X-Git-Url: http://plrg.eecs.uci.edu/git/?a=blobdiff_plain;f=lib%2FTarget%2FARM%2FARMSubtarget.cpp;h=3b0d787528dd607ef46fc3db8177eb4c995301c7;hb=765aab84d619c67a9a44d97b78c696d3d9bd4fb9;hp=547c69a6fd9a82129843cfb419a80ce7663e46fc;hpb=0aba46f4cda778036196ca721193b49bdc03a13d;p=oota-llvm.git diff --git a/lib/Target/ARM/ARMSubtarget.cpp b/lib/Target/ARM/ARMSubtarget.cpp index 547c69a6fd9..3b0d787528d 100644 --- a/lib/Target/ARM/ARMSubtarget.cpp +++ b/lib/Target/ARM/ARMSubtarget.cpp @@ -12,21 +12,33 @@ //===----------------------------------------------------------------------===// #include "ARMSubtarget.h" -#include "ARMBaseInstrInfo.h" -#include "ARMBaseRegisterInfo.h" +#include "ARMFrameLowering.h" +#include "ARMISelLowering.h" +#include "ARMInstrInfo.h" +#include "ARMMachineFunctionInfo.h" +#include "ARMSelectionDAGInfo.h" +#include "ARMSubtarget.h" +#include "ARMTargetMachine.h" +#include "Thumb1FrameLowering.h" +#include "Thumb1InstrInfo.h" +#include "Thumb2InstrInfo.h" +#include "llvm/CodeGen/MachineRegisterInfo.h" #include "llvm/IR/Attributes.h" -#include "llvm/IR/GlobalValue.h" #include "llvm/IR/Function.h" +#include "llvm/IR/GlobalValue.h" #include "llvm/Support/CommandLine.h" #include "llvm/Target/TargetInstrInfo.h" #include "llvm/Target/TargetOptions.h" +#include "llvm/Target/TargetRegisterInfo.h" + +using namespace llvm; + +#define DEBUG_TYPE "arm-subtarget" #define GET_SUBTARGETINFO_TARGET_DESC #define GET_SUBTARGETINFO_CTOR #include "ARMGenSubtargetInfo.inc" -using namespace llvm; - static cl::opt ReserveR9("arm-reserve-r9", cl::Hidden, cl::desc("Reserve R9, making it unavailable as GPR")); @@ -38,11 +50,13 @@ static cl::opt UseFusedMulOps("arm-use-mulops", cl::init(true), cl::Hidden); +namespace { enum AlignMode { DefaultAlign, StrictAlign, NoStrictAlign }; +} static cl::opt Align(cl::desc("Load/store alignment support"), @@ -74,26 +88,48 @@ IT(cl::desc("IT block support"), cl::Hidden, cl::init(DefaultIT), "Allow IT blocks based on ARMv7"), clEnumValEnd)); -ARMSubtarget::ARMSubtarget(const std::string &TT, const std::string &CPU, - const std::string &FS, const TargetOptions &Options) - : ARMGenSubtargetInfo(TT, CPU, FS) - , ARMProcFamily(Others) - , ARMProcClass(None) - , stackAlignment(4) - , CPUString(CPU) - , TargetTriple(TT) - , Options(Options) - , TargetABI(ARM_ABI_UNKNOWN) { +/// initializeSubtargetDependencies - Initializes using a CPU and feature string +/// so that we can use initializer lists for subtarget initialization. +ARMSubtarget &ARMSubtarget::initializeSubtargetDependencies(StringRef CPU, + StringRef FS) { initializeEnvironment(); - resetSubtargetFeatures(CPU, FS); + initSubtargetFeatures(CPU, FS); + return *this; +} + +ARMFrameLowering *ARMSubtarget::initializeFrameLowering(StringRef CPU, + StringRef FS) { + ARMSubtarget &STI = initializeSubtargetDependencies(CPU, FS); + if (STI.isThumb1Only()) + return (ARMFrameLowering *)new Thumb1FrameLowering(STI); + + return new ARMFrameLowering(STI); } +ARMSubtarget::ARMSubtarget(const std::string &TT, const std::string &CPU, + const std::string &FS, + const ARMBaseTargetMachine &TM, bool IsLittle) + : ARMGenSubtargetInfo(TT, CPU, FS), ARMProcFamily(Others), + ARMProcClass(None), stackAlignment(4), CPUString(CPU), IsLittle(IsLittle), + TargetTriple(TT), Options(TM.Options), TM(TM), + TSInfo(*TM.getDataLayout()), + FrameLowering(initializeFrameLowering(CPU, FS)), + // At this point initializeSubtargetDependencies has been called so + // we can query directly. + InstrInfo(isThumb1Only() + ? (ARMBaseInstrInfo *)new Thumb1InstrInfo(*this) + : !isThumb() + ? (ARMBaseInstrInfo *)new ARMInstrInfo(*this) + : (ARMBaseInstrInfo *)new Thumb2InstrInfo(*this)), + TLInfo(TM, *this) {} + void ARMSubtarget::initializeEnvironment() { HasV4TOps = false; HasV5TOps = false; HasV5TEOps = false; HasV6Ops = false; HasV6MOps = false; + HasV6KOps = false; HasV6T2Ops = false; HasV7Ops = false; HasV8Ops = false; @@ -102,7 +138,6 @@ void ARMSubtarget::initializeEnvironment() { HasVFPv4 = false; HasFPARMv8 = false; HasNEON = false; - MinSize = false; UseNEONForSinglePrecisionFP = false; UseMulOps = UseFusedMulOps; SlowFPVMLx = false; @@ -111,7 +146,6 @@ void ARMSubtarget::initializeEnvironment() { InThumbMode = false; HasThumb2 = false; NoARM = false; - PostRAScheduler = false; IsR9Reserved = ReserveR9; UseMovt = false; SupportsTailCall = false; @@ -132,34 +166,16 @@ void ARMSubtarget::initializeEnvironment() { HasTrustZone = false; HasCrypto = false; HasCRC = false; + HasZeroCycleZeroing = false; AllowsUnalignedMem = false; Thumb2DSP = false; UseNaClTrap = false; UnsafeFPMath = false; } -void ARMSubtarget::resetSubtargetFeatures(const MachineFunction *MF) { - AttributeSet FnAttrs = MF->getFunction()->getAttributes(); - Attribute CPUAttr = FnAttrs.getAttribute(AttributeSet::FunctionIndex, - "target-cpu"); - Attribute FSAttr = FnAttrs.getAttribute(AttributeSet::FunctionIndex, - "target-features"); - std::string CPU = - !CPUAttr.hasAttribute(Attribute::None) ?CPUAttr.getValueAsString() : ""; - std::string FS = - !FSAttr.hasAttribute(Attribute::None) ? FSAttr.getValueAsString() : ""; - if (!FS.empty()) { - initializeEnvironment(); - resetSubtargetFeatures(CPU, FS); - } - - MinSize = - FnAttrs.hasAttribute(AttributeSet::FunctionIndex, Attribute::MinSize); -} - -void ARMSubtarget::resetSubtargetFeatures(StringRef CPU, StringRef FS) { +void ARMSubtarget::initSubtargetFeatures(StringRef CPU, StringRef FS) { if (CPUString.empty()) { - if (isTargetIOS() && TargetTriple.getArchName().endswith("v7s")) + if (isTargetDarwin() && TargetTriple.getArchName().endswith("v7s")) // Default to the Swift CPU when targeting armv7s/thumbv7s. CPUString = "swift"; else @@ -169,8 +185,8 @@ void ARMSubtarget::resetSubtargetFeatures(StringRef CPU, StringRef FS) { // Insert the architecture feature derived from the target triple into the // feature string. This is important for setting features that are implied // based on the architecture version. - std::string ArchFS = ARM_MC::ParseARMTriple(TargetTriple.getTriple(), - CPUString); + std::string ArchFS = + ARM_MC::ParseARMTriple(TargetTriple.getTriple(), CPUString); if (!FS.empty()) { if (!ArchFS.empty()) ArchFS = ArchFS + "," + FS.str(); @@ -189,66 +205,51 @@ void ARMSubtarget::resetSubtargetFeatures(StringRef CPU, StringRef FS) { // Initialize scheduling itinerary for the specified CPU. InstrItins = getInstrItineraryForCPU(CPUString); - if (TargetABI == ARM_ABI_UNKNOWN) { - switch (TargetTriple.getEnvironment()) { - case Triple::Android: - case Triple::EABI: - case Triple::EABIHF: - case Triple::GNUEABI: - case Triple::GNUEABIHF: - case Triple::MachO: - TargetABI = ARM_ABI_AAPCS; - break; - default: - if (isTargetIOS() && isMClass()) - TargetABI = ARM_ABI_AAPCS; - else - TargetABI = ARM_ABI_APCS; - break; - } - } + // FIXME: this is invalid for WindowsCE + if (isTargetWindows()) + NoARM = true; if (isAAPCS_ABI()) stackAlignment = 8; + if (isTargetNaCl()) + stackAlignment = 16; UseMovt = hasV6T2Ops() && ArmUseMOVT; if (isTargetMachO()) { - IsR9Reserved = ReserveR9 | !HasV6Ops; + IsR9Reserved = ReserveR9 || !HasV6Ops; SupportsTailCall = !isTargetIOS() || !getTargetTriple().isOSVersionLT(5, 0); - } else + } else { IsR9Reserved = ReserveR9; + SupportsTailCall = !isThumb1Only(); + } - if (!isThumb() || hasThumb2()) - PostRAScheduler = true; - - switch (Align) { - case DefaultAlign: - // Assume pre-ARMv6 doesn't support unaligned accesses. - // - // ARMv6 may or may not support unaligned accesses depending on the - // SCTLR.U bit, which is architecture-specific. We assume ARMv6 - // Darwin targets support unaligned accesses, and others don't. - // - // ARMv7 always has SCTLR.U set to 1, but it has a new SCTLR.A bit - // which raises an alignment fault on unaligned accesses. Linux - // defaults this bit to 0 and handles it as a system-wide (not - // per-process) setting. It