X-Git-Url: http://plrg.eecs.uci.edu/git/?p=oota-llvm.git;a=blobdiff_plain;f=lib%2FTarget%2FAArch64%2FAArch64InstrFormats.td;h=5eef82153e39660e6a8744ea9e9b874b53265484;hp=106f2158909635cfa4aca6d86cdaebff21965cd4;hb=ce8e2a0d91724b07d3a262c6005e5705a3b7837e;hpb=a3b8caf8a739fe19ab27259dc95a9771c11df16f

diff --git a/lib/Target/AArch64/AArch64InstrFormats.td b/lib/Target/AArch64/AArch64InstrFormats.td
index 106f2158909..5eef82153e3 100644
--- a/lib/Target/AArch64/AArch64InstrFormats.td
+++ b/lib/Target/AArch64/AArch64InstrFormats.td
@@ -248,6 +248,12 @@ def simm7s16 : Operand<i32> {
   let PrintMethod = "printImmScale<16>";
 }
 
+def am_indexed7s8   : ComplexPattern<i64, 2, "SelectAddrModeIndexed7S8", []>;
+def am_indexed7s16  : ComplexPattern<i64, 2, "SelectAddrModeIndexed7S16", []>;
+def am_indexed7s32  : ComplexPattern<i64, 2, "SelectAddrModeIndexed7S32", []>;
+def am_indexed7s64  : ComplexPattern<i64, 2, "SelectAddrModeIndexed7S64", []>;
+def am_indexed7s128 : ComplexPattern<i64, 2, "SelectAddrModeIndexed7S128", []>;
+
 class AsmImmRange<int Low, int High> : AsmOperandClass {
   let Name = "Imm" # Low # "_" # High;
   let DiagnosticType = "InvalidImm" # Low # "_" # High;
@@ -346,9 +352,11 @@ class fixedpoint_i64<ValueType FloatVT>
   let ParserMatchClass = Imm1_64Operand;
 }
 
+def fixedpoint_f16_i32 : fixedpoint_i32<f16>;
 def fixedpoint_f32_i32 : fixedpoint_i32<f32>;
 def fixedpoint_f64_i32 : fixedpoint_i32<f64>;
 
+def fixedpoint_f16_i64 : fixedpoint_i64<f16>;
 def fixedpoint_f32_i64 : fixedpoint_i64<f32>;
 def fixedpoint_f64_i64 : fixedpoint_i64<f64>;
 
@@ -402,6 +410,7 @@ def vecshiftR64Narrow : Operand<i32>, ImmLeaf<i32, [{
   let ParserMatchClass = Imm1_32Operand;
 }
 
+def Imm0_1Operand : AsmImmRange<0, 1>;
 def Imm0_7Operand : AsmImmRange<0, 7>;
 def Imm0_15Operand : AsmImmRange<0, 15>;
 def Imm0_31Operand : AsmImmRange<0, 31>;
@@ -441,11 +450,11 @@ def vecshiftL64 : Operand<i32>, ImmLeaf<i32, [{
 // instructions for splatting repeating bit patterns across the immediate.
 def logical_imm32_XFORM : SDNodeXForm<imm, [{
   uint64_t enc = AArch64_AM::encodeLogicalImmediate(N->getZExtValue(), 32);
-  return CurDAG->getTargetConstant(enc, MVT::i32);
+  return CurDAG->getTargetConstant(enc, SDLoc(N), MVT::i32);
 }]>;
 def logical_imm64_XFORM : SDNodeXForm<imm, [{
   uint64_t enc = AArch64_AM::encodeLogicalImmediate(N->getZExtValue(), 64);
-  return CurDAG->getTargetConstant(enc, MVT::i32);
+  return CurDAG->getTargetConstant(enc, SDLoc(N), MVT::i32);
 }]>;
 
 let DiagnosticType = "LogicalSecondSource" in {
@@ -525,6 +534,20 @@ def imm0_31 : Operand<i64>, ImmLeaf<i64, [{
   let ParserMatchClass = Imm0_31Operand;
 }
 
+// True if the 32-bit immediate is in the range [0,31]
+def imm32_0_31 : Operand<i32>, ImmLeaf<i32, [{
+  return ((uint64_t)Imm) < 32;
+}]> {
+  let ParserMatchClass = Imm0_31Operand;
+}
+
+// imm0_1 predicate - True if the immediate is in the range [0,1]
+def imm0_1 : Operand<i64>, ImmLeaf<i64, [{
+  return ((uint64_t)Imm) < 2;
+}]> {
+  let ParserMatchClass = Imm0_1Operand;
+}
+
 // imm0_15 predicate - True if the immediate is in the range [0,15]
 def imm0_15 : Operand<i64>, ImmLeaf<i64, [{
   return ((uint64_t)Imm) < 16;
@@ -542,7 +565,9 @@ def imm0_7 : Operand<i64>, ImmLeaf<i64, [{
 // imm32_0_15 predicate - True if the 32-bit immediate is in the range [0,15]
 def imm32_0_15 : Operand<i32>, ImmLeaf<i32, [{
   return ((uint32_t)Imm) < 16;
-}]>;
+}]> {
+  let ParserMatchClass = Imm0_15Operand;
+}
 
 // An arithmetic shifter operand:
 //  {7-6} - shift type: 00 = lsl, 01 = lsr, 10 = asr
@@ -614,10 +639,15 @@ def move_vec_shift : Operand<i32> {
   let ParserMatchClass = MoveVecShifterOperand;
 }
 
-def AddSubImmOperand : AsmOperandClass {
-  let Name = "AddSubImm";
-  let ParserMethod = "tryParseAddSubImm";
-  let DiagnosticType = "AddSubSecondSource";
+let DiagnosticType = "AddSubSecondSource" in {
+  def AddSubImmOperand : AsmOperandClass {
+    let Name = "AddSubImm";
+    let ParserMethod = "tryParseAddSubImm";
+  }
+  def AddSubImmNegOperand : AsmOperandClass {
+    let Name = "AddSubImmNeg";
+    let ParserMethod = "tryParseAddSubImm";
+  }
 }
 // An ADD/SUB immediate shifter operand:
 //  second operand:
@@ -631,8 +661,17 @@ class addsub_shifted_imm<ValueType Ty>
   let MIOperandInfo = (ops i32imm, i32imm);
 }
 
+class addsub_shifted_imm_neg<ValueType Ty>
+    : Operand<Ty> {
+  let EncoderMethod = "getAddSubImmOpValue";
+  let ParserMatchClass = AddSubImmNegOperand;
+  let MIOperandInfo = (ops i32imm, i32imm);
+}
+
 def addsub_shifted_imm32 : addsub_shifted_imm<i32>;
 def addsub_shifted_imm64 : addsub_shifted_imm<i64>;
+def addsub_shifted_imm32_neg : addsub_shifted_imm_neg<i32>;
+def addsub_shifted_imm64_neg : addsub_shifted_imm_neg<i64>;
 
 class neg_addsub_shifted_imm<ValueType Ty>
     : Operand<Ty>, ComplexPattern<Ty, 2, "SelectNegArithImmed", [imm]> {
@@ -676,13 +715,24 @@ class arith_extended_reg32to64<ValueType Ty> : Operand<Ty>,
 }
 
