X-Git-Url: http://demsky.eecs.uci.edu/git/?a=blobdiff_plain;f=lib%2FTarget%2FX86%2FX86RegisterInfo.td;h=a5e3fc25d267664a866e9b922dac905378514e30;hb=163e83d69a22916b34c06ae50a1540e90b82becb;hp=c40dcb935500b08d1b1be185bf90cca936d1bd1c;hpb=a3231ba237b1a5113a71253eab50c6b7a5239132;p=oota-llvm.git diff --git a/lib/Target/X86/X86RegisterInfo.td b/lib/Target/X86/X86RegisterInfo.td index c40dcb93550..a5e3fc25d26 100644 --- a/lib/Target/X86/X86RegisterInfo.td +++ b/lib/Target/X86/X86RegisterInfo.td @@ -1,9 +1,9 @@ -//===- X86RegisterInfo.td - Describe the X86 Register File ------*- C++ -*-===// +//===- X86RegisterInfo.td - Describe the X86 Register File --*- tablegen -*-==// // // The LLVM Compiler Infrastructure // -// This file was developed by the LLVM research group 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. // //===----------------------------------------------------------------------===// // @@ -18,193 +18,243 @@ // let Namespace = "X86" in { + // Subregister indices. + def sub_8bit : SubRegIndex; + def sub_8bit_hi : SubRegIndex; + def sub_16bit : SubRegIndex; + def sub_32bit : SubRegIndex; + + def sub_ss : SubRegIndex; + def sub_sd : SubRegIndex; + def sub_xmm : SubRegIndex; + + // In the register alias definitions below, we define which registers alias // which others. We only specify which registers the small registers alias, // because the register file generator is smart enough to figure out that // AL aliases AX if we tell it that AX aliased AL (for example). - // FIXME: X86-64 have different Dwarf numbers. + // Dwarf numbering is different for 32-bit and 64-bit, and there are + // variations by target as well. Currently the first entry is for X86-64, + // second - for EH on X86-32/Darwin and third is 'generic' one (X86-32/Linux + // and debug information on X86-32/Darwin) + // 8-bit registers // Low registers - def AL : Register<"AL">, DwarfRegNum<0>; - def CL : Register<"CL">, DwarfRegNum<1>; - def DL : Register<"DL">, DwarfRegNum<2>; - def BL : Register<"BL">, DwarfRegNum<3>; + def AL : Register<"al">, DwarfRegNum<[0, 0, 0]>; + def DL : Register<"dl">, DwarfRegNum<[1, 2, 2]>; + def CL : Register<"cl">, DwarfRegNum<[2, 1, 1]>; + def BL : Register<"bl">, DwarfRegNum<[3, 3, 3]>; // X86-64 only - def SIL : Register<"SIL">, DwarfRegNum<4>; - def DIL : Register<"DIL">, DwarfRegNum<5>; - def BPL : Register<"BPL">, DwarfRegNum<6>; - def SPL : Register<"SPL">, DwarfRegNum<7>; - def R8B : Register<"R8B">, DwarfRegNum<8>; - def R9B : Register<"R9B">, DwarfRegNum<9>; - def R10B : Register<"R10B">, DwarfRegNum<10>; - def R11B : Register<"R11B">, DwarfRegNum<11>; - def R12B : Register<"R12B">, DwarfRegNum<12>; - def R13B : Register<"R13B">, DwarfRegNum<13>; - def R14B : Register<"R14B">, DwarfRegNum<14>; - def R15B : Register<"R15B">, DwarfRegNum<15>; - - // High registers X86-32 only - def AH : Register<"AH">, DwarfRegNum<0>; - def CH : Register<"CH">, DwarfRegNum<1>; - def DH : Register<"DH">, DwarfRegNum<2>; - def BH : Register<"BH">, DwarfRegNum<3>; + def SIL : Register<"sil">, DwarfRegNum<[4, 6, 6]>; + def DIL : Register<"dil">, DwarfRegNum<[5, 7, 7]>; + def BPL : Register<"bpl">, DwarfRegNum<[6, 4, 5]>; + def SPL : Register<"spl">, DwarfRegNum<[7, 5, 4]>; + def R8B : Register<"r8b">, DwarfRegNum<[8, -2, -2]>; + def R9B : Register<"r9b">, DwarfRegNum<[9, -2, -2]>; + def R10B : Register<"r10b">, DwarfRegNum<[10, -2, -2]>; + def R11B : Register<"r11b">, DwarfRegNum<[11, -2, -2]>; + def R12B : Register<"r12b">, DwarfRegNum<[12, -2, -2]>; + def R13B : Register<"r13b">, DwarfRegNum<[13, -2, -2]>; + def R14B : Register<"r14b">, DwarfRegNum<[14, -2, -2]>; + def R15B : Register<"r15b">, DwarfRegNum<[15, -2, -2]>; + + // High registers. On x86-64, these cannot be used in any instruction + // with a REX prefix. + def AH : Register<"ah">, DwarfRegNum<[0, 0, 0]>; + def DH : Register<"dh">, DwarfRegNum<[1, 2, 2]>; + def CH : Register<"ch">, DwarfRegNum<[2, 1, 1]>; + def BH : Register<"bh">, DwarfRegNum<[3, 3, 3]>; // 16-bit registers - def AX : RegisterWithSubRegs<"AX", [AH,AL]>, DwarfRegNum<0>; - def CX : RegisterWithSubRegs<"CX", [CH,CL]>, DwarfRegNum<1>; - def DX : RegisterWithSubRegs<"DX", [DH,DL]>, DwarfRegNum<2>; - def BX : RegisterWithSubRegs<"BX", [BH,BL]>, DwarfRegNum<3>; - def SP : RegisterWithSubRegs<"SP", [SPL]>, DwarfRegNum<4>; - def BP : RegisterWithSubRegs<"BP", [BPL]>, DwarfRegNum<5>; - def SI : RegisterWithSubRegs<"SI", [SIL]>, DwarfRegNum<6>; - def DI : RegisterWithSubRegs<"DI", [DIL]>, DwarfRegNum<7>; - def IP : Register<"IP">, DwarfRegNum<8>; + let SubRegIndices = [sub_8bit, sub_8bit_hi] in { + def AX : RegisterWithSubRegs<"ax", [AL,AH]>, DwarfRegNum<[0, 0, 0]>; + def DX : RegisterWithSubRegs<"dx", [DL,DH]>, DwarfRegNum<[1, 2, 2]>; + def CX : RegisterWithSubRegs<"cx", [CL,CH]>, DwarfRegNum<[2, 1, 1]>; + def BX : RegisterWithSubRegs<"bx", [BL,BH]>, DwarfRegNum<[3, 3, 3]>; + } + let SubRegIndices = [sub_8bit] in { + def SI : RegisterWithSubRegs<"si", [SIL]>, DwarfRegNum<[4, 6, 6]>; + def DI : RegisterWithSubRegs<"di", [DIL]>, DwarfRegNum<[5, 7, 7]>; + def BP : RegisterWithSubRegs<"bp", [BPL]>, DwarfRegNum<[6, 4, 5]>; + def SP : RegisterWithSubRegs<"sp", [SPL]>, DwarfRegNum<[7, 5, 4]>; + } + def IP : Register<"ip">, DwarfRegNum<[16]>; // X86-64 only - def R8W : RegisterWithSubRegs<"R8W", [R8B]>, DwarfRegNum<8>; - def R9W : RegisterWithSubRegs<"R9W", [R9B]>, DwarfRegNum<9>; - def R10W : RegisterWithSubRegs<"R10W", [R10B]>, DwarfRegNum<10>; - def R11W : RegisterWithSubRegs<"R11W", [R11B]>, DwarfRegNum<11>; - def R12W : RegisterWithSubRegs<"R12W", [R12B]>, DwarfRegNum<12>; - def R13W : RegisterWithSubRegs<"R13W", [R13B]>, DwarfRegNum<13>; - def R14W : RegisterWithSubRegs<"R14W", [R14B]>, DwarfRegNum<14>; - def R15W : RegisterWithSubRegs<"R15W", [R15B]>, DwarfRegNum<15>; - + let SubRegIndices = [sub_8bit] in { + def R8W : RegisterWithSubRegs<"r8w", [R8B]>, DwarfRegNum<[8, -2, -2]>; + def R9W : RegisterWithSubRegs<"r9w", [R9B]>, DwarfRegNum<[9, -2, -2]>; + def R10W : RegisterWithSubRegs<"r10w", [R10B]>, DwarfRegNum<[10, -2, -2]>; + def R11W : RegisterWithSubRegs<"r11w", [R11B]>, DwarfRegNum<[11, -2, -2]>; + def R12W : RegisterWithSubRegs<"r12w", [R12B]>, DwarfRegNum<[12, -2, -2]>; + def R13W : RegisterWithSubRegs<"r13w", [R13B]>, DwarfRegNum<[13, -2, -2]>; + def R14W : RegisterWithSubRegs<"r14w", [R14B]>, DwarfRegNum<[14, -2, -2]>; + def R15W : RegisterWithSubRegs<"r15w", [R15B]>, DwarfRegNum<[15, -2, -2]>; + } // 32-bit registers - def EAX : RegisterWithSubRegs<"EAX", [AX]>, DwarfRegNum<0>; - def ECX : RegisterWithSubRegs<"ECX", [CX]>, DwarfRegNum<1>; - def EDX : RegisterWithSubRegs<"EDX", [DX]>, DwarfRegNum<2>; - def EBX : RegisterWithSubRegs<"EBX", [BX]>, DwarfRegNum<3>; - def ESP : RegisterWithSubRegs<"ESP", [SP]>, DwarfRegNum<4>; - def EBP : RegisterWithSubRegs<"EBP", [BP]>, DwarfRegNum<5>; - def ESI : RegisterWithSubRegs<"ESI", [SI]>, DwarfRegNum<6>; - def EDI : RegisterWithSubRegs<"EDI", [DI]>, DwarfRegNum<7>; - def EIP : RegisterWithSubRegs<"EIP", [IP]>, DwarfRegNum<8>; + let SubRegIndices = [sub_16bit] in { + def EAX : RegisterWithSubRegs<"eax", [AX]>, DwarfRegNum<[0, 0, 0]>; + def EDX : RegisterWithSubRegs<"edx", [DX]>, DwarfRegNum<[1, 2, 2]>; + def ECX : RegisterWithSubRegs<"ecx", [CX]>, DwarfRegNum<[2, 1, 1]>; + def EBX : RegisterWithSubRegs<"ebx", [BX]>, DwarfRegNum<[3, 3, 3]>; + def ESI : RegisterWithSubRegs<"esi", [SI]>, DwarfRegNum<[4, 6, 6]>; + def EDI : RegisterWithSubRegs<"edi", [DI]>, DwarfRegNum<[5, 7, 7]>; + def EBP : RegisterWithSubRegs<"ebp", [BP]>, DwarfRegNum<[6, 4, 5]>; + def ESP : RegisterWithSubRegs<"esp", [SP]>, DwarfRegNum<[7, 5, 4]>; + def EIP : RegisterWithSubRegs<"eip", [IP]>, DwarfRegNum<[16, 8, 8]>; // X86-64 only - def R8D : RegisterWithSubRegs<"R8D", [R8W]>, DwarfRegNum<8>; - def R9D : RegisterWithSubRegs<"R9D", [R9W]>, DwarfRegNum<9>; - def R10D : RegisterWithSubRegs<"R10D", [R10W]>, DwarfRegNum<10>; - def R11D : RegisterWithSubRegs<"R11D", [R11W]>, DwarfRegNum<11>; - def R12D : RegisterWithSubRegs<"R12D", [R12W]>, DwarfRegNum<12>; - def R13D : RegisterWithSubRegs<"R13D", [R13W]>, DwarfRegNum<13>; - def R14D : RegisterWithSubRegs<"R14D", [R14W]>, DwarfRegNum<14>; - def R15D : RegisterWithSubRegs<"R15D", [R15W]>, DwarfRegNum<15>; + def R8D : RegisterWithSubRegs<"r8d", [R8W]>, DwarfRegNum<[8, -2, -2]>; + def R9D : RegisterWithSubRegs<"r9d", [R9W]>, DwarfRegNum<[9, -2, -2]>; + def R10D : RegisterWithSubRegs<"r10d", [R10W]>, DwarfRegNum<[10, -2, -2]>; + def R11D : RegisterWithSubRegs<"r11d", [R11W]>, DwarfRegNum<[11, -2, -2]>; + def R12D : RegisterWithSubRegs<"r12d", [R12W]>, DwarfRegNum<[12, -2, -2]>; + def R13D : RegisterWithSubRegs<"r13d", [R13W]>, DwarfRegNum<[13, -2, -2]>; + def R14D : RegisterWithSubRegs<"r14d", [R14W]>, DwarfRegNum<[14, -2, -2]>; + def R15D : RegisterWithSubRegs<"r15d", [R15W]>, DwarfRegNum<[15, -2, -2]>; + } // 64-bit registers, X86-64 only - def RAX : RegisterWithSubRegs<"RAX", [EAX]>, DwarfRegNum<0>; - def RDX : RegisterWithSubRegs<"RDX", [EDX]>, DwarfRegNum<1>; - def RCX : RegisterWithSubRegs<"RCX", [ECX]>, DwarfRegNum<2>; - def RBX : RegisterWithSubRegs<"RBX", [EBX]>, DwarfRegNum<3>; - def RSI : RegisterWithSubRegs<"RSI", [ESI]>, DwarfRegNum<4>; - def RDI : RegisterWithSubRegs<"RDI", [EDI]>, DwarfRegNum<5>; - def RBP : RegisterWithSubRegs<"RBP", [EBP]>, DwarfRegNum<6>; - def RSP : RegisterWithSubRegs<"RSP", [ESP]>, DwarfRegNum<7>; - - def R8 : RegisterWithSubRegs<"R8", [R8D]>, DwarfRegNum<8>; - def R9 : RegisterWithSubRegs<"R9", [R9D]>, DwarfRegNum<9>; - def R10 : RegisterWithSubRegs<"R10", [R10D]>, DwarfRegNum<10>; - def R11 : RegisterWithSubRegs<"R11", [R11D]>, DwarfRegNum<11>; - def R12 : RegisterWithSubRegs<"R12", [R12D]>, DwarfRegNum<12>; - def R13 : RegisterWithSubRegs<"R13", [R13D]>, DwarfRegNum<13>; - def R14 : RegisterWithSubRegs<"R14", [R14D]>, DwarfRegNum<14>; - def R15 : RegisterWithSubRegs<"R15", [R15D]>, DwarfRegNum<15>; - def RIP : RegisterWithSubRegs<"RIP", [EIP]>, DwarfRegNum<16>; + let SubRegIndices = [sub_32bit] in { + def RAX : RegisterWithSubRegs<"rax", [EAX]>, DwarfRegNum<[0, -2, -2]>; + def RDX : RegisterWithSubRegs<"rdx", [EDX]>, DwarfRegNum<[1, -2, -2]>; + def RCX : RegisterWithSubRegs<"rcx", [ECX]>, DwarfRegNum<[2, -2, -2]>; + def RBX : RegisterWithSubRegs<"rbx", [EBX]>, DwarfRegNum<[3, -2, -2]>; + def RSI : RegisterWithSubRegs<"rsi", [ESI]>, DwarfRegNum<[4, -2, -2]>; + def RDI : RegisterWithSubRegs<"rdi", [EDI]>, DwarfRegNum<[5, -2, -2]>; + def RBP : RegisterWithSubRegs<"rbp", [EBP]>, DwarfRegNum<[6, -2, -2]>; + def RSP : RegisterWithSubRegs<"rsp", [ESP]>, DwarfRegNum<[7, -2, -2]>; + + def R8 : RegisterWithSubRegs<"r8", [R8D]>, DwarfRegNum<[8, -2, -2]>; + def R9 : RegisterWithSubRegs<"r9", [R9D]>, DwarfRegNum<[9, -2, -2]>; + def R10 : RegisterWithSubRegs<"r10", [R10D]>, DwarfRegNum<[10, -2, -2]>; + def R11 : RegisterWithSubRegs<"r11", [R11D]>, DwarfRegNum<[11, -2, -2]>; + def R12 : RegisterWithSubRegs<"r12", [R12D]>, DwarfRegNum<[12, -2, -2]>; + def R13 : RegisterWithSubRegs<"r13", [R13D]>, DwarfRegNum<[13, -2, -2]>; + def R14 : RegisterWithSubRegs<"r14", [R14D]>, DwarfRegNum<[14, -2, -2]>; + def R15 : RegisterWithSubRegs<"r15", [R15D]>, DwarfRegNum<[15, -2, -2]>; + def RIP : RegisterWithSubRegs<"rip", [EIP]>, DwarfRegNum<[16, -2, -2]>; + } // MMX Registers. These are actually aliased to ST0 .. ST7 - def MM0 : Register<"MM0">, DwarfRegNum<29>; - def MM1 : Register<"MM1">, DwarfRegNum<30>; - def MM2 : Register<"MM2">, DwarfRegNum<31>; - def MM3 : Register<"MM3">, DwarfRegNum<32>; - def MM4 : Register<"MM4">, DwarfRegNum<33>; - def MM5 : Register<"MM5">, DwarfRegNum<34>; - def MM6 : Register<"MM6">, DwarfRegNum<35>; - def MM7 : Register<"MM7">, DwarfRegNum<36>; - + def MM0 : Register<"mm0">, DwarfRegNum<[41, 29, 29]>; + def MM1 : Register<"mm1">, DwarfRegNum<[42, 30, 30]>; + def MM2 : Register<"mm2">, DwarfRegNum<[43, 31, 31]>; + def MM3 : Register<"mm3">, DwarfRegNum<[44, 32, 32]>; + def MM4 : Register<"mm4">, DwarfRegNum<[45, 33, 33]>; + def MM5 : Register<"mm5">, DwarfRegNum<[46, 34, 34]>; + def MM6 : Register<"mm6">, DwarfRegNum<[47, 35, 35]>; + def MM7 : Register<"mm7">, DwarfRegNum<[48, 36, 36]>; + // Pseudo Floating Point registers - def FP0 : Register<"FP0">, DwarfRegNum<-1>; - def FP1 : Register<"FP1">, DwarfRegNum<-1>; - def FP2 : Register<"FP2">, DwarfRegNum<-1>; - def FP3 : Register<"FP3">, DwarfRegNum<-1>; - def FP4 : Register<"FP4">, DwarfRegNum<-1>; - def FP5 : Register<"FP5">, DwarfRegNum<-1>; - def FP6 : Register<"FP6">, DwarfRegNum<-1>; - - // XMM Registers, used by the various SSE instruction set extensions - def XMM0: Register<"XMM0">, DwarfRegNum<17>; - def XMM1: Register<"XMM1">, DwarfRegNum<18>; - def XMM2: Register<"XMM2">, DwarfRegNum<19>; - def XMM3: Register<"XMM3">, DwarfRegNum<20>; - def XMM4: Register<"XMM4">, DwarfRegNum<21>; - def XMM5: Register<"XMM5">, DwarfRegNum<22>; - def XMM6: Register<"XMM6">, DwarfRegNum<23>; - def XMM7: Register<"XMM7">, DwarfRegNum<24>; + def FP0 : Register<"fp0">; + def FP1 : Register<"fp1">; + def FP2 : Register<"fp2">; + def FP3 : Register<"fp3">; + def FP4 : Register<"fp4">; + def FP5 : Register<"fp5">; + def FP6 : Register<"fp6">; + + // XMM Registers, used by the various SSE instruction set extensions. + // The sub_ss and sub_sd subregs are the same registers with another regclass. + let CompositeIndices = [(sub_ss), (sub_sd)] in { + def XMM0: Register<"xmm0">, DwarfRegNum<[17, 21, 21]>; + def XMM1: Register<"xmm1">, DwarfRegNum<[18, 22, 22]>; + def XMM2: Register<"xmm2">, DwarfRegNum<[19, 23, 23]>; + def XMM3: Register<"xmm3">, DwarfRegNum<[20, 24, 24]>; + def XMM4: Register<"xmm4">, DwarfRegNum<[21, 25, 25]>; + def XMM5: Register<"xmm5">, DwarfRegNum<[22, 26, 26]>; + def XMM6: Register<"xmm6">, DwarfRegNum<[23, 27, 27]>; + def XMM7: Register<"xmm7">, DwarfRegNum<[24, 28, 28]>; // X86-64 only - def XMM8: Register<"XMM8">, DwarfRegNum<25>; - def XMM9: Register<"XMM9">, DwarfRegNum<26>; - def XMM10: Register<"XMM10">, DwarfRegNum<27>; - def XMM11: Register<"XMM11">, DwarfRegNum<28>; - def XMM12: Register<"XMM12">, DwarfRegNum<29>; - def XMM13: Register<"XMM13">, DwarfRegNum<30>; - def XMM14: Register<"XMM14">, DwarfRegNum<31>; - def XMM15: Register<"XMM15">, DwarfRegNum<32>; + def XMM8: Register<"xmm8">, DwarfRegNum<[25, -2, -2]>; + def XMM9: Register<"xmm9">, DwarfRegNum<[26, -2, -2]>; + def XMM10: Register<"xmm10">, DwarfRegNum<[27, -2, -2]>; + def XMM11: Register<"xmm11">, DwarfRegNum<[28, -2, -2]>; + def XMM12: Register<"xmm12">, DwarfRegNum<[29, -2, -2]>; + def XMM13: Register<"xmm13">, DwarfRegNum<[30, -2, -2]>; + def XMM14: Register<"xmm14">, DwarfRegNum<[31, -2, -2]>; + def XMM15: Register<"xmm15">, DwarfRegNum<[32, -2, -2]>; + } + + // YMM Registers, used by AVX instructions + let SubRegIndices = [sub_xmm] in { + def YMM0: RegisterWithSubRegs<"ymm0", [XMM0]>, DwarfRegNum<[17, 21, 21]>; + def YMM1: RegisterWithSubRegs<"ymm1", [XMM1]>, DwarfRegNum<[18, 22, 22]>; + def YMM2: RegisterWithSubRegs<"ymm2", [XMM2]>, DwarfRegNum<[19, 23, 23]>; + def YMM3: RegisterWithSubRegs<"ymm3", [XMM3]>, DwarfRegNum<[20, 24, 24]>; + def YMM4: RegisterWithSubRegs<"ymm4", [XMM4]>, DwarfRegNum<[21, 25, 25]>; + def YMM5: RegisterWithSubRegs<"ymm5", [XMM5]>, DwarfRegNum<[22, 26, 26]>; + def YMM6: RegisterWithSubRegs<"ymm6", [XMM6]>, DwarfRegNum<[23, 27, 27]>; + def YMM7: RegisterWithSubRegs<"ymm7", [XMM7]>, DwarfRegNum<[24, 28, 28]>; + def YMM8: RegisterWithSubRegs<"ymm8", [XMM8]>, DwarfRegNum<[25, -2, -2]>; + def YMM9: RegisterWithSubRegs<"ymm9", [XMM9]>, DwarfRegNum<[26, -2, -2]>; + def YMM10: RegisterWithSubRegs<"ymm10", [XMM10]>, DwarfRegNum<[27, -2, -2]>; + def YMM11: RegisterWithSubRegs<"ymm11", [XMM11]>, DwarfRegNum<[28, -2, -2]>; + def YMM12: RegisterWithSubRegs<"ymm12", [XMM12]>, DwarfRegNum<[29, -2, -2]>; + def YMM13: RegisterWithSubRegs<"ymm13", [XMM13]>, DwarfRegNum<[30, -2, -2]>; + def YMM14: RegisterWithSubRegs<"ymm14", [XMM14]>, DwarfRegNum<[31, -2, -2]>; + def YMM15: RegisterWithSubRegs<"ymm15", [XMM15]>, DwarfRegNum<[32, -2, -2]>; + } // Floating point stack registers - def ST0 : Register<"ST(0)">, DwarfRegNum<11>; - def ST1 : Register<"ST(1)">, DwarfRegNum<12>; - def ST2 : Register<"ST(2)">, DwarfRegNum<13>; - def ST3 : Register<"ST(3)">, DwarfRegNum<14>; - def ST4 : Register<"ST(4)">, DwarfRegNum<15>; - def ST5 : Register<"ST(5)">, DwarfRegNum<16>; - def ST6 : Register<"ST(6)">, DwarfRegNum<17>; - def ST7 : Register<"ST(7)">, DwarfRegNum<18>; + def ST0 : Register<"st(0)">, DwarfRegNum<[33, 12, 11]>; + def ST1 : Register<"st(1)">, DwarfRegNum<[34, 13, 12]>; + def ST2 : Register<"st(2)">, DwarfRegNum<[35, 14, 13]>; + def ST3 : Register<"st(3)">, DwarfRegNum<[36, 15, 14]>; + def ST4 : Register<"st(4)">, DwarfRegNum<[37, 16, 15]>; + def ST5 : Register<"st(5)">, DwarfRegNum<[38, 17, 16]>; + def ST6 : Register<"st(6)">, DwarfRegNum<[39, 18, 17]>; + def ST7 : Register<"st(7)">, DwarfRegNum<[40, 19, 18]>; + + // Status flags register + def EFLAGS : Register<"flags">; + + // Segment registers + def CS : Register<"cs">; + def DS : Register<"ds">; + def SS : Register<"ss">; + def ES : Register<"es">; + def FS : Register<"fs">; + def GS : Register<"gs">; + + // Debug registers + def DR0 : Register<"dr0">; + def DR1 : Register<"dr1">; + def DR2 : Register<"dr2">; + def DR3 : Register<"dr3">; + def DR4 : Register<"dr4">; + def DR5 : Register<"dr5">; + def DR6 : Register<"dr6">; + def DR7 : Register<"dr7">; + + // Control registers + def CR0 : Register<"cr0">; + def CR1 : Register<"cr1">; + def CR2 : Register<"cr2">; + def CR3 : Register<"cr3">; + def CR4 : Register<"cr4">; + def CR5 : Register<"cr5">; + def CR6 : Register<"cr6">; + def CR7 : Register<"cr7">; + def CR8 : Register<"cr8">; + def CR9 : Register<"cr9">; + def CR10 : Register<"cr10">; + def CR11 : Register<"cr11">; + def CR12 : Register<"cr12">; + def CR13 : Register<"cr13">; + def CR14 : Register<"cr14">; + def CR15 : Register<"cr15">; + + // Pseudo index registers + def EIZ : Register<"eiz">; + def RIZ : Register<"riz">; } -//===----------------------------------------------------------------------===// -// Subregister Set Definitions... now that we have all of the pieces, define the -// sub registers for each register. -// - -def : SubRegSet<1, [AX, CX, DX, BX, SP, BP, SI, DI, - R8W, R9W, R10W, R11W, R12W, R13W, R14W, R15W], - [AL, CL, DL, BL, SPL, BPL, SIL, DIL, - R8B, R9B, R10B, R11B, R12B, R13B, R14B, R15B]>; - -// It's unclear if this subreg set is safe, given that not all registers -// in the class have an 'H' subreg. -// def : SubRegSet<2, [AX, CX, DX, BX], -// [AH, CH, DH, BH]>; - -def : SubRegSet<1, [EAX, ECX, EDX, EBX, ESP, EBP, ESI, EDI, - R8D, R9D, R10D, R11D, R12D, R13D, R14D, R15D], - [AL, CL, DL, BL, SPL, BPL, SIL, DIL, - R8B, R9B, R10B, R11B, R12B, R13B, R14B, R15B]>; - -def : SubRegSet<2, [EAX, ECX, EDX, EBX, ESP, EBP, ESI, EDI, - R8D, R9D, R10D, R11D, R12D, R13D, R14D, R15D], - [AX, CX, DX, BX, SP, BP, SI, DI, - R8W, R9W, R10W, R11W, R12W, R13W, R14W, R15W]>; - - -def : SubRegSet<1, [RAX, RCX, RDX, RBX, RSP, RBP, RSI, RDI, - R8, R9, R10, R11, R12, R13, R14, R15], - [AL, CL, DL, BL, SPL, BPL, SIL, DIL, - R8B, R9B, R10B, R11B, R12B, R13B, R14B, R15B]>; - -def : SubRegSet<2, [RAX, RCX, RDX, RBX, RSP, RBP, RSI, RDI, - R8, R9, R10, R11, R12, R13, R14, R15], - [AX, CX, DX, BX, SP, BP, SI, DI, - R8W, R9W, R10W, R11W, R12W, R13W, R14W, R15W]>; - -def : SubRegSet<3, [RAX, RCX, RDX, RBX, RSP, RBP, RSI, RDI, - R8, R9, R10, R11, R12, R13, R14, R15], - [EAX, ECX, EDX, EBX, ESP, EBP, ESI, EDI, - R8D, R9D, R10D, R11D, R12D, R13D, R14D, R15D]>; - //===----------------------------------------------------------------------===// // Register Class Definitions... now that we have all of the pieces, define the // top-level register classes. The order specified in the register list is @@ -215,191 +265,170 @@ def : SubRegSet<3, [RAX, RCX, RDX, RBX, RSP, RBP, RSI, RDI, // R12, R13, R14, and R15 for X86-64) are callee-save registers. // In 64-mode, there are 12 additional i8 registers, SIL, DIL, BPL, SPL, and // R8B, ... R15B. -// FIXME: Allow AH, CH, DH, BH in 64-mode for non-REX instructions, +// Allocate R12 and R13 last, as these require an extra byte when +// encoded in x86_64 instructions. +// FIXME: Allow AH, CH, DH, BH to be used as general-purpose registers in +// 64-bit mode. The main complication is that they cannot be encoded in an +// instruction requiring a REX prefix, while SIL, DIL, BPL, R8D, etc. +// require a REX prefix. For example, "addb %ah, %dil" and "movzbl %ah, %r8d" +// cannot be encoded. def GR8 : RegisterClass<"X86", [i8], 8, - [AL, CL, DL, BL, AH, CH, DH, BH, SIL, DIL, BPL, SPL, - R8B, R9B, R10B, R11B, R12B, R13B, R14B, R15B]> { + [AL, CL, DL, AH, CH, DH, BL, BH, SIL, DIL, BPL, SPL, + R8B, R9B, R10B, R11B, R14B, R15B, R12B, R13B]> { let MethodProtos = [{ iterator allocation_order_begin(const MachineFunction &MF) const; iterator allocation_order_end(const MachineFunction &MF) const; }]; let MethodBodies = [{ - // Does the function dedicate RBP / EBP to being a frame ptr? - // If so, don't allocate SPL or BPL. - static const unsigned X86_GR8_AO_64_fp[] = - {X86::AL, X86::CL, X86::DL, X86::SIL, X86::DIL, - X86::R8B, X86::R9B, X86::R10B, X86::R11B, - X86::BL, X86::R14B, X86::R15B, X86::R12B, X86::R13B}; - // If not, just don't allocate SPL. - static const unsigned X86_GR8_AO_64[] = - {X86::AL, X86::CL, X86::DL, X86::SIL, X86::DIL, - X86::R8B, X86::R9B, X86::R10B, X86::R11B, - X86::BL, X86::R14B, X86::R15B, X86::R12B, X86::R13B, X86::BPL}; - // In 32-mode, none of the 8-bit registers aliases EBP or ESP. - static const unsigned X86_GR8_AO_32[] = - {X86::AL, X86::CL, X86::DL, X86::AH, X86::CH, X86::DH, X86::BL, X86::BH}; + static const unsigned X86_GR8_AO_64[] = { + X86::AL, X86::CL, X86::DL, X86::SIL, X86::DIL, + X86::R8B, X86::R9B, X86::R10B, X86::R11B, + X86::BL, X86::R14B, X86::R15B, X86::R12B, X86::R13B, X86::BPL + }; GR8Class::iterator GR8Class::allocation_order_begin(const MachineFunction &MF) const { const TargetMachine &TM = MF.getTarget(); - const MRegisterInfo *RI = TM.getRegisterInfo(); const X86Subtarget &Subtarget = TM.getSubtarget(); - if (!Subtarget.is64Bit()) - return X86_GR8_AO_32; - else if (RI->hasFP(MF)) - return X86_GR8_AO_64_fp; - else + if (Subtarget.is64Bit()) return X86_GR8_AO_64; + else + return begin(); } GR8Class::iterator GR8Class::allocation_order_end(const MachineFunction &MF) const { const TargetMachine &TM = MF.getTarget(); - const MRegisterInfo *RI = TM.getRegisterInfo(); + const TargetFrameInfo *TFI = TM.getFrameInfo(); const X86Subtarget &Subtarget = TM.getSubtarget(); + const X86MachineFunctionInfo *MFI = MF.getInfo(); + // Does the function dedicate RBP / EBP to being a frame ptr? if (!Subtarget.is64Bit()) - return X86_GR8_AO_32 + (sizeof(X86_GR8_AO_32) / sizeof(unsigned)); - else if (RI->hasFP(MF)) - return X86_GR8_AO_64_fp + (sizeof(X86_GR8_AO_64_fp) / sizeof(unsigned)); + // In 32-mode, none of the 8-bit registers aliases EBP or ESP. + return begin() + 8; + else if (TFI->hasFP(MF) || MFI->getReserveFP()) + // If so, don't allocate SPL or BPL. + return array_endof(X86_GR8_AO_64) - 1; else - return X86_GR8_AO_64 + (sizeof(X86_GR8_AO_64) / sizeof(unsigned)); + // If not, just don't allocate SPL. + return array_endof(X86_GR8_AO_64); } }]; } - def GR16 : RegisterClass<"X86", [i16], 16, [AX, CX, DX, SI, DI, BX, BP, SP, - R8W, R9W, R10W, R11W, R12W, R13W, R14W, R15W]> { - let SubRegClassList = [GR8]; + R8W, R9W, R10W, R11W, R14W, R15W, R12W, R13W]> { + let SubRegClasses = [(GR8 sub_8bit, sub_8bit_hi)]; let MethodProtos = [{ iterator allocation_order_begin(const MachineFunction &MF) const; iterator allocation_order_end(const MachineFunction &MF) const; }]; let MethodBodies = [{ - // Does the function dedicate RBP / EBP to being a frame ptr? - // If so, don't allocate SP or BP. - static const unsigned X86_GR16_AO_64_fp[] = - {X86::AX, X86::CX, X86::DX, X86::SI, X86::DI, - X86::R8W, X86::R9W, X86::R10W, X86::R11W, - X86::BX, X86::R14W, X86::R15W, X86::R12W, X86::R13W}; - static const unsigned X86_GR16_AO_32_fp[] = - {X86::AX, X86::CX, X86::DX, X86::SI, X86::DI, X86::BX}; - // If not, just don't allocate SPL. - static const unsigned X86_GR16_AO_64[] = - {X86::AX, X86::CX, X86::DX, X86::SI, X86::DI, - X86::R8W, X86::R9W, X86::R10W, X86::R11W, - X86::BX, X86::R14W, X86::R15W, X86::R12W, X86::R13W, X86::BP}; - static const unsigned X86_GR16_AO_32[] = - {X86::AX, X86::CX, X86::DX, X86::SI, X86::DI, X86::BX, X86::BP}; + static const unsigned X86_GR16_AO_64[] = { + X86::AX, X86::CX, X86::DX, X86::SI, X86::DI, + X86::R8W, X86::R9W, X86::R10W, X86::R11W, + X86::BX, X86::R14W, X86::R15W, X86::R12W, X86::R13W, X86::BP + }; GR16Class::iterator GR16Class::allocation_order_begin(const MachineFunction &MF) const { const TargetMachine &TM = MF.getTarget(); - const MRegisterInfo *RI = TM.getRegisterInfo(); const X86Subtarget &Subtarget = TM.getSubtarget(); - if (Subtarget.is64Bit()) { - if (RI->hasFP(MF)) - return X86_GR16_AO_64_fp; - else - return X86_GR16_AO_64; - } else { - if (RI->hasFP(MF)) - return X86_GR16_AO_32_fp; - else - return X86_GR16_AO_32; - } + if (Subtarget.is64Bit()) + return X86_GR16_AO_64; + else + return begin(); } GR16Class::iterator GR16Class::allocation_order_end(const MachineFunction &MF) const { const TargetMachine &TM = MF.getTarget(); - const MRegisterInfo *RI = TM.getRegisterInfo(); + const TargetFrameInfo *TFI = TM.getFrameInfo(); const X86Subtarget &Subtarget = TM.getSubtarget(); + const X86MachineFunctionInfo *MFI = MF.getInfo(); if (Subtarget.is64Bit()) { - if (RI->hasFP(MF)) - return X86_GR16_AO_64_fp+(sizeof(X86_GR16_AO_64_fp)/sizeof(unsigned)); + // Does the function dedicate RBP to being a frame ptr? + if (TFI->hasFP(MF) || MFI->getReserveFP()) + // If so, don't allocate SP or BP. + return array_endof(X86_GR16_AO_64) - 1; else - return X86_GR16_AO_64 + (sizeof(X86_GR16_AO_64) / sizeof(unsigned)); + // If not, just don't allocate SP. + return array_endof(X86_GR16_AO_64); } else { - if (RI->hasFP(MF)) - return X86_GR16_AO_32_fp+(sizeof(X86_GR16_AO_32_fp)/sizeof(unsigned)); + // Does the function dedicate EBP to being a frame ptr? + if (TFI->hasFP(MF) || MFI->getReserveFP()) + // If so, don't allocate SP or BP. + return begin() + 6; else - return X86_GR16_AO_32 + (sizeof(X86_GR16_AO_32) / sizeof(unsigned)); + // If not, just don't allocate SP. + return begin() + 7; } } }]; } - -def GR32 : RegisterClass<"X86", [i32], 32, +def GR32 : RegisterClass<"X86", [i32], 32, [EAX, ECX, EDX, ESI, EDI, EBX, EBP, ESP, - R8D, R9D, R10D, R11D, R12D, R13D, R14D, R15D]> { - let SubRegClassList = [GR8, GR16]; + R8D, R9D, R10D, R11D, R14D, R15D, R12D, R13D]> { + let SubRegClasses = [(GR8 sub_8bit, sub_8bit_hi), (GR16 sub_16bit)]; let MethodProtos = [{ iterator allocation_order_begin(const MachineFunction &MF) const; iterator allocation_order_end(const MachineFunction &MF) const; }]; let MethodBodies = [{ - // Does the function dedicate RBP / EBP to being a frame ptr? - // If so, don't allocate ESP or EBP. - static const unsigned X86_GR32_AO_64_fp[] = - {X86::EAX, X86::ECX, X86::EDX, X86::ESI, X86::EDI, - X86::R8D, X86::R9D, X86::R10D, X86::R11D, - X86::EBX, X86::R14D, X86::R15D, X86::R12D, X86::R13D}; - static const unsigned X86_GR32_AO_32_fp[] = - {X86::EAX, X86::ECX, X86::EDX, X86::ESI, X86::EDI, X86::EBX}; - // If not, just don't allocate SPL. - static const unsigned X86_GR32_AO_64[] = - {X86::EAX, X86::ECX, X86::EDX, X86::ESI, X86::EDI, - X86::R8D, X86::R9D, X86::R10D, X86::R11D, - X86::EBX, X86::R14D, X86::R15D, X86::R12D, X86::R13D, X86::EBP}; - static