); const uint32_t *Mask = TRI->getCallPreservedMask(CLI.CallConv); assert(Mask && "Missing call preserved mask for calling convention"); - if (Subtarget->inMips16HardFloat()) { + if (Subtarget.inMips16HardFloat()) { if (GlobalAddressSDNode *G = dyn_cast(CLI.Callee)) { llvm::StringRef Sym = G->getGlobal()->getName(); Function *F = G->getGlobal()->getParent()->getFunction(Sym); - if (F->hasFnAttribute("__Mips16RetHelper")) { + if (F && F->hasFnAttribute("__Mips16RetHelper")) { Mask = MipsRegisterInfo::getMips16RetHelperMask(); } } @@ -2332,21 +2582,20 @@ MipsTargetLowering::LowerCall(TargetLowering::CallLoweringInfo &CLI, MachineFunction &MF = DAG.getMachineFunction(); MachineFrameInfo *MFI = MF.getFrameInfo(); - const TargetFrameLowering *TFL = MF.getTarget().getFrameLowering(); + const TargetFrameLowering *TFL = MF.getSubtarget().getFrameLowering(); + MipsFunctionInfo *FuncInfo = MF.getInfo(); bool IsPIC = getTargetMachine().getRelocationModel() == Reloc::PIC_; // Analyze operands of the call, assigning locations to each operand. SmallVector ArgLocs; - CCState CCInfo(CallConv, IsVarArg, DAG.getMachineFunction(), - getTargetMachine(), ArgLocs, *DAG.getContext()); - MipsCC::SpecialCallingConvType SpecialCallingConv = - getSpecialCallingConv(Callee); - MipsCC MipsCCInfo(CallConv, IsO32, Subtarget->isFP64bit(), CCInfo, - SpecialCallingConv); + MipsCCState CCInfo(CallConv, IsVarArg, DAG.getMachineFunction(), ArgLocs, + *DAG.getContext()); + MipsCC MipsCCInfo(CallConv, Subtarget, CCInfo); - MipsCCInfo.analyzeCallOperands(Outs, IsVarArg, - Subtarget->mipsSEUsesSoftFloat(), - Callee.getNode(), CLI.Args); + CCInfo.PreAnalyzeCallOperandsForF128_(Outs, CLI.getArgs(), Callee.getNode()); + MipsCCInfo.analyzeCallOperands(Outs, IsVarArg, Subtarget.abiUsesSoftFloat(), + Callee.getNode(), CLI.getArgs(), CCInfo); + CCInfo.ClearOriginalArgWasF128(); // Get a count of how many bytes are to be pushed on the stack. unsigned NextStackOffset = CCInfo.getNextStackOffset(); @@ -2357,6 +2606,10 @@ MipsTargetLowering::LowerCall(TargetLowering::CallLoweringInfo &CLI, isEligibleForTailCallOptimization(MipsCCInfo, NextStackOffset, *MF.getInfo()); + if (!IsTailCall && CLI.CS && CLI.CS->isMustTailCall()) + report_fatal_error("failed to perform tail call elimination on a call " + "site marked musttail"); + if (IsTailCall) ++NumTailCalls; @@ -2370,9 +2623,9 @@ MipsTargetLowering::LowerCall(TargetLowering::CallLoweringInfo &CLI, if (!IsTailCall) Chain = DAG.getCALLSEQ_START(Chain, NextStackOffsetVal, DL); - SDValue StackPtr = DAG.getCopyFromReg(Chain, DL, - IsN64 ? Mips::SP_64 : Mips::SP, - getPointerTy()); + SDValue StackPtr = DAG.getCopyFromReg( + Chain, DL, Subtarget.isABI_N64() ? Mips::SP_64 : Mips::SP, + getPointerTy()); // With EABI is it possible to have 16 args on registers. std::deque< std::pair > RegsToPass; @@ -2394,7 +2647,7 @@ MipsTargetLowering::LowerCall(TargetLowering::CallLoweringInfo &CLI, assert(!IsTailCall && "Do not tail-call optimize if there is a byval argument."); passByValArg(Chain, DL, RegsToPass, MemOpChains, StackPtr, MFI, DAG, Arg, - MipsCCInfo, *ByValArg, Flags, Subtarget->isLittle()); + MipsCCInfo, *ByValArg, Flags, Subtarget.isLittle()); ++ByValArg; continue; } @@ -2413,7 +2666,7 @@ MipsTargetLowering::LowerCall(TargetLowering::CallLoweringInfo &CLI, Arg, DAG.getConstant(0, MVT::i32)); SDValue Hi = DAG.getNode(MipsISD::ExtractElementF64, DL, MVT::i32, Arg, DAG.getConstant(1, MVT::i32)); - if (!Subtarget->isLittle()) + if (!Subtarget.isLittle()) std::swap(Lo, Hi); unsigned LocRegLo = VA.getLocReg(); unsigned LocRegHigh = getNextIntArgReg(LocRegLo); @@ -2423,6 +2676,9 @@ MipsTargetLowering::LowerCall(TargetLowering::CallLoweringInfo &CLI, } } break; + case CCValAssign::BCvt: + Arg = DAG.getNode(ISD::BITCAST, DL, LocVT, Arg); + break; case CCValAssign::SExt: Arg = DAG.getNode(ISD::SIGN_EXTEND, DL, LocVT, Arg); break; @@ -2453,41 +2709,57 @@ MipsTargetLowering::LowerCall(TargetLowering::CallLoweringInfo &CLI, // Transform all store nodes into one single node because all store // nodes are independent of each other. if (!MemOpChains.empty()) - Chain = DAG.getNode(ISD::TokenFactor, DL, MVT::Other, - &MemOpChains[0], MemOpChains.size()); + Chain = DAG.getNode(ISD::TokenFactor, DL, MVT::Other, MemOpChains); // If the callee is a GlobalAddress/ExternalSymbol node (quite common, every // direct call is) turn it into a TargetGlobalAddress/TargetExternalSymbol // node so that legalize doesn't hack it. - bool IsPICCall = (IsN64 || IsPIC); // true if calls are translated to jalr $25 - bool GlobalOrExternal = false, InternalLinkage = false; + bool IsPICCall = + (Subtarget.isABI_N64() || IsPIC); // true if calls are translated to + // jalr $25 + bool GlobalOrExternal = false, InternalLinkage = false, IsCallReloc = false; SDValue CalleeLo; + EVT Ty = Callee.getValueType(); if (GlobalAddressSDNode *G = dyn_cast(Callee)) { if (IsPICCall) { - InternalLinkage = G->getGlobal()->hasInternalLinkage(); + const GlobalValue *Val = G->getGlobal(); + InternalLinkage = Val->hasInternalLinkage(); if (InternalLinkage) - Callee = getAddrLocal(Callee, DAG, HasMips64); - else if (LargeGOT) - Callee = getAddrGlobalLargeGOT(Callee, DAG, MipsII::MO_CALL_HI16, - MipsII::MO_CALL_LO16); - else - Callee = getAddrGlobal(Callee, DAG, MipsII::MO_GOT_CALL); + Callee = getAddrLocal(G, Ty, DAG, + Subtarget.isABI_N32() || Subtarget.isABI_N64()); + else if (LargeGOT) { + Callee = getAddrGlobalLargeGOT(G, Ty, DAG, MipsII::MO_CALL_HI16, + MipsII::MO_CALL_LO16, Chain, + FuncInfo->callPtrInfo(Val)); + IsCallReloc = true; + } else { + Callee = getAddrGlobal(G, Ty, DAG, MipsII::MO_GOT_CALL, Chain, + FuncInfo->callPtrInfo(Val)); + IsCallReloc = true; + } } else Callee = DAG.getTargetGlobalAddress(G->getGlobal(), DL, getPointerTy(), 0, MipsII::MO_NO_FLAG); GlobalOrExternal = true; } else if (ExternalSymbolSDNode *S = dyn_cast(Callee)) { - if (!IsN64 && !IsPIC) // !N64 && static - Callee = DAG.getTargetExternalSymbol(S->getSymbol(), getPointerTy(), + const char *Sym = S->getSymbol(); + + if (!Subtarget.isABI_N64() && !IsPIC) // !N64 && static + Callee = DAG.getTargetExternalSymbol(Sym, getPointerTy(), MipsII::MO_NO_FLAG); - else if (LargeGOT) - Callee = getAddrGlobalLargeGOT(Callee, DAG, MipsII::MO_CALL_HI16, - MipsII::MO_CALL_LO16); - else // N64 || PIC - Callee = getAddrGlobal(Callee, DAG, MipsII::MO_GOT_CALL); + else if (LargeGOT) { + Callee = getAddrGlobalLargeGOT(S, Ty, DAG, MipsII::MO_CALL_HI16, + MipsII::MO_CALL_LO16, Chain, + FuncInfo->callPtrInfo(Sym)); + IsCallReloc = true; + } else { // N64 || PIC + Callee = getAddrGlobal(S, Ty, DAG, MipsII::MO_GOT_CALL, Chain, + FuncInfo->callPtrInfo(Sym)); + IsCallReloc = true; + } GlobalOrExternal = true; } @@ -2496,12 +2768,12 @@ MipsTargetLowering::LowerCall(TargetLowering::CallLoweringInfo &CLI, SDVTList NodeTys = DAG.getVTList(MVT::Other, MVT::Glue); getOpndList(Ops, RegsToPass, IsPICCall, GlobalOrExternal, InternalLinkage, - CLI, Callee, Chain); + IsCallReloc, CLI, Callee, Chain); if (IsTailCall) - return DAG.getNode(MipsISD::TailCall, DL, MVT::Other, &Ops[0], Ops.size()); + return DAG.getNode(MipsISD::TailCall, DL, MVT::Other, Ops); - Chain = DAG.getNode(MipsISD::JmpLink, DL, NodeTys, &Ops[0], Ops.size()); + Chain = DAG.getNode(MipsISD::JmpLink, DL, NodeTys, Ops); SDValue InFlag = Chain.getValue(1); // Create the CALLSEQ_END node. @@ -2511,38 +2783,68 @@ MipsTargetLowering::LowerCall(TargetLowering::CallLoweringInfo &CLI, // Handle result values, copying them out of physregs into vregs that we // return. - return LowerCallResult(Chain, InFlag, CallConv, IsVarArg, - Ins, DL, DAG, InVals, CLI.Callee.getNode(), CLI.RetTy); + return LowerCallResult(Chain, InFlag, CallConv, IsVarArg, Ins, DL, DAG, + InVals, CLI); } /// LowerCallResult - Lower the result values of a call into the /// appropriate copies out of appropriate physical registers. -SDValue -MipsTargetLowering::LowerCallResult(SDValue Chain, SDValue InFlag, - CallingConv::ID CallConv, bool IsVarArg, - const SmallVectorImpl &Ins, - SDLoc DL, SelectionDAG &DAG, - SmallVectorImpl &InVals, - const SDNode *CallNode, - const Type *RetTy) const { +SDValue MipsTargetLowering::LowerCallResult( + SDValue Chain, SDValue InFlag, CallingConv::ID CallConv, bool IsVarArg, + const SmallVectorImpl &Ins, SDLoc DL, SelectionDAG &DAG, + SmallVectorImpl &InVals, + TargetLowering::CallLoweringInfo &CLI) const { // Assign locations to each value returned by this call. SmallVector RVLocs; - CCState CCInfo(CallConv, IsVarArg, DAG.getMachineFunction(), - getTargetMachine(), RVLocs, *DAG.getContext()); - MipsCC MipsCCInfo(CallConv, IsO32, Subtarget->isFP64bit(), CCInfo); - - MipsCCInfo.analyzeCallResult(Ins, Subtarget->mipsSEUsesSoftFloat(), - CallNode, RetTy); + MipsCCState CCInfo(CallConv, IsVarArg, DAG.getMachineFunction(), RVLocs, + *DAG.getContext()); + CCInfo.AnalyzeCallResult(Ins, RetCC_Mips, CLI); // Copy all of the result registers out of their specified physreg. for (unsigned i = 0; i != RVLocs.size(); ++i) { + CCValAssign &VA = RVLocs[i]; + assert(VA.isRegLoc() && "Can only return in registers!"); + SDValue Val = DAG.getCopyFromReg(Chain, DL, RVLocs[i].getLocReg(), RVLocs[i].getLocVT(), InFlag); Chain = Val.getValue(1); InFlag = Val.getValue(2); - if (RVLocs[i].getValVT() != RVLocs[i].getLocVT()) - Val = DAG.getNode(ISD::BITCAST, DL, RVLocs[i].getValVT(), Val); + if (VA.isUpperBitsInLoc()) { + unsigned ValSizeInBits = Ins[i].ArgVT.getSizeInBits(); + unsigned LocSizeInBits = VA.getLocVT().getSizeInBits(); + unsigned Shift = + VA.getLocInfo() == CCValAssign::ZExtUpper ? ISD::SRL : ISD::SRA; + Val = DAG.getNode( + Shift, DL, VA.getLocVT(), Val, + DAG.getConstant(LocSizeInBits - ValSizeInBits, VA.getLocVT())); + } + + switch (VA.getLocInfo()) { + default: + llvm_unreachable("Unknown loc info!"); + case CCValAssign::Full: + break; + case CCValAssign::BCvt: + Val = DAG.getNode(ISD::BITCAST, DL, VA.getValVT(), Val); + break; + case CCValAssign::AExt: + case CCValAssign::AExtUpper: + Val = DAG.getNode(ISD::TRUNCATE, DL, VA.getValVT(), Val); + break; + case CCValAssign::ZExt: + case CCValAssign::ZExtUpper: + Val = DAG.getNode(ISD::AssertZext, DL, VA.getLocVT(), Val, + DAG.getValueType(VA.getValVT())); + Val = DAG.getNode(ISD::TRUNCATE, DL, VA.getValVT(), Val); + break; + case CCValAssign::SExt: + case CCValAssign::SExtUpper: + Val = DAG.getNode(ISD::AssertSext, DL, VA.getLocVT(), Val, + DAG.getValueType(VA.getValVT())); + Val = DAG.getNode(ISD::TRUNCATE, DL, VA.getValVT(), Val); + break; + } InVals.push_back(Val); } @@ -2574,14 +2876,16 @@ MipsTargetLowering::LowerFormalArguments(SDValue Chain, // Assign locations to all of the incoming arguments. SmallVector ArgLocs; - CCState CCInfo(CallConv, IsVarArg, DAG.getMachineFunction(), - getTargetMachine(), ArgLocs, *DAG.getContext()); - MipsCC MipsCCInfo(CallConv, IsO32, Subtarget->isFP64bit(), CCInfo); + MipsCCState CCInfo(CallConv, IsVarArg, DAG.getMachineFunction(), ArgLocs, + *DAG.getContext()); + MipsCC MipsCCInfo(CallConv, Subtarget, CCInfo); Function::const_arg_iterator FuncArg = DAG.getMachineFunction().getFunction()->arg_begin(); - bool UseSoftFloat = Subtarget->mipsSEUsesSoftFloat(); + bool UseSoftFloat = Subtarget.abiUsesSoftFloat(); - MipsCCInfo.analyzeFormalArguments(Ins, UseSoftFloat, FuncArg); + CCInfo.PreAnalyzeFormalArgumentsForF128_(Ins); + MipsCCInfo.analyzeFormalArguments(Ins, UseSoftFloat, CCInfo); + CCInfo.ClearOriginalArgWasF128(); MipsFI->setFormalArgInfo(CCInfo.getNextStackOffset(), MipsCCInfo.hasByValArg()); @@ -2608,22 +2912,9 @@ MipsTargetLowering::LowerFormalArguments(SDValue Chain, // Arguments stored on registers if (IsRegLoc) { - EVT RegVT = VA.getLocVT(); + MVT RegVT = VA.getLocVT(); unsigned ArgReg = VA.getLocReg(); - const TargetRegisterClass *RC; - - if (RegVT == MVT::i32) - RC = Subtarget->inMips16Mode()? &Mips::CPU16RegsRegClass : - &Mips::GPR32RegClass; - else if (RegVT == MVT::i64) - RC = &Mips::GPR64RegClass; - else if (RegVT == MVT::f32) - RC = &Mips::FGR32RegClass; - else if (RegVT == MVT::f64) - RC = Subtarget->isFP64bit() ? &Mips::FGR64RegClass : - &Mips::AFGR64RegClass; - else - llvm_unreachable("RegVT not supported by FormalArguments Lowering"); + const TargetRegisterClass *RC = getRegClassFor(RegVT); // Transform the arguments stored on // physical registers into virtual ones @@ -2633,16 +2924,24 @@ MipsTargetLowering::LowerFormalArguments(SDValue Chain, // If this is an 8 or 16-bit value, it has been passed promoted // to 32 bits. Insert an assert[sz]ext to capture this, then // truncate to the right size. - if (VA.getLocInfo() != CCValAssign::Full) { - unsigned Opcode = 0; - if (VA.getLocInfo() == CCValAssign::SExt) - Opcode = ISD::AssertSext; - else if (VA.getLocInfo() == CCValAssign::ZExt) - Opcode = ISD::AssertZext; - if (Opcode) - ArgValue = DAG.getNode(Opcode, DL, RegVT, ArgValue, - DAG.getValueType(ValVT)); + switch (VA.getLocInfo()) { + default: + llvm_unreachable("Unknown loc info!"); + case CCValAssign::Full: + break; + case CCValAssign::SExt: + ArgValue = DAG.getNode(ISD::AssertSext, DL, RegVT, ArgValue, + DAG.getValueType(ValVT)); ArgValue = DAG.getNode(ISD::TRUNCATE, DL, ValVT, ArgValue); + break; + case CCValAssign::ZExt: + ArgValue = DAG.getNode(ISD::AssertZext, DL, RegVT, ArgValue, + DAG.getValueType(ValVT)); + ArgValue = DAG.getNode(ISD::TRUNCATE, DL, ValVT, ArgValue); + break; + case CCValAssign::BCvt: + ArgValue = DAG.getNode(ISD::BITCAST, DL, ValVT, ArgValue); + break; } // Handle floating point arguments passed in integer registers and @@ -2651,11 +2950,12 @@ MipsTargetLowering::LowerFormalArguments(SDValue Chain, (RegVT == MVT::i64 && ValVT == MVT::f64) || (RegVT == MVT::f64 && ValVT == MVT::i64)) ArgValue = DAG.getNode(ISD::BITCAST, DL, ValVT, ArgValue); - else if (IsO32 && RegVT == MVT::i32 && ValVT == MVT::f64) { + else if (Subtarget.isABI_O32() && RegVT == MVT::i32 && + ValVT == MVT::f64) { unsigned Reg2 = addLiveIn(DAG.getMachineFunction(), getNextIntArgReg(ArgReg), RC); SDValue ArgValue2 = DAG.getCopyFromReg(Chain, DL, Reg2, RegVT); - if (!Subtarget->isLittle()) + if (!Subtarget.isLittle()) std::swap(ArgValue, ArgValue2); ArgValue = DAG.getNode(MipsISD::BuildPairF64, DL, MVT::f64, ArgValue, ArgValue2); @@ -2673,35 +2973,39 @@ MipsTargetLowering::LowerFormalArguments(SDValue Chain, // Create load nodes to retrieve arguments from the stack SDValue FIN = DAG.getFrameIndex(FI, getPointerTy()); - InVals.push_back(DAG.getLoad(ValVT, DL, Chain, FIN, - MachinePointerInfo::getFixedStack(FI), - false, false, false, 0)); + SDValue Load = DAG.getLoad(ValVT, DL, Chain, FIN, + MachinePointerInfo::getFixedStack(FI), + false, false, false, 0); + InVals.push_back(Load); + OutChains.push_back(Load.getValue(1)); } } - // The mips ABIs for returning structs by value requires that we copy - // the sret argument into $v0 for the return. Save the argument into - // a virtual register so that we can access it from the return points. - if (DAG.getMachineFunction().getFunction()->hasStructRetAttr()) { - unsigned Reg = MipsFI->getSRetReturnReg(); - if (!Reg) { - Reg = MF.getRegInfo(). - createVirtualRegister(getRegClassFor(IsN64 ? MVT::i64 : MVT::i32)); - MipsFI->setSRetReturnReg(Reg); + for (unsigned i = 0, e = ArgLocs.size(); i != e; ++i) { + // The mips ABIs for returning structs by value requires that we copy + // the sret argument into $v0 for the return. Save the argument into + // a virtual register so that we can access it from the return points. + if (Ins[i].Flags.isSRet()) { + unsigned Reg = MipsFI->getSRetReturnReg(); + if (!Reg) { + Reg = MF.getRegInfo().createVirtualRegister( + getRegClassFor(Subtarget.isABI_N64() ? MVT::i64 : MVT::i32)); + MipsFI->setSRetReturnReg(Reg); + } + SDValue Copy = DAG.getCopyToReg(DAG.getEntryNode(), DL, Reg, InVals[i]); + Chain = DAG.getNode(ISD::TokenFactor, DL, MVT::Other, Copy, Chain); + break; } - SDValue Copy = DAG.getCopyToReg(DAG.getEntryNode(), DL, Reg, InVals[0]); - Chain = DAG.getNode(ISD::TokenFactor, DL, MVT::Other, Copy, Chain); } if (IsVarArg) - writeVarArgRegs(OutChains, MipsCCInfo, Chain, DL, DAG); + writeVarArgRegs(OutChains, MipsCCInfo, Chain, DL, DAG, CCInfo); // All stores are grouped in one node to allow the matching between // the size of Ins and InVals. This only happens when on varg functions if (!OutChains.empty()) { OutChains.push_back(Chain); - Chain = DAG.getNode(ISD::TokenFactor, DL, MVT::Other, - &OutChains[0], OutChains.size()); + Chain = DAG.getNode(ISD::TokenFactor, DL, MVT::Other, OutChains); } return Chain; @@ -2717,8 +3021,7 @@ MipsTargetLowering::CanLowerReturn(CallingConv::ID CallConv, const SmallVectorImpl &Outs, LLVMContext &Context) const { SmallVector RVLocs; - CCState CCInfo(CallConv, IsVarArg, MF, getTargetMachine(), - RVLocs, Context); + MipsCCState CCInfo(CallConv, IsVarArg, MF, RVLocs, Context); return CCInfo.CheckReturn(Outs, RetCC_Mips); } @@ -2734,13 +3037,11 @@ MipsTargetLowering::LowerReturn(SDValue Chain, MachineFunction &MF = DAG.getMachineFunction(); // CCState - Info about the registers and stack slot. - CCState CCInfo(CallConv, IsVarArg, MF, getTargetMachine(), RVLocs, - *DAG.getContext()); - MipsCC MipsCCInfo(CallConv, IsO32, Subtarget->isFP64bit(), CCInfo); + MipsCCState CCInfo(CallConv, IsVarArg, MF, RVLocs, *DAG.getContext()); + MipsCC MipsCCInfo(CallConv, Subtarget, CCInfo); // Analyze return values. - MipsCCInfo.analyzeReturn(Outs, Subtarget->mipsSEUsesSoftFloat(), - MF.getFunction()->getReturnType()); + CCInfo.AnalyzeReturn(Outs, RetCC_Mips); SDValue Flag; SmallVector RetOps(1, Chain); @@ -2750,9 +3051,43 @@ MipsTargetLowering::LowerReturn(SDValue Chain, SDValue Val = OutVals[i]; CCValAssign &VA = RVLocs[i]; assert(VA.isRegLoc() && "Can only return in registers!"); + bool