let PrintMethod = "printNEONModImmOperand";
let ParserMatchClass = nImmVMOVI32AsmOperand;
}
+def nImmVMOVF32 : Operand<i32> {
+ let PrintMethod = "printFPImmOperand";
+ let ParserMatchClass = FPImmOperand;
+}
def nImmSplatI64AsmOperand : AsmOperandClass { let Name = "NEONi64splat"; }
def nImmSplatI64 : Operand<i32> {
let PrintMethod = "printNEONModImmOperand";
let MIOperandInfo = (ops i32imm);
}
+// Register list of one D register.
+def VecListOneDAsmOperand : AsmOperandClass {
+ let Name = "VecListOneD";
+ let ParserMethod = "parseVectorList";
+ let RenderMethod = "addVecListOperands";
+}
+def VecListOneD : RegisterOperand<DPR, "printVectorListOne"> {
+ let ParserMatchClass = VecListOneDAsmOperand;
+}
+// Register list of two sequential D registers.
+def VecListTwoDAsmOperand : AsmOperandClass {
+ let Name = "VecListTwoD";
+ let ParserMethod = "parseVectorList";
+ let RenderMethod = "addVecListOperands";
+}
+def VecListTwoD : RegisterOperand<DPR, "printVectorListTwo"> {
+ let ParserMatchClass = VecListTwoDAsmOperand;
+}
+// Register list of three sequential D registers.
+def VecListThreeDAsmOperand : AsmOperandClass {
+ let Name = "VecListThreeD";
+ let ParserMethod = "parseVectorList";
+ let RenderMethod = "addVecListOperands";
+}
+def VecListThreeD : RegisterOperand<DPR, "printVectorListThree"> {
+ let ParserMatchClass = VecListThreeDAsmOperand;
+}
+// Register list of four sequential D registers.
+def VecListFourDAsmOperand : AsmOperandClass {
+ let Name = "VecListFourD";
+ let ParserMethod = "parseVectorList";
+ let RenderMethod = "addVecListOperands";
+}
+def VecListFourD : RegisterOperand<DPR, "printVectorListFour"> {
+ let ParserMatchClass = VecListFourDAsmOperand;
+}
+// Register list of two D registers spaced by 2 (two sequential Q registers).
+def VecListTwoQAsmOperand : AsmOperandClass {
+ let Name = "VecListTwoQ";
+ let ParserMethod = "parseVectorList";
+ let RenderMethod = "addVecListOperands";
+}
+def VecListTwoQ : RegisterOperand<DPR, "printVectorListTwo"> {
+ let ParserMatchClass = VecListTwoQAsmOperand;
+}
+
+// Register list of one D register, with "all lanes" subscripting.
+def VecListOneDAllLanesAsmOperand : AsmOperandClass {
+ let Name = "VecListOneDAllLanes";
+ let ParserMethod = "parseVectorList";
+ let RenderMethod = "addVecListOperands";
+}
+def VecListOneDAllLanes : RegisterOperand<DPR, "printVectorListOneAllLanes"> {
+ let ParserMatchClass = VecListOneDAllLanesAsmOperand;
+}
+// Register list of two D registers, with "all lanes" subscripting.
+def VecListTwoDAllLanesAsmOperand : AsmOperandClass {
+ let Name = "VecListTwoDAllLanes";
+ let ParserMethod = "parseVectorList";
+ let RenderMethod = "addVecListOperands";
+}
+def VecListTwoDAllLanes : RegisterOperand<DPR, "printVectorListTwoAllLanes"> {
+ let ParserMatchClass = VecListTwoDAllLanesAsmOperand;
+}
+
+// Register list of one D register, with byte lane subscripting.
+def VecListOneDByteIndexAsmOperand : AsmOperandClass {
+ let Name = "VecListOneDByteIndexed";
+ let ParserMethod = "parseVectorList";
+ let RenderMethod = "addVecListIndexedOperands";
+}
+def VecListOneDByteIndexed : Operand<i32> {
+ let ParserMatchClass = VecListOneDByteIndexAsmOperand;
+ let MIOperandInfo = (ops DPR:$Vd, i32imm:$idx);
+}
+
//===----------------------------------------------------------------------===//
// NEON-specific DAG Nodes.
//===----------------------------------------------------------------------===//
def SDTARMVMOVIMM : SDTypeProfile<1, 1, [SDTCisVec<0>, SDTCisVT<1, i32>]>;
def NEONvmovImm : SDNode<"ARMISD::VMOVIMM", SDTARMVMOVIMM>;
def NEONvmvnImm : SDNode<"ARMISD::VMVNIMM", SDTARMVMOVIMM>;
+def NEONvmovFPImm : SDNode<"ARMISD::VMOVFPIMM", SDTARMVMOVIMM>;
def SDTARMVORRIMM : SDTypeProfile<1, 2, [SDTCisVec<0>, SDTCisSameAs<0, 1>,
SDTCisVT<2, i32>]>;
: PseudoNLdSt<(outs QPR:$dst, GPR:$wb),
(ins addrmode6:$addr, am6offset:$offset), itin,
"$addr.addr = $wb">;
+class VLDQWBfixedPseudo<InstrItinClass itin>
+ : PseudoNLdSt<(outs QPR:$dst, GPR:$wb),
+ (ins addrmode6:$addr), itin,
+ "$addr.addr = $wb">;
+class VLDQWBregisterPseudo<InstrItinClass itin>
+ : PseudoNLdSt<(outs QPR:$dst, GPR:$wb),
+ (ins addrmode6:$addr, rGPR:$offset), itin,
+ "$addr.addr = $wb">;
class VLDQQPseudo<InstrItinClass itin>
: PseudoNLdSt<(outs QQPR:$dst), (ins addrmode6:$addr), itin, "">;
class VLDQQWBPseudo<InstrItinClass itin>
// VLD1 : Vector Load (multiple single elements)
class VLD1D<bits<4> op7_4, string Dt>
- : NLdSt<0,0b10,0b0111,op7_4, (outs DPR:$Vd),
+ : NLdSt<0,0b10,0b0111,op7_4, (outs VecListOneD:$Vd),
(ins addrmode6:$Rn), IIC_VLD1,
- "vld1", Dt, "\\{$Vd\\}, $Rn", "", []> {
+ "vld1", Dt, "$Vd, $Rn", "", []> {
let Rm = 0b1111;
let Inst{4} = Rn{4};
let DecoderMethod = "DecodeVLDInstruction";
}
class VLD1Q<bits<4> op7_4, string Dt>
- : NLdSt<0,0b10,0b1010,op7_4, (outs DPR:$Vd, DPR:$dst2),
+ : NLdSt<0,0b10,0b1010,op7_4, (outs VecListTwoD:$Vd),
(ins addrmode6:$Rn), IIC_VLD1x2,
- "vld1", Dt, "\\{$Vd, $dst2\\}, $Rn", "", []> {
+ "vld1", Dt, "$Vd, $Rn", "", []> {
let Rm = 0b1111;
let Inst{5-4} = Rn{5-4};
let DecoderMethod = "DecodeVLDInstruction";
def VLD1q64Pseudo : VLDQPseudo<IIC_VLD1x2>;
// ...with address register writeback:
-class VLD1DWB<bits<4> op7_4, string Dt>
- : NLdSt<0,0b10,0b0111,op7_4, (outs DPR:$Vd, GPR:$wb),
- (ins addrmode6:$Rn, am6offset:$Rm), IIC_VLD1u,
- "vld1", Dt, "\\{$Vd\\}, $Rn$Rm",
- "$Rn.addr = $wb", []> {
- let Inst{4} = Rn{4};
- let DecoderMethod = "DecodeVLDInstruction";
+multiclass VLD1DWB<bits<4> op7_4, string Dt> {
+ def _fixed : NLdSt<0,0b10, 0b0111,op7_4, (outs VecListOneD:$Vd, GPR:$wb),
+ (ins addrmode6:$Rn), IIC_VLD1u,
+ "vld1", Dt, "$Vd, $Rn!",
+ "$Rn.addr = $wb", []> {
+ let Rm = 0b1101; // NLdSt will assign to the right encoding bits.
+ let Inst{4} = Rn{4};
+ let DecoderMethod = "DecodeVLDInstruction";
+ let AsmMatchConverter = "cvtVLDwbFixed";
+ }
+ def _register : NLdSt<0,0b10,0b0111,op7_4, (outs VecListOneD:$Vd, GPR:$wb),
+ (ins addrmode6:$Rn, rGPR:$Rm), IIC_VLD1u,
+ "vld1", Dt, "$Vd, $Rn, $Rm",
+ "$Rn.addr = $wb", []> {
+ let Inst{4} = Rn{4};
+ let DecoderMethod = "DecodeVLDInstruction";
+ let AsmMatchConverter = "cvtVLDwbRegister";
+ }
}
-class VLD1QWB<bits<4> op7_4, string Dt>
- : NLdSt<0,0b10,0b1010,op7_4, (outs DPR:$Vd, DPR:$dst2, GPR:$wb),
- (ins addrmode6:$Rn, am6offset:$Rm), IIC_VLD1x2u,
- "vld1", Dt, "\\{$Vd, $dst2\\}, $Rn$Rm",
- "$Rn.addr = $wb", []> {
- let Inst{5-4} = Rn{5-4};
- let DecoderMethod = "DecodeVLDInstruction";
+multiclass VLD1QWB<bits<4> op7_4, string Dt> {
+ def _fixed : NLdSt<0,0b10,0b1010,op7_4, (outs VecListTwoD:$Vd, GPR:$wb),
+ (ins addrmode6:$Rn), IIC_VLD1x2u,
+ "vld1", Dt, "$Vd, $Rn!",
+ "$Rn.addr = $wb", []> {
+ let Rm = 0b1101; // NLdSt will assign to the right encoding bits.
