def sub_16bit : SubRegIndex<16>;
def sub_32bit : SubRegIndex<32>;
def sub_xmm : SubRegIndex<128>;
+ def sub_ymm : SubRegIndex<256>;
}
//===----------------------------------------------------------------------===//
def FP4 : X86Reg<"fp4", 0>;
def FP5 : X86Reg<"fp5", 0>;
def FP6 : X86Reg<"fp6", 0>;
+def FP7 : X86Reg<"fp7", 0>;
// XMM Registers, used by the various SSE instruction set extensions.
def XMM0: X86Reg<"xmm0", 0>, DwarfRegNum<[17, 21, 21]>;
def XMM13: X86Reg<"xmm13", 13>, DwarfRegNum<[30, -2, -2]>;
def XMM14: X86Reg<"xmm14", 14>, DwarfRegNum<[31, -2, -2]>;
def XMM15: X86Reg<"xmm15", 15>, DwarfRegNum<[32, -2, -2]>;
+
+def XMM16: X86Reg<"xmm16", 16>, DwarfRegNum<[60, -2, -2]>;
+def XMM17: X86Reg<"xmm17", 17>, DwarfRegNum<[61, -2, -2]>;
+def XMM18: X86Reg<"xmm18", 18>, DwarfRegNum<[62, -2, -2]>;
+def XMM19: X86Reg<"xmm19", 19>, DwarfRegNum<[63, -2, -2]>;
+def XMM20: X86Reg<"xmm20", 20>, DwarfRegNum<[64, -2, -2]>;
+def XMM21: X86Reg<"xmm21", 21>, DwarfRegNum<[65, -2, -2]>;
+def XMM22: X86Reg<"xmm22", 22>, DwarfRegNum<[66, -2, -2]>;
+def XMM23: X86Reg<"xmm23", 23>, DwarfRegNum<[67, -2, -2]>;
+def XMM24: X86Reg<"xmm24", 24>, DwarfRegNum<[68, -2, -2]>;
+def XMM25: X86Reg<"xmm25", 25>, DwarfRegNum<[69, -2, -2]>;
+def XMM26: X86Reg<"xmm26", 26>, DwarfRegNum<[70, -2, -2]>;
+def XMM27: X86Reg<"xmm27", 27>, DwarfRegNum<[71, -2, -2]>;
+def XMM28: X86Reg<"xmm28", 28>, DwarfRegNum<[72, -2, -2]>;
+def XMM29: X86Reg<"xmm29", 29>, DwarfRegNum<[73, -2, -2]>;
+def XMM30: X86Reg<"xmm30", 30>, DwarfRegNum<[74, -2, -2]>;
+def XMM31: X86Reg<"xmm31", 31>, DwarfRegNum<[75, -2, -2]>;
+
} // CostPerUse
-// YMM Registers, used by AVX instructions
+// YMM0-15 registers, used by AVX instructions and
+// YMM16-31 registers, used by AVX-512 instructions.