 // Floating-point immediate.
+def fpimm16 : Operand<f16>,
+              PatLeaf<(f16 fpimm), [{
+      return AArch64_AM::getFP16Imm(N->getValueAPF()) != -1;
+    }], SDNodeXForm<fpimm, [{
+      APFloat InVal = N->getValueAPF();
+      uint32_t enc = AArch64_AM::getFP16Imm(InVal);
+      return CurDAG->getTargetConstant(enc, SDLoc(N), MVT::i32);
+    }]>> {
+  let ParserMatchClass = FPImmOperand;
+  let PrintMethod = "printFPImmOperand";
+}
 def fpimm32 : Operand<f32>,
               PatLeaf<(f32 fpimm), [{
       return AArch64_AM::getFP32Imm(N->getValueAPF()) != -1;
     }], SDNodeXForm<fpimm, [{
       APFloat InVal = N->getValueAPF();
       uint32_t enc = AArch64_AM::getFP32Imm(InVal);
-      return CurDAG->getTargetConstant(enc, MVT::i32);
+      return CurDAG->getTargetConstant(enc, SDLoc(N), MVT::i32);
     }]>> {
   let ParserMatchClass = FPImmOperand;
   let PrintMethod = "printFPImmOperand";
@@ -693,7 +743,7 @@ def fpimm64 : Operand<f64>,
     }], SDNodeXForm<fpimm, [{
       APFloat InVal = N->getValueAPF();
       uint32_t enc = AArch64_AM::getFP64Imm(InVal);
-      return CurDAG->getTargetConstant(enc, MVT::i32);
+      return CurDAG->getTargetConstant(enc, SDLoc(N), MVT::i32);
     }]>> {
   let ParserMatchClass = FPImmOperand;
   let PrintMethod = "printFPImmOperand";
@@ -768,7 +818,7 @@ def simdimmtype10 : Operand<i32>,
       uint32_t enc = AArch64_AM::encodeAdvSIMDModImmType10(N->getValueAPF()
                                                            .bitcastToAPInt()
                                                            .getZExtValue());
-      return CurDAG->getTargetConstant(enc, MVT::i32);
+      return CurDAG->getTargetConstant(enc, SDLoc(N), MVT::i32);
     }]>> {
   let ParserMatchClass = SIMDImmType10Operand;
   let PrintMethod = "printSIMDType10Operand";
@@ -808,7 +858,7 @@ class RtSystemI<bit L, dag oops, dag iops, string asm, string operands>
 // model patterns with sufficiently fine granularity
 let mayStore = 1, mayLoad = 1, hasSideEffects = 1 in
   class HintI<string mnemonic>
-      : SimpleSystemI<0, (ins imm0_127:$imm), mnemonic#" $imm", "",
+      : SimpleSystemI<0, (ins imm0_127:$imm), mnemonic#"\t$imm", "",
                       [(int_aarch64_hint imm0_127:$imm)]>,
         Sched<[WriteHint]> {
     bits <7> imm;
@@ -843,7 +893,7 @@ def MRSSystemRegisterOperand : AsmOperandClass {
   let ParserMethod = "tryParseSysReg";
   let DiagnosticType = "MRS";
 }
-// concatenation of 1, op0, op1, CRn, CRm, op2. 16-bit immediate.
+// concatenation of op0, op1, CRn, CRm, op2. 16-bit immediate.
 def mrs_sysreg_op : Operand<i32> {
   let ParserMatchClass = MRSSystemRegisterOperand;
   let DecoderMethod = "DecodeMRSSystemRegister";
@@ -861,11 +911,29 @@ def msr_sysreg_op : Operand<i32> {
   let PrintMethod = "printMSRSystemRegister";
 }
 
+def PSBHintOperand : AsmOperandClass {
+  let Name = "PSBHint";
+  let ParserMethod = "tryParsePSBHint";
+}
+def psbhint_op : Operand<i32> {
+  let ParserMatchClass = PSBHintOperand;
+  let PrintMethod = "printPSBHintOp";
+  let MCOperandPredicate = [{
+    // Check, if operand is valid, to fix exhaustive aliasing in disassembly.
+    // "psb" is an alias to "hint" only for certain values of CRm:Op2 fields.
+    if (!MCOp.isImm())
+      return false;
+    bool ValidNamed;
+    (void)AArch64PSBHint::PSBHintMapper().toString(MCOp.getImm(),
+      STI.getFeatureBits(), ValidNamed);
+    return ValidNamed;
+  }];
+}
+
 class MRSI : RtSystemI<1, (outs GPR64:$Rt), (ins mrs_sysreg_op:$systemreg),
                        "mrs", "\t$Rt, $systemreg"> {
-  bits<15> systemreg;
-  let Inst{20} = 1;
-  let Inst{19-5} = systemreg;
+  bits<16> systemreg;
+  let Inst{20-5} = systemreg;
 }
 
 // FIXME: Some of these def NZCV, others don't. Best way to model that?
@@ -873,24 +941,23 @@ class MRSI : RtSystemI<1, (outs GPR64:$Rt), (ins mrs_sysreg_op:$systemreg),
 // would do it, but feels like overkill at this point.
 class MSRI : RtSystemI<0, (outs), (ins msr_sysreg_op:$systemreg, GPR64:$Rt),
                        "msr", "\t$systemreg, $Rt"> {
-  bits<15> systemreg;
-  let Inst{20} = 1;
-  let Inst{19-5} = systemreg;
+  bits<16> systemreg;
+  let Inst{20-5} = systemreg;
 }
 
-def SystemPStateFieldOperand : AsmOperandClass {
-  let Name = "SystemPStateField";
+def SystemPStateFieldWithImm0_15Operand : AsmOperandClass {
+  let Name = "SystemPStateFieldWithImm0_15";
   let ParserMethod = "tryParseSysReg";
 }
-def pstatefield_op : Operand<i32> {
-  let ParserMatchClass = SystemPStateFieldOperand;
+def pstatefield4_op : Operand<i32> {
+  let ParserMatchClass = SystemPStateFieldWithImm0_15Operand;
   let PrintMethod = "printSystemPStateField";
 }
 
 let Defs = [NZCV] in
-class MSRpstateI
-  : SimpleSystemI<0, (ins pstatefield_op:$pstate_field, imm0_15:$imm),
-                  "msr", "\t$pstate_field, $imm">,
+class MSRpstateImm0_15
+  : SimpleSystemI<0, (ins pstatefield4_op:$pstatefield, imm0_15:$imm),
+                  "msr", "\t$pstatefield, $imm">,
     Sched<[WriteSys]> {
   bits<6> pstatefield;
   bits<4> imm;
@@ -901,6 +968,37 @@ class MSRpstateI
   let Inst{7-5} = pstatefield{2-0};
 
   let DecoderMethod = "DecodeSystemPStateInstruction";
+  // MSRpstateI aliases with MSRI. When the MSRpstateI decoder method returns
+  // Fail the decoder should attempt to decode the instruction as MSRI.
+  let hasCompleteDecoder = 0;
+}
+
+def SystemPStateFieldWithImm0_1Operand : AsmOperandClass {
+  let Name = "SystemPStateFieldWithImm0_1";
+  let ParserMethod = "tryParseSysReg";
+}
+def pstatefield1_op : Operand<i32> {
+  let ParserMatchClass = SystemPStateFieldWithImm0_1Operand;
+  let PrintMethod = "printSystemPStateField";
+}
+
+let Defs = [NZCV] in
+class MSRpstateImm0_1
+  : SimpleSystemI<0, (ins pstatefield1_op:$pstatefield, imm0_1:$imm),
+                  "msr", "\t$pstatefield, $imm">,
+    Sched<[WriteSys]> {
+  bits<6> pstatefield;
+  bit imm;
+  let Inst{20-19} = 0b00;
+  let Inst{18-16} = pstatefield{5-3};
+  let Inst{15-9} = 0b0100000;
+  let Inst{8} = imm;
+  let Inst{7-5} = pstatefield{2-0};
+
+  let DecoderMethod = "DecodeSystemPStateInstruction";
+  // MSRpstateI aliases with MSRI. When the MSRpstateI decoder method returns
+  // Fail the decoder should attempt to decode the instruction as MSRI.
+  let hasCompleteDecoder = 0;
 }
 