+ let Inst{5-4} = Rn{5-4};
+ let DecoderMethod = "DecodeVLDInstruction";
+ let AsmMatchConverter = "cvtVLDwbFixed";
+ }
+ def _register : NLdSt<0,0b10,0b1010,op7_4, (outs VecListTwoD:$Vd, GPR:$wb),
+ (ins addrmode6:$Rn, rGPR:$Rm), IIC_VLD1x2u,
+ "vld1", Dt, "$Vd, $Rn, $Rm",
+ "$Rn.addr = $wb", []> {
+ let Inst{5-4} = Rn{5-4};
+ let DecoderMethod = "DecodeVLDInstruction";
+ let AsmMatchConverter = "cvtVLDwbRegister";
+ }
}
-def VLD1d8_UPD : VLD1DWB<{0,0,0,?}, "8">;
-def VLD1d16_UPD : VLD1DWB<{0,1,0,?}, "16">;
-def VLD1d32_UPD : VLD1DWB<{1,0,0,?}, "32">;
-def VLD1d64_UPD : VLD1DWB<{1,1,0,?}, "64">;
-
-def VLD1q8_UPD : VLD1QWB<{0,0,?,?}, "8">;
-def VLD1q16_UPD : VLD1QWB<{0,1,?,?}, "16">;
-def VLD1q32_UPD : VLD1QWB<{1,0,?,?}, "32">;
-def VLD1q64_UPD : VLD1QWB<{1,1,?,?}, "64">;
-
-def VLD1q8Pseudo_UPD : VLDQWBPseudo<IIC_VLD1x2u>;
-def VLD1q16Pseudo_UPD : VLDQWBPseudo<IIC_VLD1x2u>;
-def VLD1q32Pseudo_UPD : VLDQWBPseudo<IIC_VLD1x2u>;
-def VLD1q64Pseudo_UPD : VLDQWBPseudo<IIC_VLD1x2u>;
-
-// ...with 3 registers (some of these are only for the disassembler):
+defm VLD1d8wb : VLD1DWB<{0,0,0,?}, "8">;
+defm VLD1d16wb : VLD1DWB<{0,1,0,?}, "16">;
+defm VLD1d32wb : VLD1DWB<{1,0,0,?}, "32">;
+defm VLD1d64wb : VLD1DWB<{1,1,0,?}, "64">;
+defm VLD1q8wb : VLD1QWB<{0,0,?,?}, "8">;
+defm VLD1q16wb : VLD1QWB<{0,1,?,?}, "16">;
+defm VLD1q32wb : VLD1QWB<{1,0,?,?}, "32">;
+defm VLD1q64wb : VLD1QWB<{1,1,?,?}, "64">;
+
+def VLD1q8PseudoWB_fixed : VLDQWBfixedPseudo<IIC_VLD1x2u>;
+def VLD1q16PseudoWB_fixed : VLDQWBfixedPseudo<IIC_VLD1x2u>;
+def VLD1q32PseudoWB_fixed : VLDQWBfixedPseudo<IIC_VLD1x2u>;
+def VLD1q64PseudoWB_fixed : VLDQWBfixedPseudo<IIC_VLD1x2u>;
+def VLD1q8PseudoWB_register : VLDQWBregisterPseudo<IIC_VLD1x2u>;
+def VLD1q16PseudoWB_register : VLDQWBregisterPseudo<IIC_VLD1x2u>;
+def VLD1q32PseudoWB_register : VLDQWBregisterPseudo<IIC_VLD1x2u>;
+def VLD1q64PseudoWB_register : VLDQWBregisterPseudo<IIC_VLD1x2u>;
+
+// ...with 3 registers
class VLD1D3<bits<4> op7_4, string Dt>
- : NLdSt<0,0b10,0b0110,op7_4, (outs DPR:$Vd, DPR:$dst2, DPR:$dst3),
+ : NLdSt<0,0b10,0b0110,op7_4, (outs VecListThreeD:$Vd),
(ins addrmode6:$Rn), IIC_VLD1x3, "vld1", Dt,
- "\\{$Vd, $dst2, $dst3\\}, $Rn", "", []> {
+ "$Vd, $Rn", "", []> {
let Rm = 0b1111;
let Inst{4} = Rn{4};
let DecoderMethod = "DecodeVLDInstruction";
}
-class VLD1D3WB<bits<4> op7_4, string Dt>
- : NLdSt<0,0b10,0b0110,op7_4, (outs DPR:$Vd, DPR:$dst2, DPR:$dst3, GPR:$wb),
- (ins addrmode6:$Rn, am6offset:$Rm), IIC_VLD1x3u, "vld1", Dt,
- "\\{$Vd, $dst2, $dst3\\}, $Rn$Rm", "$Rn.addr = $wb", []> {
- let Inst{4} = Rn{4};
- let DecoderMethod = "DecodeVLDInstruction";
+multiclass VLD1D3WB<bits<4> op7_4, string Dt> {
+ def _fixed : NLdSt<0,0b10,0b0110, op7_4, (outs VecListThreeD:$Vd, GPR:$wb),
+ (ins addrmode6:$Rn), IIC_VLD1x2u,
+ "vld1", Dt, "$Vd, $Rn!",
+ "$Rn.addr = $wb", []> {
+ let Rm = 0b1101; // NLdSt will assign to the right encoding bits.
+ let Inst{4} = Rn{4};
+ let DecoderMethod = "DecodeVLDInstruction";
+ let AsmMatchConverter = "cvtVLDwbFixed";
+ }
+ def _register : NLdSt<0,0b10,0b0110,op7_4, (outs VecListThreeD:$Vd, GPR:$wb),
+ (ins addrmode6:$Rn, rGPR:$Rm), IIC_VLD1x2u,
+ "vld1", Dt, "$Vd, $Rn, $Rm",
+ "$Rn.addr = $wb", []> {
+ let Inst{4} = Rn{4};
+ let DecoderMethod = "DecodeVLDInstruction";
+ let AsmMatchConverter = "cvtVLDwbRegister";
+ }
}
def VLD1d8T : VLD1D3<{0,0,0,?}, "8">;
def VLD1d32T : VLD1D3<{1,0,0,?}, "32">;
def VLD1d64T : VLD1D3<{1,1,0,?}, "64">;
-def VLD1d8T_UPD : VLD1D3WB<{0,0,0,?}, "8">;
-def VLD1d16T_UPD : VLD1D3WB<{0,1,0,?}, "16">;
-def VLD1d32T_UPD : VLD1D3WB<{1,0,0,?}, "32">;
-def VLD1d64T_UPD : VLD1D3WB<{1,1,0,?}, "64">;
+defm VLD1d8Twb : VLD1D3WB<{0,0,0,?}, "8">;
+defm VLD1d16Twb : VLD1D3WB<{0,1,0,?}, "16">;
+defm VLD1d32Twb : VLD1D3WB<{1,0,0,?}, "32">;
+defm VLD1d64Twb : VLD1D3WB<{1,1,0,?}, "64">;
-def VLD1d64TPseudo : VLDQQPseudo<IIC_VLD1x3>;
-def VLD1d64TPseudo_UPD : VLDQQWBPseudo<IIC_VLD1x3u>;
+def VLD1d64TPseudo : VLDQQPseudo<IIC_VLD1x3>;
-// ...with 4 registers (some of these are only for the disassembler):
+// ...with 4 registers
class VLD1D4<bits<4> op7_4, string Dt>
- : NLdSt<0,0b10,0b0010,op7_4,(outs DPR:$Vd, DPR:$dst2, DPR:$dst3, DPR:$dst4),
+ : NLdSt<0, 0b10, 0b0010, op7_4, (outs VecListFourD:$Vd),
(ins addrmode6:$Rn), IIC_VLD1x4, "vld1", Dt,
- "\\{$Vd, $dst2, $dst3, $dst4\\}, $Rn", "", []> {
+ "$Vd, $Rn", "", []> {
let Rm = 0b1111;
let Inst{5-4} = Rn{5-4};
let DecoderMethod = "DecodeVLDInstruction";
}
-class VLD1D4WB<bits<4> op7_4, string Dt>
- : NLdSt<0,0b10,0b0010,op7_4,
- (outs DPR:$Vd, DPR:$dst2, DPR:$dst3, DPR:$dst4, GPR:$wb),
- (ins addrmode6:$Rn, am6offset:$Rm), IIC_VLD1x4u, "vld1", Dt,
- "\\{$Vd, $dst2, $dst3, $dst4\\}, $Rn$Rm", "$Rn.addr = $wb",
- []> {
- let Inst{5-4} = Rn{5-4};
- let DecoderMethod = "DecodeVLDInstruction";
+multiclass VLD1D4WB<bits<4> op7_4, string Dt> {
+ def _fixed : NLdSt<0,0b10,0b0010, op7_4, (outs VecListFourD:$Vd, GPR:$wb),
+ (ins addrmode6:$Rn), IIC_VLD1x2u,
+ "vld1", Dt, "$Vd, $Rn!",
+ "$Rn.addr = $wb", []> {
+ let Rm = 0b1101; // NLdSt will assign to the right encoding bits.
+ let Inst{5-4} = Rn{5-4};
+ let DecoderMethod = "DecodeVLDInstruction";
+ let AsmMatchConverter = "cvtVLDwbFixed";
+ }
+ def _register : NLdSt<0,0b10,0b0010,op7_4, (outs VecListFourD:$Vd, GPR:$wb),
+ (ins addrmode6:$Rn, rGPR:$Rm), IIC_VLD1x2u,
+ "vld1", Dt, "$Vd, $Rn, $Rm",
+ "$Rn.addr = $wb", []> {
+ let Inst{5-4} = Rn{5-4};
+ let DecoderMethod = "DecodeVLDInstruction";
+ let AsmMatchConverter = "cvtVLDwbRegister";
+ }
}
def VLD1d8Q : VLD1D4<{0,0,?,?}, "8">;
def VLD1d32Q : VLD1D4<{1,0,?,?}, "32">;
def VLD1d64Q : VLD1D4<{1,1,?,?}, "64">;
-def VLD1d8Q_UPD : VLD1D4WB<{0,0,?,?}, "8">;
-def VLD1d16Q_UPD : VLD1D4WB<{0,1,?,?}, "16">;
-def VLD1d32Q_UPD : VLD1D4WB<{1,0,?,?}, "32">;
-def VLD1d64Q_UPD : VLD1D4WB<{1,1,?,?}, "64">;
+defm VLD1d8Qwb : VLD1D4WB<{0,0,?,?}, "8">;
+defm VLD1d16Qwb : VLD1D4WB<{0,1,?,?}, "16">;
+defm VLD1d32Qwb : VLD1D4WB<{1,0,?,?}, "32">;
+defm VLD1d64Qwb : VLD1D4WB<{1,1,?,?}, "64">;
-def VLD1d64QPseudo : VLDQQPseudo<IIC_VLD1x4>;
-def VLD1d64QPseudo_UPD : VLDQQWBPseudo<IIC_VLD1x4u>;
+def VLD1d64QPseudo : VLDQQPseudo<IIC_VLD1x4>;
// VLD2 : Vector Load (multiple 2-element structures)
-class VLD2D<bits<4> op11_8, bits<4> op7_4, string Dt>
- : NLdSt<0, 0b10, op11_8, op7_4, (outs DPR:$Vd, DPR:$dst2),
+class VLD2D<bits<4> op11_8, bits<4> op7_4, string Dt, RegisterOperand VdTy>
+ : NLdSt<0, 0b10, op11_8, op7_4, (outs VdTy:$Vd),
(ins addrmode6:$Rn), IIC_VLD2,
- "vld2", Dt, "\\{$Vd, $dst2\\}, $Rn", "", []> {
+ "vld2", Dt, "$Vd, $Rn", "", []> {
let Rm = 0b1111;
let Inst{5-4} = Rn{5-4};
let DecoderMethod = "DecodeVLDInstruction";
}