UseUpperBits = false; + + switch (VA.getLocInfo()) { + default: + llvm_unreachable("Unknown loc info!"); + case CCValAssign::Full: + break; + case CCValAssign::BCvt: + Val = DAG.getNode(ISD::BITCAST, DL, VA.getLocVT(), Val); + break; + case CCValAssign::AExtUpper: + UseUpperBits = true; + // Fallthrough + case CCValAssign::AExt: + Val = DAG.getNode(ISD::ANY_EXTEND, DL, VA.getLocVT(), Val); + break; + case CCValAssign::ZExtUpper: + UseUpperBits = true; + // Fallthrough + case CCValAssign::ZExt: + Val = DAG.getNode(ISD::ZERO_EXTEND, DL, VA.getLocVT(), Val); + break; + case CCValAssign::SExtUpper: + UseUpperBits = true; + // Fallthrough + case CCValAssign::SExt: + Val = DAG.getNode(ISD::SIGN_EXTEND, DL, VA.getLocVT(), Val); + break; + } - if (RVLocs[i].getValVT() != RVLocs[i].getLocVT()) - Val = DAG.getNode(ISD::BITCAST, DL, RVLocs[i].getLocVT(), Val); + if (UseUpperBits) { + unsigned ValSizeInBits = Outs[i].ArgVT.getSizeInBits(); + unsigned LocSizeInBits = VA.getLocVT().getSizeInBits(); + Val = DAG.getNode( + ISD::SHL, DL, VA.getLocVT(), Val, + DAG.getConstant(LocSizeInBits - ValSizeInBits, VA.getLocVT())); + } Chain = DAG.getCopyToReg(Chain, DL, VA.getLocReg(), Val, Flag); @@ -2772,7 +3107,7 @@ MipsTargetLowering::LowerReturn(SDValue Chain, if (!Reg) llvm_unreachable("sret virtual register not created in the entry block"); SDValue Val = DAG.getCopyFromReg(Chain, DL, Reg, getPointerTy()); - unsigned V0 = IsN64 ? Mips::V0_64 : Mips::V0; + unsigned V0 = Subtarget.isABI_N64() ? Mips::V0_64 : Mips::V0; Chain = DAG.getCopyToReg(Chain, DL, V0, Val, Flag); Flag = Chain.getValue(1); @@ -2786,7 +3121,7 @@ MipsTargetLowering::LowerReturn(SDValue Chain, RetOps.push_back(Flag); // Return on Mips is always a "jr $ra" - return DAG.getNode(MipsISD::Ret, DL, MVT::Other, &RetOps[0], RetOps.size()); + return DAG.getNode(MipsISD::Ret, DL, MVT::Other, RetOps); } //===----------------------------------------------------------------------===// @@ -2798,7 +3133,7 @@ MipsTargetLowering::LowerReturn(SDValue Chain, MipsTargetLowering::ConstraintType MipsTargetLowering:: getConstraintType(const std::string &Constraint) const { - // Mips specific constrainy + // Mips specific constraints // GCC config/mips/constraints.md // // 'd' : An address register. Equivalent to r @@ -2836,7 +3171,7 @@ MipsTargetLowering::getSingleConstraintMatchWeight( Value *CallOperandVal = info.CallOperandVal; // If we don't have a value, we can't do a match, // but allow it at the lowest weight. - if (CallOperandVal == NULL) + if (!CallOperandVal) return CW_Default; Type *type = CallOperandVal->getType(); // Look at the constraint type. @@ -2849,16 +3184,19 @@ MipsTargetLowering::getSingleConstraintMatchWeight( if (type->isIntegerTy()) weight = CW_Register; break; - case 'f': - if (type->isFloatTy()) + case 'f': // FPU or MSA register + if (Subtarget.hasMSA() && type->isVectorTy() && + cast(type)->getBitWidth() == 128) + weight = CW_Register; + else if (type->isFloatTy()) weight = CW_Register; break; case 'c': // $25 for indirect jumps case 'l': // lo register case 'x': // hilo register pair - if (type->isIntegerTy()) + if (type->isIntegerTy()) weight = CW_SpecificReg; - break; + break; case 'I': // signed 16 bit immediate case 'J': // integer zero case 'K': // unsigned 16 bit immediate @@ -2881,7 +3219,7 @@ MipsTargetLowering::getSingleConstraintMatchWeight( /// that is returned indicates whether parsing was successful. The second flag /// is true if the numeric part exists. static std::pair -parsePhysicalReg(const StringRef &C, std::string &Prefix, +parsePhysicalReg(StringRef C, std::string &Prefix, unsigned long long &Reg) { if (C.front() != '{' || C.back() != '}') return std::make_pair(false, false); @@ -2902,8 +3240,9 @@ parsePhysicalReg(const StringRef &C, std::string &Prefix, } std::pair MipsTargetLowering:: -parseRegForInlineAsmConstraint(const StringRef &C, MVT VT) const { - const TargetRegisterInfo *TRI = getTargetMachine().getRegisterInfo(); +parseRegForInlineAsmConstraint(StringRef C, MVT VT) const { + const TargetRegisterInfo *TRI = + getTargetMachine().getSubtargetImpl()->getRegisterInfo(); const TargetRegisterClass *RC; std::string Prefix; unsigned long long Reg; @@ -2911,35 +3250,60 @@ parseRegForInlineAsmConstraint(const StringRef &C, MVT VT) const { std::pair R = parsePhysicalReg(C, Prefix, Reg); if (!R.first) - return std::make_pair((unsigned)0, (const TargetRegisterClass*)0); + return std::make_pair(0U, nullptr); if ((Prefix == "hi" || Prefix == "lo")) { // Parse hi/lo. // No numeric characters follow "hi" or "lo". if (R.second) - return std::make_pair((unsigned)0, (const TargetRegisterClass*)0); + return std::make_pair(0U, nullptr); RC = TRI->getRegClass(Prefix == "hi" ? Mips::HI32RegClassID : Mips::LO32RegClassID); return std::make_pair(*(RC->begin()), RC); + } else if (Prefix.compare(0, 4, "$msa") == 0) { + // Parse $msa(ir|csr|access|save|modify|request|map|unmap) + + // No numeric characters follow the name. + if (R.second) + return std::make_pair(0U, nullptr); + + Reg = StringSwitch(Prefix) + .Case("$msair", Mips::MSAIR) + .Case("$msacsr", Mips::MSACSR) + .Case("$msaaccess", Mips::MSAAccess) + .Case("$msasave", Mips::MSASave) + .Case("$msamodify", Mips::MSAModify) + .Case("$msarequest", Mips::MSARequest) + .Case("$msamap", Mips::MSAMap) + .Case("$msaunmap", Mips::MSAUnmap) + .Default(0); + + if (!Reg) + return std::make_pair(0U, nullptr); + + RC = TRI->getRegClass(Mips::MSACtrlRegClassID); + return std::make_pair(Reg, RC); } if (!R.second) - return std::make_pair((unsigned)0, (const TargetRegisterClass*)0); + return std::make_pair(0U, nullptr); if (Prefix == "$f") { // Parse $f0-$f31. // If the size of FP registers is 64-bit or Reg is an even number, select // the 64-bit register class. Otherwise, select the 32-bit register class. if (VT == MVT::Other) - VT = (Subtarget->isFP64bit() || !