 // SYS and SYSL generic system instructions.
@@ -1329,7 +1427,7 @@ multiclass Shift<bits<2> shift_type, string asm, SDNode OpNode> {
 }
 
 class ShiftAlias<string asm, Instruction inst, RegisterClass regtype>
-    : InstAlias<asm#" $dst, $src1, $src2",
+    : InstAlias<asm#"\t$dst, $src1, $src2",
                 (inst regtype:$dst, regtype:$src1, regtype:$src2), 0>;
 
 class BaseMulAccum<bit isSub, bits<3> opc, RegisterClass multype,
@@ -1351,14 +1449,15 @@ class BaseMulAccum<bit isSub, bits<3> opc, RegisterClass multype,
 }
 
 multiclass MulAccum<bit isSub, string asm, SDNode AccNode> {
+  // MADD/MSUB generation is decided by MachineCombiner.cpp
   def Wrrr : BaseMulAccum<isSub, 0b000, GPR32, GPR32, asm,
-      [(set GPR32:$Rd, (AccNode GPR32:$Ra, (mul GPR32:$Rn, GPR32:$Rm)))]>,
+      [/*(set GPR32:$Rd, (AccNode GPR32:$Ra, (mul GPR32:$Rn, GPR32:$Rm)))*/]>,
       Sched<[WriteIM32, ReadIM, ReadIM, ReadIMA]> {
     let Inst{31} = 0;
   }
 
   def Xrrr : BaseMulAccum<isSub, 0b000, GPR64, GPR64, asm,
-      [(set GPR64:$Rd, (AccNode GPR64:$Ra, (mul GPR64:$Rn, GPR64:$Rm)))]>,
+      [/*(set GPR64:$Rd, (AccNode GPR64:$Ra, (mul GPR64:$Rn, GPR64:$Rm)))*/]>,
       Sched<[WriteIM64, ReadIM, ReadIM, ReadIMA]> {
     let Inst{31} = 1;
   }
@@ -1394,13 +1493,13 @@ class MulHi<bits<3> opc, string asm, SDNode OpNode>
 }
 
 class MulAccumWAlias<string asm, Instruction inst>
-    : InstAlias<asm#" $dst, $src1, $src2",
+    : InstAlias<asm#"\t$dst, $src1, $src2",
                 (inst GPR32:$dst, GPR32:$src1, GPR32:$src2, WZR)>;
 class MulAccumXAlias<string asm, Instruction inst>
-    : InstAlias<asm#" $dst, $src1, $src2",
+    : InstAlias<asm#"\t$dst, $src1, $src2",
                 (inst GPR64:$dst, GPR64:$src1, GPR64:$src2, XZR)>;
 class WideMulAccumAlias<string asm, Instruction inst>
-    : InstAlias<asm#" $dst, $src1, $src2",
+    : InstAlias<asm#"\t$dst, $src1, $src2",
                 (inst GPR64:$dst, GPR32:$src1, GPR32:$src2, XZR)>;
 
 class BaseCRC32<bit sf, bits<2> sz, bit C, RegisterClass StreamReg,
@@ -1630,18 +1729,24 @@ class BaseAddSubEReg64<bit isSub, bit setFlags, RegisterClass dstRegtype,
 class AddSubRegAlias<string asm, Instruction inst, RegisterClass dstRegtype,
                      RegisterClass src1Regtype, RegisterClass src2Regtype,
                      int shiftExt>
-    : InstAlias<asm#" $dst, $src1, $src2",
+    : InstAlias<asm#"\t$dst, $src1, $src2",
                 (inst dstRegtype:$dst, src1Regtype:$src1, src2Regtype:$src2,
                       shiftExt)>;
 
-multiclass AddSub<bit isSub, string mnemonic,
+multiclass AddSub<bit isSub, string mnemonic, string alias,
                   SDPatternOperator OpNode = null_frag> {
   let hasSideEffects = 0, isReMaterializable = 1, isAsCheapAsAMove = 1 in {
   // Add/Subtract immediate
+  // Increase the weight of the immediate variant to try to match it before
+  // the extended register variant.
+  // We used to match the register variant before the immediate when the
+  // register argument could be implicitly zero-extended.
+  let AddedComplexity = 6 in
   def Wri  : BaseAddSubImm<isSub, 0, GPR32sp, GPR32sp, addsub_shifted_imm32,
                            mnemonic, OpNode> {
     let Inst{31} = 0;
   }
+  let AddedComplexity = 6 in
   def Xri  : BaseAddSubImm<isSub, 0, GPR64sp, GPR64sp, addsub_shifted_imm64,
                            mnemonic, OpNode> {
     let Inst{31} = 1;
@@ -1681,6 +1786,14 @@ multiclass AddSub<bit isSub, string mnemonic,
     let Inst{31} = 1;
   }
 
+  // add Rd, Rb, -imm -> sub Rd, Rn, imm
+  def : InstAlias<alias#"\t$Rd, $Rn, $imm",
+                  (!cast<Instruction>(NAME # "Wri") GPR32sp:$Rd, GPR32sp:$Rn,
+                      addsub_shifted_imm32_neg:$imm), 0>;
+  def : InstAlias<alias#"\t$Rd, $Rn, $imm",
+                  (!cast<Instruction>(NAME # "Xri") GPR64sp:$Rd, GPR64sp:$Rn,
+                       addsub_shifted_imm64_neg:$imm), 0>;
+
   // Register/register aliases with no shift when SP is not used.
   def : AddSubRegAlias<mnemonic, !cast<Instruction>(NAME#"Wrs"),
                        GPR32, GPR32, GPR32, 0>;
@@ -1701,7 +1814,8 @@ multiclass AddSub<bit isSub, string mnemonic,
                        GPR64sp, GPR64sponly, GPR64, 24>; // UXTX #0
 }
 
-multiclass AddSubS<bit isSub, string mnemonic, SDNode OpNode, string cmp> {
+multiclass AddSubS<bit isSub, string mnemonic, SDNode OpNode, string cmp,
+                   string alias, string cmpAlias> {
   let isCompare = 1, Defs = [NZCV] in {
   // Add/Subtract immediate
   def Wri  : BaseAddSubImm<isSub, 1, GPR32, GPR32sp, addsub_shifted_imm32,
@@ -1747,30 +1861,44 @@ multiclass AddSubS<bit isSub, string mnemonic, SDNode OpNode, string cmp> {
   }
   } // Defs = [NZCV]
 