let SubRegIndices = [sub_xmm] in {
-def YMM0: X86Reg<"ymm0", 0, [XMM0]>, DwarfRegAlias<XMM0>;
-def YMM1: X86Reg<"ymm1", 1, [XMM1]>, DwarfRegAlias<XMM1>;
-def YMM2: X86Reg<"ymm2", 2, [XMM2]>, DwarfRegAlias<XMM2>;
-def YMM3: X86Reg<"ymm3", 3, [XMM3]>, DwarfRegAlias<XMM3>;
-def YMM4: X86Reg<"ymm4", 4, [XMM4]>, DwarfRegAlias<XMM4>;
-def YMM5: X86Reg<"ymm5", 5, [XMM5]>, DwarfRegAlias<XMM5>;
-def YMM6: X86Reg<"ymm6", 6, [XMM6]>, DwarfRegAlias<XMM6>;
-def YMM7: X86Reg<"ymm7", 7, [XMM7]>, DwarfRegAlias<XMM7>;
-def YMM8: X86Reg<"ymm8", 8, [XMM8]>, DwarfRegAlias<XMM8>;
-def YMM9: X86Reg<"ymm9", 9, [XMM9]>, DwarfRegAlias<XMM9>;
-def YMM10: X86Reg<"ymm10", 10, [XMM10]>, DwarfRegAlias<XMM10>;
-def YMM11: X86Reg<"ymm11", 11, [XMM11]>, DwarfRegAlias<XMM11>;
-def YMM12: X86Reg<"ymm12", 12, [XMM12]>, DwarfRegAlias<XMM12>;
-def YMM13: X86Reg<"ymm13", 13, [XMM13]>, DwarfRegAlias<XMM13>;
-def YMM14: X86Reg<"ymm14", 14, [XMM14]>, DwarfRegAlias<XMM14>;
-def YMM15: X86Reg<"ymm15", 15, [XMM15]>, DwarfRegAlias<XMM15>;
+ foreach Index = 0-31 in {
+ def YMM#Index : X86Reg<"ymm"#Index, Index, [!cast<X86Reg>("XMM"#Index)]>,
+ DwarfRegAlias<!cast<X86Reg>("XMM"#Index)>;
+ }
}
-class STRegister<string n, bits<16> Enc, list<Register> A> : X86Reg<n, Enc> {
- let Aliases = A;
+// ZMM Registers, used by AVX-512 instructions.
+let SubRegIndices = [sub_ymm] in {
+ foreach Index = 0-31 in {
+ def ZMM#Index : X86Reg<"zmm"#Index, Index, [!cast<X86Reg>("YMM"#Index)]>,
+ DwarfRegAlias<!cast<X86Reg>("XMM"#Index)>;
+ }
}
+// Mask Registers, used by AVX-512 instructions.
+def K0 : X86Reg<"k0", 0>, DwarfRegNum<[118, -2, -2]>;
+def K1 : X86Reg<"k1", 1>, DwarfRegNum<[119, -2, -2]>;
+def K2 : X86Reg<"k2", 2>, DwarfRegNum<[120, -2, -2]>;
+def K3 : X86Reg<"k3", 3>, DwarfRegNum<[121, -2, -2]>;
+def K4 : X86Reg<"k4", 4>, DwarfRegNum<[122, -2, -2]>;
+def K5 : X86Reg<"k5", 5>, DwarfRegNum<[123, -2, -2]>;
+def K6 : X86Reg<"k6", 6>, DwarfRegNum<[124, -2, -2]>;
+def K7 : X86Reg<"k7", 7>, DwarfRegNum<[125, -2, -2]>;
+
// Floating point stack registers. These don't map one-to-one to the FP
// pseudo registers, but we still mark them as aliasing FP registers. That
// way both kinds can be live without exceeding the stack depth. ST registers
// are only live around inline assembly.