-class VLD2Q<bits<4> op7_4, string Dt>
+class VLD2Q<bits<4> op7_4, string Dt, RegisterOperand VdTy>
: NLdSt<0, 0b10, 0b0011, op7_4,
- (outs DPR:$Vd, DPR:$dst2, DPR:$dst3, DPR:$dst4),
+ (outs VdTy:$Vd),
(ins addrmode6:$Rn), IIC_VLD2x2,
- "vld2", Dt, "\\{$Vd, $dst2, $dst3, $dst4\\}, $Rn", "", []> {
+ "vld2", Dt, "$Vd, $Rn", "", []> {
let Rm = 0b1111;
let Inst{5-4} = Rn{5-4};
let DecoderMethod = "DecodeVLDInstruction";
}
-def VLD2d8 : VLD2D<0b1000, {0,0,?,?}, "8">;
-def VLD2d16 : VLD2D<0b1000, {0,1,?,?}, "16">;
-def VLD2d32 : VLD2D<0b1000, {1,0,?,?}, "32">;
+def VLD2d8 : VLD2D<0b1000, {0,0,?,?}, "8", VecListTwoD>;
+def VLD2d16 : VLD2D<0b1000, {0,1,?,?}, "16", VecListTwoD>;
+def VLD2d32 : VLD2D<0b1000, {1,0,?,?}, "32", VecListTwoD>;
-def VLD2q8 : VLD2Q<{0,0,?,?}, "8">;
-def VLD2q16 : VLD2Q<{0,1,?,?}, "16">;
-def VLD2q32 : VLD2Q<{1,0,?,?}, "32">;
+def VLD2q8 : VLD2Q<{0,0,?,?}, "8", VecListFourD>;
+def VLD2q16 : VLD2Q<{0,1,?,?}, "16", VecListFourD>;
+def VLD2q32 : VLD2Q<{1,0,?,?}, "32", VecListFourD>;
def VLD2d8Pseudo : VLDQPseudo<IIC_VLD2>;
def VLD2d16Pseudo : VLDQPseudo<IIC_VLD2>;
def VLD2q32Pseudo : VLDQQPseudo<IIC_VLD2x2>;
// ...with address register writeback:
-class VLD2DWB<bits<4> op11_8, bits<4> op7_4, string Dt>
- : NLdSt<0, 0b10, op11_8, op7_4, (outs DPR:$Vd, DPR:$dst2, GPR:$wb),
+class VLD2DWB<bits<4> op11_8, bits<4> op7_4, string Dt, RegisterOperand VdTy>
+ : NLdSt<0, 0b10, op11_8, op7_4, (outs VdTy:$Vd, GPR:$wb),
(ins addrmode6:$Rn, am6offset:$Rm), IIC_VLD2u,
- "vld2", Dt, "\\{$Vd, $dst2\\}, $Rn$Rm",
+ "vld2", Dt, "$Vd, $Rn$Rm",
"$Rn.addr = $wb", []> {
let Inst{5-4} = Rn{5-4};
let DecoderMethod = "DecodeVLDInstruction";
}
-class VLD2QWB<bits<4> op7_4, string Dt>
+class VLD2QWB<bits<4> op7_4, string Dt, RegisterOperand VdTy>
: NLdSt<0, 0b10, 0b0011, op7_4,
- (outs DPR:$Vd, DPR:$dst2, DPR:$dst3, DPR:$dst4, GPR:$wb),
+ (outs VdTy:$Vd, GPR:$wb),
(ins addrmode6:$Rn, am6offset:$Rm), IIC_VLD2x2u,
- "vld2", Dt, "\\{$Vd, $dst2, $dst3, $dst4\\}, $Rn$Rm",
+ "vld2", Dt, "$Vd, $Rn$Rm",
"$Rn.addr = $wb", []> {
let Inst{5-4} = Rn{5-4};
let DecoderMethod = "DecodeVLDInstruction";
}
-def VLD2d8_UPD : VLD2DWB<0b1000, {0,0,?,?}, "8">;
-def VLD2d16_UPD : VLD2DWB<0b1000, {0,1,?,?}, "16">;
-def VLD2d32_UPD : VLD2DWB<0b1000, {1,0,?,?}, "32">;
+def VLD2d8_UPD : VLD2DWB<0b1000, {0,0,?,?}, "8", VecListTwoD>;
+def VLD2d16_UPD : VLD2DWB<0b1000, {0,1,?,?}, "16", VecListTwoD>;
+def VLD2d32_UPD : VLD2DWB<0b1000, {1,0,?,?}, "32", VecListTwoD>;
-def VLD2q8_UPD : VLD2QWB<{0,0,?,?}, "8">;
-def VLD2q16_UPD : VLD2QWB<{0,1,?,?}, "16">;
-def VLD2q32_UPD : VLD2QWB<{1,0,?,?}, "32">;
+def VLD2q8_UPD : VLD2QWB<{0,0,?,?}, "8", VecListFourD>;
+def VLD2q16_UPD : VLD2QWB<{0,1,?,?}, "16", VecListFourD>;
+def VLD2q32_UPD : VLD2QWB<{1,0,?,?}, "32", VecListFourD>;
def VLD2d8Pseudo_UPD : VLDQWBPseudo<IIC_VLD2u>;
def VLD2d16Pseudo_UPD : VLDQWBPseudo<IIC_VLD2u>;
def VLD2q16Pseudo_UPD : VLDQQWBPseudo<IIC_VLD2x2u>;
def VLD2q32Pseudo_UPD : VLDQQWBPseudo<IIC_VLD2x2u>;
-// ...with double-spaced registers (for disassembly only):
-def VLD2b8 : VLD2D<0b1001, {0,0,?,?}, "8">;
-def VLD2b16 : VLD2D<0b1001, {0,1,?,?}, "16">;
-def VLD2b32 : VLD2D<0b1001, {1,0,?,?}, "32">;
-def VLD2b8_UPD : VLD2DWB<0b1001, {0,0,?,?}, "8">;
-def VLD2b16_UPD : VLD2DWB<0b1001, {0,1,?,?}, "16">;
-def VLD2b32_UPD : VLD2DWB<0b1001, {1,0,?,?}, "32">;
+// ...with double-spaced registers
+def VLD2b8 : VLD2D<0b1001, {0,0,?,?}, "8", VecListTwoQ>;
+def VLD2b16 : VLD2D<0b1001, {0,1,?,?}, "16", VecListTwoQ>;
+def VLD2b32 : VLD2D<0b1001, {1,0,?,?}, "32", VecListTwoQ>;
+def VLD2b8_UPD : VLD2DWB<0b1001, {0,0,?,?}, "8", VecListTwoQ>;
+def VLD2b16_UPD : VLD2DWB<0b1001, {0,1,?,?}, "16", VecListTwoQ>;
+def VLD2b32_UPD : VLD2DWB<0b1001, {1,0,?,?}, "32", VecListTwoQ>;
// VLD3 : Vector Load (multiple 3-element structures)
class VLD3D<bits<4> op11_8, bits<4> op7_4, string Dt>
// VLD1DUP : Vector Load (single element to all lanes)
class VLD1DUP<bits<4> op7_4, string Dt, ValueType Ty, PatFrag LoadOp>
- : NLdSt<1, 0b10, 0b1100, op7_4, (outs DPR:$Vd), (ins addrmode6dup:$Rn),
- IIC_VLD1dup, "vld1", Dt, "\\{$Vd[]\\}, $Rn", "",
- [(set DPR:$Vd, (Ty (NEONvdup (i32 (LoadOp addrmode6dup:$Rn)))))]> {
+ : NLdSt<1, 0b10, 0b1100, op7_4, (outs VecListOneDAllLanes:$Vd),
+ (ins addrmode6dup:$Rn),
+ IIC_VLD1dup, "vld1", Dt, "$Vd, $Rn", "",
+ [(set VecListOneDAllLanes:$Vd,
+ (Ty (NEONvdup (i32 (LoadOp addrmode6dup:$Rn)))))]> {
let Rm = 0b1111;
let Inst{4} = Rn{4};
let DecoderMethod = "DecodeVLD1DupInstruction";
let mayLoad = 1, neverHasSideEffects = 1, hasExtraDefRegAllocReq = 1 in {
class VLD1QDUP<bits<4> op7_4, string Dt>
- : NLdSt<1, 0b10, 0b1100, op7_4, (outs DPR:$Vd, DPR:$dst2),
+ : NLdSt<1, 0b10, 0b1100, op7_4, (outs VecListTwoDAllLanes:$Vd),
(ins addrmode6dup:$Rn), IIC_VLD1dup,
- "vld1", Dt, "\\{$Vd[], $dst2[]\\}, $Rn", "", []> {
+ "vld1", Dt, "$Vd, $Rn", "", []> {
let Rm = 0b1111;
let Inst{4} = Rn{4};
let DecoderMethod = "DecodeVLD1DupInstruction";
def VLD1DUPq32 : VLD1QDUP<{1,0,1,?}, "32">;
// ...with address register writeback:
-class VLD1DUPWB<bits<4> op7_4, string Dt>
- : NLdSt<1, 0b10, 0b1100, op7_4, (outs DPR:$Vd, GPR:$wb),
- (ins addrmode6dup:$Rn, am6offset:$Rm), IIC_VLD1dupu,
- "vld1", Dt, "\\{$Vd[]\\}, $Rn$Rm", "$Rn.addr = $wb", []> {
- let Inst{4} = Rn{4};
- let DecoderMethod = "DecodeVLD1DupInstruction";
+multiclass VLD1DUPWB<bits<4> op7_4, string Dt> {
+ def _fixed : NLdSt<1, 0b10, 0b1100, op7_4,
+ (outs VecListOneDAllLanes:$Vd, GPR:$wb),
+ (ins addrmode6dup:$Rn), IIC_VLD1dupu,
+ "vld1", Dt, "$Vd, $Rn!",
+ "$Rn.addr = $wb", []> {
+ let Rm = 0b1101; // NLdSt will assign to the right encoding bits.
+ let Inst{4} = Rn{4};
+ let DecoderMethod = "DecodeVLD1DupInstruction";
+ let AsmMatchConverter = "cvtVLDwbFixed";
+ }
+ def _register : NLdSt<1, 0b10, 0b1100, op7_4,
+ (outs VecListOneDAllLanes:$Vd, GPR:$wb),
+ (ins addrmode6dup:$Rn, rGPR:$Rm), IIC_VLD1dupu,
+ "vld1", Dt, "$Vd, $Rn, $Rm",
+ "$Rn.addr = $wb", []> {
+ let Inst{4} = Rn{4};
+ let DecoderMethod = "DecodeVLD1DupInstruction";
+ let AsmMatchConverter = "cvtVLDwbRegister";
+ }
}
-class VLD1QDUPWB<bits<4> op7_4, string Dt>
- : NLdSt<1, 0b10, 0b1100, op7_4, (outs DPR:$Vd, DPR:$dst2, GPR:$wb),
- (ins addrmode6dup:$Rn, am6offset:$Rm), IIC_VLD1dupu,
- "vld1", Dt, "\\{$Vd[], $dst2[]\\}, $Rn$Rm", "$Rn.addr = $wb", []> {
- let Inst{4} = Rn{4};
- let DecoderMethod = "DecodeVLD1DupInstruction";
+multiclass VLD1QDUPWB<bits<4> op7_4, string Dt> {
+ def _fixed : NLdSt<1, 0b10, 0b1100, op7_4,
+ (outs VecListTwoDAllLanes:$Vd, GPR:$wb),
+ (ins addrmode6dup:$Rn), IIC_VLD1dupu,
+ "vld1", Dt, "$Vd, $Rn!",
+ "$Rn.addr = $wb", []> {
+ let Rm = 0b1101; // NLdSt will assign to the right encoding bits.