+  // Support negative immediates, e.g. adds Rd, Rn, -imm -> subs Rd, Rn, imm
+  def : InstAlias<alias#"\t$Rd, $Rn, $imm",
+                  (!cast<Instruction>(NAME # "Wri") GPR32:$Rd, GPR32sp:$Rn,
+                      addsub_shifted_imm32_neg:$imm), 0>;
+  def : InstAlias<alias#"\t$Rd, $Rn, $imm",
+                  (!cast<Instruction>(NAME # "Xri") GPR64:$Rd, GPR64sp:$Rn,
+                       addsub_shifted_imm64_neg:$imm), 0>;
+
   // Compare aliases
-  def : InstAlias<cmp#" $src, $imm", (!cast<Instruction>(NAME#"Wri")
+  def : InstAlias<cmp#"\t$src, $imm", (!cast<Instruction>(NAME#"Wri")
                   WZR, GPR32sp:$src, addsub_shifted_imm32:$imm), 5>;
-  def : InstAlias<cmp#" $src, $imm", (!cast<Instruction>(NAME#"Xri")
+  def : InstAlias<cmp#"\t$src, $imm", (!cast<Instruction>(NAME#"Xri")
                   XZR, GPR64sp:$src, addsub_shifted_imm64:$imm), 5>;
-  def : InstAlias<cmp#" $src1, $src2$sh", (!cast<Instruction>(NAME#"Wrx")
+  def : InstAlias<cmp#"\t$src1, $src2$sh", (!cast<Instruction>(NAME#"Wrx")
                   WZR, GPR32sp:$src1, GPR32:$src2, arith_extend:$sh), 4>;
-  def : InstAlias<cmp#" $src1, $src2$sh", (!cast<Instruction>(NAME#"Xrx")
+  def : InstAlias<cmp#"\t$src1, $src2$sh", (!cast<Instruction>(NAME#"Xrx")
                   XZR, GPR64sp:$src1, GPR32:$src2, arith_extend:$sh), 4>;
-  def : InstAlias<cmp#" $src1, $src2$sh", (!cast<Instruction>(NAME#"Xrx64")
+  def : InstAlias<cmp#"\t$src1, $src2$sh", (!cast<Instruction>(NAME#"Xrx64")
                   XZR, GPR64sp:$src1, GPR64:$src2, arith_extendlsl64:$sh), 4>;
-  def : InstAlias<cmp#" $src1, $src2$sh", (!cast<Instruction>(NAME#"Wrs")
+  def : InstAlias<cmp#"\t$src1, $src2$sh", (!cast<Instruction>(NAME#"Wrs")
                   WZR, GPR32:$src1, GPR32:$src2, arith_shift32:$sh), 4>;
-  def : InstAlias<cmp#" $src1, $src2$sh", (!cast<Instruction>(NAME#"Xrs")
+  def : InstAlias<cmp#"\t$src1, $src2$sh", (!cast<Instruction>(NAME#"Xrs")
                   XZR, GPR64:$src1, GPR64:$src2, arith_shift64:$sh), 4>;
 
+  // Support negative immediates, e.g. cmp Rn, -imm -> cmn Rn, imm
+  def : InstAlias<cmpAlias#"\t$src, $imm", (!cast<Instruction>(NAME#"Wri")
+                  WZR, GPR32sp:$src, addsub_shifted_imm32_neg:$imm), 0>;
+  def : InstAlias<cmpAlias#"\t$src, $imm", (!cast<Instruction>(NAME#"Xri")
+                  XZR, GPR64sp:$src, addsub_shifted_imm64_neg:$imm), 0>;
+
   // Compare shorthands
-  def : InstAlias<cmp#" $src1, $src2", (!cast<Instruction>(NAME#"Wrs")
+  def : InstAlias<cmp#"\t$src1, $src2", (!cast<Instruction>(NAME#"Wrs")
                   WZR, GPR32:$src1, GPR32:$src2, 0), 5>;
-  def : InstAlias<cmp#" $src1, $src2", (!cast<Instruction>(NAME#"Xrs")
+  def : InstAlias<cmp#"\t$src1, $src2", (!cast<Instruction>(NAME#"Xrs")
                   XZR, GPR64:$src1, GPR64:$src2, 0), 5>;
-  def : InstAlias<cmp#" $src1, $src2", (!cast<Instruction>(NAME#"Wrx")
+  def : InstAlias<cmp#"\t$src1, $src2", (!cast<Instruction>(NAME#"Wrx")
                   WZR, GPR32sponly:$src1, GPR32:$src2, 16), 5>;
-  def : InstAlias<cmp#" $src1, $src2", (!cast<Instruction>(NAME#"Xrx64")
+  def : InstAlias<cmp#"\t$src1, $src2", (!cast<Instruction>(NAME#"Xrx64")
                   XZR, GPR64sponly:$src1, GPR64:$src2, 24), 5>;
 
   // Register/register aliases with no shift when SP is not used.
@@ -1956,7 +2084,7 @@ class BaseLogicalSReg<bits<2> opc, bit N, RegisterClass regtype,
 
 // Aliases for register+register logical instructions.
 class LogicalRegAlias<string asm, Instruction inst, RegisterClass regtype>
-    : InstAlias<asm#" $dst, $src1, $src2",
+    : InstAlias<asm#"\t$dst, $src1, $src2",
                 (inst regtype:$dst, regtype:$src1, regtype:$src2, 0)>;
 
 multiclass LogicalImm<bits<2> opc, string mnemonic, SDNode OpNode,
@@ -1975,10 +2103,10 @@ multiclass LogicalImm<bits<2> opc, string mnemonic, SDNode OpNode,
     let Inst{31} = 1;
   }
 
-  def : InstAlias<Alias # " $Rd, $Rn, $imm",
+  def : InstAlias<Alias # "\t$Rd, $Rn, $imm",
                   (!cast<Instruction>(NAME # "Wri") GPR32sp:$Rd, GPR32:$Rn,
                       logical_imm32_not:$imm), 0>;
-  def : InstAlias<Alias # " $Rd, $Rn, $imm",
+  def : InstAlias<Alias # "\t$Rd, $Rn, $imm",
                   (!cast<Instruction>(NAME # "Xri") GPR64sp:$Rd, GPR64:$Rn,
                        logical_imm64_not:$imm), 0>;
 }
@@ -1997,10 +2125,10 @@ multiclass LogicalImmS<bits<2> opc, string mnemonic, SDNode OpNode,
   }
   } // end Defs = [NZCV]
 
-  def : InstAlias<Alias # " $Rd, $Rn, $imm",
+  def : InstAlias<Alias # "\t$Rd, $Rn, $imm",
                   (!cast<Instruction>(NAME # "Wri") GPR32:$Rd, GPR32:$Rn,
                       logical_imm32_not:$imm), 0>;
-  def : InstAlias<Alias # " $Rd, $Rn, $imm",
+  def : InstAlias<Alias # "\t$Rd, $Rn, $imm",
                   (!cast<Instruction>(NAME # "Xri") GPR64:$Rd, GPR64:$Rn,
                        logical_imm64_not:$imm), 0>;
 }
@@ -2063,9 +2191,12 @@ multiclass LogicalRegS<bits<2> opc, bit N, string mnemonic,
 //---
 
 let mayLoad = 0, mayStore = 0, hasSideEffects = 0 in
-class BaseCondSetFlagsImm<bit op, RegisterClass regtype, string asm>
-    : I<(outs), (ins regtype:$Rn, imm0_31:$imm, imm0_15:$nzcv, ccode:$cond),
-         asm, "\t$Rn, $imm, $nzcv, $cond", "", []>,
+class BaseCondComparisonImm<bit op, RegisterClass regtype, ImmLeaf immtype,
+                            string mnemonic, SDNode OpNode>
+    : I<(outs), (ins regtype:$Rn, immtype:$imm, imm32_0_15:$nzcv, ccode:$cond),
+         mnemonic, "\t$Rn, $imm, $nzcv, $cond", "",
+         [(set NZCV, (OpNode regtype:$Rn, immtype:$imm, (i32 imm:$nzcv),
+                             (i32 imm:$cond), NZCV))]>,
       Sched<[WriteI, ReadI]> {
   let Uses = [NZCV];
   let Defs = [NZCV];
@@ -2085,19 +2216,13 @@ class BaseCondSetFlagsImm<bit op, RegisterClass regtype, string asm>
   let Inst{3-0}   = nzcv;
 }
 