(Reg % 2)) ? MVT::f64 : MVT::f32; + VT = (Subtarget.isFP64bit() || !(Reg % 2)) ? MVT::f64 : MVT::f32; - RC= getRegClassFor(VT); + RC = getRegClassFor(VT); if (RC == &Mips::AFGR64RegClass) { assert(Reg % 2 == 0); Reg >>= 1; } - } else if (Prefix == "$fcc") { // Parse $fcc0-$fcc7. + } else if (Prefix == "$fcc") // Parse $fcc0-$fcc7. RC = TRI->getRegClass(Mips::FCCRegClassID); + else if (Prefix == "$w") { // Parse $w0-$w31. + RC = getRegClassFor((VT == MVT::Other) ? MVT::v16i8 : VT); } else { // Parse $0-$31. assert(Prefix == "$"); RC = getRegClassFor((VT == MVT::Other) ? MVT::i32 : VT); @@ -2961,21 +3325,29 @@ getRegForInlineAsmConstraint(const std::string &Constraint, MVT VT) const case 'y': // Same as 'r'. Exists for compatibility. case 'r': if (VT == MVT::i32 || VT == MVT::i16 || VT == MVT::i8) { - if (Subtarget->inMips16Mode()) + if (Subtarget.inMips16Mode()) return std::make_pair(0U, &Mips::CPU16RegsRegClass); return std::make_pair(0U, &Mips::GPR32RegClass); } - if (VT == MVT::i64 && !HasMips64) + if (VT == MVT::i64 && !Subtarget.isGP64bit()) return std::make_pair(0U, &Mips::GPR32RegClass); - if (VT == MVT::i64 && HasMips64) + if (VT == MVT::i64 && Subtarget.isGP64bit()) return std::make_pair(0U, &Mips::GPR64RegClass); // This will generate an error message - return std::make_pair(0u, static_cast(0)); - case 'f': - if (VT == MVT::f32) + return std::make_pair(0U, nullptr); + case 'f': // FPU or MSA register + if (VT == MVT::v16i8) + return std::make_pair(0U, &Mips::MSA128BRegClass); + else if (VT == MVT::v8i16 || VT == MVT::v8f16) + return std::make_pair(0U, &Mips::MSA128HRegClass); + else if (VT == MVT::v4i32 || VT == MVT::v4f32) + return std::make_pair(0U, &Mips::MSA128WRegClass); + else if (VT == MVT::v2i64 || VT == MVT::v2f64) + return std::make_pair(0U, &Mips::MSA128DRegClass); + else if (VT == MVT::f32) return std::make_pair(0U, &Mips::FGR32RegClass); - if ((VT == MVT::f64) && (!Subtarget->isSingleFloat())) { - if (Subtarget->isFP64bit()) + else if ((VT == MVT::f64) && (!Subtarget.isSingleFloat())) { + if (Subtarget.isFP64bit()) return std::make_pair(0U, &Mips::FGR64RegClass); return std::make_pair(0U, &Mips::AFGR64RegClass); } @@ -2992,7 +3364,7 @@ getRegForInlineAsmConstraint(const std::string &Constraint, MVT VT) const case 'x': // register suitable for indirect jump // Fixme: Not triggering the use of both hi and low // This will generate an error message - return std::make_pair(0u, static_cast(0)); + return std::make_pair(0U, nullptr); } } @@ -3011,7 +3383,7 @@ void MipsTargetLowering::LowerAsmOperandForConstraint(SDValue Op, std::string &Constraint, std::vector&Ops, SelectionDAG &DAG) const { - SDValue Result(0, 0); + SDValue Result; // Only support length 1 constraints for now. if (Constraint.length() > 1) return; @@ -3100,8 +3472,8 @@ void MipsTargetLowering::LowerAsmOperandForConstraint(SDValue Op, TargetLowering::LowerAsmOperandForConstraint(Op, Constraint, Ops, DAG); } -bool -MipsTargetLowering::isLegalAddressingMode(const AddrMode &AM, Type *Ty) const { +bool MipsTargetLowering::isLegalAddressingMode(const AddrMode &AM, + Type *Ty) const { // No global is ever allowed as a base. if (AM.BaseGV) return false; @@ -3131,7 +3503,7 @@ EVT MipsTargetLowering::getOptimalMemOpType(uint64_t Size, unsigned DstAlign, bool IsMemset, bool ZeroMemset, bool MemcpyStrSrc, MachineFunction &MF) const { - if (Subtarget->hasMips64()) + if (Subtarget.hasMips64()) return MVT::i64; return MVT::i32; @@ -3146,7 +3518,7 @@ bool MipsTargetLowering::isFPImmLegal(const APFloat &Imm, EVT VT) const { } unsigned MipsTargetLowering::getJumpTableEncoding() const { - if (IsN64) + if (Subtarget.isABI_N64()) return MachineJumpTableInfo::EK_GPRel64BlockAddress; return TargetLowering::getJumpTableEncoding(); @@ -3165,25 +3537,29 @@ static bool isF128SoftLibCall(const char *CallSym) { "log10l", "log2l", "logl", "nearbyintl", "powl", "rintl", "sinl", "sqrtl", "truncl"}; - const char * const *End = LibCalls + array_lengthof(LibCalls); + const char *const *End = LibCalls + array_lengthof(LibCalls); // Check that LibCalls is sorted alphabetically. MipsTargetLowering::LTStr Comp; #ifndef NDEBUG - for (const char * const *I = LibCalls; I < End - 1; ++I) + for (const char *const *I = LibCalls; I < End - 1; ++I) assert(Comp(*I, *(I + 1))); #endif return std::binary_search(LibCalls, End, CallSym, Comp); } -/// This function returns true if Ty is fp128 or i128 which was originally a -/// fp128. +/// This function returns true if Ty is fp128, {f128} or i128 which was +/// originally a fp128. static bool originalTypeIsF128(const Type *Ty, const SDNode *CallNode) { if (Ty->isFP128Ty()) return true; + if (Ty->isStructTy() && Ty->getStructNumElements() == 1 && + Ty->getStructElementType(0)->isFP128Ty()) + return true; + const ExternalSymbolSDNode *ES = dyn_cast_or_null(CallNode); @@ -3193,14 +3569,15 @@ static bool originalTypeIsF128(const Type *Ty, const SDNode *CallNode) { } MipsTargetLowering::MipsCC::SpecialCallingConvType - MipsTargetLowering::getSpecialCallingConv(SDValue Callee) const { +MipsTargetLowering::MipsCC::getSpecialCallingConv(const SDNode *Callee) const { MipsCC::SpecialCallingConvType SpecialCallingConv = - MipsCC::NoSpecialCallingConv;; - if (Subtarget->inMips16HardFloat()) { - if (GlobalAddressSDNode *G = dyn_cast(Callee)) { + MipsCC::NoSpecialCallingConv; + if (Subtarget.inMips16HardFloat()) { + if (const GlobalAddressSDNode *G = + dyn_cast(Callee)) { llvm::StringRef Sym = G->getGlobal()->getName(); Function *F = G->getGlobal()->getParent()->getFunction(Sym); - if (F->hasFnAttribute("__Mips16RetHelper")) { + if (F && F->hasFnAttribute("__Mips16RetHelper")) { SpecialCallingConv = MipsCC::Mips16RetHelperConv; } } @@ -3208,25 +3585,28 @@ MipsTargetLowering::MipsCC::SpecialCallingConvType return SpecialCallingConv; } -MipsTargetLowering::MipsCC::MipsCC( - CallingConv::ID CC, bool IsO32_, bool IsFP64_, CCState &Info, - MipsCC::SpecialCallingConvType