-//=====---- X86Subtarget.h - Define Subtarget for the X86 -----*- C++ -*--====//
+//===-- X86Subtarget.h - Define Subtarget for the X86 ----------*- C++ -*--===//
//
// The LLVM Compiler Infrastructure
//
-// This file was developed by Nate Begeman and is distributed under the
-// University of Illinois Open Source License. See LICENSE.TXT for details.
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
-// 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/Target/TargetSubtarget.h"
-
+#include "llvm/CallingConv.h"
+#include "llvm/ADT/Triple.h"
+#include "llvm/Target/TargetSubtargetInfo.h"
#include <string>
+#define GET_SUBTARGETINFO_HEADER
+#include "X86GenSubtargetInfo.inc"
+
namespace llvm {
-class Module;
class GlobalValue;
+class StringRef;
class TargetMachine;
-
-namespace PICStyle {
+
+/// PICStyles - The X86 backend supports a number of different styles of PIC.
+///
+namespace PICStyles {
enum Style {
- Stub, GOT, RIPRel, WinPIC, None
+ StubPIC, // Used on i386-darwin in -fPIC mode.
+ StubDynamicNoPIC, // Used on i386-darwin in -mdynamic-no-pic mode.
+ GOT, // Used on many 32-bit unices in -fPIC mode.
+ RIPRel, // Used on X86-64 when not in -static mode.
+ None // Set when in -static mode (not PIC or DynamicNoPIC mode).
};
}
-class X86Subtarget : public TargetSubtarget {
-public:
- enum AsmWriterFlavorTy {
- ATT, Intel, Unset
- };
+class X86Subtarget : public X86GenSubtargetInfo {
protected:
enum X86SSEEnum {
- NoMMXSSE, MMX, SSE1, SSE2, SSE3
+ NoMMXSSE, MMX, SSE1, SSE2, SSE3, SSSE3, SSE41, SSE42, AVX, AVX2
};
enum X863DNowEnum {
NoThreeDNow, ThreeDNow, ThreeDNowA
};
- /// AsmFlavor - Which x86 asm dialect to use.
- AsmWriterFlavorTy AsmFlavor;
+ enum X86ProcFamilyEnum {
+ Others, IntelAtom
+ };
+
+ /// X86ProcFamily - X86 processor family: Intel Atom, and others
+ X86ProcFamilyEnum X86ProcFamily;
/// PICStyle - Which PIC style to use
- PICStyle::Style PICStyle;
-
- /// X86SSELevel - MMX, SSE1, SSE2, SSE3, or none supported.
+ ///
+ PICStyles::Style PICStyle;
+
+ /// X86SSELevel - MMX, SSE1, SSE2, SSE3, SSSE3, SSE41, SSE42, or
+ /// none supported.
X86SSEEnum X86SSELevel;
/// X863DNowLevel - 3DNow or 3DNow Athlon, or none supported.
+ ///
X863DNowEnum X863DNowLevel;
+ /// HasCMov - True if this processor has conditional move instructions
+ /// (generally pentium pro+).
+ bool HasCMov;
+
/// HasX86_64 - True if the processor supports X86-64 instructions.
+ ///
bool HasX86_64;
+ /// HasPOPCNT - True if the processor supports POPCNT.
+ bool HasPOPCNT;
+
+ /// HasSSE4A - True if the processor supports SSE4A instructions.
+ bool HasSSE4A;
+
+ /// HasAES - Target has AES instructions
+ bool HasAES;
+
+ /// HasPCLMUL - Target has carry-less multiplication
+ bool HasPCLMUL;
+
+ /// 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;
+
+ /// IsUAMemFast - True if unaligned memory access is fast.
+ bool IsUAMemFast;
+
+ /// HasVectorUAMem - True if SIMD operations can have unaligned memory
+ /// 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;
- /// Min. memset / memcpy size that is turned into rep/movs, rep/stos ops.
- unsigned MinRepStrSizeThreshold;
+ /// Max. memset / memcpy size that is turned into rep/movs, rep/stos ops.
+ ///
+ unsigned MaxInlineSizeThreshold;
+
+ /// 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 module
- /// pointer size is 64 bit.
- bool Is64Bit;
+ /// In64BitMode - True if compiling for 64-bit, false for 32-bit.
+ bool In64BitMode;
public:
- enum {
- isELF, isCygwin, isDarwin, isWindows, isMingw
- } TargetType;
/// This constructor initializes the data members to match that
- /// of the specified module.
+ /// of the specified triple.
///
- X86Subtarget(const Module &M, 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
/// function for this subtarget.
unsigned getStackAlignment() const { return stackAlignment; }
- /// getMinRepStrSizeThreshold - Returns the minimum memset / memcpy size
- /// required to turn the operation into a X86 rep/movs or rep/stos
- /// instruction. This is only used if the src / dst alignment is not DWORD
- /// aligned.
- unsigned getMinRepStrSizeThreshold() const { return MinRepStrSizeThreshold; }
+ /// getMaxInlineSizeThreshold - Returns the maximum memset / memcpy size
+ /// that still makes it profitable to inline the call.
+ unsigned getMaxInlineSizeThreshold() const { return MaxInlineSizeThreshold; }
/// ParseSubtargetFeatures - Parses features string setting specified
/// subtarget options. Definition of function is auto generated by tblgen.