const unsigned X86_GR32_AO_32[] = - {X86::EAX, X86::ECX, X86::EDX, X86::ESI, X86::EDI, X86::EBX, X86::EBP}; + static const unsigned X86_GR32_AO_64[] = { + X86::EAX, X86::ECX, X86::EDX, X86::ESI, X86::EDI, + X86::R8D, X86::R9D, X86::R10D, X86::R11D, + X86::EBX, X86::R14D, X86::R15D, X86::R12D, X86::R13D, X86::EBP + }; GR32Class::iterator GR32Class::allocation_order_begin(const MachineFunction &MF) const { const TargetMachine &TM = MF.getTarget(); - const MRegisterInfo *RI = TM.getRegisterInfo(); const X86Subtarget &Subtarget = TM.getSubtarget(); - if (Subtarget.is64Bit()) { - if (RI->hasFP(MF)) - return X86_GR32_AO_64_fp; - else - return X86_GR32_AO_64; - } else { - if (RI->hasFP(MF)) - return X86_GR32_AO_32_fp; - else - return X86_GR32_AO_32; - } + if (Subtarget.is64Bit()) + return X86_GR32_AO_64; + else + return begin(); } GR32Class::iterator GR32Class::allocation_order_end(const MachineFunction &MF) const { const TargetMachine &TM = MF.getTarget(); - const MRegisterInfo *RI = TM.getRegisterInfo(); + const TargetFrameInfo *TFI = TM.getFrameInfo(); const X86Subtarget &Subtarget = TM.getSubtarget(); + const X86MachineFunctionInfo *MFI = MF.getInfo(); if (Subtarget.is64Bit()) { - if (RI->hasFP(MF)) - return X86_GR32_AO_64_fp+(sizeof(X86_GR32_AO_64_fp)/sizeof(unsigned)); + // Does the function dedicate RBP to being a frame ptr? + if (TFI->hasFP(MF) || MFI->getReserveFP()) + // If so, don't allocate ESP or EBP. + return array_endof(X86_GR32_AO_64) - 1; else - return X86_GR32_AO_64 + (sizeof(X86_GR32_AO_64) / sizeof(unsigned)); + // If not, just don't allocate ESP. + return array_endof(X86_GR32_AO_64); } else { - if (RI->hasFP(MF)) - return X86_GR32_AO_32_fp+(sizeof(X86_GR32_AO_32_fp)/sizeof(unsigned)); + // Does the function dedicate EBP to being a frame ptr? + if (TFI->hasFP(MF) || MFI->getReserveFP()) + // If so, don't allocate ESP or EBP. + return begin() + 6; else - return X86_GR32_AO_32 + (sizeof(X86_GR32_AO_32) / sizeof(unsigned)); + // If not, just don't allocate ESP. + return begin() + 7; } } }]; } - -def GR64 : RegisterClass<"X86", [i64], 64, +// GR64 - 64-bit GPRs. This oddly includes RIP, which isn't accurate, since +// RIP isn't really a register and it can't be used anywhere except in an +// address, but it doesn't cause trouble. +def GR64 : RegisterClass<"X86", [i64], 64, [RAX, RCX, RDX, RSI, RDI, R8, R9, R10, R11, - RBX, R14, R15, R12, R13, RBP, RSP]> { - let SubRegClassList = [GR8, GR16, GR32]; + RBX, R14, R15, R12, R13, RBP, RSP, RIP]> { + let SubRegClasses = [(GR8 sub_8bit, sub_8bit_hi), + (GR16 sub_16bit), + (GR32 sub_32bit)]; let MethodProtos = [{ iterator allocation_order_end(const MachineFunction &MF) const; }]; @@ -407,23 +436,286 @@ def GR64 : RegisterClass<"X86", [i64], 64, GR64Class::iterator GR64Class::allocation_order_end(const MachineFunction &MF) const { const TargetMachine &TM = MF.getTarget(); - const MRegisterInfo *RI = TM.getRegisterInfo(); - if (RI->hasFP(MF)) // Does the function dedicate RBP to being a frame ptr? - return end()-2; // If so, don't allocate RSP or RBP + const TargetFrameInfo *TFI = TM.getFrameInfo(); + const X86Subtarget &Subtarget = TM.getSubtarget(); + const X86MachineFunctionInfo *MFI = MF.getInfo(); + if (!Subtarget.is64Bit()) + return begin(); // None of these are allocatable in 32-bit. + // Does the function dedicate RBP to being a frame ptr? + if (TFI->hasFP(MF) || MFI->getReserveFP()) + return end()-3; // If so, don't allocate RIP, RSP or RBP else - return end()-1; // If not, just don't allocate RSP + return end()-2; // If not, just don't allocate RIP or RSP } }]; } +// Segment registers for use by MOV instructions (and others) that have a +// segment register as one operand. Always contain a 16-bit segment +// descriptor. +def SEGMENT_REG : RegisterClass<"X86", [i16], 16, [CS, DS, SS, ES, FS, GS]>; + +// Debug registers. +def DEBUG_REG : RegisterClass<"X86", [i32], 32, + [DR0, DR1, DR2, DR3, DR4, DR5, DR6, DR7]>; + +// Control registers. +def CONTROL_REG : RegisterClass<"X86", [i64], 64, + [CR0, CR1, CR2, CR3, CR4, CR5, CR6, CR7, CR8, + CR9, CR10, CR11, CR12, CR13, CR14, CR15]>; + +// GR8_ABCD_L, GR8_ABCD_H, GR16_ABCD, GR32_ABCD, GR64_ABCD - Subclasses of +// GR8, GR16, GR32, and GR64 which contain just the "a" "b", "c", and "d" +// registers. On x86-32, GR16_ABCD and GR32_ABCD are classes for registers +// that support 8-bit subreg operations. On x86-64, GR16_ABCD, GR32_ABCD, +// and GR64_ABCD are classes for registers that support 8-bit h-register +// operations. +def GR8_ABCD_L : RegisterClass<"X86", [i8], 8, [AL, CL, DL, BL]>; +def GR8_ABCD_H : RegisterClass<"X86", [i8], 8, [AH, CH, DH, BH]>; +def GR16_ABCD : RegisterClass<"X86", [i16], 16, [AX, CX, DX, BX]> { + let SubRegClasses = [(GR8_ABCD_L sub_8bit), (GR8_ABCD_H sub_8bit_hi)]; +} +def GR32_ABCD : RegisterClass<"X86", [i32], 32, [EAX, ECX, EDX, EBX]> { + let SubRegClasses = [(GR8_ABCD_L sub_8bit), + (GR8_ABCD_H sub_8bit_hi), + (GR16_ABCD sub_16bit)]; +} +def GR64_ABCD : RegisterClass<"X86", [i64], 64, [RAX, RCX, RDX, RBX]> { + let SubRegClasses = [(GR8_ABCD_L sub_8bit), + (GR8_ABCD_H sub_8bit_hi), + (GR16_ABCD sub_16bit), + (GR32_ABCD sub_32bit)]; +} +def GR32_TC : RegisterClass<"X86", [i32], 32, [EAX, ECX, EDX]> { + let SubRegClasses = [(GR8 sub_8bit, sub_8bit_hi), (GR16 sub_16bit)]; +} +def GR64_TC : RegisterClass<"X86", [i64], 64, [RAX, RCX, RDX, RSI, RDI, + R8, R9, R11]> { + let SubRegClasses = [(GR8 