SpecialCallingConv_) - : CCInfo(Info), CallConv(CC), IsO32(IsO32_), IsFP64(IsFP64_), - SpecialCallingConv(SpecialCallingConv_){ +MipsTargetLowering::MipsCC::MipsCC(CallingConv::ID CC, + const MipsSubtarget &Subtarget_, + CCState &Info) + : CallConv(CC), Subtarget(Subtarget_) { // Pre-allocate reserved argument area. - CCInfo.AllocateStack(reservedArgArea(), 1); + Info.AllocateStack(reservedArgArea(), 1); } - -void MipsTargetLowering::MipsCC:: -analyzeCallOperands(const SmallVectorImpl &Args, - bool IsVarArg, bool IsSoftFloat, const SDNode *CallNode, - std::vector &FuncArgs) { +void MipsTargetLowering::MipsCC::analyzeCallOperands( + const SmallVectorImpl &Args, bool IsVarArg, + bool IsSoftFloat, const SDNode *CallNode, + std::vector &FuncArgs, CCState &State) { + MipsCC::SpecialCallingConvType SpecialCallingConv = + getSpecialCallingConv(CallNode); assert((CallConv != CallingConv::Fast || !IsVarArg) && "CallingConv::Fast shouldn't be used for vararg functions."); unsigned NumOpnds = Args.size(); - llvm::CCAssignFn *FixedFn = fixedArgFn(), *VarFn = varArgFn(); + llvm::CCAssignFn *FixedFn = CC_Mips_FixedArg; + if (CallConv != CallingConv::Fast && + SpecialCallingConv == Mips16RetHelperConv) + FixedFn = CC_Mips16RetHelper; for (unsigned I = 0; I != NumOpnds; ++I) { MVT ArgVT = Args[I].VT; @@ -3234,177 +3614,118 @@ analyzeCallOperands(const SmallVectorImpl &Args, bool R; if (ArgFlags.isByVal()) { - handleByValArg(I, ArgVT, ArgVT, CCValAssign::Full, ArgFlags); + handleByValArg(I, ArgVT, ArgVT, CCValAssign::Full, ArgFlags, State); continue; } if (IsVarArg && !Args[I].IsFixed) - R = VarFn(I, ArgVT, ArgVT, CCValAssign::Full, ArgFlags, CCInfo); - else { - MVT RegVT = getRegVT(ArgVT, FuncArgs[Args[I].OrigArgIndex].Ty, CallNode, - IsSoftFloat); - R = FixedFn(I, ArgVT, RegVT, CCValAssign::Full, ArgFlags, CCInfo); - } + R = CC_Mips_VarArg(I, ArgVT, ArgVT, CCValAssign::Full, ArgFlags, State); + else + R = FixedFn(I, ArgVT, ArgVT, CCValAssign::Full, ArgFlags, State); if (R) { #ifndef NDEBUG dbgs() << "Call operand #" << I << " has unhandled type " << EVT(ArgVT).getEVTString(); #endif - llvm_unreachable(0); + llvm_unreachable(nullptr); } } } -void MipsTargetLowering::MipsCC:: -analyzeFormalArguments(const SmallVectorImpl &Args, - bool IsSoftFloat, Function::const_arg_iterator FuncArg) { +void MipsTargetLowering::MipsCC::analyzeFormalArguments( + const SmallVectorImpl &Args, bool IsSoftFloat, + CCState &State) { unsigned NumArgs = Args.size(); - llvm::CCAssignFn *FixedFn = fixedArgFn(); - unsigned CurArgIdx = 0; for (unsigned I = 0; I != NumArgs; ++I) { MVT ArgVT = Args[I].VT; ISD::ArgFlagsTy ArgFlags = Args[I].Flags; - std::advance(FuncArg, Args[I].OrigArgIndex - CurArgIdx); - CurArgIdx = Args[I].OrigArgIndex; if (ArgFlags.isByVal()) { - handleByValArg(I, ArgVT, ArgVT, CCValAssign::Full, ArgFlags); + handleByValArg(I, ArgVT, ArgVT, CCValAssign::Full, ArgFlags, State); continue; } - MVT RegVT = getRegVT(ArgVT, FuncArg->getType(), 0, IsSoftFloat); - - if (!FixedFn(I, ArgVT, RegVT, CCValAssign::Full, ArgFlags, CCInfo)) + if (!CC_Mips_FixedArg(I, ArgVT, ArgVT, CCValAssign::Full, ArgFlags, State)) continue; #ifndef NDEBUG dbgs() << "Formal Arg #" << I << " has unhandled type " << EVT(ArgVT).getEVTString(); #endif - llvm_unreachable(0); + llvm_unreachable(nullptr); } } -template -void MipsTargetLowering::MipsCC:: -analyzeReturn(const SmallVectorImpl &RetVals, bool IsSoftFloat, - const SDNode *CallNode, const Type *RetTy) const { - CCAssignFn *Fn; - - if (IsSoftFloat && originalTypeIsF128(RetTy, CallNode)) - Fn = RetCC_F128Soft; - else - Fn = RetCC_Mips; - - for (unsigned I = 0, E = RetVals.size(); I < E; ++I) { - MVT VT = RetVals[I].VT; - ISD::ArgFlagsTy Flags = RetVals[I].Flags; - MVT RegVT = this->getRegVT(VT, RetTy, CallNode, IsSoftFloat); - - if (Fn(I, VT, RegVT, CCValAssign::Full, Flags, this->CCInfo)) { -#ifndef NDEBUG - dbgs() << "Call result #" << I << " has unhandled type " - << EVT(VT).getEVTString() << '\n'; -#endif - llvm_unreachable(0); - } - } -} - -void MipsTargetLowering::MipsCC:: -analyzeCallResult(const SmallVectorImpl &Ins, bool IsSoftFloat, - const SDNode *CallNode, const Type *RetTy) const { - analyzeReturn(Ins, IsSoftFloat, CallNode, RetTy); -} - -void MipsTargetLowering::MipsCC:: -analyzeReturn(const SmallVectorImpl &Outs, bool IsSoftFloat, - const Type *RetTy) const { - analyzeReturn(Outs, IsSoftFloat, 0, RetTy); -} - -void -MipsTargetLowering::MipsCC::handleByValArg(unsigned ValNo, MVT ValVT, - MVT LocVT, - CCValAssign::LocInfo LocInfo, - ISD::ArgFlagsTy ArgFlags) { +void MipsTargetLowering::MipsCC::handleByValArg(unsigned ValNo, MVT ValVT, + MVT LocVT, + CCValAssign::LocInfo LocInfo, + ISD::ArgFlagsTy ArgFlags, + CCState &State) { assert(ArgFlags.getByValSize() && "Byval argument's size shouldn't be 0."); struct ByValArgInfo ByVal; - unsigned RegSize = regSize(); - unsigned ByValSize = RoundUpToAlignment(ArgFlags.getByValSize(), RegSize); - unsigned Align = std::min(std::max(ArgFlags.getByValAlign(), RegSize), - RegSize * 2); + unsigned RegSizeInBytes = Subtarget.getGPRSizeInBytes(); + unsigned ByValSize = + RoundUpToAlignment(ArgFlags.getByValSize(), RegSizeInBytes); + unsigned Align = std::min(std::max(ArgFlags.getByValAlign(), RegSizeInBytes), + RegSizeInBytes * 2); if (useRegsForByval()) - allocateRegs(ByVal, ByValSize, Align); + allocateRegs(ByVal, ByValSize, Align, State); // Allocate space on caller's stack. - ByVal.Address = CCInfo.AllocateStack(ByValSize - RegSize * ByVal.NumRegs, - Align); - CCInfo.addLoc(CCValAssign::getMem(ValNo, ValVT, ByVal.Address, LocVT, - LocInfo)); + ByVal.Address = + State.AllocateStack(ByValSize - RegSizeInBytes * ByVal.NumRegs, Align); + State.addLoc( + CCValAssign::getMem(ValNo, ValVT, ByVal.Address, LocVT, LocInfo)); ByValArgs.push_back(ByVal); } -unsigned MipsTargetLowering::MipsCC::numIntArgRegs() const { - return IsO32 ? array_lengthof(O32IntRegs) : array_lengthof(Mips64IntRegs); -} - unsigned MipsTargetLowering::MipsCC::reservedArgArea() const { - return (IsO32 && (CallConv != CallingConv::Fast)) ? 