-def ST0 : STRegister<"st(0)", 0, []>, DwarfRegNum<[33, 12, 11]>;
-def ST1 : STRegister<"st(1)", 1, [FP6]>, DwarfRegNum<[34, 13, 12]>;
-def ST2 : STRegister<"st(2)", 2, [FP5]>, DwarfRegNum<[35, 14, 13]>;
-def ST3 : STRegister<"st(3)", 3, [FP4]>, DwarfRegNum<[36, 15, 14]>;
-def ST4 : STRegister<"st(4)", 4, [FP3]>, DwarfRegNum<[37, 16, 15]>;
-def ST5 : STRegister<"st(5)", 5, [FP2]>, DwarfRegNum<[38, 17, 16]>;
-def ST6 : STRegister<"st(6)", 6, [FP1]>, DwarfRegNum<[39, 18, 17]>;
-def ST7 : STRegister<"st(7)", 7, [FP0]>, DwarfRegNum<[40, 19, 18]>;
+def ST0 : X86Reg<"st(0)", 0>, DwarfRegNum<[33, 12, 11]>;
+def ST1 : X86Reg<"st(1)", 1>, DwarfRegNum<[34, 13, 12]>;
+def ST2 : X86Reg<"st(2)", 2>, DwarfRegNum<[35, 14, 13]>;
+def ST3 : X86Reg<"st(3)", 3>, DwarfRegNum<[36, 15, 14]>;
+def ST4 : X86Reg<"st(4)", 4>, DwarfRegNum<[37, 16, 15]>;
+def ST5 : X86Reg<"st(5)", 5>, DwarfRegNum<[38, 17, 16]>;
+def ST6 : X86Reg<"st(6)", 6>, DwarfRegNum<[39, 18, 17]>;
+def ST7 : X86Reg<"st(7)", 7>, DwarfRegNum<[40, 19, 18]>;
// Floating-point status word
def FPSW : X86Reg<"fpsw", 0>;
def GS : X86Reg<"gs", 5>;
// Debug registers
-def DR0 : X86Reg<"dr0", 0>;
-def DR1 : X86Reg<"dr1", 1>;
-def DR2 : X86Reg<"dr2", 2>;
-def DR3 : X86Reg<"dr3", 3>;
-def DR4 : X86Reg<"dr4", 4>;
-def DR5 : X86Reg<"dr5", 5>;
-def DR6 : X86Reg<"dr6", 6>;
-def DR7 : X86Reg<"dr7", 7>;
+def DR0 : X86Reg<"dr0", 0>;
+def DR1 : X86Reg<"dr1", 1>;
+def DR2 : X86Reg<"dr2", 2>;
+def DR3 : X86Reg<"dr3", 3>;
+def DR4 : X86Reg<"dr4", 4>;
+def DR5 : X86Reg<"dr5", 5>;
+def DR6 : X86Reg<"dr6", 6>;
+def DR7 : X86Reg<"dr7", 7>;
+def DR8 : X86Reg<"dr8", 8>;
+def DR9 : X86Reg<"dr9", 9>;
+def DR10 : X86Reg<"dr10", 10>;
+def DR11 : X86Reg<"dr11", 11>;
+def DR12 : X86Reg<"dr12", 12>;
+def DR13 : X86Reg<"dr13", 13>;
+def DR14 : X86Reg<"dr14", 14>;
+def DR15 : X86Reg<"dr15", 15>;
// Control registers
def CR0 : X86Reg<"cr0", 0>;
def EIZ : X86Reg<"eiz", 4>;
def RIZ : X86Reg<"riz", 4>;
+// Bound registers, used in MPX instructions
+def BND0 : X86Reg<"bnd0", 0>;
+def BND1 : X86Reg<"bnd1", 1>;
+def BND2 : X86Reg<"bnd2", 2>;
+def BND3 : X86Reg<"bnd3", 3>;
//===----------------------------------------------------------------------===//
// Register Class Definitions... now that we have all of the pieces, define the
R8B, R9B, R10B, R11B, R14B, R15B, R12B, R13B)> {
let AltOrders = [(sub GR8, AH, BH, CH, DH)];
let AltOrderSelect = [{
- return MF.getTarget().getSubtarget<X86Subtarget>().is64Bit();
+ return MF.getSubtarget<X86Subtarget>().is64Bit();
}];
}
def GR64_TC : RegisterClass<"X86", [i64], 64, (add RAX, RCX, RDX, RSI, RDI,
R8, R9, R11, RIP)>;
def GR64_TCW64 : RegisterClass<"X86", [i64], 64, (add RAX, RCX, RDX,
- R8, R9, R11)>;
+ R8, R9, R10, R11, RIP)>;
// GR8_NOREX - GR8 registers which do not require a REX prefix.
def GR8_NOREX : RegisterClass<"X86", [i8], 8,
(add AL, CL, DL, AH, CH, DH, BL, BH)> {
let AltOrders = [(sub GR8_NOREX, AH, BH, CH, DH)];
let AltOrderSelect = [{
- return MF.getTarget().getSubtarget<X86Subtarget>().is64Bit();
+ return MF.getSubtarget<X86Subtarget>().is64Bit();
}];
}
// GR16_NOREX - GR16 registers which do not require a REX prefix.