is therefore safe to assume that ARMv7+ - // Linux targets support unaligned accesses. The same goes for NaCl. - // - // The above behavior is consistent with GCC. - AllowsUnalignedMem = ( - (hasV7Ops() && (isTargetLinux() || isTargetNaCl())) || - (hasV6Ops() && isTargetMachO())); - break; - case StrictAlign: - AllowsUnalignedMem = false; - break; - case NoStrictAlign: - AllowsUnalignedMem = true; - break; + if (Align == DefaultAlign) { + // Assume pre-ARMv6 doesn't support unaligned accesses. + // + // ARMv6 may or may not support unaligned accesses depending on the + // SCTLR.U bit, which is architecture-specific. We assume ARMv6 + // Darwin and NetBSD targets support unaligned accesses, and others don't. + // + // ARMv7 always has SCTLR.U set to 1, but it has a new SCTLR.A bit + // which raises an alignment fault on unaligned accesses. Linux + // defaults this bit to 0 and handles it as a system-wide (not + // per-process) setting. It is therefore safe to assume that ARMv7+ + // Linux targets support unaligned accesses. The same goes for NaCl. + // + // The above behavior is consistent with GCC. + AllowsUnalignedMem = + (hasV7Ops() && (isTargetLinux() || isTargetNaCl() || + isTargetNetBSD())) || + (hasV6Ops() && (isTargetMachO() || isTargetNetBSD())); + } else { + AllowsUnalignedMem = !(Align == StrictAlign); } + // No v6M core supports unaligned memory access (v6M ARM ARM A3.2) + if (isV6M()) + AllowsUnalignedMem = false; + switch (IT) { case DefaultIT: RestrictIT = hasV8Ops() ? true : false; @@ -268,6 +269,15 @@ void ARMSubtarget::resetSubtargetFeatures(StringRef CPU, StringRef FS) { UseNEONForSinglePrecisionFP = true; } +bool ARMSubtarget::isAPCS_ABI() const { + assert(TM.TargetABI != ARMBaseTargetMachine::ARM_ABI_UNKNOWN); + return TM.TargetABI == ARMBaseTargetMachine::ARM_ABI_APCS; +} +bool ARMSubtarget::isAAPCS_ABI() const { + assert(TM.TargetABI != ARMBaseTargetMachine::ARM_ABI_UNKNOWN); + return TM.TargetABI == ARMBaseTargetMachine::ARM_ABI_AAPCS; +} + /// GVIsIndirectSymbol - true if the GV will be accessed via an indirect symbol. bool ARMSubtarget::GVIsIndirectSymbol(const GlobalValue *GV, @@ -275,11 +285,7 @@ ARMSubtarget::GVIsIndirectSymbol(const GlobalValue *GV, if (RelocM == Reloc::Static) return false; - // Materializable GVs (in JIT lazy compilation mode) do not require an extra - // load from stub. - bool isDecl = GV->hasAvailableExternallyLinkage(); - if (GV->isDeclaration() && !GV->isMaterializable()) - isDecl = true; + bool isDecl = GV->isDeclarationForLinker(); if (!isTargetMachO()) { // Extra load is needed for all externally visible. @@ -322,18 +328,26 @@ ARMSubtarget::GVIsIndirectSymbol(const GlobalValue *GV, } unsigned ARMSubtarget::getMispredictionPenalty() const { - return SchedModel->MispredictPenalty; + return SchedModel.MispredictPenalty; } bool ARMSubtarget::hasSinCos() const { - return getTargetTriple().getOS() == Triple::IOS && - !getTargetTriple().isOSVersionLT(7, 0); + return getTargetTriple().isiOS() && !getTargetTriple().isOSVersionLT(7, 0); +} + +// This overrides the PostRAScheduler bit in the SchedModel for any CPU. +bool ARMSubtarget::enablePostMachineScheduler() const { + return (!isThumb() || hasThumb2()); +} + +bool ARMSubtarget::enableAtomicExpand() const { + return hasAnyDataBarrier() && !isThumb1Only(); } -bool ARMSubtarget::enablePostRAScheduler( - CodeGenOpt::Level OptLevel, - TargetSubtargetInfo::AntiDepBreakMode& Mode, - RegClassVector& CriticalPathRCs) const { - Mode = TargetSubtargetInfo::ANTIDEP_NONE; - return PostRAScheduler && OptLevel >= CodeGenOpt::Default; +bool ARMSubtarget::useMovt(const MachineFunction &MF) const { + // NOTE Windows on ARM needs to use mov.w/mov.t pairs to materialise 32-bit + // immediates as it is inherently position independent, and may be out of + // range otherwise. + return UseMovt && (isTargetWindows() || + !MF.getFunction()->hasFnAttribute(Attribute::MinSize)); }