-multiclass CondSetFlagsImm<bit op, string asm> {
-  def Wi : BaseCondSetFlagsImm<op, GPR32, asm> {
-    let Inst{31} = 0;
-  }
-  def Xi : BaseCondSetFlagsImm<op, GPR64, asm> {
-    let Inst{31} = 1;
-  }
-}
-
 let mayLoad = 0, mayStore = 0, hasSideEffects = 0 in
-class BaseCondSetFlagsReg<bit op, RegisterClass regtype, string asm>
-    : I<(outs), (ins regtype:$Rn, regtype:$Rm, imm0_15:$nzcv, ccode:$cond),
-         asm, "\t$Rn, $Rm, $nzcv, $cond", "", []>,
+class BaseCondComparisonReg<bit op, RegisterClass regtype, string mnemonic,
+                            SDNode OpNode>
+    : I<(outs), (ins regtype:$Rn, regtype:$Rm, imm32_0_15:$nzcv, ccode:$cond),
+         mnemonic, "\t$Rn, $Rm, $nzcv, $cond", "",
+         [(set NZCV, (OpNode regtype:$Rn, regtype:$Rm, (i32 imm:$nzcv),
+                             (i32 imm:$cond), NZCV))]>,
       Sched<[WriteI, ReadI, ReadI]> {
   let Uses = [NZCV];
   let Defs = [NZCV];
@@ -2117,11 +2242,19 @@ class BaseCondSetFlagsReg<bit op, RegisterClass regtype, string asm>
   let Inst{3-0}   = nzcv;
 }
 
-multiclass CondSetFlagsReg<bit op, string asm> {
-  def Wr : BaseCondSetFlagsReg<op, GPR32, asm> {
+multiclass CondComparison<bit op, string mnemonic, SDNode OpNode> {
+  // immediate operand variants
+  def Wi : BaseCondComparisonImm<op, GPR32, imm32_0_31, mnemonic, OpNode> {
     let Inst{31} = 0;
   }
-  def Xr : BaseCondSetFlagsReg<op, GPR64, asm> {
+  def Xi : BaseCondComparisonImm<op, GPR64, imm0_31, mnemonic, OpNode> {
+    let Inst{31} = 1;
+  }
+  // register operand variants
+  def Wr : BaseCondComparisonReg<op, GPR32, mnemonic, OpNode> {
+    let Inst{31} = 0;
+  }
+  def Xr : BaseCondComparisonReg<op, GPR64, mnemonic, OpNode> {
     let Inst{31} = 1;
   }
 }
@@ -2187,7 +2320,8 @@ class BaseCondSelectOp<bit op, bits<2> op2, RegisterClass regtype, string asm,
 
 def inv_cond_XFORM : SDNodeXForm<imm, [{
   AArch64CC::CondCode CC = static_cast<AArch64CC::CondCode>(N->getZExtValue());
-  return CurDAG->getTargetConstant(AArch64CC::getInvertedCondCode(CC), MVT::i32);
+  return CurDAG->getTargetConstant(AArch64CC::getInvertedCondCode(CC), SDLoc(N),
+                                   MVT::i32);
 }]>;
 
 multiclass CondSelectOp<bit op, bits<2> op2, string asm, PatFrag frag> {
@@ -2285,7 +2419,7 @@ multiclass LoadUI<bits<2> sz, bit V, bits<2> opc, RegisterClass regtype,
                            asm, pattern>,
            Sched<[WriteLD]>;
 
-  def : InstAlias<asm # " $Rt, [$Rn]",
+  def : InstAlias<asm # "\t$Rt, [$Rn]",
                   (!cast<Instruction>(NAME # "ui") regtype:$Rt, GPR64sp:$Rn, 0)>;
 }
 
@@ -2297,7 +2431,7 @@ multiclass StoreUI<bits<2> sz, bit V, bits<2> opc, RegisterClass regtype,
                            asm, pattern>,
            Sched<[WriteST]>;
 
-  def : InstAlias<asm # " $Rt, [$Rn]",
+  def : InstAlias<asm # "\t$Rt, [$Rn]",
                   (!cast<Instruction>(NAME # "ui") regtype:$Rt, GPR64sp:$Rn, 0)>;
 }
 
@@ -2465,7 +2599,7 @@ class LoadStore8RO<bits<2> sz, bit V, bits<2> opc, RegisterClass regtype,
 }
 
 class ROInstAlias<string asm, RegisterClass regtype, Instruction INST>
-  : InstAlias<asm # " $Rt, [$Rn, $Rm]",
+  : InstAlias<asm # "\t$Rt, [$Rn, $Rm]",
               (INST regtype:$Rt, GPR64sp:$Rn, GPR64:$Rm, 0, 0)>;
 
 multiclass Load8RO<bits<2> sz, bit V, bits<2> opc, RegisterClass regtype,
@@ -2891,7 +3025,7 @@ multiclass LoadUnscaled<bits<2> sz, bit V, bits<2> opc, RegisterClass regtype,
                                (ins GPR64sp:$Rn, simm9:$offset), asm, pattern>,
           Sched<[WriteLD]>;
 
-  def : InstAlias<asm # " $Rt, [$Rn]",
+  def : InstAlias<asm # "\t$Rt, [$Rn]",
                   (!cast<Instruction>(NAME # "i") regtype:$Rt, GPR64sp:$Rn, 0)>;
 }
 
@@ -2903,7 +3037,7 @@ multiclass StoreUnscaled<bits<2> sz, bit V, bits<2> opc, RegisterClass regtype,
                                asm, pattern>,
           Sched<[WriteST]>;
 
-  def : InstAlias<asm # " $Rt, [$Rn]",
+  def : InstAlias<asm # "\t$Rt, [$Rn]",
                   (!cast<Instruction>(NAME # "i") regtype:$Rt, GPR64sp:$Rn, 0)>;
 }
 
@@ -2915,7 +3049,7 @@ multiclass PrefetchUnscaled<bits<2> sz, bit V, bits<2> opc, string asm,
                                asm, pat>,
           Sched<[WriteLD]>;
 
-  def : InstAlias<asm # " $Rt, [$Rn]",
+  def : InstAlias<asm # "\t$Rt, [$Rn]",
                   (!cast<Instruction>(NAME # "i") prfop:$Rt, GPR64sp:$Rn, 0)>;
 }
 
@@ -2950,7 +3084,7 @@ multiclass LoadUnprivileged<bits<2> sz, bit V, bits<2> opc,
                                     (ins GPR64sp:$Rn, simm9:$offset), asm>,
           Sched<[WriteLD]>;
 
-  def : InstAlias<asm # " $Rt, [$Rn]",
+  def : InstAlias<asm # "\t$Rt, [$Rn]",
                   (!cast<Instruction>(NAME # "i") regtype:$Rt, GPR64sp:$Rn, 0)>;
 }
 
@@ -2962,7 +3096,7 @@ multiclass StoreUnprivileged<bits<2> sz, bit V, bits<2> opc,
                                  asm>,
           Sched<[WriteST]>;
 
-  def : InstAlias<asm # " $Rt, [$Rn]",
+  def : InstAlias<asm # "\t$Rt, [$Rn]",
                   (!cast<Instruction>(NAME # "i") regtype:$Rt, GPR64sp:$Rn, 0)>;
 }
 