+ let Inst{4} = Rn{4};
+ let DecoderMethod = "DecodeVLD1DupInstruction";
+ let AsmMatchConverter = "cvtVLDwbFixed";
+ }
+ def _register : NLdSt<1, 0b10, 0b1100, op7_4,
+ (outs VecListTwoDAllLanes:$Vd, GPR:$wb),
+ (ins addrmode6dup:$Rn, rGPR:$Rm), IIC_VLD1dupu,
+ "vld1", Dt, "$Vd, $Rn, $Rm",
+ "$Rn.addr = $wb", []> {
+ let Inst{4} = Rn{4};
+ let DecoderMethod = "DecodeVLD1DupInstruction";
+ let AsmMatchConverter = "cvtVLDwbRegister";
+ }
}
-def VLD1DUPd8_UPD : VLD1DUPWB<{0,0,0,0}, "8">;
-def VLD1DUPd16_UPD : VLD1DUPWB<{0,1,0,?}, "16">;
-def VLD1DUPd32_UPD : VLD1DUPWB<{1,0,0,?}, "32">;
+defm VLD1DUPd8wb : VLD1DUPWB<{0,0,0,0}, "8">;
+defm VLD1DUPd16wb : VLD1DUPWB<{0,1,0,?}, "16">;
+defm VLD1DUPd32wb : VLD1DUPWB<{1,0,0,?}, "32">;
-def VLD1DUPq8_UPD : VLD1QDUPWB<{0,0,1,0}, "8">;
-def VLD1DUPq16_UPD : VLD1QDUPWB<{0,1,1,?}, "16">;
-def VLD1DUPq32_UPD : VLD1QDUPWB<{1,0,1,?}, "32">;
+defm VLD1DUPq8wb : VLD1QDUPWB<{0,0,1,0}, "8">;
+defm VLD1DUPq16wb : VLD1QDUPWB<{0,1,1,?}, "16">;
+defm VLD1DUPq32wb : VLD1QDUPWB<{1,0,1,?}, "32">;
-def VLD1DUPq8Pseudo_UPD : VLDQWBPseudo<IIC_VLD1dupu>;
-def VLD1DUPq16Pseudo_UPD : VLDQWBPseudo<IIC_VLD1dupu>;
-def VLD1DUPq32Pseudo_UPD : VLDQWBPseudo<IIC_VLD1dupu>;
+def VLD1DUPq8PseudoWB_fixed : VLDQWBfixedPseudo<IIC_VLD1dupu>;
+def VLD1DUPq16PseudoWB_fixed : VLDQWBfixedPseudo<IIC_VLD1dupu>;
+def VLD1DUPq32PseudoWB_fixed : VLDQWBfixedPseudo<IIC_VLD1dupu>;
+def VLD1DUPq8PseudoWB_register : VLDQWBregisterPseudo<IIC_VLD1dupu>;
+def VLD1DUPq16PseudoWB_register : VLDQWBregisterPseudo<IIC_VLD1dupu>;
+def VLD1DUPq32PseudoWB_register : VLDQWBregisterPseudo<IIC_VLD1dupu>;
// VLD2DUP : Vector Load (single 2-element structure to all lanes)
class VLD2DUP<bits<4> op7_4, string Dt>
: PseudoNLdSt<(outs GPR:$wb),
(ins addrmode6:$addr, am6offset:$offset, QPR:$src), itin,
"$addr.addr = $wb">;
+class VSTQWBfixedPseudo<InstrItinClass itin>
+ : PseudoNLdSt<(outs GPR:$wb),
+ (ins addrmode6:$addr, QPR:$src), itin,
+ "$addr.addr = $wb">;
+class VSTQWBregisterPseudo<InstrItinClass itin>
+ : PseudoNLdSt<(outs GPR:$wb),
+ (ins addrmode6:$addr, rGPR:$offset, QPR:$src), itin,
+ "$addr.addr = $wb">;
class VSTQQPseudo<InstrItinClass itin>
: PseudoNLdSt<(outs), (ins addrmode6:$addr, QQPR:$src), itin, "">;
class VSTQQWBPseudo<InstrItinClass itin>
// VST1 : Vector Store (multiple single elements)
class VST1D<bits<4> op7_4, string Dt>
- : NLdSt<0,0b00,0b0111,op7_4, (outs), (ins addrmode6:$Rn, DPR:$Vd),
- IIC_VST1, "vst1", Dt, "\\{$Vd\\}, $Rn", "", []> {
+ : NLdSt<0,0b00,0b0111,op7_4, (outs), (ins addrmode6:$Rn, VecListOneD:$Vd),
+ IIC_VST1, "vst1", Dt, "$Vd, $Rn", "", []> {
let Rm = 0b1111;
let Inst{4} = Rn{4};
let DecoderMethod = "DecodeVSTInstruction";
}
class VST1Q<bits<4> op7_4, string Dt>
- : NLdSt<0,0b00,0b1010,op7_4, (outs),
- (ins addrmode6:$Rn, DPR:$Vd, DPR:$src2), IIC_VST1x2,
- "vst1", Dt, "\\{$Vd, $src2\\}, $Rn", "", []> {
+ : NLdSt<0,0b00,0b1010,op7_4, (outs), (ins addrmode6:$Rn, VecListTwoD:$Vd),
+ IIC_VST1x2, "vst1", Dt, "$Vd, $Rn", "", []> {
let Rm = 0b1111;
let Inst{5-4} = Rn{5-4};
let DecoderMethod = "DecodeVSTInstruction";
def VST1q64Pseudo : VSTQPseudo<IIC_VST1x2>;
// ...with address register writeback:
-class VST1DWB<bits<4> op7_4, string Dt>
- : NLdSt<0, 0b00, 0b0111, op7_4, (outs GPR:$wb),
- (ins addrmode6:$Rn, am6offset:$Rm, DPR:$Vd), IIC_VST1u,
- "vst1", Dt, "\\{$Vd\\}, $Rn$Rm", "$Rn.addr = $wb", []> {
- let Inst{4} = Rn{4};
- let DecoderMethod = "DecodeVSTInstruction";
+multiclass VST1DWB<bits<4> op7_4, string Dt> {
+ def _fixed : NLdSt<0,0b00, 0b0111,op7_4, (outs GPR:$wb),
+ (ins addrmode6:$Rn, VecListOneD:$Vd), IIC_VLD1u,
+ "vst1", Dt, "$Vd, $Rn!",
+ "$Rn.addr = $wb", []> {
+ let Rm = 0b1101; // NLdSt will assign to the right encoding bits.
+ let Inst{4} = Rn{4};
+ let DecoderMethod = "DecodeVSTInstruction";
+ let AsmMatchConverter = "cvtVSTwbFixed";
+ }
+ def _register : NLdSt<0,0b00,0b0111,op7_4, (outs GPR:$wb),
+ (ins addrmode6:$Rn, rGPR:$Rm, VecListOneD:$Vd),
+ IIC_VLD1u,
+ "vst1", Dt, "$Vd, $Rn, $Rm",
+ "$Rn.addr = $wb", []> {
+ let Inst{4} = Rn{4};
+ let DecoderMethod = "DecodeVSTInstruction";
+ let AsmMatchConverter = "cvtVSTwbRegister";
+ }
}
-class VST1QWB<bits<4> op7_4, string Dt>
- : NLdSt<0, 0b00, 0b1010, op7_4, (outs GPR:$wb),
- (ins addrmode6:$Rn, am6offset:$Rm, DPR:$Vd, DPR:$src2),
- IIC_VST1x2u, "vst1", Dt, "\\{$Vd, $src2\\}, $Rn$Rm",
- "$Rn.addr = $wb", []> {
- let Inst{5-4} = Rn{5-4};
- let DecoderMethod = "DecodeVSTInstruction";
+multiclass VST1QWB<bits<4> op7_4, string Dt> {
+ def _fixed : NLdSt<0,0b00,0b1010,op7_4, (outs GPR:$wb),
+ (ins addrmode6:$Rn, VecListTwoD:$Vd), IIC_VLD1x2u,
+ "vst1", Dt, "$Vd, $Rn!",
+ "$Rn.addr = $wb", []> {
+ let Rm = 0b1101; // NLdSt will assign to the right encoding bits.
+ let Inst{5-4} = Rn{5-4};
+ let DecoderMethod = "DecodeVSTInstruction";
+ let AsmMatchConverter = "cvtVSTwbFixed";
+ }
+ def _register : NLdSt<0,0b00,0b1010,op7_4, (outs GPR:$wb),
+ (ins addrmode6:$Rn, rGPR:$Rm, VecListTwoD:$Vd),
+ IIC_VLD1x2u,
+ "vst1", Dt, "$Vd, $Rn, $Rm",
+ "$Rn.addr = $wb", []> {
+ let Inst{5-4} = Rn{5-4};
+ let DecoderMethod = "DecodeVSTInstruction";
+ let AsmMatchConverter = "cvtVSTwbRegister";
+ }
}
-def VST1d8_UPD : VST1DWB<{0,0,0,?}, "8">;
-def VST1d16_UPD : VST1DWB<{0,1,0,?}, "16">;
-def VST1d32_UPD : VST1DWB<{1,0,0,?}, "32">;
-def VST1d64_UPD : VST1DWB<{1,1,0,?}, "64">;
+defm VST1d8wb : VST1DWB<{0,0,0,?}, "8">;
+defm VST1d16wb : VST1DWB<{0,1,0,?}, "16">;
+defm VST1d32wb : VST1DWB<{1,0,0,?}, "32">;
+defm VST1d64wb : VST1DWB<{1,1,0,?}, "64">;
-def VST1q8_UPD : VST1QWB<{0,0,?,?}, "8">;
-def VST1q16_UPD : VST1QWB<{0,1,?,?}, "16">;
-def VST1q32_UPD : VST1QWB<{1,0,?,?}, "32">;
-def VST1q64_UPD : VST1QWB<{1,1,?,?}, "64">;
+defm VST1q8wb : VST1QWB<{0,0,?,?}, "8">;
+defm VST1q16wb : VST1QWB<{0,1,?,?}, "16">;
+defm VST1q32wb : VST1QWB<{1,0,?,?}, "32">;
+defm VST1q64wb : VST1QWB<{1,1,?,?}, "64">;
-def VST1q8Pseudo_UPD : VSTQWBPseudo<IIC_VST1x2u>;
-def VST1q16Pseudo_UPD : VSTQWBPseudo<IIC_VST1x2u>;
-def VST1q32Pseudo_UPD : VSTQWBPseudo<IIC_VST1x2u>;
-def VST1q64Pseudo_UPD : VSTQWBPseudo<IIC_VST1x2u>;
+def VST1q8PseudoWB_fixed : VSTQWBfixedPseudo<IIC_VST1x2u>;
+def VST1q16PseudoWB_fixed : VSTQWBfixedPseudo<IIC_VST1x2u>;
+def VST1q32PseudoWB_fixed : VSTQWBfixedPseudo<IIC_VST1x2u>;
+def VST1q64PseudoWB_fixed : VSTQWBfixedPseudo<IIC_VST1x2u>;
+def VST1q8PseudoWB_register : VSTQWBregisterPseudo<IIC_VST1x2u>;
+def VST1q16PseudoWB_register : VSTQWBregisterPseudo<IIC_VST1x2u>;
+def VST1q32PseudoWB_register : VSTQWBregisterPseudo<IIC_VST1x2u>;
+def VST1q64PseudoWB_register : VSTQWBregisterPseudo<IIC_VST1x2u>;
-// ...with 3 registers (some of these are only for the disassembler):
+// ...with 3 registers
class VST1D3<bits<4> op7_4, string Dt>
: NLdSt<0, 0b00, 0b0110, op7_4, (outs),
- (ins addrmode6:$Rn, DPR:$Vd, DPR:$src2, DPR:$src3),
- IIC_VST1x3, "vst1", Dt, "\\{$Vd, $src2, $src3\\}, $Rn", "", []> {
+ (ins addrmode6:$Rn, VecListThreeD:$Vd),
+ IIC_VST1x3, "vst1", Dt, "$Vd, $Rn", "", []> {
let Rm = 0b1111;
let Inst{4} = Rn{4};
let DecoderMethod = "DecodeVSTInstruction";
}
-class VST1D3WB<bits<4> op7_4, string Dt>
- : NLdSt<0, 0b00, 0b0110, op7_4, (outs GPR:$wb),
- (ins addrmode6:$Rn, am6offset:$Rm,
- DPR:$Vd, DPR:$src2, DPR:$src3),
- IIC_VST1x3u, "vst1", Dt, "\\{$Vd, $src2, $src3\\}, $Rn$Rm",
- "$Rn.addr = $wb", []> {
- let Inst{4} = Rn{4};
- let DecoderMethod = "DecodeVSTInstruction";
+multiclass VST1D3WB<bits<4> op7_4, string Dt> {
+ def _fixed : NLdSt<0,0b00,0b0110,op7_4, (outs GPR:$wb),
+ (ins addrmode6:$Rn, VecListThreeD:$Vd), IIC_VLD1x3u,
+ "vst1", Dt, "$Vd, $Rn!",
+ "$Rn.addr = $wb", []> {
+ let Rm = 0b1101; // NLdSt will assign to the right encoding bits.