16 : 0; + return (Subtarget.isABI_O32() && (CallConv != CallingConv::Fast)) ? 16 : 0; } -const uint16_t *MipsTargetLowering::MipsCC::intArgRegs() const { - return IsO32 ? O32IntRegs : Mips64IntRegs; +const ArrayRef MipsTargetLowering::MipsCC::intArgRegs() const { + if (Subtarget.isABI_O32()) + return makeArrayRef(O32IntRegs); + return makeArrayRef(Mips64IntRegs); } -llvm::CCAssignFn *MipsTargetLowering::MipsCC::fixedArgFn() const { - if (CallConv == CallingConv::Fast) - return CC_Mips_FastCC; - - if (SpecialCallingConv == Mips16RetHelperConv) - return CC_Mips16RetHelper; - return IsO32 ? (IsFP64 ? CC_MipsO32_FP64 : CC_MipsO32_FP32) : CC_MipsN; -} - -llvm::CCAssignFn *MipsTargetLowering::MipsCC::varArgFn() const { - return IsO32 ? (IsFP64 ? CC_MipsO32_FP64 : CC_MipsO32_FP32) : CC_MipsN_VarArg; -} - -const uint16_t *MipsTargetLowering::MipsCC::shadowRegs() const { - return IsO32 ? O32IntRegs : Mips64DPRegs; +const MCPhysReg *MipsTargetLowering::MipsCC::shadowRegs() const { + return Subtarget.isABI_O32() ? O32IntRegs : Mips64DPRegs; } void MipsTargetLowering::MipsCC::allocateRegs(ByValArgInfo &ByVal, unsigned ByValSize, - unsigned Align) { - unsigned RegSize = regSize(), NumIntArgRegs = numIntArgRegs(); - const uint16_t *IntArgRegs = intArgRegs(), *ShadowRegs = shadowRegs(); - assert(!(ByValSize % RegSize) && !(Align % RegSize) && + unsigned Align, CCState &State) { + unsigned RegSizeInBytes = Subtarget.getGPRSizeInBytes(); + const ArrayRef IntArgRegs = intArgRegs(); + const MCPhysReg *ShadowRegs = shadowRegs(); + assert(!(ByValSize % RegSizeInBytes) && !(Align % RegSizeInBytes) && "Byval argument's size and alignment should be a multiple of" - "RegSize."); + "RegSizeInBytes."); - ByVal.FirstIdx = CCInfo.getFirstUnallocated(IntArgRegs, NumIntArgRegs); + ByVal.FirstIdx = + State.getFirstUnallocated(IntArgRegs.data(), IntArgRegs.size()); - // If Align > RegSize, the first arg register must be even. - if ((Align > RegSize) && (ByVal.FirstIdx % 2)) { - CCInfo.AllocateReg(IntArgRegs[ByVal.FirstIdx], ShadowRegs[ByVal.FirstIdx]); + // If Align > RegSizeInBytes, the first arg register must be even. + if ((Align > RegSizeInBytes) && (ByVal.FirstIdx % 2)) { + State.AllocateReg(IntArgRegs[ByVal.FirstIdx], ShadowRegs[ByVal.FirstIdx]); ++ByVal.FirstIdx; } // Mark the registers allocated. - for (unsigned I = ByVal.FirstIdx; ByValSize && (I < NumIntArgRegs); - ByValSize -= RegSize, ++I, ++ByVal.NumRegs) - CCInfo.AllocateReg(IntArgRegs[I], ShadowRegs[I]); + for (unsigned I = ByVal.FirstIdx; ByValSize && (I < IntArgRegs.size()); + ByValSize -= RegSizeInBytes, ++I, ++ByVal.NumRegs) + State.AllocateReg(IntArgRegs[I], ShadowRegs[I]); } MVT MipsTargetLowering::MipsCC::getRegVT(MVT VT, const Type *OrigTy, const SDNode *CallNode, bool IsSoftFloat) const { - if (IsSoftFloat || IsO32) + if (IsSoftFloat || Subtarget.isABI_O32()) return VT; // Check if the original type was fp128. @@ -3423,13 +3744,15 @@ copyByValRegs(SDValue Chain, SDLoc DL, std::vector &OutChains, const MipsCC &CC, const ByValArgInfo &ByVal) const { MachineFunction &MF = DAG.getMachineFunction(); MachineFrameInfo *MFI = MF.getFrameInfo(); - unsigned RegAreaSize = ByVal.NumRegs * CC.regSize(); + unsigned GPRSizeInBytes = Subtarget.getGPRSizeInBytes(); + unsigned RegAreaSize = ByVal.NumRegs * GPRSizeInBytes; unsigned FrameObjSize = std::max(Flags.getByValSize(), RegAreaSize); int FrameObjOffset; if (RegAreaSize) - FrameObjOffset = (int)CC.reservedArgArea() - - (int)((CC.numIntArgRegs() - ByVal.FirstIdx) * CC.regSize()); + FrameObjOffset = + (int)CC.reservedArgArea() - + (int)((CC.intArgRegs().size() - ByVal.FirstIdx) * GPRSizeInBytes); else FrameObjOffset = ByVal.Address; @@ -3443,13 +3766,13 @@ copyByValRegs(SDValue Chain, SDLoc DL, std::vector &OutChains, return; // Copy arg registers. - MVT RegTy = MVT::getIntegerVT(CC.regSize() * 8); + MVT RegTy = MVT::getIntegerVT(GPRSizeInBytes * 8); const TargetRegisterClass *RC = getRegClassFor(RegTy); for (unsigned I = 0; I < ByVal.NumRegs; ++I) { unsigned ArgReg = CC.intArgRegs()[ByVal.FirstIdx + I]; unsigned VReg = addLiveIn(MF, ArgReg, RC); - unsigned Offset = I * CC.regSize(); + unsigned Offset = I * GPRSizeInBytes; SDValue StorePtr = DAG.getNode(ISD::ADD, DL, PtrTy, FIN, DAG.getConstant(Offset, PtrTy)); SDValue Store = DAG.getStore(Chain, DL, DAG.getRegister(VReg, RegTy), @@ -3467,21 +3790,22 @@ passByValArg(SDValue Chain, SDLoc DL, MachineFrameInfo *MFI, SelectionDAG &DAG, SDValue Arg, const MipsCC &CC, const ByValArgInfo &ByVal, const ISD::ArgFlagsTy &Flags, bool isLittle) const { - unsigned ByValSize = Flags.getByValSize(); - unsigned Offset = 0; // Offset in # of bytes from the beginning of struct. - unsigned RegSize = CC.regSize(); - unsigned Alignment = std::min(Flags.getByValAlign(), RegSize); - EVT PtrTy = getPointerTy(), RegTy = MVT::getIntegerVT(RegSize * 8); + unsigned ByValSizeInBytes = Flags.getByValSize(); + unsigned OffsetInBytes = 0; // From beginning of struct + unsigned RegSizeInBytes = Subtarget.getGPRSizeInBytes(); + unsigned Alignment = std::min(Flags.getByValAlign(), RegSizeInBytes); + EVT PtrTy = getPointerTy(), RegTy = MVT::getIntegerVT(RegSizeInBytes * 8); if (ByVal.NumRegs) { - const uint16_t *ArgRegs = CC.intArgRegs(); - bool LeftoverBytes = (ByVal.NumRegs * RegSize > ByValSize); + const ArrayRef ArgRegs = CC.intArgRegs(); + bool LeftoverBytes = (ByVal.NumRegs * RegSizeInBytes > ByValSizeInBytes); unsigned I = 0; // Copy words to registers. - for (; I < ByVal.NumRegs - LeftoverBytes; ++I, Offset += RegSize) { + for (; I < ByVal.NumRegs - LeftoverBytes; + ++I, OffsetInBytes += RegSizeInBytes) { SDValue LoadPtr = DAG.getNode(ISD::ADD, DL, PtrTy, Arg, - DAG.getConstant(Offset, PtrTy)); + DAG.getConstant(OffsetInBytes, PtrTy)); SDValue