diff options
Diffstat (limited to 'lib/Target/PowerPC')
-rw-r--r-- | lib/Target/PowerPC/InstPrinter/PPCInstPrinter.cpp | 11 | ||||
-rw-r--r-- | lib/Target/PowerPC/MCTargetDesc/PPCELFObjectWriter.cpp | 25 | ||||
-rw-r--r-- | lib/Target/PowerPC/MCTargetDesc/PPCMCTargetDesc.cpp | 17 | ||||
-rw-r--r-- | lib/Target/PowerPC/PPC.h | 7 | ||||
-rw-r--r-- | lib/Target/PowerPC/PPCAsmPrinter.cpp | 162 | ||||
-rw-r--r-- | lib/Target/PowerPC/PPCFastISel.cpp | 33 | ||||
-rw-r--r-- | lib/Target/PowerPC/PPCFrameLowering.cpp | 39 | ||||
-rw-r--r-- | lib/Target/PowerPC/PPCISelDAGToDAG.cpp | 25 | ||||
-rw-r--r-- | lib/Target/PowerPC/PPCISelLowering.cpp | 163 | ||||
-rw-r--r-- | lib/Target/PowerPC/PPCISelLowering.h | 18 | ||||
-rw-r--r-- | lib/Target/PowerPC/PPCInstr64Bit.td | 19 | ||||
-rw-r--r-- | lib/Target/PowerPC/PPCInstrInfo.td | 48 | ||||
-rw-r--r-- | lib/Target/PowerPC/PPCMCInstLower.cpp | 6 | ||||
-rw-r--r-- | lib/Target/PowerPC/PPCRegisterInfo.td | 11 |
14 files changed, 385 insertions, 199 deletions
diff --git a/lib/Target/PowerPC/InstPrinter/PPCInstPrinter.cpp b/lib/Target/PowerPC/InstPrinter/PPCInstPrinter.cpp index 7279b091b34c..628819de407e 100644 --- a/lib/Target/PowerPC/InstPrinter/PPCInstPrinter.cpp +++ b/lib/Target/PowerPC/InstPrinter/PPCInstPrinter.cpp @@ -17,6 +17,7 @@ #include "llvm/MC/MCExpr.h" #include "llvm/MC/MCInst.h" #include "llvm/MC/MCInstrInfo.h" +#include "llvm/MC/MCSymbol.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/raw_ostream.h" #include "llvm/Target/TargetOpcodes.h" @@ -260,7 +261,7 @@ void PPCInstPrinter::printAbsBranchOperand(const MCInst *MI, unsigned OpNo, if (!MI->getOperand(OpNo).isImm()) return printOperand(MI, OpNo, O); - O << (int)MI->getOperand(OpNo).getImm()*4; + O << SignExtend32<32>((unsigned)MI->getOperand(OpNo).getImm() << 2); } @@ -308,10 +309,16 @@ void PPCInstPrinter::printMemRegReg(const MCInst *MI, unsigned OpNo, void PPCInstPrinter::printTLSCall(const MCInst *MI, unsigned OpNo, raw_ostream &O) { - printBranchOperand(MI, OpNo, O); + // On PPC64, VariantKind is VK_None, but on PPC32, it's VK_PLT, and it must + // come at the _end_ of the expression. + const MCOperand &Op = MI->getOperand(OpNo); + const MCSymbolRefExpr &refExp = cast<MCSymbolRefExpr>(*Op.getExpr()); + O << refExp.getSymbol().getName(); O << '('; printOperand(MI, OpNo+1, O); O << ')'; + if (refExp.getKind() != MCSymbolRefExpr::VK_None) + O << '@' << MCSymbolRefExpr::getVariantKindName(refExp.getKind()); } diff --git a/lib/Target/PowerPC/MCTargetDesc/PPCELFObjectWriter.cpp b/lib/Target/PowerPC/MCTargetDesc/PPCELFObjectWriter.cpp index e93e95fc0751..ca813176bd5a 100644 --- a/lib/Target/PowerPC/MCTargetDesc/PPCELFObjectWriter.cpp +++ b/lib/Target/PowerPC/MCTargetDesc/PPCELFObjectWriter.cpp @@ -236,7 +236,10 @@ unsigned PPCELFObjectWriter::getRelocTypeInner(const MCValue &Target, Type = ELF::R_PPC64_DTPREL16_HIGHESTA; break; case MCSymbolRefExpr::VK_PPC_GOT_TLSGD: - Type = ELF::R_PPC64_GOT_TLSGD16; + if (is64Bit()) + Type = ELF::R_PPC64_GOT_TLSGD16; + else + Type = ELF::R_PPC_GOT_TLSGD16; break; case MCSymbolRefExpr::VK_PPC_GOT_TLSGD_LO: Type = ELF::R_PPC64_GOT_TLSGD16_LO; @@ -248,7 +251,10 @@ unsigned PPCELFObjectWriter::getRelocTypeInner(const MCValue &Target, Type = ELF::R_PPC64_GOT_TLSGD16_HA; break; case MCSymbolRefExpr::VK_PPC_GOT_TLSLD: - Type = ELF::R_PPC64_GOT_TLSLD16; + if (is64Bit()) + Type = ELF::R_PPC64_GOT_TLSLD16; + else + Type = ELF::R_PPC_GOT_TLSLD16; break; case MCSymbolRefExpr::VK_PPC_GOT_TLSLD_LO: Type = ELF::R_PPC64_GOT_TLSLD16_LO; @@ -344,13 +350,22 @@ unsigned PPCELFObjectWriter::getRelocTypeInner(const MCValue &Target, switch (Modifier) { default: llvm_unreachable("Unsupported Modifier"); case MCSymbolRefExpr::VK_PPC_TLSGD: - Type = ELF::R_PPC64_TLSGD; + if (is64Bit()) + Type = ELF::R_PPC64_TLSGD; + else + Type = ELF::R_PPC_TLSGD; break; case MCSymbolRefExpr::VK_PPC_TLSLD: - Type = ELF::R_PPC64_TLSLD; + if (is64Bit()) + Type = ELF::R_PPC64_TLSLD; + else + Type = ELF::R_PPC_TLSLD; break; case MCSymbolRefExpr::VK_PPC_TLS: - Type = ELF::R_PPC64_TLS; + if (is64Bit()) + Type = ELF::R_PPC64_TLS; + else + Type = ELF::R_PPC_TLS; break; } break; diff --git a/lib/Target/PowerPC/MCTargetDesc/PPCMCTargetDesc.cpp b/lib/Target/PowerPC/MCTargetDesc/PPCMCTargetDesc.cpp index 4c6780ff75a7..1bb5173e1937 100644 --- a/lib/Target/PowerPC/MCTargetDesc/PPCMCTargetDesc.cpp +++ b/lib/Target/PowerPC/MCTargetDesc/PPCMCTargetDesc.cpp @@ -184,6 +184,23 @@ public: if ((Flags & ELF::EF_PPC64_ABI) == 0) MCA.setELFHeaderEFlags(Flags | 2); } + void emitAssignment(MCSymbol *Symbol, const MCExpr *Value) override { + // When encoding an assignment to set symbol A to symbol B, also copy + // the st_other bits encoding the local entry point offset. + if (Value->getKind() != MCExpr::SymbolRef) + return; + const MCSymbol &RhsSym = + static_cast<const MCSymbolRefExpr *>(Value)->getSymbol(); + MCSymbolData &Data = getStreamer().getOrCreateSymbolData(&RhsSym); + MCSymbolData &SymbolData = getStreamer().getOrCreateSymbolData(Symbol); + // The "other" values are stored in the last 6 bits of the second byte. + // The traditional defines for STO values assume the full byte and thus + // the shift to pack it. + unsigned Other = MCELF::getOther(SymbolData) << 2; + Other &= ~ELF::STO_PPC64_LOCAL_MASK; + Other |= (MCELF::getOther(Data) << 2) & ELF::STO_PPC64_LOCAL_MASK; + MCELF::setOther(SymbolData, Other >> 2); + } }; class PPCTargetMachOStreamer : public PPCTargetStreamer { diff --git a/lib/Target/PowerPC/PPC.h b/lib/Target/PowerPC/PPC.h index ba5fa4f79b4e..38840e624a53 100644 --- a/lib/Target/PowerPC/PPC.h +++ b/lib/Target/PowerPC/PPC.h @@ -96,7 +96,12 @@ namespace llvm { MO_TOC_LO = 7 << 4, // Symbol for VK_PPC_TLS fixup attached to an ADD instruction - MO_TLS = 8 << 4 + MO_TLS = 8 << 4, + + // Symbols for VK_PPC_TLSGD and VK_PPC_TLSLD in __tls_get_addr + // call sequences. + MO_TLSLD = 9 << 4, + MO_TLSGD = 10 << 4 }; } // end namespace PPCII diff --git a/lib/Target/PowerPC/PPCAsmPrinter.cpp b/lib/Target/PowerPC/PPCAsmPrinter.cpp index 6f67c598c754..8357665d30db 100644 --- a/lib/Target/PowerPC/PPCAsmPrinter.cpp +++ b/lib/Target/PowerPC/PPCAsmPrinter.cpp @@ -373,7 +373,7 @@ void PPCAsmPrinter::EmitInstruction(const MachineInstr *MI) { const MachineOperand &MO = MI->getOperand(1); // Map symbol -> label of TOC entry - assert(MO.isGlobal() || MO.isCPI() || MO.isJTI()); + assert(MO.isGlobal() || MO.isCPI() || MO.isJTI() || MO.isBlockAddress()); MCSymbol *MOSymbol = nullptr; if (MO.isGlobal()) MOSymbol = getSymbol(MO.getGlobal()); @@ -381,6 +381,8 @@ void PPCAsmPrinter::EmitInstruction(const MachineInstr *MI) { MOSymbol = GetCPISymbol(MO.getIndex()); else if (MO.isJTI()) MOSymbol = GetJTISymbol(MO.getIndex()); + else if (MO.isBlockAddress()) + MOSymbol = GetBlockAddressSymbol(MO.getBlockAddress()); MCSymbol *TOCEntry = lookUpOrCreateTOCEntry(MOSymbol); @@ -397,6 +399,7 @@ void PPCAsmPrinter::EmitInstruction(const MachineInstr *MI) { } case PPC::LDtocJTI: case PPC::LDtocCPT: + case PPC::LDtocBA: case PPC::LDtoc: { // Transform %X3 = LDtoc <ga:@min1>, %X2 LowerPPCMachineInstrToMCInst(MI, TmpInst, *this, Subtarget.isDarwin()); @@ -407,7 +410,7 @@ void PPCAsmPrinter::EmitInstruction(const MachineInstr *MI) { const MachineOperand &MO = MI->getOperand(1); // Map symbol -> label of TOC entry - assert(MO.isGlobal() || MO.isCPI() || MO.isJTI()); + assert(MO.isGlobal() || MO.isCPI() || MO.isJTI() || MO.isBlockAddress()); MCSymbol *MOSymbol = nullptr; if (MO.isGlobal()) MOSymbol = getSymbol(MO.getGlobal()); @@ -415,6 +418,8 @@ void PPCAsmPrinter::EmitInstruction(const MachineInstr *MI) { MOSymbol = GetCPISymbol(MO.getIndex()); else if (MO.isJTI()) MOSymbol = GetJTISymbol(MO.getIndex()); + else if (MO.isBlockAddress()) + MOSymbol = GetBlockAddressSymbol(MO.getBlockAddress()); MCSymbol *TOCEntry = lookUpOrCreateTOCEntry(MOSymbol); @@ -436,7 +441,8 @@ void PPCAsmPrinter::EmitInstruction(const MachineInstr *MI) { // reference the symbol directly. TmpInst.setOpcode(PPC::ADDIS8); const MachineOperand &MO = MI->getOperand(2); - assert((MO.isGlobal() || MO.isCPI() || MO.isJTI()) && + assert((MO.isGlobal() || MO.isCPI() || MO.isJTI() || + MO.isBlockAddress()) && "Invalid operand for ADDIStocHA!"); MCSymbol *MOSymbol = nullptr; bool IsExternal = false; @@ -456,9 +462,12 @@ void PPCAsmPrinter::EmitInstruction(const MachineInstr *MI) { MOSymbol = GetCPISymbol(MO.getIndex()); else if (MO.isJTI()) MOSymbol = GetJTISymbol(MO.getIndex()); + else if (MO.isBlockAddress()) + MOSymbol = GetBlockAddressSymbol(MO.getBlockAddress()); if (IsExternal || IsNonLocalFunction || IsCommon || IsAvailExt || - MO.isJTI() || TM.getCodeModel() == CodeModel::Large) + MO.isJTI() || MO.isBlockAddress() || + TM.getCodeModel() == CodeModel::Large) MOSymbol = lookUpOrCreateTOCEntry(MOSymbol); const MCExpr *Exp = @@ -477,12 +486,17 @@ void PPCAsmPrinter::EmitInstruction(const MachineInstr *MI) { // associated TOC entry. Otherwise reference the symbol directly. TmpInst.setOpcode(PPC::LD); const MachineOperand &MO = MI->getOperand(1); - assert((MO.isGlobal() || MO.isJTI() || MO.isCPI()) && + assert((MO.isGlobal() || MO.isCPI() || MO.isJTI() || + MO.isBlockAddress()) && "Invalid operand for LDtocL!"); MCSymbol *MOSymbol = nullptr; if (MO.isJTI()) MOSymbol = lookUpOrCreateTOCEntry(GetJTISymbol(MO.getIndex())); + else if (MO.isBlockAddress()) { + MOSymbol = GetBlockAddressSymbol(MO.getBlockAddress()); + MOSymbol = lookUpOrCreateTOCEntry(MOSymbol); + } else if (MO.isCPI()) { MOSymbol = GetCPISymbol(MO.getIndex()); if (TM.getCodeModel() == CodeModel::Large) @@ -573,6 +587,34 @@ void PPCAsmPrinter::EmitInstruction(const MachineInstr *MI) { return; } + case PPC::PPC32PICGOT: { + MCSymbol *GOTSymbol = OutContext.GetOrCreateSymbol(StringRef("_GLOBAL_OFFSET_TABLE_")); + MCSymbol *GOTRef = OutContext.CreateTempSymbol(); + MCSymbol *NextInstr = OutContext.CreateTempSymbol(); + + EmitToStreamer(OutStreamer, MCInstBuilder(PPC::BL) + // FIXME: We would like an efficient form for this, so we don't have to do + // a lot of extra uniquing. + .addExpr(MCSymbolRefExpr::Create(NextInstr, OutContext))); + const MCExpr *OffsExpr = + MCBinaryExpr::CreateSub(MCSymbolRefExpr::Create(GOTSymbol, OutContext), + MCSymbolRefExpr::Create(GOTRef, OutContext), + OutContext); + OutStreamer.EmitLabel(GOTRef); + OutStreamer.EmitValue(OffsExpr, 4); + OutStreamer.EmitLabel(NextInstr); + EmitToStreamer(OutStreamer, MCInstBuilder(PPC::MFLR) + .addReg(MI->getOperand(0).getReg())); + EmitToStreamer(OutStreamer, MCInstBuilder(PPC::LWZ) + .addReg(MI->getOperand(1).getReg()) + .addImm(0) + .addReg(MI->getOperand(0).getReg())); + EmitToStreamer(OutStreamer, MCInstBuilder(PPC::ADD4) + .addReg(MI->getOperand(0).getReg()) + .addReg(MI->getOperand(1).getReg()) + .addReg(MI->getOperand(0).getReg())); + return; + } case PPC::PPC32GOT: { MCSymbol *GOTSymbol = OutContext.GetOrCreateSymbol(StringRef("_GLOBAL_OFFSET_TABLE_")); const MCExpr *SymGotTlsL = @@ -606,40 +648,25 @@ void PPCAsmPrinter::EmitInstruction(const MachineInstr *MI) { .addExpr(SymGotTlsGD)); return; } - case PPC::ADDItlsgdL: { + case PPC::ADDItlsgdL: // Transform: %Xd = ADDItlsgdL %Xs, <ga:@sym> // Into: %Xd = ADDI8 %Xs, sym@got@tlsgd@l - assert(Subtarget.isPPC64() && "Not supported for 32-bit PowerPC"); + case PPC::ADDItlsgdL32: { + // Transform: %Rd = ADDItlsgdL32 %Rs, <ga:@sym> + // Into: %Rd = ADDI %Rs, sym@got@tlsgd const MachineOperand &MO = MI->getOperand(2); const GlobalValue *GValue = MO.getGlobal(); MCSymbol *MOSymbol = getSymbol(GValue); const MCExpr *SymGotTlsGD = - MCSymbolRefExpr::Create(MOSymbol, MCSymbolRefExpr::VK_PPC_GOT_TLSGD_LO, + MCSymbolRefExpr::Create(MOSymbol, Subtarget.isPPC64() ? + MCSymbolRefExpr::VK_PPC_GOT_TLSGD_LO : + MCSymbolRefExpr::VK_PPC_GOT_TLSGD, OutContext); - EmitToStreamer(OutStreamer, MCInstBuilder(PPC::ADDI8) - .addReg(MI->getOperand(0).getReg()) - .addReg(MI->getOperand(1).getReg()) - .addExpr(SymGotTlsGD)); - return; - } - case PPC::GETtlsADDR: { - // Transform: %X3 = GETtlsADDR %X3, <ga:@sym> - // Into: BL8_NOP_TLS __tls_get_addr(sym@tlsgd) - assert(Subtarget.isPPC64() && "Not supported for 32-bit PowerPC"); - - StringRef Name = "__tls_get_addr"; - MCSymbol *TlsGetAddr = OutContext.GetOrCreateSymbol(Name); - const MCSymbolRefExpr *TlsRef = - MCSymbolRefExpr::Create(TlsGetAddr, MCSymbolRefExpr::VK_None, OutContext); - const MachineOperand &MO = MI->getOperand(2); - const GlobalValue *GValue = MO.getGlobal(); - MCSymbol *MOSymbol = getSymbol(GValue); - const MCExpr *SymVar = - MCSymbolRefExpr::Create(MOSymbol, MCSymbolRefExpr::VK_PPC_TLSGD, - OutContext); - EmitToStreamer(OutStreamer, MCInstBuilder(PPC::BL8_NOP_TLS) - .addExpr(TlsRef) - .addExpr(SymVar)); + EmitToStreamer(OutStreamer, + MCInstBuilder(Subtarget.isPPC64() ? PPC::ADDI8 : PPC::ADDI) + .addReg(MI->getOperand(0).getReg()) + .addReg(MI->getOperand(1).getReg()) + .addExpr(SymGotTlsGD)); return; } case PPC::ADDIStlsldHA: { @@ -658,72 +685,63 @@ void PPCAsmPrinter::EmitInstruction(const MachineInstr *MI) { .addExpr(SymGotTlsLD)); return; } - case PPC::ADDItlsldL: { + case PPC::ADDItlsldL: // Transform: %Xd = ADDItlsldL %Xs, <ga:@sym> // Into: %Xd = ADDI8 %Xs, sym@got@tlsld@l - assert(Subtarget.isPPC64() && "Not supported for 32-bit PowerPC"); + case PPC::ADDItlsldL32: { + // Transform: %Rd = ADDItlsldL32 %Rs, <ga:@sym> + // Into: %Rd = ADDI %Rs, sym@got@tlsld const MachineOperand &MO = MI->getOperand(2); const GlobalValue *GValue = MO.getGlobal(); MCSymbol *MOSymbol = getSymbol(GValue); const MCExpr *SymGotTlsLD = - MCSymbolRefExpr::Create(MOSymbol, MCSymbolRefExpr::VK_PPC_GOT_TLSLD_LO, + MCSymbolRefExpr::Create(MOSymbol, Subtarget.isPPC64() ? + MCSymbolRefExpr::VK_PPC_GOT_TLSLD_LO : + MCSymbolRefExpr::VK_PPC_GOT_TLSLD, OutContext); - EmitToStreamer(OutStreamer, MCInstBuilder(PPC::ADDI8) - .addReg(MI->getOperand(0).getReg()) - .addReg(MI->getOperand(1).getReg()) - .addExpr(SymGotTlsLD)); + EmitToStreamer(OutStreamer, + MCInstBuilder(Subtarget.isPPC64() ? PPC::ADDI8 : PPC::ADDI) + .addReg(MI->getOperand(0).getReg()) + .addReg(MI->getOperand(1).getReg()) + .addExpr(SymGotTlsLD)); return; } - case PPC::GETtlsldADDR: { - // Transform: %X3 = GETtlsldADDR %X3, <ga:@sym> - // Into: BL8_NOP_TLS __tls_get_addr(sym@tlsld) - assert(Subtarget.isPPC64() && "Not supported for 32-bit PowerPC"); - - StringRef Name = "__tls_get_addr"; - MCSymbol *TlsGetAddr = OutContext.GetOrCreateSymbol(Name); - const MCSymbolRefExpr *TlsRef = - MCSymbolRefExpr::Create(TlsGetAddr, MCSymbolRefExpr::VK_None, OutContext); - const MachineOperand &MO = MI->getOperand(2); - const GlobalValue *GValue = MO.getGlobal(); - MCSymbol *MOSymbol = getSymbol(GValue); - const MCExpr *SymVar = - MCSymbolRefExpr::Create(MOSymbol, MCSymbolRefExpr::VK_PPC_TLSLD, - OutContext); - EmitToStreamer(OutStreamer, MCInstBuilder(PPC::BL8_NOP_TLS) - .addExpr(TlsRef) - .addExpr(SymVar)); - return; - } - case PPC::ADDISdtprelHA: { + case PPC::ADDISdtprelHA: // Transform: %Xd = ADDISdtprelHA %X3, <ga:@sym> // Into: %Xd = ADDIS8 %X3, sym@dtprel@ha - assert(Subtarget.isPPC64() && "Not supported for 32-bit PowerPC"); + case PPC::ADDISdtprelHA32: { + // Transform: %Rd = ADDISdtprelHA32 %R3, <ga:@sym> + // Into: %Rd = ADDIS %R3, sym@dtprel@ha const MachineOperand &MO = MI->getOperand(2); const GlobalValue *GValue = MO.getGlobal(); MCSymbol *MOSymbol = getSymbol(GValue); const MCExpr *SymDtprel = MCSymbolRefExpr::Create(MOSymbol, MCSymbolRefExpr::VK_PPC_DTPREL_HA, OutContext); - EmitToStreamer(OutStreamer, MCInstBuilder(PPC::ADDIS8) - .addReg(MI->getOperand(0).getReg()) - .addReg(PPC::X3) - .addExpr(SymDtprel)); + EmitToStreamer(OutStreamer, + MCInstBuilder(Subtarget.isPPC64() ? PPC::ADDIS8 : PPC::ADDIS) + .addReg(MI->getOperand(0).getReg()) + .addReg(Subtarget.isPPC64() ? PPC::X3 : PPC::R3) + .addExpr(SymDtprel)); return; } - case PPC::ADDIdtprelL: { + case PPC::ADDIdtprelL: // Transform: %Xd = ADDIdtprelL %Xs, <ga:@sym> // Into: %Xd = ADDI8 %Xs, sym@dtprel@l - assert(Subtarget.isPPC64() && "Not supported for 32-bit PowerPC"); + case PPC::ADDIdtprelL32: { + // Transform: %Rd = ADDIdtprelL32 %Rs, <ga:@sym> + // Into: %Rd = ADDI %Rs, sym@dtprel@l const MachineOperand &MO = MI->getOperand(2); const GlobalValue *GValue = MO.getGlobal(); MCSymbol *MOSymbol = getSymbol(GValue); const MCExpr *SymDtprel = MCSymbolRefExpr::Create(MOSymbol, MCSymbolRefExpr::VK_PPC_DTPREL_LO, OutContext); - EmitToStreamer(OutStreamer, MCInstBuilder(PPC::ADDI8) - .addReg(MI->getOperand(0).getReg()) - .addReg(MI->getOperand(1).getReg()) - .addExpr(SymDtprel)); + EmitToStreamer(OutStreamer, + MCInstBuilder(Subtarget.isPPC64() ? PPC::ADDI8 : PPC::ADDI) + .addReg(MI->getOperand(0).getReg()) + .addReg(MI->getOperand(1).getReg()) + .addExpr(SymDtprel)); return; } case PPC::MFOCRF: @@ -903,7 +921,7 @@ bool PPCLinuxAsmPrinter::doFinalization(Module &M) { for (MapVector<MCSymbol*, MCSymbol*>::iterator I = TOC.begin(), E = TOC.end(); I != E; ++I) { OutStreamer.EmitLabel(I->second); - MCSymbol *S = OutContext.GetOrCreateSymbol(I->first->getName()); + MCSymbol *S = I->first; if (isPPC64) TS.emitTCEntry(*S); else diff --git a/lib/Target/PowerPC/PPCFastISel.cpp b/lib/Target/PowerPC/PPCFastISel.cpp index 2e524d604789..9c55a2900389 100644 --- a/lib/Target/PowerPC/PPCFastISel.cpp +++ b/lib/Target/PowerPC/PPCFastISel.cpp @@ -153,7 +153,7 @@ class PPCFastISel final : public FastISel { unsigned DestReg, bool IsZExt); unsigned PPCMaterializeFP(const ConstantFP *CFP, MVT VT); unsigned PPCMaterializeGV(const GlobalValue *GV, MVT VT); - unsigned PPCMaterializeInt(const Constant *C, MVT VT); + unsigned PPCMaterializeInt(const Constant *C, MVT VT, bool UseSExt = true); unsigned PPCMaterialize32BitInt(int64_t Imm, const TargetRegisterClass *RC); unsigned PPCMaterialize64BitInt(int64_t Imm, @@ -865,7 +865,7 @@ bool PPCFastISel::SelectFPTrunc(const Instruction *I) { } // Move an i32 or i64 value in a GPR to an f64 value in an FPR. -// FIXME: When direct register moves are implemented (see PowerISA 2.08), +// FIXME: When direct register moves are implemented (see PowerISA 2.07), // those should be used instead of moving via a stack slot when the // subtarget permits. // FIXME: The code here is sloppy for the 4-byte case. Can use a 4-byte @@ -898,10 +898,10 @@ unsigned PPCFastISel::PPCMoveToFPReg(MVT SrcVT, unsigned SrcReg, if (SrcVT == MVT::i32) { if (!IsSigned) { LoadOpc = PPC::LFIWZX; - Addr.Offset = 4; + Addr.Offset = (PPCSubTarget->isLittleEndian()) ? 