X-Git-Url: http://plrg.eecs.uci.edu/git/?a=blobdiff_plain;f=lib%2FTarget%2FX86%2FX86Subtarget.h;h=f723c66b729cc9a1e482684c26b37a53be7a3039;hb=7bbd6e366b39157445cc921024a987e61ea68c00;hp=7dcfe546a58ff69d80ea71baef08188482301b8d;hpb=2cf711f1e69dc1d00d57814478c8eb12d97ce3a8;p=oota-llvm.git diff --git a/lib/Target/X86/X86Subtarget.h b/lib/Target/X86/X86Subtarget.h index 7dcfe546a58..f723c66b729 100644 --- a/lib/Target/X86/X86Subtarget.h +++ b/lib/Target/X86/X86Subtarget.h @@ -1,4 +1,4 @@ -//=====---- X86Subtarget.h - Define Subtarget for the X86 -----*- C++ -*--====// +//===-- X86Subtarget.h - Define Subtarget for the X86 ----------*- C++ -*--===// // // The LLVM Compiler Infrastructure // @@ -7,20 +7,24 @@ // //===----------------------------------------------------------------------===// // -// This file declares the X86 specific subclass of TargetSubtarget. +// This file declares the X86 specific subclass of TargetSubtargetInfo. // //===----------------------------------------------------------------------===// #ifndef X86SUBTARGET_H #define X86SUBTARGET_H -#include "llvm/ADT/Triple.h" -#include "llvm/Target/TargetSubtarget.h" #include "llvm/CallingConv.h" +#include "llvm/ADT/Triple.h" +#include "llvm/Target/TargetSubtargetInfo.h" #include +#define GET_SUBTARGETINFO_HEADER +#include "X86GenSubtargetInfo.inc" + namespace llvm { class GlobalValue; +class StringRef; class TargetMachine; /// PICStyles - The X86 backend supports a number of different styles of PIC. @@ -35,16 +39,23 @@ enum Style { }; } -class X86Subtarget : public TargetSubtarget { +class X86Subtarget : public X86GenSubtargetInfo { protected: enum X86SSEEnum { - NoMMXSSE, MMX, SSE1, SSE2, SSE3, SSSE3, SSE41, SSE42 + NoMMXSSE, MMX, SSE1, SSE2, SSE3, SSSE3, SSE41, SSE42, AVX, AVX2 }; enum X863DNowEnum { NoThreeDNow, ThreeDNow, ThreeDNowA }; + enum X86ProcFamilyEnum { + Others, IntelAtom + }; + + /// X86ProcFamily - X86 processor family: Intel Atom, and others + X86ProcFamilyEnum X86ProcFamily; + /// PICStyle - Which PIC style to use /// PICStyles::Style PICStyle; @@ -71,21 +82,42 @@ protected: /// HasSSE4A - True if the processor supports SSE4A instructions. bool HasSSE4A; - /// HasAVX - Target has AVX instructions - bool HasAVX; - /// HasAES - Target has AES instructions bool HasAES; - /// HasCLMUL - Target has carry-less multiplication - bool HasCLMUL; + /// HasPCLMUL - Target has carry-less multiplication + bool HasPCLMUL; - /// HasFMA3 - Target has 3-operand fused multiply-add - bool HasFMA3; + /// HasFMA - Target has 3-operand fused multiply-add + bool HasFMA; /// HasFMA4 - Target has 4-operand fused multiply-add bool HasFMA4; + /// HasXOP - Target has XOP instructions + bool HasXOP; + + /// HasMOVBE - True if the processor has the MOVBE instruction. + bool HasMOVBE; + + /// HasRDRAND - True if the processor has the RDRAND instruction. + bool HasRDRAND; + + /// HasF16C - Processor has 16-bit floating point conversion instructions. + bool HasF16C; + + /// HasFSGSBase - Processor has FS/GS base insturctions. + bool HasFSGSBase; + + /// HasLZCNT - Processor has LZCNT instruction. + bool HasLZCNT; + + /// HasBMI - Processor has BMI1 instructions. + bool HasBMI; + + /// HasBMI2 - Processor has BMI2 instructions. + bool HasBMI2; + /// IsBTMemSlow - True if BT (bit test) of memory instructions are slow. bool IsBTMemSlow; @@ -96,6 +128,21 @@ protected: /// operands. This may require setting a feature bit in the processor. bool HasVectorUAMem; + /// HasCmpxchg16b - True if this processor has the CMPXCHG16B instruction; + /// this is true for most x86-64 chips, but not the first AMD chips. + bool