sub_8bit, sub_8bit_hi), + (GR16 sub_16bit), + (GR32_TC sub_32bit)]; +} + +// GR8_NOREX - GR8 registers which do not require a REX prefix. +def GR8_NOREX : RegisterClass<"X86", [i8], 8, + [AL, CL, DL, AH, CH, DH, BL, BH]> { + let MethodProtos = [{ + iterator allocation_order_begin(const MachineFunction &MF) const; + iterator allocation_order_end(const MachineFunction &MF) const; + }]; + let MethodBodies = [{ + // In 64-bit mode, it's not safe to blindly allocate H registers. + static const unsigned X86_GR8_NOREX_AO_64[] = { + X86::AL, X86::CL, X86::DL, X86::BL + }; + + GR8_NOREXClass::iterator + GR8_NOREXClass::allocation_order_begin(const MachineFunction &MF) const { + const TargetMachine &TM = MF.getTarget(); + const X86Subtarget &Subtarget = TM.getSubtarget(); + if (Subtarget.is64Bit()) + return X86_GR8_NOREX_AO_64; + else + return begin(); + } -// GR16, GR32 subclasses which contain registers that have GR8 sub-registers. -// These should only be used for 32-bit mode. -def GR16_ : RegisterClass<"X86", [i16], 16, [AX, CX, DX, BX]> { - let SubRegClassList = [GR8]; + GR8_NOREXClass::iterator + GR8_NOREXClass::allocation_order_end(const MachineFunction &MF) const { + const TargetMachine &TM = MF.getTarget(); + const X86Subtarget &Subtarget = TM.getSubtarget(); + if (Subtarget.is64Bit()) + return array_endof(X86_GR8_NOREX_AO_64); + else + return end(); + } + }]; } -def GR32_ : RegisterClass<"X86", [i32], 32, [EAX, ECX, EDX, EBX]> { - let SubRegClassList = [GR8, GR16]; +// GR16_NOREX - GR16 registers which do not require a REX prefix. +def GR16_NOREX : RegisterClass<"X86", [i16], 16, + [AX, CX, DX, SI, DI, BX, BP, SP]> { + let SubRegClasses = [(GR8_NOREX sub_8bit, sub_8bit_hi)]; + let MethodProtos = [{ + iterator allocation_order_end(const MachineFunction &MF) const; + }]; + let MethodBodies = [{ + GR16_NOREXClass::iterator + GR16_NOREXClass::allocation_order_end(const MachineFunction &MF) const { + const TargetMachine &TM = MF.getTarget(); + const TargetFrameInfo *TFI = TM.getFrameInfo(); + const X86MachineFunctionInfo *MFI = MF.getInfo(); + // Does the function dedicate RBP / EBP to being a frame ptr? + if (TFI->hasFP(MF) || MFI->getReserveFP()) + // If so, don't allocate SP or BP. + return end() - 2; + else + // If not, just don't allocate SP. + return end() - 1; + } + }]; +} +// GR32_NOREX - GR32 registers which do not require a REX prefix. +def GR32_NOREX : RegisterClass<"X86", [i32], 32, + [EAX, ECX, EDX, ESI, EDI, EBX, EBP, ESP]> { + let SubRegClasses = [(GR8_NOREX sub_8bit, sub_8bit_hi), + (GR16_NOREX sub_16bit)]; + let MethodProtos = [{ + iterator allocation_order_end(const MachineFunction &MF) const; + }]; + let MethodBodies = [{ + GR32_NOREXClass::iterator + GR32_NOREXClass::allocation_order_end(const MachineFunction &MF) const { + const TargetMachine &TM = MF.getTarget(); + const TargetFrameInfo *TFI = TM.getFrameInfo(); + const X86MachineFunctionInfo *MFI = MF.getInfo(); + // Does the function dedicate RBP / EBP to being a frame ptr? + if (TFI->hasFP(MF) || MFI->getReserveFP()) + // If so, don't allocate ESP or EBP. + return end() - 2; + else + // If not, just don't allocate ESP. + return end() - 1; + } + }]; +} +// GR64_NOREX - GR64 registers which do not require a REX prefix. +def GR64_NOREX : RegisterClass<"X86", [i64], 64, + [RAX, RCX, RDX, RSI, RDI, RBX, RBP, RSP, RIP]> { + let SubRegClasses = [(GR8_NOREX sub_8bit, sub_8bit_hi), + (GR16_NOREX sub_16bit), + (GR32_NOREX sub_32bit)]; + let MethodProtos = [{ + iterator allocation_order_end(const MachineFunction &MF) const; + }]; + let MethodBodies = [{ + GR64_NOREXClass::iterator + GR64_NOREXClass::allocation_order_end(const MachineFunction &MF) const { + const TargetMachine &TM = MF.getTarget(); + const TargetFrameInfo *TFI = TM.getFrameInfo(); + const X86MachineFunctionInfo *MFI = MF.getInfo(); + // Does the function dedicate RBP to being a frame ptr? + if (TFI->hasFP(MF) || MFI->getReserveFP()) + // If so, don't allocate RIP, RSP or RBP. + return end() - 3; + else + // If not, just don't allocate RIP or RSP. + return end() - 2; + } + }]; +} + +// GR32_NOSP - GR32 registers except ESP. +def GR32_NOSP : RegisterClass<"X86", [i32], 32, + [EAX, ECX, EDX, ESI, EDI, EBX, EBP, + R8D, R9D, R10D, R11D, R14D, R15D, R12D, R13D]> { + let SubRegClasses = [(GR8 sub_8bit, sub_8bit_hi), (GR16 sub_16bit)]; + let MethodProtos = [{ + iterator allocation_order_begin(const MachineFunction &MF) const; + iterator allocation_order_end(const MachineFunction &MF) const; + }]; + let MethodBodies = [{ + static const unsigned X86_GR32_NOSP_AO_64[] = { + X86::EAX, X86::ECX, X86::EDX, X86::ESI, X86::EDI, + X86::R8D, X86::R9D, X86::R10D, X86::R11D, + X86::EBX, X86::R14D, X86::R15D, X86::R12D, X86::R13D, X86::EBP + }; + + GR32_NOSPClass::iterator + GR32_NOSPClass::allocation_order_begin(const MachineFunction &MF) const { + const TargetMachine &TM = MF.getTarget(); + const X86Subtarget &Subtarget = TM.getSubtarget(); + if (Subtarget.is64Bit()) + return X86_GR32_NOSP_AO_64; + else + return begin(); + } + + GR32_NOSPClass::iterator + GR32_NOSPClass::allocation_order_end(const MachineFunction &MF) const { + const TargetMachine &TM = MF.getTarget(); + const TargetFrameInfo *TFI = TM.getFrameInfo(); + const X86Subtarget &Subtarget = TM.getSubtarget(); + const X86MachineFunctionInfo *MFI = MF.getInfo(); + if (Subtarget.is64Bit()) { + // Does the function dedicate RBP to being a frame ptr? + if (TFI->hasFP(MF) || MFI->getReserveFP()) + // If so, don't allocate EBP. + return