0 : 4; } else if (PPCSubTarget->hasLFIWAX()) { LoadOpc = PPC::LFIWAX; - Addr.Offset = 4; + Addr.Offset = (PPCSubTarget->isLittleEndian()) ? 0 : 4; } } @@ -985,7 +985,7 @@ bool PPCFastISel::SelectIToFP(const Instruction *I, bool IsSigned) { // Move the floating-point value in SrcReg into an integer destination // register, and return the register (or zero if we can't handle it). -// FIXME: When direct register moves are implemented (see PowerISA 2.08), +// FIXME: When direct register moves are implemented (see PowerISA 2.07), // those should be used instead of moving via a stack slot when the // subtarget permits. unsigned PPCFastISel::PPCMoveToIntReg(const Instruction *I, MVT VT, @@ -1548,13 +1548,23 @@ bool PPCFastISel::SelectRet(const Instruction *I) { // Special case for returning a constant integer of any size. // Materialize the constant as an i64 and copy it to the return - // register. This avoids an unnecessary extend or truncate. + // register. We still need to worry about properly extending the sign. E.g: + // If the constant has only one bit, it means it is a boolean. Therefore + // we can't use PPCMaterializeInt because it extends the sign which will + // cause negations of the returned value to be incorrect as they are + // implemented as the flip of the least significant bit. if (isa<ConstantInt>(*RV)) { const Constant *C = cast<Constant>(RV); - unsigned SrcReg = PPCMaterializeInt(C, MVT::i64); - unsigned RetReg = ValLocs[0].getLocReg(); + + CCValAssign &VA = ValLocs[0]; + + unsigned RetReg = VA.getLocReg(); + unsigned SrcReg = PPCMaterializeInt(C, MVT::i64, + VA.getLocInfo() == CCValAssign::SExt); + BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, - TII.get(TargetOpcode::COPY), RetReg).addReg(SrcReg); + TII.get(TargetOpcode::COPY), RetReg).addReg(SrcReg); + RetRegs.push_back(RetReg); } else { @@ -2014,7 +2024,8 @@ unsigned PPCFastISel::PPCMaterialize64BitInt(int64_t Imm, // Materialize an integer constant into a register, and return // the register number (or zero if we failed to handle it). -unsigned PPCFastISel::PPCMaterializeInt(const Constant *C, MVT VT) { +unsigned PPCFastISel::PPCMaterializeInt(const Constant *C, MVT VT, + bool UseSExt) { // If we're using CR bit registers for i1 values, handle that as a special // case first. if (VT == MVT::i1 && PPCSubTarget->useCRBits()) { @@ -2038,7 +2049,7 @@ unsigned PPCFastISel::PPCMaterializeInt(const Constant *C, MVT VT) { unsigned Opc = (VT == MVT::i64) ? PPC::LI8 : PPC::LI; unsigned ImmReg = createResultReg(RC); BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(Opc), ImmReg) - .addImm(CI->getSExtValue()); + .addImm( (UseSExt) ? CI->getSExtValue() : CI->getZExtValue() ); return ImmReg; } diff --git a/lib/Target/PowerPC/PPCFrameLowering.cpp b/lib/Target/PowerPC/PPCFrameLowering.cpp index b2577a9c7cf7..2a02dfca6ae6 100644 --- a/lib/Target/PowerPC/PPCFrameLowering.cpp +++ b/lib/Target/PowerPC/PPCFrameLowering.cpp @@ -505,7 +505,7 @@ void PPCFrameLowering::emitPrologue(MachineFunction &MF) const { MachineModuleInfo &MMI = MF.getMMI(); const MCRegisterInfo *MRI = MMI.getContext().getRegisterInfo(); DebugLoc dl; - bool needsFrameMoves = MMI.hasDebugInfo() || + bool needsCFI = MMI.hasDebugInfo() || MF.getFunction()->needsUnwindTableEntry(); bool isPIC = MF.getTarget().getRelocationModel() == Reloc::PIC_; @@ -726,17 +726,28 @@ void PPCFrameLowering::emitPrologue(MachineFunction &MF) const { .addReg(ScratchReg); } - // Add the "machine moves" for the instructions we generated above, but in - // reverse order. - if (needsFrameMoves) { - // Show update of SP. - assert(NegFrameSize); - unsigned CFIIndex = MMI.addFrameInst( - MCCFIInstruction::createDefCfaOffset(nullptr, NegFrameSize)); + // Add Call Frame Information for the instructions we generated above. + if (needsCFI) { + unsigned CFIIndex; + + if (HasBP) { + // Define CFA in terms of BP. Do this in preference to using FP/SP, + // because if the stack needed aligning then CFA won't be at a fixed + // offset from FP/SP. + unsigned Reg = MRI->getDwarfRegNum(BPReg, true); + CFIIndex = MMI.addFrameInst( + MCCFIInstruction::createDefCfaRegister(nullptr, Reg)); + } else { + // Adjust the definition of CFA to account for the change in SP. + assert(NegFrameSize); + CFIIndex = MMI.addFrameInst( + MCCFIInstruction::createDefCfaOffset(nullptr, NegFrameSize)); + } BuildMI(MBB, MBBI, dl, TII.get(TargetOpcode::CFI_INSTRUCTION)) .addCFIIndex(CFIIndex); if (HasFP) { + // Describe where FP was saved, at a fixed offset from CFA. unsigned Reg = MRI->getDwarfRegNum(FPReg, true); CFIIndex = MMI.addFrameInst( MCCFIInstruction::createOffset(nullptr, Reg, FPOffset)); @@ -745,6 +756,7 @@ void PPCFrameLowering::emitPrologue(MachineFunction &MF) const { } if (HasBP) { + // Describe where BP was saved, at a fixed offset from CFA. unsigned Reg = MRI->getDwarfRegNum(BPReg, true); CFIIndex = MMI.addFrameInst( MCCFIInstruction::createOffset(nullptr, Reg, BPOffset)); @@ -753,6 +765,7 @@ void PPCFrameLowering::emitPrologue(MachineFunction &MF) const { } if (MustSaveLR) { + // Describe where LR was saved, at a fixed offset from CFA. unsigned Reg = MRI->getDwarfRegNum(LRReg, true); CFIIndex = MMI.addFrameInst( MCCFIInstruction::createOffset(nullptr, Reg, LROffset)); @@ -767,8 +780,9 @@ void PPCFrameLowering::emitPrologue(MachineFunction &MF) const { .addReg(SPReg) .addReg(SPReg); - if (needsFrameMoves) { - // Mark effective beginning of when frame pointer is ready. + if (!HasBP && needsCFI) { + // Change the definition of CFA from SP+offset to FP+offset, because SP + // will change at every alloca. unsigned Reg = MRI->getDwarfRegNum(FPReg, true); unsigned CFIIndex = MMI.addFrameInst( MCCFIInstruction::createDefCfaRegister(nullptr, Reg)); @@ -778,8 +792,9 @@ void PPCFrameLowering::emitPrologue(MachineFunction &MF) const { } } - if (needsFrameMoves) { - // Add callee saved registers to move list. + if (needsCFI) { + // Describe where callee saved registers were saved, at fixed offsets from + // CFA. const std::vector<CalleeSavedInfo> &CSI = MFI->getCalleeSavedInfo(); for (unsigned I = 0, E = CSI.size(); I != E; ++I) { unsigned Reg = CSI[I].getReg(); diff --git a/lib/Target/PowerPC/PPCISelDAGToDAG.cpp b/lib/Target/PowerPC/PPCISelDAGToDAG.cpp index 536c2824fb7a..57960598c83f 100644 --- a/lib/Target/PowerPC/PPCISelDAGToDAG.cpp +++ b/lib/Target/PowerPC/PPCISelDAGToDAG.cpp @@ -172,10 +172,20 @@ namespace { /// a register. The case of adding a (possibly relocatable) constant to a /// register can be improved, but it is wrong to substitute Reg+Reg for /// Reg in an asm, because the load or store opcode would have to change. - bool SelectInlineAsmMemoryOperand(const SDValue &Op, + bool SelectInlineAsmMemoryOperand(const SDValue &Op, char ConstraintCode, std::vector<SDValue> &OutOps) override { - OutOps.push_back(Op); + // We need to make sure that this one operand does not end up in r0 + // (because we might end up lowering this as 0(%op)). + const TargetRegisterInfo *TRI = TM.getRegisterInfo(); + const TargetRegisterClass *TRC = TRI->getPointerRegClass(*MF, /*Kind=*/1); + SDValue RC = CurDAG->getTargetConstant(TRC->getID(), MVT::i32); + SDValue NewOp = + SDValue(CurDAG->getMachineNode(TargetOpcode::COPY_TO_REGCLASS, + SDLoc(Op), Op.getValueType(), + Op, RC), 0); + + OutOps.push_back(NewOp); return false; } @@ -1439,7 +1449,7 @@ SDNode *PPCDAGToDAGISel::Select(SDNode *N) { // For medium and large code model, we generate two instructions as // described below. Otherwise we allow SelectCodeCommon to handle this, - // selecting one of LDtoc, LDtocJTI, and LDtocCPT. + // selecting one of LDtoc, LDtocJTI, LDtocCPT, and LDtocBA. CodeModel::Model CModel = TM.getCodeModel(); if (CModel != CodeModel::Medium && CModel != CodeModel::Large) break; @@ -1456,7 +1466,8 @@ SDNode *PPCDAGToDAGISel::Select(SDNode *N) { SDNode *Tmp = CurDAG->getMachineNode(PPC::ADDIStocHA, dl, MVT::i64, TOCbase, GA); - if (isa<JumpTableSDNode>(GA) || CModel == CodeModel::Large) + if (isa<JumpTableSDNode>(GA) || isa<BlockAddressSDNode>(GA) || + CModel == CodeModel::Large) return CurDAG->getMachineNode(PPC::LDtocL, dl, MVT::i64, GA, SDValue(Tmp, 0)); @@ -1473,6 +1484,12 @@ SDNode *PPCDAGToDAGISel::Select(SDNode *N) { return CurDAG->getMachineNode(PPC::ADDItocL, dl, MVT::i64, SDValue(Tmp, 0), GA); } + case PPCISD::PPC32_PICGOT: { + // Generate a PIC-safe GOT reference. + assert(!PPCSubTarget->isPPC64() && PPCSubTarget->isSVR4ABI() && + "PPCISD::PPC32_PICGOT is only supported for 32-bit SVR4"); + return CurDAG->SelectNodeTo(N, PPC::PPC32PICGOT, PPCLowering->getPointerTy(), MVT::i32); + } case PPCISD::VADD_SPLAT: { // This expands into one of three sequences, depending on whether // the first operand is odd or even, positive or negative. diff --git a/lib/Target/PowerPC/PPCISelLowering.cpp b/lib/Target/PowerPC/PPCISelLowering.cpp index 708d36f6f978..0d37f50cd17b 100644 --- a/lib/Target/PowerPC/PPCISelLowering.cpp +++ b/lib/Target/PowerPC/PPCISelLowering.cpp @@ -781,6 +781,8 @@ const char *PPCTargetLowering::getTargetNodeName(unsigned Opcode) const { case PPCISD::SHL: return "PPCISD::SHL"; case PPCISD::CALL: return "PPCISD::CALL"; case PPCISD::CALL_NOP: return "PPCISD::CALL_NOP"; + case PPCISD::CALL_TLS: return "PPCISD::CALL_TLS"; + case PPCISD::CALL_NOP_TLS: return "PPCISD::CALL_NOP_TLS"; case PPCISD::MTCTR: return "PPCISD::MTCTR"; case PPCISD::BCTRL: return "PPCISD::BCTRL"; case PPCISD::RET_FLAG: return "PPCISD::RET_FLAG"; @@ -810,10 +812,8 @@ const char *PPCTargetLowering::getTargetNodeName(unsigned Opcode) const { case PPCISD::ADD_TLS: return "PPCISD::ADD_TLS"; case PPCISD::ADDIS_TLSGD_HA: return "PPCISD::ADDIS_TLSGD_HA"; case PPCISD::ADDI_TLSGD_L: return "PPCISD::ADDI_TLSGD_L"; - case PPCISD::GET_TLS_ADDR: return "PPCISD::GET_TLS_ADDR"; case PPCISD::ADDIS_TLSLD_HA: return "PPCISD::ADDIS_TLSLD_HA"; case PPCISD::ADDI_TLSLD_L: return "PPCISD::ADDI_TLSLD_L"; - case PPCISD::GET_TLSLD_ADDR: return "PPCISD::GET_TLSLD_ADDR"; case PPCISD::ADDIS_DTPREL_HA: return "PPCISD::ADDIS_DTPREL_HA"; case PPCISD::ADDI_DTPREL_L: return "PPCISD::ADDI_DTPREL_L"; case PPCISD::VADD_SPLAT: return "PPCISD::VADD_SPLAT"; @@ -1631,8 +1631,16 @@ SDValue PPCTargetLowering::LowerJumpTable(SDValue Op, SelectionDAG &DAG) const { SDValue PPCTargetLowering::LowerBlockAddress(SDValue Op, SelectionDAG &DAG) const { EVT PtrVT = Op.getValueType(); + BlockAddressSDNode *BASDN = cast<BlockAddressSDNode>(Op); + const BlockAddress *BA = BASDN->getBlockAddress(); - const BlockAddress *BA = cast<BlockAddressSDNode>(Op)->getBlockAddress(); + // 64-bit SVR4 ABI code is always position-independent. + // The actual BlockAddress is stored in the TOC. + if (Subtarget.isSVR4ABI() && Subtarget.isPPC64()) { + SDValue GA = DAG.getTargetBlockAddress(BA, PtrVT, BASDN->getOffset()); + return DAG.getNode(PPCISD::TOC_ENTRY, SDLoc(BASDN), MVT::i64, GA, + DAG.getRegister(PPC::X2, MVT::i64)); + } unsigned MOHiFlag, MOLoFlag; bool isPIC = GetLabelAccessInfo(DAG.getTarget(), MOHiFlag, MOLoFlag); @@ -1641,6 +1649,27 @@ SDValue PPCTargetLowering::LowerBlockAddress(SDValue Op, return LowerLabelRef(TgtBAHi, TgtBALo, isPIC, DAG); } +// Generate a call to __tls_get_addr for the given GOT entry Op. +std::pair<SDValue,SDValue> +PPCTargetLowering::lowerTLSCall(SDValue Op, SDLoc dl, + SelectionDAG &DAG) const { + + Type *IntPtrTy = getDataLayout()->getIntPtrType(*DAG.getContext()); + TargetLowering::ArgListTy Args; + TargetLowering::ArgListEntry Entry; + Entry.Node = Op; + Entry.Ty = IntPtrTy; + Args.push_back(Entry); + + TargetLowering::CallLoweringInfo CLI(DAG); + CLI.setDebugLoc(dl).setChain(DAG.getEntryNode()) + .setCallee(CallingConv::C, IntPtrTy, + DAG.getTargetExternalSymbol("__tls_get_addr", getPointerTy()), + std::move(Args), 0); + + return LowerCallTo(CLI); +} + SDValue PPCTargetLowering::LowerGlobalTLSAddress(SDValue Op, SelectionDAG &DAG) const { @@ -1684,50 +1713,40 @@ SDValue PPCTargetLowering::LowerGlobalTLSAddress(SDValue Op, } if (Model == TLSModel::GeneralDynamic) { - SDValue TGA = DAG.getTargetGlobalAddress(GV, dl, PtrVT, 0, 0); - SDValue GOTReg = DAG.getRegister(PPC::X2, MVT::i64); - SDValue GOTEntryHi = DAG.getNode(PPCISD::ADDIS_TLSGD_HA, dl, PtrVT, - GOTReg, TGA); + SDValue TGA = DAG.getTargetGlobalAddress(GV, dl, PtrVT, 