+ let Inst{5-4} = Rn{5-4};
+ let DecoderMethod = "DecodeVSTInstruction";
+ let AsmMatchConverter = "cvtVSTwbFixed";
+ }
+ def _register : NLdSt<0,0b00,0b0110,op7_4, (outs GPR:$wb),
+ (ins addrmode6:$Rn, rGPR:$Rm, VecListThreeD:$Vd),
+ IIC_VLD1x3u,
+ "vst1", Dt, "$Vd, $Rn, $Rm",
+ "$Rn.addr = $wb", []> {
+ let Inst{5-4} = Rn{5-4};
+ let DecoderMethod = "DecodeVSTInstruction";
+ let AsmMatchConverter = "cvtVSTwbRegister";
+ }
}
-def VST1d8T : VST1D3<{0,0,0,?}, "8">;
-def VST1d16T : VST1D3<{0,1,0,?}, "16">;
-def VST1d32T : VST1D3<{1,0,0,?}, "32">;
-def VST1d64T : VST1D3<{1,1,0,?}, "64">;
+def VST1d8T : VST1D3<{0,0,0,?}, "8">;
+def VST1d16T : VST1D3<{0,1,0,?}, "16">;
+def VST1d32T : VST1D3<{1,0,0,?}, "32">;
+def VST1d64T : VST1D3<{1,1,0,?}, "64">;
-def VST1d8T_UPD : VST1D3WB<{0,0,0,?}, "8">;
-def VST1d16T_UPD : VST1D3WB<{0,1,0,?}, "16">;
-def VST1d32T_UPD : VST1D3WB<{1,0,0,?}, "32">;
-def VST1d64T_UPD : VST1D3WB<{1,1,0,?}, "64">;
+defm VST1d8Twb : VST1D3WB<{0,0,0,?}, "8">;
+defm VST1d16Twb : VST1D3WB<{0,1,0,?}, "16">;
+defm VST1d32Twb : VST1D3WB<{1,0,0,?}, "32">;
+defm VST1d64Twb : VST1D3WB<{1,1,0,?}, "64">;
-def VST1d64TPseudo : VSTQQPseudo<IIC_VST1x3>;
-def VST1d64TPseudo_UPD : VSTQQWBPseudo<IIC_VST1x3u>;
+def VST1d64TPseudo : VSTQQPseudo<IIC_VST1x3>;
+def VST1d64TPseudoWB_fixed : VSTQQWBPseudo<IIC_VST1x3u>;
+def VST1d64TPseudoWB_register : VSTQQWBPseudo<IIC_VST1x3u>;
-// ...with 4 registers (some of these are only for the disassembler):
+// ...with 4 registers
class VST1D4<bits<4> op7_4, string Dt>
: NLdSt<0, 0b00, 0b0010, op7_4, (outs),
- (ins addrmode6:$Rn, DPR:$Vd, DPR:$src2, DPR:$src3, DPR:$src4),
- IIC_VST1x4, "vst1", Dt, "\\{$Vd, $src2, $src3, $src4\\}, $Rn", "",
+ (ins addrmode6:$Rn, VecListFourD:$Vd),
+ IIC_VST1x4, "vst1", Dt, "$Vd, $Rn", "",
[]> {
let Rm = 0b1111;
let Inst{5-4} = Rn{5-4};
let DecoderMethod = "DecodeVSTInstruction";
}
-class VST1D4WB<bits<4> op7_4, string Dt>
- : NLdSt<0, 0b00, 0b0010, op7_4, (outs GPR:$wb),
- (ins addrmode6:$Rn, am6offset:$Rm,
- DPR:$Vd, DPR:$src2, DPR:$src3, DPR:$src4), IIC_VST1x4u,
- "vst1", Dt, "\\{$Vd, $src2, $src3, $src4\\}, $Rn$Rm",
- "$Rn.addr = $wb", []> {
- let Inst{5-4} = Rn{5-4};
- let DecoderMethod = "DecodeVSTInstruction";
+multiclass VST1D4WB<bits<4> op7_4, string Dt> {
+ def _fixed : NLdSt<0,0b00,0b0010,op7_4, (outs GPR:$wb),
+ (ins addrmode6:$Rn, VecListFourD:$Vd), IIC_VLD1x4u,
+ "vst1", Dt, "$Vd, $Rn!",
+ "$Rn.addr = $wb", []> {
+ let Rm = 0b1101; // NLdSt will assign to the right encoding bits.
+ let Inst{5-4} = Rn{5-4};
+ let DecoderMethod = "DecodeVSTInstruction";
+ let AsmMatchConverter = "cvtVSTwbFixed";
+ }
+ def _register : NLdSt<0,0b00,0b0010,op7_4, (outs GPR:$wb),
+ (ins addrmode6:$Rn, rGPR:$Rm, VecListFourD:$Vd),
+ IIC_VLD1x4u,
+ "vst1", Dt, "$Vd, $Rn, $Rm",
+ "$Rn.addr = $wb", []> {
+ let Inst{5-4} = Rn{5-4};
+ let DecoderMethod = "DecodeVSTInstruction";
+ let AsmMatchConverter = "cvtVSTwbRegister";
+ }
}
-def VST1d8Q : VST1D4<{0,0,?,?}, "8">;
-def VST1d16Q : VST1D4<{0,1,?,?}, "16">;
-def VST1d32Q : VST1D4<{1,0,?,?}, "32">;
-def VST1d64Q : VST1D4<{1,1,?,?}, "64">;
+def VST1d8Q : VST1D4<{0,0,?,?}, "8">;
+def VST1d16Q : VST1D4<{0,1,?,?}, "16">;
+def VST1d32Q : VST1D4<{1,0,?,?}, "32">;
+def VST1d64Q : VST1D4<{1,1,?,?}, "64">;
-def VST1d8Q_UPD : VST1D4WB<{0,0,?,?}, "8">;
-def VST1d16Q_UPD : VST1D4WB<{0,1,?,?}, "16">;
-def VST1d32Q_UPD : VST1D4WB<{1,0,?,?}, "32">;
-def VST1d64Q_UPD : VST1D4WB<{1,1,?,?}, "64">;
+defm VST1d8Qwb : VST1D4WB<{0,0,?,?}, "8">;
+defm VST1d16Qwb : VST1D4WB<{0,1,?,?}, "16">;
+defm VST1d32Qwb : VST1D4WB<{1,0,?,?}, "32">;
+defm VST1d64Qwb : VST1D4WB<{1,1,?,?}, "64">;
-def VST1d64QPseudo : VSTQQPseudo<IIC_VST1x4>;
-def VST1d64QPseudo_UPD : VSTQQWBPseudo<IIC_VST1x4u>;
+def VST1d64QPseudo : VSTQQPseudo<IIC_VST1x4>;
+def VST1d64QPseudoWB_fixed : VSTQQWBPseudo<IIC_VST1x4u>;
+def VST1d64QPseudoWB_register : VSTQQWBPseudo<IIC_VST1x4u>;
// VST2 : Vector Store (multiple 2-element structures)
class VST2D<bits<4> op11_8, bits<4> op7_4, string Dt>
def VST2q16Pseudo_UPD : VSTQQWBPseudo<IIC_VST2x2u>;
def VST2q32Pseudo_UPD : VSTQQWBPseudo<IIC_VST2x2u>;
-// ...with double-spaced registers (for disassembly only):
+// ...with double-spaced registers
def VST2b8 : VST2D<0b1001, {0,0,?,?}, "8">;
def VST2b16 : VST2D<0b1001, {0,1,?,?}, "16">;
def VST2b32 : VST2D<0b1001, {1,0,?,?}, "32">;
InstrItinClass itin, string OpcodeStr, string Dt,
ValueType ResTy, ValueType OpTy, SDNode ShOp>
: N3VLane32<1, 1, op21_20, op11_8, 1, 0,
- (outs QPR:$Vd), (ins QPR:$Vn, DPR_VFP2:$Vm, nohash_imm:$lane),
- NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm[$lane]", "",
+ (outs QPR:$Vd), (ins QPR:$Vn, DPR_VFP2:$Vm, VectorIndex32:$lane),
+ NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "",
[(set (ResTy QPR:$Vd),
(ResTy (ShOp (ResTy QPR:$Vn),
(ResTy (NEONvduplane (OpTy DPR_VFP2:$Vm),
class N3VQSL16<bits<2> op21_20, bits<4> op11_8, string OpcodeStr, string Dt,
ValueType ResTy, ValueType OpTy, SDNode ShOp>
: N3VLane16<1, 1, op21_20, op11_8, 1, 0,
- (outs QPR:$Vd), (ins QPR:$Vn, DPR_8:$Vm, nohash_imm:$lane),
- NVMulSLFrm, IIC_VMULi16Q, OpcodeStr, Dt,"$Vd, $Vn, $Vm[$lane]","",
+ (outs QPR:$Vd), (ins QPR:$Vn, DPR_8:$Vm, VectorIndex16:$lane),
+ NVMulSLFrm, IIC_VMULi16Q, OpcodeStr, Dt,"$Vd, $Vn, $Vm$lane", "",
[(set (ResTy QPR:$Vd),
(ResTy (ShOp (ResTy QPR:$Vn),
(ResTy (NEONvduplane (OpTy DPR_8:$Vm),
string OpcodeStr, string Dt,
ValueType ResTy, ValueType OpTy, Intrinsic IntOp>
: N3VLane32<1, 1, op21_20, op11_8, 1, 0,
- (outs QPR:$Vd), (ins QPR:$Vn, DPR_VFP2:$Vm, nohash_imm:$lane),
- NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm[$lane]", "",
+ (outs QPR:$Vd), (ins QPR:$Vn, DPR_VFP2:$Vm, VectorIndex32:$lane),
+ NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "",
[(set (ResTy QPR:$Vd),
(ResTy (IntOp (ResTy QPR:$Vn),
(ResTy (NEONvduplane (OpTy DPR_VFP2:$Vm),
string OpcodeStr, string Dt,
ValueType ResTy, ValueType OpTy, Intrinsic IntOp>
: N3VLane16<1, 1, op21_20, op11_8, 1, 0,
- (outs QPR:$Vd), (ins QPR:$Vn, DPR_8:$Vm, nohash_imm:$lane),
- NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm[$lane]", "",
+ (outs QPR:$Vd), (ins QPR:$Vn, DPR_8:$Vm, VectorIndex16:$lane),
+ NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "",
[(set (ResTy QPR:$Vd),
(ResTy (IntOp (ResTy QPR:$Vn),
(ResTy (NEONvduplane (OpTy DPR_8:$Vm),
InstrItinClass itin, string OpcodeStr, string Dt,
ValueType TyQ, ValueType TyD, SDNode MulOp, SDNode OpNode>