@@ -2997,7 +3131,7 @@ class LoadPreIdx<bits<2> sz, bit V, bits<2> opc, RegisterClass regtype,
     : BaseLoadStorePreIdx<sz, V, opc,
                      (outs GPR64sp:$wback, regtype:$Rt),
                      (ins GPR64sp:$Rn, simm9:$offset), asm,
-                     "$Rn = $wback", []>,
+                     "$Rn = $wback,@earlyclobber $wback", []>,
       Sched<[WriteLD, WriteAdr]>;
 
 let mayStore = 1, mayLoad = 0 in
@@ -3006,7 +3140,7 @@ class StorePreIdx<bits<2> sz, bit V, bits<2> opc, RegisterClass regtype,
     : BaseLoadStorePreIdx<sz, V, opc,
                       (outs GPR64sp:$wback),
                       (ins regtype:$Rt, GPR64sp:$Rn, simm9:$offset),
-                      asm, "$Rn = $wback",
+                      asm, "$Rn = $wback,@earlyclobber $wback",
       [(set GPR64sp:$wback,
             (storeop (Ty regtype:$Rt), GPR64sp:$Rn, simm9:$offset))]>,
       Sched<[WriteAdr, WriteST]>;
@@ -3016,7 +3150,6 @@ class StorePreIdx<bits<2> sz, bit V, bits<2> opc, RegisterClass regtype,
 // Load/store post-indexed
 //---
 
-// (pre-index) load/stores.
 class BaseLoadStorePostIdx<bits<2> sz, bit V, bits<2> opc, dag oops, dag iops,
                           string asm, string cstr, list<dag> pat>
     : I<oops, iops, asm, "\t$Rt, [$Rn], $offset", cstr, pat> {
@@ -3044,7 +3177,7 @@ class LoadPostIdx<bits<2> sz, bit V, bits<2> opc, RegisterClass regtype,
     : BaseLoadStorePostIdx<sz, V, opc,
                       (outs GPR64sp:$wback, regtype:$Rt),
                       (ins GPR64sp:$Rn, simm9:$offset),
-                      asm, "$Rn = $wback", []>,
+                      asm, "$Rn = $wback,@earlyclobber $wback", []>,
       Sched<[WriteLD, WriteI]>;
 
 let mayStore = 1, mayLoad = 0 in
@@ -3053,7 +3186,7 @@ class StorePostIdx<bits<2> sz, bit V, bits<2> opc, RegisterClass regtype,
     : BaseLoadStorePostIdx<sz, V, opc,
                       (outs GPR64sp:$wback),
                       (ins regtype:$Rt, GPR64sp:$Rn, simm9:$offset),
-                       asm, "$Rn = $wback",
+                       asm, "$Rn = $wback,@earlyclobber $wback",
       [(set GPR64sp:$wback,
             (storeop (Ty regtype:$Rt), GPR64sp:$Rn, simm9:$offset))]>,
     Sched<[WriteAdr, WriteST, ReadAdrBase]>;
@@ -3094,7 +3227,7 @@ multiclass LoadPairOffset<bits<2> opc, bit V, RegisterClass regtype,
                                   (ins GPR64sp:$Rn, indextype:$offset), asm>,
           Sched<[WriteLD, WriteLDHi]>;
 
-  def : InstAlias<asm # " $Rt, $Rt2, [$Rn]",
+  def : InstAlias<asm # "\t$Rt, $Rt2, [$Rn]",
                   (!cast<Instruction>(NAME # "i") regtype:$Rt, regtype:$Rt2,
                                                   GPR64sp:$Rn, 0)>;
 }
@@ -3109,7 +3242,7 @@ multiclass StorePairOffset<bits<2> opc, bit V, RegisterClass regtype,
                                   asm>,
           Sched<[WriteSTP]>;
 
-  def : InstAlias<asm # " $Rt, $Rt2, [$Rn]",
+  def : InstAlias<asm # "\t$Rt, $Rt2, [$Rn]",
                   (!cast<Instruction>(NAME # "i") regtype:$Rt, regtype:$Rt2,
                                                   GPR64sp:$Rn, 0)>;
 }
@@ -3117,7 +3250,7 @@ multiclass StorePairOffset<bits<2> opc, bit V, RegisterClass regtype,
 // (pre-indexed)
 class BaseLoadStorePairPreIdx<bits<2> opc, bit V, bit L, dag oops, dag iops,
                               string asm>
-    : I<oops, iops, asm, "\t$Rt, $Rt2, [$Rn, $offset]!", "$Rn = $wback", []> {
+    : I<oops, iops, asm, "\t$Rt, $Rt2, [$Rn, $offset]!", "$Rn = $wback,@earlyclobber $wback", []> {
   bits<5> Rt;
   bits<5> Rt2;
   bits<5> Rn;
@@ -3158,7 +3291,7 @@ class StorePairPreIdx<bits<2> opc, bit V, RegisterClass regtype,
 
 class BaseLoadStorePairPostIdx<bits<2> opc, bit V, bit L, dag oops, dag iops,
                               string asm>
-    : I<oops, iops, asm, "\t$Rt, $Rt2, [$Rn], $offset", "$Rn = $wback", []> {
+    : I<oops, iops, asm, "\t$Rt, $Rt2, [$Rn], $offset", "$Rn = $wback,@earlyclobber $wback", []> {
   bits<5> Rt;
   bits<5> Rt2;
   bits<5> Rn;
@@ -3188,8 +3321,8 @@ class LoadPairPostIdx<bits<2> opc, bit V, RegisterClass regtype,
 let mayStore = 1, mayLoad = 0 in
 class StorePairPostIdx<bits<2> opc, bit V, RegisterClass regtype,
                        Operand idxtype, string asm>
-    : BaseLoadStorePairPostIdx<opc, V, 0, (outs),
-                             (ins GPR64sp:$wback, regtype:$Rt, regtype:$Rt2,
+    : BaseLoadStorePairPostIdx<opc, V, 0, (outs GPR64sp:$wback),
+                             (ins regtype:$Rt, regtype:$Rt2,
                                   GPR64sp:$Rn, idxtype:$offset),
                              asm>,
       Sched<[WriteAdr, WriteSTP]>;
@@ -3284,6 +3417,10 @@ class LoadStoreExclusiveSimple<bits<2> sz, bit o2, bit L, bit o1, bit o0,
     : BaseLoadStoreExclusive<sz, o2, L, o1, o0, oops, iops, asm, operands> {
   bits<5> Rt;
   bits<5> Rn;
+  let Inst{20-16} = 0b11111;
+  let Unpredictable{20-16} = 0b11111;
+  let Inst{14-10} = 0b11111;
+  let Unpredictable{14-10} = 0b11111;
   let Inst{9-5} = Rn;
   let Inst{4-0} = Rt;
 
@@ -3431,6 +3568,20 @@ class BaseFPToInteger<bits<2> type, bits<2> rmode, bits<3> opcode,
 
 multiclass FPToIntegerUnscaled<bits<2> rmode, bits<3> opcode, string asm,
            SDPatternOperator OpN> {
+  // Unscaled half-precision to 32-bit
+  def UWHr : BaseFPToIntegerUnscaled<0b11, rmode, opcode, FPR16, GPR32, asm,
+                                     [(set GPR32:$Rd, (OpN FPR16:$Rn))]> {
+    let Inst{31} = 0; // 32-bit GPR flag
+    let Predicates = [HasFullFP16];
+  }
+
+  // Unscaled half-precision to 64-bit
+  def UXHr : BaseFPToIntegerUnscaled<0b11, rmode, opcode, FPR16, GPR64, asm,
+                                     [(set GPR64:$Rd, (OpN FPR16:$Rn))]> {
+    let Inst{31} = 1; // 64-bit GPR flag
+    let Predicates = [HasFullFP16];
+  }
+
   // Unscaled single-precision to 32-bit
   def UWSr : BaseFPToIntegerUnscaled<0b00, rmode, opcode, FPR32, GPR32, asm,
                                      [(set GPR32:$Rd, (OpN FPR32:$Rn))]> {
@@ -3458,6 +3609,25 @@ multiclass FPToIntegerUnscaled<bits<2> rmode, bits<3> opcode, string asm,
 