0, + PPCII::MO_TLSGD); + SDValue GOTPtr; + if (is64bit) { + SDValue GOTReg = DAG.getRegister(PPC::X2, MVT::i64); + GOTPtr = DAG.getNode(PPCISD::ADDIS_TLSGD_HA, dl, PtrVT, + GOTReg, TGA); + } else { + GOTPtr = DAG.getNode(PPCISD::PPC32_PICGOT, dl, PtrVT); + } SDValue GOTEntry = DAG.getNode(PPCISD::ADDI_TLSGD_L, dl, PtrVT, - GOTEntryHi, TGA); - - // We need a chain node, and don't have one handy. The underlying - // call has no side effects, so using the function entry node - // suffices. - SDValue Chain = DAG.getEntryNode(); - Chain = DAG.getCopyToReg(Chain, dl, PPC::X3, GOTEntry); - SDValue ParmReg = DAG.getRegister(PPC::X3, MVT::i64); - SDValue TLSAddr = DAG.getNode(PPCISD::GET_TLS_ADDR, dl, - PtrVT, ParmReg, TGA); - // The return value from GET_TLS_ADDR really is in X3 already, but - // some hacks are needed here to tie everything together. The extra - // copies dissolve during subsequent transforms. - Chain = DAG.getCopyToReg(Chain, dl, PPC::X3, TLSAddr); - return DAG.getCopyFromReg(Chain, dl, PPC::X3, PtrVT); + GOTPtr, TGA); + std::pair<SDValue, SDValue> CallResult = lowerTLSCall(GOTEntry, dl, DAG); + return CallResult.first; } if (Model == TLSModel::LocalDynamic) { - SDValue TGA = DAG.getTargetGlobalAddress(GV, dl, PtrVT, 0, 0); - SDValue GOTReg = DAG.getRegister(PPC::X2, MVT::i64); - SDValue GOTEntryHi = DAG.getNode(PPCISD::ADDIS_TLSLD_HA, dl, PtrVT, - GOTReg, TGA); + SDValue TGA = DAG.getTargetGlobalAddress(GV, dl, PtrVT, 0, + PPCII::MO_TLSLD); + SDValue GOTPtr; + if (is64bit) { + SDValue GOTReg = DAG.getRegister(PPC::X2, MVT::i64); + GOTPtr = DAG.getNode(PPCISD::ADDIS_TLSLD_HA, dl, PtrVT, + GOTReg, TGA); + } else { + GOTPtr = DAG.getNode(PPCISD::PPC32_PICGOT, dl, PtrVT); + } SDValue GOTEntry = DAG.getNode(PPCISD::ADDI_TLSLD_L, dl, PtrVT, - GOTEntryHi, TGA); - - // We need a chain node, and don't have one handy. The underlying - // call has no side effects, so using the function entry node - // suffices. - SDValue Chain = DAG.getEntryNode(); - Chain = DAG.getCopyToReg(Chain, dl, PPC::X3, GOTEntry); - SDValue ParmReg = DAG.getRegister(PPC::X3, MVT::i64); - SDValue TLSAddr = DAG.getNode(PPCISD::GET_TLSLD_ADDR, dl, - PtrVT, ParmReg, TGA); - // The return value from GET_TLSLD_ADDR really is in X3 already, but - // some hacks are needed here to tie everything together. The extra - // copies dissolve during subsequent transforms. - Chain = DAG.getCopyToReg(Chain, dl, PPC::X3, TLSAddr); + GOTPtr, TGA); + std::pair<SDValue, SDValue> CallResult = lowerTLSCall(GOTEntry, dl, DAG); + SDValue TLSAddr = CallResult.first; + SDValue Chain = CallResult.second; SDValue DtvOffsetHi = DAG.getNode(PPCISD::ADDIS_DTPREL_HA, dl, PtrVT, - Chain, ParmReg, TGA); + Chain, TLSAddr, TGA); return DAG.getNode(PPCISD::ADDI_DTPREL_L, dl, PtrVT, DtvOffsetHi, TGA); } @@ -2676,7 +2695,7 @@ PPCTargetLowering::LowerFormalArguments_64SVR4( int FI; if (HasParameterArea || ArgSize + ArgOffset > LinkageSize + Num_GPR_Regs * PtrByteSize) - FI = MFI->CreateFixedObject(ArgSize, ArgOffset, true); + FI = MFI->CreateFixedObject(ArgSize, ArgOffset, false); else FI = MFI->CreateStackObject(ArgSize, Align, false); SDValue FIN = DAG.getFrameIndex(FI, PtrVT); @@ -3042,7 +3061,7 @@ PPCTargetLowering::LowerFormalArguments_Darwin( CurArgOffset = CurArgOffset + (4 - ObjSize); } // The value of the object is its address. - int FI = MFI->CreateFixedObject(ObjSize, CurArgOffset, true); + int FI = MFI->CreateFixedObject(ObjSize, CurArgOffset, false); SDValue FIN = DAG.getFrameIndex(FI, PtrVT); InVals.push_back(FIN); if (ObjSize==1 || ObjSize==2) { @@ -3690,6 +3709,23 @@ unsigned PrepareCall(SelectionDAG &DAG, SDValue &Callee, SDValue &InFlag, if (Callee.getNode()) { Ops.push_back(Chain); Ops.push_back(Callee); + + // If this is a call to __tls_get_addr, find the symbol whose address + // is to be taken and add it to the list. This will be used to + // generate __tls_get_addr(<sym>@tlsgd) or __tls_get_addr(<sym>@tlsld). + // We find the symbol by walking the chain to the CopyFromReg, walking + // back from the CopyFromReg to the ADDI_TLSGD_L or ADDI_TLSLD_L, and + // pulling the symbol from that node. + if (ExternalSymbolSDNode *S = dyn_cast<ExternalSymbolSDNode>(Callee)) + if (!strcmp(S->getSymbol(), "__tls_get_addr")) { + assert(!needIndirectCall && "Indirect call to __tls_get_addr???"); + SDNode *AddI = Chain.getNode()->getOperand(2).getNode(); + SDValue TGTAddr = AddI->getOperand(1); + assert(TGTAddr.getNode()->getOpcode() == ISD::TargetGlobalTLSAddress && + "Didn't find target global TLS address where we expected one"); + Ops.push_back(TGTAddr); + CallOpc = PPCISD::CALL_TLS; + } } // If this is a tail call add stack pointer delta. if (isTailCall) @@ -3841,7 +3877,9 @@ PPCTargetLowering::FinishCall(CallingConv::ID CallConv, SDLoc dl, DAG.getTarget().getRelocationModel() == Reloc::PIC_)) { // Otherwise insert NOP for non-local calls. CallOpc = PPCISD::CALL_NOP; - } + } else if (CallOpc == PPCISD::CALL_TLS) + // For 64-bit SVR4, TLS calls are always non-local. + CallOpc = PPCISD::CALL_NOP_TLS; } Chain = DAG.getNode(CallOpc, dl, NodeTys, Ops); @@ -8936,6 +8974,12 @@ PPCTargetLowering::getRegForInlineAsmConstraint(const std::string &Constraint, &PPC::G8RCRegClass); } + // GCC accepts 'cc' as an alias for 'cr0', and we need to do the same. + if (!R.second && StringRef("{cc}").equals_lower(Constraint)) { + R.first = PPC::CR0; + R.second = &PPC::CRRCRegClass; + } + return R; } @@ -8964,37 +9008,42 @@ void PPCTargetLowering::LowerAsmOperandForConstraint(SDValue Op, case 'P': { ConstantSDNode *CST = dyn_cast<ConstantSDNode>(Op); if (!CST) return; // Must be an immediate to match. - unsigned Value = CST->getZExtValue(); + int64_t Value = CST->getSExtValue(); + EVT TCVT = MVT::i64; // All constants taken to be 64 bits so that negative + // numbers are printed as such. switch (Letter) { default: llvm_unreachable("Unknown constraint letter!"); case 'I': // "I" is a signed 16-bit constant. - if ((short)Value == (int)Value) - Result = DAG.getTargetConstant(Value, Op.getValueType()); + if (isInt<16>(Value)) + Result = DAG.getTargetConstant(Value, TCVT); break; case 