HasCmpxchg16b; + + /// UseLeaForSP - True if the LEA instruction should be used for adjusting + /// the stack pointer. This is an optimization for Intel Atom processors. + bool UseLeaForSP; + + /// HasSlowDivide - True if smaller divides are significantly faster than + /// full divides and should be used when possible. + bool HasSlowDivide; + + /// PostRAScheduler - True if using post-register-allocation scheduler. + bool PostRAScheduler; + /// stackAlignment - The minimum alignment known to hold of the stack frame on /// entry to the function and which must be maintained by every function. unsigned stackAlignment; @@ -107,17 +154,21 @@ protected: /// TargetTriple - What processor and OS we're targeting. Triple TargetTriple; + /// Instruction itineraries for scheduling + InstrItineraryData InstrItins; + private: - /// Is64Bit - True if the processor supports 64-bit instructions and - /// pointer size is 64 bit. - bool Is64Bit; + /// In64BitMode - True if compiling for 64-bit, false for 32-bit. + bool In64BitMode; public: /// This constructor initializes the data members to match that /// of the specified triple. /// - X86Subtarget(const std::string &TT, const std::string &FS, bool is64Bit); + X86Subtarget(const std::string &TT, const std::string &CPU, + const std::string &FS, + unsigned StackAlignOverride, bool is64Bit); /// getStackAlignment - Returns the minimum alignment known to hold of the /// stack frame on entry to the function and which must be maintained by every @@ -130,14 +181,13 @@ public: /// ParseSubtargetFeatures - Parses features string setting specified /// subtarget options. Definition of function is auto generated by tblgen. - std::string ParseSubtargetFeatures(const std::string &FS, - const std::string &CPU); + void ParseSubtargetFeatures(StringRef CPU, StringRef FS); /// AutoDetectSubtargetFeatures - Auto-detect CPU features using CPUID /// instruction. void AutoDetectSubtargetFeatures(); - bool is64Bit() const { return Is64Bit; } + bool is64Bit() const { return In64BitMode; } PICStyles::Style getPICStyle() const { return PICStyle; } void setPICStyle(PICStyles::Style Style) { PICStyle = Style; } @@ -150,56 +200,72 @@ public: bool hasSSSE3() const { return X86SSELevel >= SSSE3; } bool hasSSE41() const { return X86SSELevel >= SSE41; } bool hasSSE42() const { return X86SSELevel >= SSE42; } + bool hasAVX() const { return X86SSELevel >= AVX; } + bool hasAVX2() const { return X86SSELevel >= AVX2; } bool hasSSE4A() const { return HasSSE4A; } bool has3DNow() const { return X863DNowLevel >= ThreeDNow; } bool has3DNowA() const { return X863DNowLevel >= ThreeDNowA; } bool hasPOPCNT() const { return HasPOPCNT; } - bool hasAVX() const { return HasAVX; } - bool hasXMM() const { return hasSSE1() || hasAVX(); } - bool hasXMMInt() const { return hasSSE2() || hasAVX(); } bool hasAES() const { return HasAES; } - bool hasCLMUL() const { return HasCLMUL; } - bool hasFMA3() const { return HasFMA3; } - bool hasFMA4() const { return HasFMA4; } + bool hasPCLMUL() const { return HasPCLMUL; } + bool hasFMA() const { return HasFMA; } + // FIXME: Favor FMA when both are enabled. Is this the right thing to do? + bool hasFMA4() const { return HasFMA4 && !HasFMA; } + bool hasXOP() const { return HasXOP; } + bool hasMOVBE() const { return HasMOVBE; } + bool hasRDRAND() const { return HasRDRAND; } + bool hasF16C() const { return HasF16C; } + bool hasFSGSBase() const { return HasFSGSBase; } + bool hasLZCNT() const { return HasLZCNT; } + bool hasBMI() const { return HasBMI; } + bool hasBMI2() const { return HasBMI2; } bool isBTMemSlow() const { return IsBTMemSlow; } bool isUnalignedMemAccessFast() const { return IsUAMemFast; } bool hasVectorUAMem() const { return HasVectorUAMem; } + bool hasCmpxchg16b() const { return HasCmpxchg16b; } + bool useLeaForSP() const { return UseLeaForSP; } + bool hasSlowDivide() const { return HasSlowDivide; } - const Triple &getTargetTriple() const { return TargetTriple; } + bool isAtom() const { return X86ProcFamily == IntelAtom; } - bool isTargetDarwin() const { return TargetTriple.getOS() == Triple::Darwin; } - bool isTargetFreeBSD() const { return TargetTriple.getOS() == Triple::FreeBSD; } - bool isTargetSolaris() const { return TargetTriple.getOS() == Triple::Solaris; } + const Triple &getTargetTriple() const { return TargetTriple; } - // ELF is a reasonably sane default and the only other X86 targets we - // support are Darwin and Windows. Just use "not those". + bool isTargetDarwin() const { return TargetTriple.isOSDarwin(); } + bool isTargetFreeBSD() const { + return TargetTriple.getOS() == Triple::FreeBSD; + } + bool isTargetSolaris() const { + return TargetTriple.getOS() == Triple::Solaris; + } bool isTargetELF() const { - return !isTargetDarwin() && !isTargetWindows() && !isTargetCygMing(); + return (TargetTriple.getEnvironment() == Triple::ELF || + TargetTriple.isOSBinFormatELF()); } bool isTargetLinux() const { return TargetTriple.getOS() == Triple::Linux; } - + bool isTargetNaCl() const { + return TargetTriple.getOS() == Triple::NativeClient; + } + bool isTargetNaCl32() const { return isTargetNaCl() && !is64Bit(); } + bool isTargetNaCl64() const { return isTargetNaCl() && is64Bit(); } bool isTargetWindows() const { return TargetTriple.getOS() == Triple::Win32; } bool isTargetMingw() const { return TargetTriple.getOS() == Triple::MinGW32; } bool isTargetCygwin() const { return TargetTriple.getOS() == Triple::Cygwin; } - bool isTargetCygMing() const { - return isTargetMingw() || isTargetCygwin(); - } - - /// isTargetCOFF - Return true if this is any COFF/Windows target variant. + bool isTargetCygMing() const { return TargetTriple.isOSCygMing(); } bool isTargetCOFF() const { - return isTargetMingw() || isTargetCygwin() || isTargetWindows(); + return (TargetTriple.getEnvironment() != Triple::ELF && + TargetTriple.isOSBinFormatCOFF()); } + bool isTargetEnvMacho() const { return TargetTriple.isEnvironmentMachO(); } bool isTargetWin64() const { - return Is64Bit && (isTargetMingw() || isTargetWindows()); - } - - bool isTargetEnvMacho() const { - return isTargetDarwin() || (TargetTriple.getEnvironment() == Triple::MachO); + // FIXME: x86_64-cygwin has not been released yet. + return In64BitMode && TargetTriple.isOSWindows(); } bool isTargetWin32() const { - return !Is64Bit && (isTargetMingw() || isTargetWindows()); + // FIXME: Cygwin is included for isTargetWin64 -- should it be included + // here too? + return !In64BitMode && (isTargetMingw() || isTargetWindows()); } bool isPICStyleSet() const { return PICStyle != PICStyles::None; } @@ -245,8 +311,16 @@ public: /// should be attempted. unsigned getSpecialAddressLatency() const; - /// IsCalleePop - Test whether a function should pop its own arguments. - bool IsCalleePop(bool isVarArg, CallingConv::ID CallConv) const; + /// enablePostRAScheduler - run for Atom optimization. + bool enablePostRAScheduler(CodeGenOpt::Level OptLevel, + TargetSubtargetInfo::AntiDepBreakMode& Mode, + RegClassVector& CriticalPathRCs) const; + + bool postRAScheduler() const { return PostRAScheduler; } + + /// getInstrItins = Return the instruction itineraries based on the + /// subtarget selection. + const InstrItineraryData &getInstrItineraryData() const { return InstrItins; } }; } // End llvm namespace