LoadVal = DAG.getLoad(RegTy, DL, Chain, LoadPtr, MachinePointerInfo(), false, false, false, Alignment); @@ -3491,38 +3815,39 @@ passByValArg(SDValue Chain, SDLoc DL, } // Return if the struct has been fully copied. - if (ByValSize == Offset) + if (ByValSizeInBytes == OffsetInBytes) return; // Copy the remainder of the byval argument with sub-word loads and shifts. if (LeftoverBytes) { - assert((ByValSize > Offset) && (ByValSize < Offset + RegSize) && - "Size of the remainder should be smaller than RegSize."); + assert((ByValSizeInBytes > OffsetInBytes) && + (ByValSizeInBytes < OffsetInBytes + RegSizeInBytes) && + "Size of the remainder should be smaller than RegSizeInBytes."); SDValue Val; - for (unsigned LoadSize = RegSize / 2, TotalSizeLoaded = 0; - Offset < ByValSize; LoadSize /= 2) { - unsigned RemSize = ByValSize - Offset; + for (unsigned LoadSizeInBytes = RegSizeInBytes / 2, TotalBytesLoaded = 0; + OffsetInBytes < ByValSizeInBytes; LoadSizeInBytes /= 2) { + unsigned RemainingSizeInBytes = ByValSizeInBytes - OffsetInBytes; - if (RemSize < LoadSize) + if (RemainingSizeInBytes < LoadSizeInBytes) continue; // Load subword. SDValue LoadPtr = DAG.getNode(ISD::ADD, DL, PtrTy, Arg, - DAG.getConstant(Offset, PtrTy)); - SDValue LoadVal = - DAG.getExtLoad(ISD::ZEXTLOAD, DL, RegTy, Chain, LoadPtr, - MachinePointerInfo(), MVT::getIntegerVT(LoadSize * 8), - false, false, Alignment); + DAG.getConstant(OffsetInBytes, PtrTy)); + SDValue LoadVal = DAG.getExtLoad( + ISD::ZEXTLOAD, DL, RegTy, Chain, LoadPtr, MachinePointerInfo(), + MVT::getIntegerVT(LoadSizeInBytes * 8), false, false, false, + Alignment); MemOpChains.push_back(LoadVal.getValue(1)); // Shift the loaded value. unsigned Shamt; if (isLittle) - Shamt = TotalSizeLoaded; + Shamt = TotalBytesLoaded * 8; else - Shamt = (RegSize - (TotalSizeLoaded + LoadSize)) * 8; + Shamt = (RegSizeInBytes - (TotalBytesLoaded + LoadSizeInBytes)) * 8; SDValue Shift = DAG.getNode(ISD::SHL, DL, RegTy, LoadVal, DAG.getConstant(Shamt, MVT::i32)); @@ -3532,9 +3857,9 @@ passByValArg(SDValue Chain, SDLoc DL, else Val = Shift; - Offset += LoadSize; - TotalSizeLoaded += LoadSize; - Alignment = std::min(Alignment, LoadSize); + OffsetInBytes += LoadSizeInBytes; + TotalBytesLoaded += LoadSizeInBytes; + Alignment = std::min(Alignment, LoadSizeInBytes); } unsigned ArgReg = ArgRegs[ByVal.FirstIdx + I]; @@ -3544,28 +3869,25 @@ passByValArg(SDValue Chain, SDLoc DL, } // Copy remainder of byval arg to it with memcpy. - unsigned MemCpySize = ByValSize - Offset; + unsigned MemCpySize = ByValSizeInBytes - OffsetInBytes; SDValue Src = DAG.getNode(ISD::ADD, DL, PtrTy, Arg, - DAG.getConstant(Offset, PtrTy)); + DAG.getConstant(OffsetInBytes, PtrTy)); SDValue Dst = DAG.getNode(ISD::ADD, DL, PtrTy, StackPtr, DAG.getIntPtrConstant(ByVal.Address)); - Chain = DAG.getMemcpy(Chain, DL, Dst, Src, - DAG.getConstant(MemCpySize, PtrTy), Alignment, - /*isVolatile=*/false, /*AlwaysInline=*/false, - MachinePointerInfo(0), MachinePointerInfo(0)); + Chain = DAG.getMemcpy(Chain, DL, Dst, Src, DAG.getConstant(MemCpySize, PtrTy), + Alignment, /*isVolatile=*/false, /*AlwaysInline=*/false, + MachinePointerInfo(), MachinePointerInfo()); MemOpChains.push_back(Chain); } -void -MipsTargetLowering::writeVarArgRegs(std::vector &OutChains, - const MipsCC &CC, SDValue Chain, - SDLoc DL, SelectionDAG &DAG) const { - unsigned NumRegs = CC.numIntArgRegs(); - const uint16_t *ArgRegs = CC.intArgRegs(); - const CCState &CCInfo = CC.getCCInfo(); - unsigned Idx = CCInfo.getFirstUnallocated(ArgRegs, NumRegs); - unsigned RegSize = CC.regSize(); - MVT RegTy = MVT::getIntegerVT(RegSize * 8); +void MipsTargetLowering::writeVarArgRegs(std::vector &OutChains, + const MipsCC &CC, SDValue Chain, + SDLoc DL, SelectionDAG &DAG, + CCState &State) const { + const ArrayRef ArgRegs = CC.intArgRegs(); + unsigned Idx = State.getFirstUnallocated(ArgRegs.data(), ArgRegs.size()); + unsigned RegSizeInBytes = Subtarget.getGPRSizeInBytes(); + MVT RegTy = MVT::getIntegerVT(RegSizeInBytes * 8); const TargetRegisterClass *RC = getRegClassFor(RegTy); MachineFunction &MF = DAG.getMachineFunction(); MachineFrameInfo *MFI = MF.getFrameInfo(); @@ -3574,29 +3896,32 @@ MipsTargetLowering::writeVarArgRegs(std::vector &OutChains, // Offset of the first variable argument from stack pointer. int VaArgOffset; - if (NumRegs == Idx) - VaArgOffset = RoundUpToAlignment(CCInfo.getNextStackOffset(), RegSize); - else + if (ArgRegs.size() == Idx) VaArgOffset = - (int)CC.reservedArgArea() - (int)(RegSize * (NumRegs - Idx)); + RoundUpToAlignment(State.getNextStackOffset(), RegSizeInBytes); + else + VaArgOffset = (int)CC.reservedArgArea() - + (int)(RegSizeInBytes * (ArgRegs.size() - Idx)); // Record the frame index of the first variable argument // which is a value necessary to VASTART. - int FI = MFI->CreateFixedObject(RegSize, VaArgOffset, true); + int FI = MFI->CreateFixedObject(RegSizeInBytes, VaArgOffset, true); MipsFI->setVarArgsFrameIndex(FI); // Copy the integer registers that have not been used for argument passing // to the argument register save area. For O32, the save area is allocated // in the caller's stack frame, while for N32/64, it is allocated in the // callee's stack frame. - for (unsigned I = Idx; I < NumRegs; ++I, VaArgOffset += RegSize) { + for (unsigned I = Idx; I < ArgRegs.size(); + ++I, VaArgOffset += RegSizeInBytes) { unsigned Reg = addLiveIn(MF, ArgRegs[I], RC); SDValue ArgValue = DAG.getCopyFromReg(Chain, DL, Reg, RegTy); - FI = MFI->CreateFixedObject(RegSize, VaArgOffset, true); + FI = MFI->CreateFixedObject(RegSizeInBytes, VaArgOffset, true); SDValue PtrOff = DAG.getFrameIndex(FI, getPointerTy()); SDValue Store = DAG.getStore(Chain, DL, ArgValue, PtrOff, MachinePointerInfo(), false, false, 0); - cast(Store.getNode())->getMemOperand()->setValue(0); + cast(Store.getNode())->getMemOperand()->setValue( + (Value *)nullptr); OutChains.push_back(Store); } }