'J': // "J" is a constant with only the high-order 16 bits nonzero. + if (isShiftedUInt<16, 16>(Value)) + Result = DAG.getTargetConstant(Value, TCVT); + break; case 'L': // "L" is a signed 16-bit constant shifted left 16 bits. - if ((short)Value == 0) - Result = DAG.getTargetConstant(Value, Op.getValueType()); + if (isShiftedInt<16, 16>(Value)) + Result = DAG.getTargetConstant(Value, TCVT); break; case 'K': // "K" is a constant with only the low-order 16 bits nonzero. - if ((Value >> 16) == 0) - Result = DAG.getTargetConstant(Value, Op.getValueType()); + if (isUInt<16>(Value)) + Result = DAG.getTargetConstant(Value, TCVT); break; case 'M': // "M" is a constant that is greater than 31. if (Value > 31) - Result = DAG.getTargetConstant(Value, Op.getValueType()); + Result = DAG.getTargetConstant(Value, TCVT); break; case 'N': // "N" is a positive constant that is an exact power of two. - if ((int)Value > 0 && isPowerOf2_32(Value)) - Result = DAG.getTargetConstant(Value, Op.getValueType()); + if (Value > 0 && isPowerOf2_64(Value)) + Result = DAG.getTargetConstant(Value, TCVT); break; case 'O': // "O" is the constant zero. if (Value == 0) - Result = DAG.getTargetConstant(Value, Op.getValueType()); + Result = DAG.getTargetConstant(Value, TCVT); break; case 'P': // "P" is a constant whose negation is a signed 16-bit constant. - if ((short)-Value == (int)-Value) - Result = DAG.getTargetConstant(Value, Op.getValueType()); + if (isInt<16>(-Value)) + Result = DAG.getTargetConstant(Value, TCVT); break; } break; diff --git a/lib/Target/PowerPC/PPCISelLowering.h b/lib/Target/PowerPC/PPCISelLowering.h index c9394dd12e7b..7179adc9704f 100644 --- a/lib/Target/PowerPC/PPCISelLowering.h +++ b/lib/Target/PowerPC/PPCISelLowering.h @@ -99,6 +99,10 @@ namespace llvm { /// SVR4 calls. CALL, CALL_NOP, + /// CALL_TLS and CALL_NOP_TLS - Versions of CALL and CALL_NOP used + /// to access TLS variables. + CALL_TLS, CALL_NOP_TLS, + /// CHAIN,FLAG = MTCTR(VAL, CHAIN[, INFLAG]) - Directly corresponds to a /// MTCTR instruction. MTCTR, @@ -181,6 +185,10 @@ namespace llvm { /// on PPC32. PPC32_GOT, + /// GPRC = address of _GLOBAL_OFFSET_TABLE_. Used by general dynamic and + /// local dynamic TLS on PPC32. + PPC32_PICGOT, + /// G8RC = ADDIS_GOT_TPREL_HA %X2, Symbol - Used by the initial-exec /// TLS model, produces an ADDIS8 instruction that adds the GOT /// base to sym\@got\@tprel\@ha. @@ -210,10 +218,6 @@ namespace llvm { /// sym\@got\@tlsgd\@l. ADDI_TLSGD_L, - /// G8RC = GET_TLS_ADDR %X3, Symbol - For the general-dynamic TLS - /// model, produces a call to __tls_get_addr(sym\@tlsgd). - GET_TLS_ADDR, - /// G8RC = ADDIS_TLSLD_HA %X2, Symbol - For the local-dynamic TLS /// model, produces an ADDIS8 instruction that adds the GOT base /// register to sym\@got\@tlsld\@ha. @@ -224,10 +228,6 @@ namespace llvm { /// sym\@got\@tlsld\@l. ADDI_TLSLD_L, - /// G8RC = GET_TLSLD_ADDR %X3, Symbol - For the local-dynamic TLS - /// model, produces a call to __tls_get_addr(sym\@tlsld). - GET_TLSLD_ADDR, - /// G8RC = ADDIS_DTPREL_HA %X3, Symbol, Chain - For the /// local-dynamic TLS model, produces an ADDIS8 instruction /// that adds X3 to sym\@dtprel\@ha. The Chain operand is needed @@ -548,6 +548,8 @@ namespace llvm { SDValue LowerFRAMEADDR(SDValue Op, SelectionDAG &DAG) const; SDValue LowerConstantPool(SDValue Op, SelectionDAG &DAG) const; SDValue LowerBlockAddress(SDValue Op, SelectionDAG &DAG) const; + std::pair<SDValue,SDValue> lowerTLSCall(SDValue Op, SDLoc dl, + SelectionDAG &DAG) const; SDValue LowerGlobalTLSAddress(SDValue Op, SelectionDAG &DAG) const; SDValue LowerGlobalAddress(SDValue Op, SelectionDAG &DAG) const; SDValue LowerJumpTable(SDValue Op, SelectionDAG &DAG) const; diff --git a/lib/Target/PowerPC/PPCInstr64Bit.td b/lib/Target/PowerPC/PPCInstr64Bit.td index 9ed384f56244..b85d9c0ab62c 100644 --- a/lib/Target/PowerPC/PPCInstr64Bit.td +++ b/lib/Target/PowerPC/PPCInstr64Bit.td @@ -188,6 +188,9 @@ def : Pat<(PPCcall (i64 texternalsym:$dst)), def : Pat<(PPCcall_nop (i64 texternalsym:$dst)), (BL8_NOP texternalsym:$dst)>; +def : Pat<(PPCcall_nop_tls texternalsym:$func, tglobaltlsaddr:$sym), + (BL8_NOP_TLS texternalsym:$func, tglobaltlsaddr:$sym)>; + // Atomic operations let usesCustomInserter = 1 in { let Defs = [CR0] in { @@ -786,7 +789,7 @@ let canFoldAsLoad = 1, PPC970_Unit = 2 in { def LD : DSForm_1<58, 0, (outs g8rc:$rD), (ins memrix:$src), "ld $rD, $src", IIC_LdStLD, [(set i64:$rD, (aligned4load ixaddr:$src))]>, isPPC64; -// The following three definitions are selected for small code model only. +// The following four definitions are selected for small code model only. // Otherwise, we need to create two instructions to form a 32-bit offset, // so we have a custom matcher for TOC_ENTRY in PPCDAGToDAGIsel::Select(). def LDtoc: Pseudo<(outs g8rc:$rD), (ins tocentry:$disp, g8rc:$reg), @@ -801,6 +804,10 @@ def LDtocCPT: Pseudo<(outs g8rc:$rD), (ins tocentry:$disp, g8rc:$reg), "#LDtocCPT", [(set i64:$rD, (PPCtoc_entry tconstpool:$disp, i64:$reg))]>, isPPC64; +def LDtocBA: Pseudo<(outs g8rc:$rD), (ins tocentry:$disp, g8rc:$reg), + "#LDtocCPT", + [(set i64:$rD, + (PPCtoc_entry tblockaddress:$disp, i64:$reg))]>, isPPC64; let hasSideEffects = 1, isCodeGenOnly = 1, RST = 2, Defs = [X2] in def LDinto_toc: DSForm_1<58, 0, (outs), (ins memrix:$src), @@ -872,11 +879,6 @@ def ADDItlsgdL : Pseudo<(outs g8rc:$rD), (ins g8rc_nox0:$reg, s16imm64:$disp), [(set i64:$rD, (PPCaddiTlsgdL i64:$reg, tglobaltlsaddr:$disp))]>, isPPC64; -def GETtlsADDR : Pseudo<(outs g8rc:$rD), (ins g8rc:$reg, tlsgd:$sym), - "#GETtlsADDR", - [(set i64:$rD, - (PPCgetTlsAddr i64:$reg, tglobaltlsaddr:$sym))]>, - isPPC64; def ADDIStlsldHA: Pseudo<(outs g8rc:$rD), (ins g8rc_nox0:$reg, s16imm64:$disp), "#ADDIStlsldHA", [(set i64:$rD, @@ -887,11 +889,6 @@ def ADDItlsldL : Pseudo<(outs g8rc:$rD), (ins g8rc_nox0:$reg, s16imm64:$disp), [(set i64:$rD, (PPCaddiTlsldL i64:$reg, tglobaltlsaddr:$disp))]>, isPPC64; -def GETtlsldADDR : Pseudo<(outs g8rc:$rD), (ins g8rc:$reg, tlsgd:$sym), - "#GETtlsldADDR", - [(set i64:$rD, - (PPCgetTlsldAddr i64:$reg, tglobaltlsaddr:$sym))]>, - isPPC64; def ADDISdtprelHA: Pseudo<(outs g8rc:$rD), (ins g8rc_nox0:$reg, s16imm64:$disp), "#ADDISdtprelHA", [(set i64:$rD, diff --git a/lib/Target/PowerPC/PPCInstrInfo.td b/lib/Target/PowerPC/PPCInstrInfo.td index 42b740f4fa46..197e751464ad 100644 --- a/lib/Target/PowerPC/PPCInstrInfo.td +++ b/lib/Target/PowerPC/PPCInstrInfo.td @@ -110,10 +110,8 @@ def PPCldGotTprelL : SDNode<"PPCISD::LD_GOT_TPREL_L", SDTIntBinOp, def PPCaddTls : SDNode<"PPCISD::ADD_TLS", SDTIntBinOp, []>; def PPCaddisTlsgdHA : SDNode<"PPCISD::ADDIS_TLSGD_HA", SDTIntBinOp>; def PPCaddiTlsgdL : SDNode<"PPCISD::ADDI_TLSGD_L", SDTIntBinOp>; -def PPCgetTlsAddr : SDNode<"PPCISD::GET_TLS_ADDR", SDTIntBinOp>; def PPCaddisTlsldHA : SDNode<"PPCISD::ADDIS_TLSLD_HA", SDTIntBinOp>; def PPCaddiTlsldL : SDNode<"PPCISD::ADDI_TLSLD_L", SDTIntBinOp>; -def PPCgetTlsldAddr : SDNode<"PPCISD::GET_TLSLD_ADDR", SDTIntBinOp>; def PPCaddisDtprelHA : SDNode<"PPCISD::ADDIS_DTPREL_HA", SDTIntBinOp, [SDNPHasChain]>; def PPCaddiDtprelL : SDNode<"PPCISD::ADDI_DTPREL_L", SDTIntBinOp>; @@ -136,9 +134,15 @@ def SDT_PPCCall : SDTypeProfile<0, -1, [SDTCisInt<0>]>; def PPCcall : SDNode<"PPCISD::CALL", SDT_PPCCall, [SDNPHasChain, SDNPOptInGlue, SDNPOutGlue, SDNPVariadic]>; +def PPCcall_tls : SDNode<"PPCISD::CALL_TLS", SDT_PPCCall, + [SDNPHasChain, SDNPOptInGlue, SDNPOutGlue, + SDNPVariadic]>; def PPCcall_nop : SDNode<"PPCISD::CALL_NOP", SDT_PPCCall, [SDNPHasChain, SDNPOptInGlue, SDNPOutGlue, SDNPVariadic]>; +def PPCcall_nop_tls : SDNode<"PPCISD::CALL_NOP_TLS", SDT_PPCCall, + [SDNPHasChain, SDNPOptInGlue, SDNPOutGlue, + SDNPVariadic]>; def PPCload : SDNode<"PPCISD::LOAD", SDTypeProfile<1, 1, []>, [SDNPHasChain, SDNPOptInGlue, SDNPOutGlue]>; def PPCload_toc : SDNode<"PPCISD::LOAD_TOC", SDTypeProfile<0, 1, []>, @@ -588,6 +592,12 @@ def tlsreg32 : Operand<i32> { let EncoderMethod = "getTLSRegEncoding"; let ParserMatchClass = PPCTLSRegOperand; } +def tlsgd32 : Operand<i32> {} +def tlscall32 : Operand<i32> { + let PrintMethod = "printTLSCall"; + let MIOperandInfo = (ops calltarget:$func, tlsgd32:$sym); + let EncoderMethod = "getTLSCallEncoding"; +} // PowerPC Predicate operand. def pred : Operand<OtherVT> { @@ -1071,6 +1081,8 @@ let isCall = 1, PPC970_Unit = 7, Defs = [LR] in { "bla $func", IIC_BrB, [(PPCcall (i32 imm:$func))]>; let isCodeGenOnly = 1 in { + def BL_TLS : IForm<18, 0, 1, (outs), (ins tlscall32:$func), + "bl $func", IIC_BrB, []>; def BCCL : BForm<16, 0, 1, (outs), (ins pred:$cond, condbrtarget:$dst), "b${cond:cc}l${cond:pm} ${cond:reg}, $dst">; def BCCLA : BForm<16, 1, 1, (outs), (ins pred:$cond, abscondbrtarget:$dst), @@ -2358,6 +2370,8 @@ def : Pat<(PPCcall (i32 tglobaladdr:$dst)), def : Pat<(PPCcall (i32 texternalsym:$dst)), (BL texternalsym:$dst)>; +def : Pat<(PPCcall_tls texternalsym:$func, tglobaltlsaddr:$sym), + (BL_TLS texternalsym:$func, tglobaltlsaddr:$sym)>; def : Pat<(PPCtc_return (i32 tglobaladdr:$dst), imm:$imm), (TCRETURNdi tglobaladdr:$dst, imm:$imm)>; @@ -2396,13 +2410,37 @@ def : Pat<(add i32:$in, (PPChi tblockaddress:$g, 0)), def PPC32GOT: Pseudo<(outs gprc:$rD), (ins), "#PPC32GOT", [(set i32:$rD, (PPCppc32GOT))]>; +// Get the _GLOBAL_OFFSET_TABLE_ in PIC mode. +// This uses two output registers, the first as the real output, the second as a +// temporary register, used internally in code generation. +def PPC32PICGOT: Pseudo<(outs gprc:$rD, gprc:$rT), (ins), "#PPC32PICGOT", + []>, NoEncode<"$rT">; + def LDgotTprelL32: Pseudo<(outs gprc:$rD), (ins s16imm:$disp, gprc_nor0:$reg), - "#LDgotTprelL32", - [(set i32:$rD, - (PPCldGotTprelL tglobaltlsaddr:$disp, i32:$reg))]>; + "#LDgotTprelL32", + [(set i32:$rD, + (PPCldGotTprelL tglobaltlsaddr:$disp, i32:$reg))]>; def : Pat<(PPCaddTls i32:$in, tglobaltlsaddr:$g), (ADD4TLS $in, tglobaltlsaddr:$g)>; +def ADDItlsgdL32 : Pseudo<(outs gprc:$rD), (ins gprc_nor0:$reg, s16imm:$disp), + "#ADDItlsgdL32", + [(set i32:$rD, + (PPCaddiTlsgdL i32:$reg, tglobaltlsaddr:$disp))]>; +def ADDItlsldL32 : Pseudo<(outs gprc:$rD), (ins gprc_nor0:$reg, s16imm:$disp), + "#ADDItlsldL32", + [(set i32:$rD, + (PPCaddiTlsldL i32:$reg, tglobaltlsaddr:$disp))]>; +def ADDIdtprelL32 : Pseudo<(outs gprc:$rD), (ins gprc_nor0:$reg, s16imm:$disp), + "#ADDIdtprelL32", + [(set i32:$rD, + (PPCaddiDtprelL i32:$reg, tglobaltlsaddr:$disp))]>; +def ADDISdtprelHA32 : Pseudo<(outs gprc:$rD), (ins gprc_nor0:$reg, s16imm:$disp), + "#ADDISdtprelHA32", + [(set i32:$rD, + (PPCaddisDtprelHA i32:$reg, + tglobaltlsaddr:$disp))]>; + // Support for Position-independent code def LWZtoc: Pseudo<(outs gprc:$rD), (ins tocentry32:$disp, gprc:$reg), "#LWZtoc", diff --git a/lib/Target/PowerPC/PPCMCInstLower.cpp b/lib/Target/PowerPC/PPCMCInstLower.cpp index 668041371780..681eda8b9ea4 100644 --- a/lib/Target/PowerPC/PPCMCInstLower.cpp +++ b/lib/Target/PowerPC/PPCMCInstLower.cpp @@ -137,6 +137,12 @@ static MCOperand GetSymbolRef(const MachineOperand &MO, const MCSymbol *Symbol, case PPCII::MO_TLS: RefKind = MCSymbolRefExpr::VK_PPC_TLS; break; + case PPCII::MO_TLSGD: + RefKind = MCSymbolRefExpr::VK_PPC_TLSGD; + break; + case PPCII::MO_TLSLD: + RefKind = MCSymbolRefExpr::VK_PPC_TLSLD; + break; } if (MO.getTargetFlags() == PPCII::MO_PLT_OR_STUB && !isDarwin) diff --git a/lib/Target/PowerPC/PPCRegisterInfo.td b/lib/Target/PowerPC/PPCRegisterInfo.td index b3d145b2cc49..0442a941f45a 100644 --- a/lib/Target/PowerPC/PPCRegisterInfo.td +++ b/lib/Target/PowerPC/PPCRegisterInfo.td @@ -188,13 +188,6 @@ def CR6 : CR<6, "cr6", [CR6LT, CR6GT, CR6EQ, CR6UN]>, DwarfRegNum<[74, 74]>; def CR7 : CR<7, "cr7", [CR7LT, CR7GT, CR7EQ, CR7UN]>, DwarfRegNum<[75, 75]>; } -// The full condition-code register. This is not modeled fully, but defined -// here primarily, for compatibility with gcc, to allow the inline asm "cc" -// clobber specification to work. -def CC : PPCReg<"cc">, DwarfRegAlias<CR0> { - let Aliases = [CR0, CR1, CR2, CR3, CR4, CR5, CR6, CR7]; -} - // Link register def LR : SPR<8, "lr">, DwarfRegNum<[-2, 65]>; //let Aliases = [LR] in @@ -308,7 +301,3 @@ def CARRYRC : RegisterClass<"PPC", [i32], 32, (add CARRY)> { let CopyCost = -1; } -def CCRC : RegisterClass<"PPC", [i32], 32, (add CC)> { - let isAllocatable = 0; -} - |