: N3VLane32<op24, 1, op21_20, op11_8, 1, 0, (outs QPR:$Vd),
- (ins QPR:$src1, DPR:$Vn, DPR_VFP2:$Vm, nohash_imm:$lane),
+ (ins QPR:$src1, DPR:$Vn, DPR_VFP2:$Vm, VectorIndex32:$lane),
NVMulSLFrm, itin,
- OpcodeStr, Dt, "$Vd, $Vn, $Vm[$lane]", "$src1 = $Vd",
+ OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "$src1 = $Vd",
[(set QPR:$Vd,
(OpNode (TyQ QPR:$src1),
(TyQ (MulOp (TyD DPR:$Vn),
InstrItinClass itin, string OpcodeStr, string Dt,
ValueType TyQ, ValueType TyD, SDNode MulOp, SDNode OpNode>
: N3VLane16<op24, 1, op21_20, op11_8, 1, 0, (outs QPR:$Vd),
- (ins QPR:$src1, DPR:$Vn, DPR_8:$Vm, nohash_imm:$lane),
+ (ins QPR:$src1, DPR:$Vn, DPR_8:$Vm, VectorIndex16:$lane),
NVMulSLFrm, itin,
- OpcodeStr, Dt, "$Vd, $Vn, $Vm[$lane]", "$src1 = $Vd",
+ OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "$src1 = $Vd",
[(set QPR:$Vd,
(OpNode (TyQ QPR:$src1),
(TyQ (MulOp (TyD DPR:$Vn),
ValueType ResTy, ValueType OpTy, Intrinsic IntOp>
: N3VLane32<op24, 1, op21_20, op11_8, 1, 0,
(outs QPR:$Vd),
- (ins QPR:$src1, DPR:$Vn, DPR_VFP2:$Vm, nohash_imm:$lane),
+ (ins QPR:$src1, DPR:$Vn, DPR_VFP2:$Vm, VectorIndex32:$lane),
NVMulSLFrm, itin,
- OpcodeStr, Dt, "$Vd, $Vn, $Vm[$lane]", "$src1 = $Vd",
+ OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "$src1 = $Vd",
[(set (ResTy QPR:$Vd),
(ResTy (IntOp (ResTy QPR:$src1),
(OpTy DPR:$Vn),
InstrItinClass itin, string OpcodeStr, string Dt,
ValueType TyQ, ValueType TyD, SDNode OpNode>
: N3VLane32<op24, 1, op21_20, op11_8, 1, 0,
- (outs QPR:$Vd), (ins DPR:$Vn, DPR_VFP2:$Vm, nohash_imm:$lane),
- NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm[$lane]", "",
+ (outs QPR:$Vd), (ins DPR:$Vn, DPR_VFP2:$Vm, VectorIndex32:$lane),
+ NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "",
[(set QPR:$Vd,
(TyQ (OpNode (TyD DPR:$Vn),
(TyD (NEONvduplane (TyD DPR_VFP2:$Vm),imm:$lane)))))]>;
InstrItinClass itin, string OpcodeStr, string Dt,
ValueType TyQ, ValueType TyD, SDNode OpNode>
: N3VLane16<op24, 1, op21_20, op11_8, 1, 0,
- (outs QPR:$Vd), (ins DPR:$Vn, DPR_8:$Vm, nohash_imm:$lane),
- NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm[$lane]", "",
+ (outs QPR:$Vd), (ins DPR:$Vn, DPR_8:$Vm, VectorIndex16:$lane),
+ NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "",
[(set QPR:$Vd,
(TyQ (OpNode (TyD DPR:$Vn),
(TyD (NEONvduplane (TyD DPR_8:$Vm), imm:$lane)))))]>;
string OpcodeStr, string Dt,
ValueType ResTy, ValueType OpTy, Intrinsic IntOp>
: N3VLane32<op24, 1, op21_20, op11_8, 1, 0,
- (outs QPR:$Vd), (ins DPR:$Vn, DPR_VFP2:$Vm, nohash_imm:$lane),
- NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm[$lane]", "",
+ (outs QPR:$Vd), (ins DPR:$Vn, DPR_VFP2:$Vm, VectorIndex32:$lane),
+ NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "",
[(set (ResTy QPR:$Vd),
(ResTy (IntOp (OpTy DPR:$Vn),
(OpTy (NEONvduplane (OpTy DPR_VFP2:$Vm),
InstrItinClass itin, string OpcodeStr, string Dt,
ValueType ResTy, ValueType OpTy, Intrinsic IntOp>
: N3VLane16<op24, 1, op21_20, op11_8, 1, 0,
- (outs QPR:$Vd), (ins DPR:$Vn, DPR_8:$Vm, nohash_imm:$lane),
- NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm[$lane]", "",
+ (outs QPR:$Vd), (ins DPR:$Vn, DPR_8:$Vm, VectorIndex16:$lane),
+ NVMulSLFrm, itin, OpcodeStr, Dt, "$Vd, $Vn, $Vm$lane", "",
[(set (ResTy QPR:$Vd),
(ResTy (IntOp (OpTy DPR:$Vn),
(OpTy (NEONvduplane (OpTy DPR_8:$Vm),
(outs DPR:$Vd), (ins DPR:$src1, DPR:$Vn, DPR:$Vm),
N3RegFrm, IIC_VBINiD,
"vbif", "$Vd, $Vn, $Vm", "$src1 = $Vd",
- [/* For disassembly only; pattern left blank */]>;
+ []>;
def VBIFq : N3VX<1, 0, 0b11, 0b0001, 1, 1,
(outs QPR:$Vd), (ins QPR:$src1, QPR:$Vn, QPR:$Vm),
N3RegFrm, IIC_VBINiQ,
"vbif", "$Vd, $Vn, $Vm", "$src1 = $Vd",
- [/* For disassembly only; pattern left blank */]>;
+ []>;
// VBIT : Vector Bitwise Insert if True
// like VBSL but with: "vbit $dst, $src2, $src1", "$src3 = $dst",
(outs DPR:$Vd), (ins DPR:$src1, DPR:$Vn, DPR:$Vm),
N3RegFrm, IIC_VBINiD,
"vbit", "$Vd, $Vn, $Vm", "$src1 = $Vd",
- [/* For disassembly only; pattern left blank */]>;
+ []>;
def VBITq : N3VX<1, 0, 0b10, 0b0001, 1, 1,
(outs QPR:$Vd), (ins QPR:$src1, QPR:$Vn, QPR:$Vm),
N3RegFrm, IIC_VBINiQ,
"vbit", "$Vd, $Vn, $Vm", "$src1 = $Vd",
- [/* For disassembly only; pattern left blank */]>;
+ []>;
// VBIT/VBIF are not yet implemented. The TwoAddress pass will not go looking
// for equivalent operations with different register constraints; it just
IIC_VCNTiQ, "vcnt", "8",
v16i8, v16i8, int_arm_neon_vcnt>;
-// Vector Swap -- for disassembly only.
+// Vector Swap
def VSWPd : N2VX<0b11, 0b11, 0b00, 0b10, 0b00000, 0, 0,
(outs DPR:$Vd), (ins DPR:$Vm), NoItinerary,
"vswp", "$Vd, $Vm", "", []>;
(VORRd DPR:$Vd, DPR:$Vm, DPR:$Vm, pred:$p)>;
def : InstAlias<"vmov${p} $Vd, $Vm",
(VORRq QPR:$Vd, QPR:$Vm, QPR:$Vm, pred:$p)>;
+defm : VFPDTAnyNoF64InstAlias<"vmov${p}", "$Vd, $Vm",
+ (VORRd DPR:$Vd, DPR:$Vm, DPR:$Vm, pred:$p)>;
+defm : VFPDTAnyNoF64InstAlias<"vmov${p}", "$Vd, $Vm",
+ (VORRq QPR:$Vd, QPR:$Vm, QPR:$Vm, pred:$p)>;
// VMOV : Vector Move (Immediate)
(ins nImmSplatI64:$SIMM), IIC_VMOVImm,
"vmov", "i64", "$Vd, $SIMM", "",
[(set QPR:$Vd, (v2i64 (NEONvmovImm timm:$SIMM)))]>;
+
+def VMOVv2f32 : N1ModImm<1, 0b000, 0b1111, 0, 0, 0, 1, (outs DPR:$Vd),
+ (ins nImmVMOVF32:$SIMM), IIC_VMOVImm,
+ "vmov", "f32", "$Vd, $SIMM", "",
+ [(set DPR:$Vd, (v2f32 (NEONvmovFPImm timm:$SIMM)))]>;
+def VMOVv4f32 : N1ModImm<1, 0b000, 0b1111, 0, 1, 0, 1, (outs QPR:$Vd),
+ (ins nImmVMOVF32:$SIMM), IIC_VMOVImm,
+ "vmov", "f32", "$Vd, $SIMM", "",
+ [(set QPR:$Vd, (v4f32 (NEONvmovFPImm timm:$SIMM)))]>;
} // isReMaterializable
// VMOV : Vector Get Lane (move scalar to ARM core register)
v4f32, v4i32, uint_to_fp>;
// VCVT : Vector Convert Between Floating-Point and Fixed-Point.
+let DecoderMethod = "DecodeVCVTD" in {
def VCVTf2xsd : N2VCvtD<0, 1, 0b1111, 0, 1, "vcvt", "s32.f32",
v2i32, v2f32, int_arm_neon_vcvtfp2fxs>;
def VCVTf2xud : N2VCvtD<1, 1, 0b1111, 0, 1, "vcvt", "u32.f32",
v2f32, v2i32, int_arm_neon_vcvtfxs2fp>;
def VCVTxu2fd : N2VCvtD<1, 1, 0b1110, 0, 1, "vcvt", "f32.u32",
v2f32, v2i32, int_arm_neon_vcvtfxu2fp>;
+}
+let DecoderMethod = "DecodeVCVTQ" in {
def VCVTf2xsq : N2VCvtQ<0, 1, 0b1111, 0, 1, "vcvt", "s32.f32",
v4i32, v4f32, int_arm_neon_vcvtfp2fxs>;
def VCVTf2xuq : N2VCvtQ<1, 1, 0b1111, 0, 1, "vcvt", "u32.f32",
v4f32, v4i32, int_arm_neon_vcvtfxs2fp>;
def VCVTxu2fq : N2VCvtQ<1, 1, 0b1110, 0, 1, "vcvt", "f32.u32",
v4f32, v4i32, int_arm_neon_vcvtfxu2fp>;
+}
// VCVT : Vector Convert Between Half-Precision and Single-Precision.