array_endof(X86_GR32_NOSP_AO_64) - 1; + else + // If not, any reg in this class is ok. + return array_endof(X86_GR32_NOSP_AO_64); + } else { + // Does the function dedicate EBP to being a frame ptr? + if (TFI->hasFP(MF) || MFI->getReserveFP()) + // If so, don't allocate EBP. + return begin() + 6; + else + // If not, any reg in this class is ok. + return begin() + 7; + } + } + }]; +} + +// GR64_NOSP - GR64 registers except RSP (and RIP). +def GR64_NOSP : RegisterClass<"X86", [i64], 64, + [RAX, RCX, RDX, RSI, RDI, R8, R9, R10, R11, + RBX, R14, R15, R12, R13, RBP]> { + let SubRegClasses = [(GR8 sub_8bit, sub_8bit_hi), + (GR16 sub_16bit), + (GR32_NOSP sub_32bit)]; + let MethodProtos = [{ + iterator allocation_order_end(const MachineFunction &MF) const; + }]; + let MethodBodies = [{ + GR64_NOSPClass::iterator + GR64_NOSPClass::allocation_order_end(const MachineFunction &MF) const { + const TargetMachine &TM = MF.getTarget(); + const TargetFrameInfo *TFI = TM.getFrameInfo(); + const X86Subtarget &Subtarget = TM.getSubtarget(); + const X86MachineFunctionInfo *MFI = MF.getInfo(); + if (!Subtarget.is64Bit()) + return begin(); // None of these are allocatable in 32-bit. + // Does the function dedicate RBP to being a frame ptr? + if (TFI->hasFP(MF) || MFI->getReserveFP()) + return end()-1; // If so, don't allocate RBP + else + return end(); // If not, any reg in this class is ok. + } + }]; +} + +// GR64_NOREX_NOSP - GR64_NOREX registers except RSP. +def GR64_NOREX_NOSP : RegisterClass<"X86", [i64], 64, + [RAX, RCX, RDX, RSI, RDI, RBX, RBP]> { + let SubRegClasses = [(GR8_NOREX sub_8bit, sub_8bit_hi), + (GR16_NOREX sub_16bit), + (GR32_NOREX sub_32bit)]; + let MethodProtos = [{ + iterator allocation_order_end(const MachineFunction &MF) const; + }]; + let MethodBodies = [{ + GR64_NOREX_NOSPClass::iterator + GR64_NOREX_NOSPClass::allocation_order_end(const MachineFunction &MF) const + { + const TargetMachine &TM = MF.getTarget(); + const TargetFrameInfo *TFI = TM.getFrameInfo(); + const X86MachineFunctionInfo *MFI = MF.getInfo(); + // Does the function dedicate RBP to being a frame ptr? + if (TFI->hasFP(MF) || MFI->getReserveFP()) + // If so, don't allocate RBP. + return end() - 1; + else + // If not, any reg in this class is ok. + return end(); + } + }]; +} + +// A class to support the 'A' assembler constraint: EAX then EDX. +def GR32_AD : RegisterClass<"X86", [i32], 32, [EAX, EDX]> { + let SubRegClasses = [(GR8_ABCD_L sub_8bit), + (GR8_ABCD_H sub_8bit_hi), + (GR16_ABCD sub_16bit)]; } // Scalar SSE2 floating point registers. @@ -474,14 +766,14 @@ def FR64 : RegisterClass<"X86", [f64], 64, // faster on common hardware. In reality, this should be controlled by a // command line option or something. -def RFP32 : RegisterClass<"X86", [f32], 32, [FP0, FP1, FP2, FP3, FP4, FP5, FP6]>; -def RFP64 : RegisterClass<"X86", [f64], 32, [FP0, FP1, FP2, FP3, FP4, FP5, FP6]>; -def RFP80 : RegisterClass<"X86", [f80], 32, [FP0, FP1, FP2, FP3, FP4, FP5, FP6]>; +def RFP32 : RegisterClass<"X86",[f32], 32, [FP0, FP1, FP2, FP3, FP4, FP5, FP6]>; +def RFP64 : RegisterClass<"X86",[f64], 32, [FP0, FP1, FP2, FP3, FP4, FP5, FP6]>; +def RFP80 : RegisterClass<"X86",[f80], 32, [FP0, FP1, FP2, FP3, FP4, FP5, FP6]>; // Floating point stack registers (these are not allocatable by the // register allocator - the floating point stackifier is responsible // for transforming FPn allocations to STn registers) -def RST : RegisterClass<"X86", [f64], 32, +def RST : RegisterClass<"X86", [f80, f64, f32], 32, [ST0, ST1, ST2, ST3, ST4, ST5, ST6, ST7]> { let MethodProtos = [{ iterator allocation_order_end(const MachineFunction &MF) const; @@ -495,12 +787,14 @@ def RST : RegisterClass<"X86", [f64], 32, } // Generic vector registers: VR64 and VR128. -def VR64 : RegisterClass<"X86", [v8i8, v4i16, v2i32, v1i64], 64, +def VR64: RegisterClass<"X86", [x86mmx], 64, [MM0, MM1, MM2, MM3, MM4, MM5, MM6, MM7]>; def VR128 : RegisterClass<"X86", [v16i8, v8i16, v4i32, v2i64, v4f32, v2f64],128, [XMM0, XMM1, XMM2, XMM3, XMM4, XMM5, XMM6, XMM7, XMM8, XMM9, XMM10, XMM11, XMM12, XMM13, XMM14, XMM15]> { + let SubRegClasses = [(FR32 sub_ss), (FR64 sub_sd)]; + let MethodProtos = [{ iterator allocation_order_end(const MachineFunction &MF) const; }]; @@ -516,3 +810,41 @@ def VR128 : RegisterClass<"X86", [v16i8, v8i16, v4i32, v2i64, v4f32, v2f64],128, } }]; } + +def VR256 : RegisterClass<"X86", [v32i8, v8i32, v4i64, v8f32, v4f64], 256, + [YMM0, YMM1, YMM2, YMM3, YMM4, YMM5, YMM6, YMM7, + YMM8, YMM9, YMM10, YMM11, + YMM12, YMM13, YMM14, YMM15]> { + let SubRegClasses = [(FR32 sub_ss), (FR64 sub_sd), (VR128 sub_xmm)]; + + let MethodProtos = [{ + iterator allocation_order_end(const MachineFunction &MF) const; + }]; + let MethodBodies = [{ + VR256Class::iterator + VR256Class::allocation_order_end(const MachineFunction &MF) const { + const TargetMachine &TM = MF.getTarget(); + const X86Subtarget &Subtarget = TM.getSubtarget(); + if (!Subtarget.is64Bit()) + return end()-8; // Only YMM0 to YMM7 are available in 32-bit mode. + else + return end(); + } + }]; +} + +// Status flags registers. +def CCR : RegisterClass<"X86", [i32], 32, [EFLAGS]> { + let CopyCost = -1; // Don't allow copying of status registers. + + // EFLAGS is not allocatable. + let MethodProtos = [{ + iterator allocation_order_end(const MachineFunction &MF) const; + }]; + let MethodBodies = [{ + CCRClass::iterator + CCRClass::allocation_order_end(const MachineFunction &MF) const { + return allocation_order_begin(MF); + } + }]; +}