 multiclass FPToIntegerScaled<bits<2> rmode, bits<3> opcode, string asm,
                              SDPatternOperator OpN> {
+  // Scaled half-precision to 32-bit
+  def SWHri : BaseFPToInteger<0b11, rmode, opcode, FPR16, GPR32,
+                              fixedpoint_f16_i32, asm,
+              [(set GPR32:$Rd, (OpN (fmul FPR16:$Rn,
+                                          fixedpoint_f16_i32:$scale)))]> {
+    let Inst{31} = 0; // 32-bit GPR flag
+    let scale{5} = 1;
+    let Predicates = [HasFullFP16];
+  }
+
+  // Scaled half-precision to 64-bit
+  def SXHri : BaseFPToInteger<0b11, rmode, opcode, FPR16, GPR64,
+                              fixedpoint_f16_i64, asm,
+              [(set GPR64:$Rd, (OpN (fmul FPR16:$Rn,
+                                          fixedpoint_f16_i64:$scale)))]> {
+    let Inst{31} = 1; // 64-bit GPR flag
+    let Predicates = [HasFullFP16];
+  }
+
   // Scaled single-precision to 32-bit
   def SWSri : BaseFPToInteger<0b00, rmode, opcode, FPR32, GPR32,
                               fixedpoint_f32_i32, asm,
@@ -3507,7 +3677,7 @@ class BaseIntegerToFP<bit isUnsigned,
   bits<5> Rd;
   bits<5> Rn;
   bits<6> scale;
-  let Inst{30-23} = 0b00111100;
+  let Inst{30-24} = 0b0011110;
   let Inst{21-17} = 0b00001;
   let Inst{16}    = isUnsigned;
   let Inst{15-10} = scale;
@@ -3524,7 +3694,7 @@ class BaseIntegerToFPUnscaled<bit isUnsigned,
   bits<5> Rd;
   bits<5> Rn;
   bits<6> scale;
-  let Inst{30-23} = 0b00111100;
+  let Inst{30-24} = 0b0011110;
   let Inst{21-17} = 0b10001;
   let Inst{16}    = isUnsigned;
   let Inst{15-10} = 0b000000;
@@ -3534,33 +3704,55 @@ class BaseIntegerToFPUnscaled<bit isUnsigned,
 
 multiclass IntegerToFP<bit isUnsigned, string asm, SDNode node> {
   // Unscaled
+  def UWHri: BaseIntegerToFPUnscaled<isUnsigned, GPR32, FPR16, f16, asm, node> {
+    let Inst{31} = 0; // 32-bit GPR flag
+    let Inst{23-22} = 0b11; // 16-bit FPR flag
+    let Predicates = [HasFullFP16];
+  }
+
   def UWSri: BaseIntegerToFPUnscaled<isUnsigned, GPR32, FPR32, f32, asm, node> {
     let Inst{31} = 0; // 32-bit GPR flag
-    let Inst{22} = 0; // 32-bit FPR flag
+    let Inst{23-22} = 0b00; // 32-bit FPR flag
   }
 
   def UWDri: BaseIntegerToFPUnscaled<isUnsigned, GPR32, FPR64, f64, asm, node> {
     let Inst{31} = 0; // 32-bit GPR flag
-    let Inst{22} = 1; // 64-bit FPR flag
+    let Inst{23-22} = 0b01; // 64-bit FPR flag
+  }
+
+  def UXHri: BaseIntegerToFPUnscaled<isUnsigned, GPR64, FPR16, f16, asm, node> {
+    let Inst{31} = 1; // 64-bit GPR flag
+    let Inst{23-22} = 0b11; // 16-bit FPR flag
+    let Predicates = [HasFullFP16];
   }
 
   def UXSri: BaseIntegerToFPUnscaled<isUnsigned, GPR64, FPR32, f32, asm, node> {
     let Inst{31} = 1; // 64-bit GPR flag
-    let Inst{22} = 0; // 32-bit FPR flag
+    let Inst{23-22} = 0b00; // 32-bit FPR flag
   }
 
   def UXDri: BaseIntegerToFPUnscaled<isUnsigned, GPR64, FPR64, f64, asm, node> {
     let Inst{31} = 1; // 64-bit GPR flag
-    let Inst{22} = 1; // 64-bit FPR flag
+    let Inst{23-22} = 0b01; // 64-bit FPR flag
   }
 
   // Scaled
+  def SWHri: BaseIntegerToFP<isUnsigned, GPR32, FPR16, fixedpoint_f16_i32, asm,
+                             [(set FPR16:$Rd,
+                                   (fdiv (node GPR32:$Rn),
+                                         fixedpoint_f16_i32:$scale))]> {
+    let Inst{31} = 0; // 32-bit GPR flag
+    let Inst{23-22} = 0b11; // 16-bit FPR flag
+    let scale{5} = 1;
+    let Predicates = [HasFullFP16];
+  }
+
   def SWSri: BaseIntegerToFP<isUnsigned, GPR32, FPR32, fixedpoint_f32_i32, asm,
                              [(set FPR32:$Rd,
                                    (fdiv (node GPR32:$Rn),
                                          fixedpoint_f32_i32:$scale))]> {
     let Inst{31} = 0; // 32-bit GPR flag
-    let Inst{22} = 0; // 32-bit FPR flag
+    let Inst{23-22} = 0b00; // 32-bit FPR flag
     let scale{5} = 1;
   }
 
@@ -3569,16 +3761,25 @@ multiclass IntegerToFP<bit isUnsigned, string asm, SDNode node> {
                                    (fdiv (node GPR32:$Rn),
                                          fixedpoint_f64_i32:$scale))]> {
     let Inst{31} = 0; // 32-bit GPR flag
-    let Inst{22} = 1; // 64-bit FPR flag
+    let Inst{23-22} = 0b01; // 64-bit FPR flag
     let scale{5} = 1;
   }
 
+  def SXHri: BaseIntegerToFP<isUnsigned, GPR64, FPR16, fixedpoint_f16_i64, asm,
+                             [(set FPR16:$Rd,
+                                   (fdiv (node GPR64:$Rn),
+                                         fixedpoint_f16_i64:$scale))]> {
+    let Inst{31} = 1; // 64-bit GPR flag
+    let Inst{23-22} = 0b11; // 16-bit FPR flag
+    let Predicates = [HasFullFP16];
+  }
+
   def SXSri: BaseIntegerToFP<isUnsigned, GPR64, FPR32, fixedpoint_f32_i64, asm,
                              [(set FPR32:$Rd,
                                    (fdiv (node GPR64:$Rn),
                                          fixedpoint_f32_i64:$scale))]> {
     let Inst{31} = 1; // 64-bit GPR flag
-    let Inst{22} = 0; // 32-bit FPR flag
+    let Inst{23-22} = 0b00; // 32-bit FPR flag
   }
 
   def SXDri: BaseIntegerToFP<isUnsigned, GPR64, FPR64, fixedpoint_f64_i64, asm,
@@ -3586,7 +3787,7 @@ multiclass IntegerToFP<bit isUnsigned, string asm, SDNode node> {
                                    (fdiv (node GPR64:$Rn),
                                          fixedpoint_f64_i64:$scale))]> {
     let Inst{31} = 1; // 64-bit GPR flag
-    let Inst{22} = 1; // 64-bit FPR flag
+    let Inst{23-22} = 0b01; // 64-bit FPR flag
   }
 }
 
@@ -3608,7 +3809,7 @@ class BaseUnscaledConversion<bits<2> rmode, bits<3> opcode,
       Sched<[WriteFCopy]> {
   bits<5> Rd;
   bits<5> Rn;
-  let Inst{30-23} = 0b00111100;
+  let Inst{30-24} = 0b0011110;
   let Inst{21}    = 1;
   let Inst{20-19} = rmode;
   let Inst{18-16} = opcode;
@@ -3658,26 +3859,49 @@ class BaseUnscaledConversionFromHigh<bits<2> rmode, bits<3> opcode,
 }
 