def VCVTf2h : N2VNInt<0b11, 0b11, 0b01, 0b10, 0b01100, 0, 0,
let DecoderMethod = "DecodeTBLInstruction" in {
def VTBL1
: N3V<1,1,0b11,0b1000,0,0, (outs DPR:$Vd),
- (ins DPR:$Vn, DPR:$Vm), NVTBLFrm, IIC_VTB1,
- "vtbl", "8", "$Vd, \\{$Vn\\}, $Vm", "",
- [(set DPR:$Vd, (v8i8 (int_arm_neon_vtbl1 DPR:$Vn, DPR:$Vm)))]>;
+ (ins VecListOneD:$Vn, DPR:$Vm), NVTBLFrm, IIC_VTB1,
+ "vtbl", "8", "$Vd, $Vn, $Vm", "",
+ [(set DPR:$Vd, (v8i8 (int_arm_neon_vtbl1 VecListOneD:$Vn, DPR:$Vm)))]>;
let hasExtraSrcRegAllocReq = 1 in {
def VTBL2
: N3V<1,1,0b11,0b1001,0,0, (outs DPR:$Vd),
// VTBX : Vector Table Extension
def VTBX1
: N3V<1,1,0b11,0b1000,1,0, (outs DPR:$Vd),
- (ins DPR:$orig, DPR:$Vn, DPR:$Vm), NVTBLFrm, IIC_VTBX1,
- "vtbx", "8", "$Vd, \\{$Vn\\}, $Vm", "$orig = $Vd",
+ (ins DPR:$orig, VecListOneD:$Vn, DPR:$Vm), NVTBLFrm, IIC_VTBX1,
+ "vtbx", "8", "$Vd, $Vn, $Vm", "$orig = $Vd",
[(set DPR:$Vd, (v8i8 (int_arm_neon_vtbx1
- DPR:$orig, DPR:$Vn, DPR:$Vm)))]>;
+ DPR:$orig, VecListOneD:$Vn, DPR:$Vm)))]>;
let hasExtraSrcRegAllocReq = 1 in {
def VTBX2
: N3V<1,1,0b11,0b1001,1,0, (outs DPR:$Vd),
def : Pat<(v2f64 (bitconvert (v8i16 QPR:$src))), (v2f64 QPR:$src)>;
def : Pat<(v2f64 (bitconvert (v16i8 QPR:$src))), (v2f64 QPR:$src)>;
def : Pat<(v2f64 (bitconvert (v4f32 QPR:$src))), (v2f64 QPR:$src)>;
+
+
+//===----------------------------------------------------------------------===//
+// Assembler aliases
+//
+
+// VAND/VEOR/VORR accept but do not require a type suffix.
+defm : VFPDTAnyInstAlias<"vand${p}", "$Vd, $Vn, $Vm",
+ (VANDd DPR:$Vd, DPR:$Vn, DPR:$Vm, pred:$p)>;
+defm : VFPDTAnyInstAlias<"vand${p}", "$Vd, $Vn, $Vm",
+ (VANDq QPR:$Vd, QPR:$Vn, QPR:$Vm, pred:$p)>;
+defm : VFPDTAnyInstAlias<"veor${p}", "$Vd, $Vn, $Vm",
+ (VEORd DPR:$Vd, DPR:$Vn, DPR:$Vm, pred:$p)>;
+defm : VFPDTAnyInstAlias<"veor${p}", "$Vd, $Vn, $Vm",
+ (VEORq QPR:$Vd, QPR:$Vn, QPR:$Vm, pred:$p)>;
+defm : VFPDTAnyInstAlias<"vorr${p}", "$Vd, $Vn, $Vm",
+ (VORRd DPR:$Vd, DPR:$Vn, DPR:$Vm, pred:$p)>;
+defm : VFPDTAnyInstAlias<"vorr${p}", "$Vd, $Vn, $Vm",
+ (VORRq QPR:$Vd, QPR:$Vn, QPR:$Vm, pred:$p)>;
+
+// VLD1 requires a size suffix, but also accepts type specific variants.
+// Load one D register.
+defm : VFPDT8ReqInstAlias<"vld1${p}", "$Vd, $Rn",
+ (VLD1d8 VecListOneD:$Vd, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vld1${p}", "$Vd, $Rn",
+ (VLD1d16 VecListOneD:$Vd, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vld1${p}", "$Vd, $Rn",
+ (VLD1d32 VecListOneD:$Vd, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vld1${p}", "$Vd, $Rn",
+ (VLD1d64 VecListOneD:$Vd, addrmode6:$Rn, pred:$p)>;
+// with writeback, fixed stride
+defm : VFPDT8ReqInstAlias<"vld1${p}", "$Vd, $Rn!",
+ (VLD1d8wb_fixed VecListOneD:$Vd, zero_reg, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vld1${p}", "$Vd, $Rn!",
+ (VLD1d16wb_fixed VecListOneD:$Vd, zero_reg, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vld1${p}", "$Vd, $Rn!",
+ (VLD1d32wb_fixed VecListOneD:$Vd, zero_reg, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vld1${p}", "$Vd, $Rn!",
+ (VLD1d64wb_fixed VecListOneD:$Vd, zero_reg, addrmode6:$Rn, pred:$p)>;
+// with writeback, register stride
+defm : VFPDT8ReqInstAlias<"vld1${p}", "$Vd, $Rn, $Rm",
+ (VLD1d8wb_register VecListOneD:$Vd, zero_reg, addrmode6:$Rn,
+ rGPR:$Rm, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vld1${p}", "$Vd, $Rn, $Rm",
+ (VLD1d16wb_register VecListOneD:$Vd, zero_reg, addrmode6:$Rn,
+ rGPR:$Rm, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vld1${p}", "$Vd, $Rn, $Rm",
+ (VLD1d32wb_register VecListOneD:$Vd, zero_reg, addrmode6:$Rn,
+ rGPR:$Rm, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vld1${p}", "$Vd, $Rn, $Rm",
+ (VLD1d64wb_register VecListOneD:$Vd, zero_reg, addrmode6:$Rn,
+ rGPR:$Rm, pred:$p)>;
+
+// Load two D registers.
+defm : VFPDT8ReqInstAlias<"vld1${p}", "$Vd, $Rn",
+ (VLD1q8 VecListTwoD:$Vd, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vld1${p}", "$Vd, $Rn",
+ (VLD1q16 VecListTwoD:$Vd, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vld1${p}", "$Vd, $Rn",
+ (VLD1q32 VecListTwoD:$Vd, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vld1${p}", "$Vd, $Rn",
+ (VLD1q64 VecListTwoD:$Vd, addrmode6:$Rn, pred:$p)>;
+// with writeback, fixed stride
+defm : VFPDT8ReqInstAlias<"vld1${p}", "$Vd, $Rn!",
+ (VLD1q8wb_fixed VecListTwoD:$Vd, zero_reg, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vld1${p}", "$Vd, $Rn!",
+ (VLD1q16wb_fixed VecListTwoD:$Vd, zero_reg, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vld1${p}", "$Vd, $Rn!",
+ (VLD1q32wb_fixed VecListTwoD:$Vd, zero_reg, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vld1${p}", "$Vd, $Rn!",
+ (VLD1q64wb_fixed VecListTwoD:$Vd, zero_reg, addrmode6:$Rn, pred:$p)>;
+// with writeback, register stride
+defm : VFPDT8ReqInstAlias<"vld1${p}", "$Vd, $Rn, $Rm",
+ (VLD1q8wb_register VecListTwoD:$Vd, zero_reg, addrmode6:$Rn,
+ rGPR:$Rm, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vld1${p}", "$Vd, $Rn, $Rm",
+ (VLD1q16wb_register VecListTwoD:$Vd, zero_reg, addrmode6:$Rn,
+ rGPR:$Rm, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vld1${p}", "$Vd, $Rn, $Rm",
+ (VLD1q32wb_register VecListTwoD:$Vd, zero_reg, addrmode6:$Rn,
+ rGPR:$Rm, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vld1${p}", "$Vd, $Rn, $Rm",
+ (VLD1q64wb_register VecListTwoD:$Vd, zero_reg, addrmode6:$Rn,
+ rGPR:$Rm, pred:$p)>;
+
+// Load three D registers.
+defm : VFPDT8ReqInstAlias<"vld1${p}", "$Vd, $Rn",
+ (VLD1d8T VecListThreeD:$Vd, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vld1${p}", "$Vd, $Rn",
+ (VLD1d16T VecListThreeD:$Vd, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vld1${p}", "$Vd, $Rn",
+ (VLD1d32T VecListThreeD:$Vd, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vld1${p}", "$Vd, $Rn",
+ (VLD1d64T VecListThreeD:$Vd, addrmode6:$Rn, pred:$p)>;
+// with writeback, fixed stride
+defm : VFPDT8ReqInstAlias<"vld1${p}", "$Vd, $Rn!",
+ (VLD1d8Twb_fixed VecListThreeD:$Vd, zero_reg,
+ addrmode6:$Rn, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vld1${p}", "$Vd, $Rn!",
+ (VLD1d16Twb_fixed VecListThreeD:$Vd, zero_reg,
+ addrmode6:$Rn, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vld1${p}", "$Vd, $Rn!",
+ (VLD1d32Twb_fixed VecListThreeD:$Vd, zero_reg,
+ addrmode6:$Rn, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vld1${p}", "$Vd, $Rn!",
+ (VLD1d64Twb_fixed VecListThreeD:$Vd, zero_reg,
+ addrmode6:$Rn, pred:$p)>;
+// with writeback, register stride
+defm : VFPDT8ReqInstAlias<"vld1${p}", "$Vd, $Rn, $Rm",
+ (VLD1d8Twb_register VecListThreeD:$Vd, zero_reg,
+ addrmode6:$Rn, rGPR:$Rm, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vld1${p}", "$Vd, $Rn, $Rm",
+ (VLD1d16Twb_register VecListThreeD:$Vd, zero_reg,
+ addrmode6:$Rn, rGPR:$Rm, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vld1${p}", "$Vd, $Rn, $Rm",
+ (VLD1d32Twb_register VecListThreeD:$Vd, zero_reg,
+ addrmode6:$Rn, rGPR:$Rm, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vld1${p}", "$Vd, $Rn, $Rm",
+ (VLD1d64Twb_register VecListThreeD:$Vd, zero_reg,
+ addrmode6:$Rn, rGPR:$Rm, pred:$p)>;
+
+
+// Load four D registers.
+defm : VFPDT8ReqInstAlias<"vld1${p}", "$Vd, $Rn",
+ (VLD1d8Q VecListFourD:$Vd, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vld1${p}", "$Vd, $Rn",
+ (VLD1d16Q VecListFourD:$Vd, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vld1${p}", "$Vd, $Rn",
+ (VLD1d32Q VecListFourD:$Vd, addrmode6:$Rn, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vld1${p}", "$Vd, $Rn",
+ (VLD1d64Q VecListFourD:$Vd, addrmode6:$Rn, pred:$p)>;
+// with writeback, fixed stride
+defm : VFPDT8ReqInstAlias<"vld1${p}", "$Vd, $Rn!",
+ (VLD1d8Qwb_fixed VecListFourD:$Vd, zero_reg,
+ addrmode6:$Rn, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vld1${p}", "$Vd, $Rn!",
+ (VLD1d16Qwb_fixed VecListFourD:$Vd, zero_reg,
+ addrmode6:$Rn, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vld1${p}", "$Vd, $Rn!",
+ (VLD1d32Qwb_fixed VecListFourD:$Vd, zero_reg,
+ addrmode6:$Rn, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vld1${p}", "$Vd, $Rn!",
+ (VLD1d64Qwb_fixed VecListFourD:$Vd, zero_reg,
+ addrmode6:$Rn, pred:$p)>;
+// with writeback, register stride
+defm : VFPDT8ReqInstAlias<"vld1${p}", "$Vd, $Rn, $Rm",
+ (VLD1d8Qwb_register VecListFourD:$Vd, zero_reg,
+ addrmode6:$Rn, rGPR:$Rm, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vld1${p}", "$Vd, $Rn, $Rm",
+ (VLD1d16Qwb_register VecListFourD:$Vd, zero_reg,
+ addrmode6:$Rn, rGPR:$Rm, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vld1${p}", "$Vd, $Rn, $Rm",
+ (VLD1d32Qwb_register VecListFourD:$Vd, zero_reg,
+ addrmode6:$Rn, rGPR:$Rm, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vld1${p}", "$Vd, $Rn, $Rm",
+ (VLD1d64Qwb_register VecListFourD:$Vd, zero_reg,
+ addrmode6:$Rn, rGPR:$Rm, pred:$p)>;
+
+// VST1 requires a size suffix, but also accepts type specific variants.