- void 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; }
- PICStyle::Style getPICStyle() const { return PICStyle; }
- void setPICStyle(PICStyle::Style Style) { PICStyle = Style; }
+ PICStyles::Style getPICStyle() const { return PICStyle; }
+ void setPICStyle(PICStyles::Style Style) { PICStyle = Style; }
+ bool hasCMov() const { return HasCMov; }
bool hasMMX() const { return X86SSELevel >= MMX; }
bool hasSSE1() const { return X86SSELevel >= SSE1; }
bool hasSSE2() const { return X86SSELevel >= SSE2; }
bool hasSSE3() const { return X86SSELevel >= SSE3; }
+ 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 hasAES() const { return HasAES; }
+ 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; }
- bool isFlavorAtt() const { return AsmFlavor == ATT; }
- bool isFlavorIntel() const { return AsmFlavor == Intel; }
-
- bool isTargetDarwin() const { return TargetType == isDarwin; }
- bool isTargetELF() const { return TargetType == isELF; }
- bool isTargetWindows() const { return TargetType == isWindows; }
- bool isTargetMingw() const { return TargetType == isMingw; }
- bool isTargetCygMing() const { return (TargetType == isMingw ||
- TargetType == isCygwin); }
- bool isTargetCygwin() const { return TargetType == isCygwin; }
-
- bool isPICStyleSet() const { return PICStyle != PICStyle::None; }
- bool isPICStyleGOT() const { return PICStyle == PICStyle::GOT; }
- bool isPICStyleStub() const { return PICStyle == PICStyle::Stub; }
- bool isPICStyleRIPRel() const { return PICStyle == PICStyle::RIPRel; }
- bool isPICStyleWinPIC() const { return PICStyle == PICStyle:: WinPIC; }
-
- /// True if accessing the GV requires an extra load. For Windows, dllimported
- /// symbols are indirect, loading the value at address GV rather then the
- /// value of GV itself. This means that the GlobalAddress must be in the base
- /// or index register of the address, not the GV offset field.
- bool GVRequiresExtraLoad(const GlobalValue* GV, const TargetMachine& TM,
- bool isDirectCall) const;
+ bool isAtom() const { return X86ProcFamily == IntelAtom; }
-};
+ const Triple &getTargetTriple() const { return TargetTriple; }
-namespace X86 {
- /// GetCpuIDAndInfo - Execute the specified cpuid and return the 4 values in
- /// the specified arguments. If we can't run cpuid on the host, return true.
- bool GetCpuIDAndInfo(unsigned value, unsigned *rEAX, unsigned *rEBX,
- unsigned *rECX, unsigned *rEDX);
-}
+ 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 (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 TargetTriple.isOSCygMing(); }
+ bool isTargetCOFF() const {
+ return (TargetTriple.getEnvironment() != Triple::ELF &&
+ TargetTriple.isOSBinFormatCOFF());
+ }
+ bool isTargetEnvMacho() const { return TargetTriple.isEnvironmentMachO(); }
+
+ bool isTargetWin64() const {
+ // FIXME: x86_64-cygwin has not been released yet.
+ return In64BitMode && TargetTriple.isOSWindows();
+ }
+
+ bool isTargetWin32() const {
+ // FIXME: Cygwin is included for isTargetWin64 -- should it be included
+ // here too?
+ return !In64BitMode && (isTargetMingw() || isTargetWindows());
+ }
+
+ bool isPICStyleSet() const { return PICStyle != PICStyles::None; }
+ bool isPICStyleGOT() const { return PICStyle == PICStyles::GOT; }
+ bool isPICStyleRIPRel() const { return PICStyle == PICStyles::RIPRel; }
+
+ bool isPICStyleStubPIC() const {
+ return PICStyle == PICStyles::StubPIC;
+ }
+
+ bool isPICStyleStubNoDynamic() const {
+ return PICStyle == PICStyles::StubDynamicNoPIC;
+ }
+ bool isPICStyleStubAny() const {
+ return PICStyle == PICStyles::StubDynamicNoPIC ||
+ PICStyle == PICStyles::StubPIC; }
+
+ /// ClassifyGlobalReference - Classify a global variable reference for the
+ /// current subtarget according to how we should reference it in a non-pcrel
+ /// context.
+ unsigned char ClassifyGlobalReference(const GlobalValue *GV,
+ const TargetMachine &TM)const;
+
+ /// ClassifyBlockAddressReference - Classify a blockaddress reference for the
+ /// current subtarget according to how we should reference it in a non-pcrel
+ /// context.
+ unsigned char ClassifyBlockAddressReference() const;
+
+ /// IsLegalToCallImmediateAddr - Return true if the subtarget allows calls
+ /// to immediate address.
+ bool IsLegalToCallImmediateAddr(const TargetMachine &TM) const;
+
+ /// This function returns the name of a function which has an interface
+ /// like the non-standard bzero function, if such a function exists on
+ /// the current subtarget and it is considered prefereable over
+ /// memset with zero passed as the second argument. Otherwise it
+ /// returns null.
+ const char *getBZeroEntry() const;
+
+ /// getSpecialAddressLatency - For targets where it is beneficial to
+ /// backschedule instructions that compute addresses, return a value
+ /// indicating the number of scheduling cycles of backscheduling that
+ /// should be attempted.
+ unsigned getSpecialAddressLatency() 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