}
// Generic vector registers: VR64 and VR128.
+// Ensure that float types are declared first - only float is legal on SSE1.
def VR64: RegisterClass<"X86", [x86mmx], 64, (sequence "MM%u", 0, 7)>;
-def VR128 : RegisterClass<"X86", [v16i8, v8i16, v4i32, v2i64, v4f32, v2f64],
+def VR128 : RegisterClass<"X86", [v4f32, v2f64, v16i8, v8i16, v4i32, v2i64],
128, (add FR32)>;
-def VR256 : RegisterClass<"X86", [v32i8, v16i16, v8i32, v4i64, v8f32, v4f64],
+def VR256 : RegisterClass<"X86", [v8f32, v4f64, v32i8, v16i16, v8i32, v4i64],
256, (sequence "YMM%u", 0, 15)>;
// Status flags registers.
let CopyCost = -1; // Don't allow copying of status registers.
let isAllocatable = 0;
}
+
+// AVX-512 vector/mask registers.
+def VR512 : RegisterClass<"X86", [v16f32, v8f64, v64i8, v32i16, v16i32, v8i64],
+ 512, (sequence "ZMM%u", 0, 31)>;
+
+// Scalar AVX-512 floating point registers.
+def FR32X : RegisterClass<"X86", [f32], 32, (sequence "XMM%u", 0, 31)>;
+
+def FR64X : RegisterClass<"X86", [f64], 64, (add FR32X)>;
+
+// Extended VR128 and VR256 for AVX-512 instructions
+def VR128X : RegisterClass<"X86", [v4f32, v2f64, v16i8, v8i16, v4i32, v2i64],
+ 128, (add FR32X)>;
+def VR256X : RegisterClass<"X86", [v8f32, v4f64, v32i8, v16i16, v8i32, v4i64],
+ 256, (sequence "YMM%u", 0, 31)>;
+
+// Mask registers
+def VK1 : RegisterClass<"X86", [i1], 8, (sequence "K%u", 0, 7)> {let Size = 8;}
+def VK2 : RegisterClass<"X86", [v2i1], 8, (add VK1)> {let Size = 8;}
+def VK4 : RegisterClass<"X86", [v4i1], 8, (add VK2)> {let Size = 8;}
+def VK8 : RegisterClass<"X86", [v8i1], 8, (add VK4)> {let Size = 8;}
+def VK16 : RegisterClass<"X86", [v16i1], 16, (add VK8)> {let Size = 16;}
+def VK32 : RegisterClass<"X86", [v32i1], 32, (add VK16)> {let Size = 32;}
+def VK64 : RegisterClass<"X86", [v64i1], 64, (add VK32)> {let Size = 64;}
+
+def VK1WM : RegisterClass<"X86", [i1], 8, (sub VK1, K0)> {let Size = 8;}
+def VK2WM : RegisterClass<"X86", [v2i1], 8, (sub VK2, K0)> {let Size = 8;}
+def VK4WM : RegisterClass<"X86", [v4i1], 8, (sub VK4, K0)> {let Size = 8;}
+def VK8WM : RegisterClass<"X86", [v8i1], 8, (sub VK8, K0)> {let Size = 8;}
+def VK16WM : RegisterClass<"X86", [v16i1], 16, (add VK8WM)> {let Size = 16;}
+def VK32WM : RegisterClass<"X86", [v32i1], 32, (add VK16WM)> {let Size = 32;}
+def VK64WM : RegisterClass<"X86", [v64i1], 64, (add VK32WM)> {let Size = 64;}
+
+// Bound registers
+def BNDR : RegisterClass<"X86", [v2i64], 128, (sequence "BND%u", 0, 3)>;