+// Store one D register.
+defm : VFPDT8ReqInstAlias<"vst1${p}", "$Vd, $Rn",
+ (VST1d8 addrmode6:$Rn, VecListOneD:$Vd, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vst1${p}", "$Vd, $Rn",
+ (VST1d16 addrmode6:$Rn, VecListOneD:$Vd, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vst1${p}", "$Vd, $Rn",
+ (VST1d32 addrmode6:$Rn, VecListOneD:$Vd, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vst1${p}", "$Vd, $Rn",
+ (VST1d64 addrmode6:$Rn, VecListOneD:$Vd, pred:$p)>;
+// with writeback, fixed stride
+defm : VFPDT8ReqInstAlias<"vst1${p}", "$Vd, $Rn!",
+ (VST1d8wb_fixed zero_reg, addrmode6:$Rn, VecListOneD:$Vd, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vst1${p}", "$Vd, $Rn!",
+ (VST1d16wb_fixed zero_reg, addrmode6:$Rn, VecListOneD:$Vd, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vst1${p}", "$Vd, $Rn!",
+ (VST1d32wb_fixed zero_reg, addrmode6:$Rn, VecListOneD:$Vd, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vst1${p}", "$Vd, $Rn!",
+ (VST1d64wb_fixed zero_reg, addrmode6:$Rn, VecListOneD:$Vd, pred:$p)>;
+// with writeback, register stride
+defm : VFPDT8ReqInstAlias<"vst1${p}", "$Vd, $Rn, $Rm",
+ (VST1d8wb_register zero_reg, addrmode6:$Rn, rGPR:$Rm,
+ VecListOneD:$Vd, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vst1${p}", "$Vd, $Rn, $Rm",
+ (VST1d16wb_register zero_reg, addrmode6:$Rn, rGPR:$Rm,
+ VecListOneD:$Vd, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vst1${p}", "$Vd, $Rn, $Rm",
+ (VST1d32wb_register zero_reg, addrmode6:$Rn, rGPR:$Rm,
+ VecListOneD:$Vd, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vst1${p}", "$Vd, $Rn, $Rm",
+ (VST1d64wb_register zero_reg, addrmode6:$Rn, rGPR:$Rm,
+ VecListOneD:$Vd, pred:$p)>;
+
+// Store two D registers.
+defm : VFPDT8ReqInstAlias<"vst1${p}", "$Vd, $Rn",
+ (VST1q8 addrmode6:$Rn, VecListTwoD:$Vd, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vst1${p}", "$Vd, $Rn",
+ (VST1q16 addrmode6:$Rn, VecListTwoD:$Vd, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vst1${p}", "$Vd, $Rn",
+ (VST1q32 addrmode6:$Rn, VecListTwoD:$Vd, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vst1${p}", "$Vd, $Rn",
+ (VST1q64 addrmode6:$Rn, VecListTwoD:$Vd, pred:$p)>;
+// with writeback, fixed stride
+defm : VFPDT8ReqInstAlias<"vst1${p}", "$Vd, $Rn!",
+ (VST1q8wb_fixed zero_reg, addrmode6:$Rn, VecListTwoD:$Vd, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vst1${p}", "$Vd, $Rn!",
+ (VST1q16wb_fixed zero_reg, addrmode6:$Rn, VecListTwoD:$Vd, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vst1${p}", "$Vd, $Rn!",
+ (VST1q32wb_fixed zero_reg, addrmode6:$Rn, VecListTwoD:$Vd, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vst1${p}", "$Vd, $Rn!",
+ (VST1q64wb_fixed zero_reg, addrmode6:$Rn, VecListTwoD:$Vd, pred:$p)>;
+// with writeback, register stride
+defm : VFPDT8ReqInstAlias<"vst1${p}", "$Vd, $Rn, $Rm",
+ (VST1q8wb_register zero_reg, addrmode6:$Rn,
+ rGPR:$Rm, VecListTwoD:$Vd, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vst1${p}", "$Vd, $Rn, $Rm",
+ (VST1q16wb_register zero_reg, addrmode6:$Rn,
+ rGPR:$Rm, VecListTwoD:$Vd, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vst1${p}", "$Vd, $Rn, $Rm",
+ (VST1q32wb_register zero_reg, addrmode6:$Rn,
+ rGPR:$Rm, VecListTwoD:$Vd, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vst1${p}", "$Vd, $Rn, $Rm",
+ (VST1q64wb_register zero_reg, addrmode6:$Rn,
+ rGPR:$Rm, VecListTwoD:$Vd, pred:$p)>;
+
+// Load three D registers.
+defm : VFPDT8ReqInstAlias<"vst1${p}", "$Vd, $Rn",
+ (VST1d8T addrmode6:$Rn, VecListThreeD:$Vd, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vst1${p}", "$Vd, $Rn",
+ (VST1d16T addrmode6:$Rn, VecListThreeD:$Vd, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vst1${p}", "$Vd, $Rn",
+ (VST1d32T addrmode6:$Rn, VecListThreeD:$Vd, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vst1${p}", "$Vd, $Rn",
+ (VST1d64T addrmode6:$Rn, VecListThreeD:$Vd, pred:$p)>;
+defm : VFPDT8ReqInstAlias<"vst1${p}", "$Vd, $Rn!",
+ (VST1d8Twb_fixed zero_reg, addrmode6:$Rn, VecListThreeD:$Vd, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vst1${p}", "$Vd, $Rn!",
+ (VST1d16Twb_fixed zero_reg, addrmode6:$Rn, VecListThreeD:$Vd, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vst1${p}", "$Vd, $Rn!",
+ (VST1d32Twb_fixed zero_reg, addrmode6:$Rn, VecListThreeD:$Vd, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vst1${p}", "$Vd, $Rn!",
+ (VST1d64Twb_fixed zero_reg, addrmode6:$Rn, VecListThreeD:$Vd, pred:$p)>;
+defm : VFPDT8ReqInstAlias<"vst1${p}", "$Vd, $Rn, $Rm",
+ (VST1d8Twb_register zero_reg, addrmode6:$Rn, rGPR:$Rm,
+ VecListThreeD:$Vd, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vst1${p}", "$Vd, $Rn, $Rm",
+ (VST1d16Twb_register zero_reg, addrmode6:$Rn, rGPR:$Rm,
+ VecListThreeD:$Vd, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vst1${p}", "$Vd, $Rn, $Rm",
+ (VST1d32Twb_register zero_reg, addrmode6:$Rn, rGPR:$Rm,
+ VecListThreeD:$Vd, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vst1${p}", "$Vd, $Rn, $Rm",
+ (VST1d64Twb_register zero_reg, addrmode6:$Rn, rGPR:$Rm,
+ VecListThreeD:$Vd, pred:$p)>;
+
+// Load four D registers.
+defm : VFPDT8ReqInstAlias<"vst1${p}", "$Vd, $Rn",
+ (VST1d8Q addrmode6:$Rn, VecListFourD:$Vd, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vst1${p}", "$Vd, $Rn",
+ (VST1d16Q addrmode6:$Rn, VecListFourD:$Vd, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vst1${p}", "$Vd, $Rn",
+ (VST1d32Q addrmode6:$Rn, VecListFourD:$Vd, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vst1${p}", "$Vd, $Rn",
+ (VST1d64Q addrmode6:$Rn, VecListFourD:$Vd, pred:$p)>;
+defm : VFPDT8ReqInstAlias<"vst1${p}", "$Vd, $Rn!",
+ (VST1d8Qwb_fixed zero_reg, addrmode6:$Rn, VecListFourD:$Vd, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vst1${p}", "$Vd, $Rn!",
+ (VST1d16Qwb_fixed zero_reg, addrmode6:$Rn, VecListFourD:$Vd, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vst1${p}", "$Vd, $Rn!",
+ (VST1d32Qwb_fixed zero_reg, addrmode6:$Rn, VecListFourD:$Vd, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vst1${p}", "$Vd, $Rn!",
+ (VST1d64Qwb_fixed zero_reg, addrmode6:$Rn, VecListFourD:$Vd, pred:$p)>;
+defm : VFPDT8ReqInstAlias<"vst1${p}", "$Vd, $Rn, $Rm",
+ (VST1d8Qwb_register zero_reg, addrmode6:$Rn, rGPR:$Rm,
+ VecListFourD:$Vd, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vst1${p}", "$Vd, $Rn, $Rm",
+ (VST1d16Qwb_register zero_reg, addrmode6:$Rn, rGPR:$Rm,
+ VecListFourD:$Vd, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vst1${p}", "$Vd, $Rn, $Rm",
+ (VST1d32Qwb_register zero_reg, addrmode6:$Rn, rGPR:$Rm,
+ VecListFourD:$Vd, pred:$p)>;
+defm : VFPDT64ReqInstAlias<"vst1${p}", "$Vd, $Rn, $Rm",
+ (VST1d64Qwb_register zero_reg, addrmode6:$Rn, rGPR:$Rm,
+ VecListFourD:$Vd, pred:$p)>;
+
+
+// VTRN instructions data type suffix aliases for more-specific types.
+defm : VFPDT8ReqInstAlias <"vtrn${p}", "$Dd, $Dm",
+ (VTRNd8 DPR:$Dd, DPR:$Dm, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vtrn${p}", "$Dd, $Dm",
+ (VTRNd16 DPR:$Dd, DPR:$Dm, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vtrn${p}", "$Dd, $Dm",
+ (VTRNd32 DPR:$Dd, DPR:$Dm, pred:$p)>;
+
+defm : VFPDT8ReqInstAlias <"vtrn${p}", "$Qd, $Qm",
+ (VTRNq8 QPR:$Qd, QPR:$Qm, pred:$p)>;
+defm : VFPDT16ReqInstAlias<"vtrn${p}", "$Qd, $Qm",
+ (VTRNq16 QPR:$Qd, QPR:$Qm, pred:$p)>;
+defm : VFPDT32ReqInstAlias<"vtrn${p}", "$Qd, $Qm",
+ (VTRNq32 QPR:$Qd, QPR:$Qm, pred:$p)>;
+
+// FIXME: Proof of concept pseudos. We want to parameterize these for all
+// the suffices we have to support.
+defm VLD1LNdAsm : NEONDT8AsmPseudoInst<"vld1${p}", "$list, $addr",
+ (ins VecListOneDByteIndexed:$list, addrmode6:$addr, pred:$p)>;
+defm VLD1LNdAsm : NEONDT16AsmPseudoInst<"vld1${p}", "$list, $addr",
+ (ins VecListOneDByteIndexed:$list, addrmode6:$addr, pred:$p)>;
+defm VLD1LNdAsm : NEONDT32AsmPseudoInst<"vld1${p}", "$list, $addr",
+ (ins VecListOneDByteIndexed:$list, addrmode6:$addr, pred:$p)>;