 
-
 multiclass UnscaledConversion<string asm> {
+  def WHr : BaseUnscaledConversion<0b00, 0b111, GPR32, FPR16, asm> {
+    let Inst{31} = 0; // 32-bit GPR flag
+    let Inst{23-22} = 0b11; // 16-bit FPR flag
+    let Predicates = [HasFullFP16];
+  }
+
+  def XHr : BaseUnscaledConversion<0b00, 0b111, GPR64, FPR16, asm> {
+    let Inst{31} = 1; // 64-bit GPR flag
+    let Inst{23-22} = 0b11; // 16-bit FPR flag
+    let Predicates = [HasFullFP16];
+  }
+
   def WSr : BaseUnscaledConversion<0b00, 0b111, GPR32, FPR32, asm> {
     let Inst{31} = 0; // 32-bit GPR flag
-    let Inst{22} = 0; // 32-bit FPR flag
+    let Inst{23-22} = 0b00; // 32-bit FPR flag
   }
 
   def XDr : BaseUnscaledConversion<0b00, 0b111, GPR64, FPR64, asm> {
     let Inst{31} = 1; // 64-bit GPR flag
-    let Inst{22} = 1; // 64-bit FPR flag
+    let Inst{23-22} = 0b01; // 64-bit FPR flag
+  }
+
+  def HWr : BaseUnscaledConversion<0b00, 0b110, FPR16, GPR32, asm> {
+    let Inst{31} = 0; // 32-bit GPR flag
+    let Inst{23-22} = 0b11; // 16-bit FPR flag
+    let Predicates = [HasFullFP16];
+  }
+
+  def HXr : BaseUnscaledConversion<0b00, 0b110, FPR16, GPR64, asm> {
+    let Inst{31} = 1; // 64-bit GPR flag
+    let Inst{23-22} = 0b11; // 16-bit FPR flag
+    let Predicates = [HasFullFP16];
   }
 
   def SWr : BaseUnscaledConversion<0b00, 0b110, FPR32, GPR32, asm> {
     let Inst{31} = 0; // 32-bit GPR flag
-    let Inst{22} = 0; // 32-bit FPR flag
+    let Inst{23-22} = 0b00; // 32-bit FPR flag
   }
 
   def DXr : BaseUnscaledConversion<0b00, 0b110, FPR64, GPR64, asm> {
     let Inst{31} = 1; // 64-bit GPR flag
-    let Inst{22} = 1; // 64-bit FPR flag
+    let Inst{23-22} = 0b01; // 64-bit FPR flag
   }
 
   def XDHighr : BaseUnscaledConversionToHigh<0b01, 0b111, GPR64, V128,
@@ -3750,7 +3974,7 @@ class BaseSingleOperandFPData<bits<4> opcode, RegisterClass regtype,
       Sched<[WriteF]> {
   bits<5> Rd;
   bits<5> Rn;
-  let Inst{31-23} = 0b000111100;
+  let Inst{31-24} = 0b00011110;
   let Inst{21-19} = 0b100;
   let Inst{18-15} = opcode;
   let Inst{14-10} = 0b10000;
@@ -3760,12 +3984,17 @@ class BaseSingleOperandFPData<bits<4> opcode, RegisterClass regtype,
 
 multiclass SingleOperandFPData<bits<4> opcode, string asm,
                                SDPatternOperator node = null_frag> {
+  def Hr : BaseSingleOperandFPData<opcode, FPR16, f16, asm, node> {
+    let Inst{23-22} = 0b11; // 16-bit size flag
+    let Predicates = [HasFullFP16];
+  }
+
   def Sr : BaseSingleOperandFPData<opcode, FPR32, f32, asm, node> {
-    let Inst{22} = 0; // 32-bit size flag
+    let Inst{23-22} = 0b00; // 32-bit size flag
   }
 
   def Dr : BaseSingleOperandFPData<opcode, FPR64, f64, asm, node> {
-    let Inst{22} = 1; // 64-bit size flag
+    let Inst{23-22} = 0b01; // 64-bit size flag
   }
 }
 
@@ -3782,7 +4011,7 @@ class BaseTwoOperandFPData<bits<4> opcode, RegisterClass regtype,
   bits<5> Rd;
   bits<5> Rn;
   bits<5> Rm;
-  let Inst{31-23} = 0b000111100;
+  let Inst{31-24} = 0b00011110;
   let Inst{21}    = 1;
   let Inst{20-16} = Rm;
   let Inst{15-12} = opcode;
@@ -3793,28 +4022,41 @@ class BaseTwoOperandFPData<bits<4> opcode, RegisterClass regtype,
 
 multiclass TwoOperandFPData<bits<4> opcode, string asm,
                             SDPatternOperator node = null_frag> {
+  def Hrr : BaseTwoOperandFPData<opcode, FPR16, asm,
+                         [(set (f16 FPR16:$Rd),
+                               (node (f16 FPR16:$Rn), (f16 FPR16:$Rm)))]> {
+    let Inst{23-22} = 0b11; // 16-bit size flag
+    let Predicates = [HasFullFP16];
+  }
+
   def Srr : BaseTwoOperandFPData<opcode, FPR32, asm,
                          [(set (f32 FPR32:$Rd),
                                (node (f32 FPR32:$Rn), (f32 FPR32:$Rm)))]> {
-    let Inst{22} = 0; // 32-bit size flag
+    let Inst{23-22} = 0b00; // 32-bit size flag
   }
 
   def Drr : BaseTwoOperandFPData<opcode, FPR64, asm,
                          [(set (f64 FPR64:$Rd),
                                (node (f64 FPR64:$Rn), (f64 FPR64:$Rm)))]> {
-    let Inst{22} = 1; // 64-bit size flag
+    let Inst{23-22} = 0b01; // 64-bit size flag
   }
 }
 
 multiclass TwoOperandFPDataNeg<bits<4> opcode, string asm, SDNode node> {
+  def Hrr : BaseTwoOperandFPData<opcode, FPR16, asm,
+                  [(set FPR16:$Rd, (fneg (node FPR16:$Rn, (f16 FPR16:$Rm))))]> {
+    let Inst{23-22} = 0b11; // 16-bit size flag
+    let Predicates = [HasFullFP16];
+  }
+
   def Srr : BaseTwoOperandFPData<opcode, FPR32, asm,
                   [(set FPR32:$Rd, (fneg (node FPR32:$Rn, (f32 FPR32:$Rm))))]> {
-    let Inst{22} = 0; // 32-bit size flag
+    let Inst{23-22} = 0b00; // 32-bit size flag
   }
 
   def Drr : BaseTwoOperandFPData<opcode, FPR64, asm,
                   [(set FPR64:$Rd, (fneg (node FPR64:$Rn, (f64 FPR64:$Rm))))]> {
-    let Inst{22} = 1; // 64-bit size flag
+    let Inst{23-22} = 0b01; // 64-bit size flag
   }
 }
 
@@ -3832,7 +4074,7 @@ class BaseThreeOperandFPData<bit isNegated, bit isSub,
   bits<5> Rn;
   bits<5> Rm;
   bits<5> Ra;
-  let Inst{31-23} = 0b000111110;
+  let Inst{31-24} = 0b00011111;
   let Inst{21}    = isNegated;
   let Inst{20-16} = Rm;
   let Inst{15}    = isSub;
@@ -3843,16 +4085,23 @@ class BaseThreeOperandFPData<bit isNegated, bit isSub,
 
 multiclass ThreeOperandFPData<bit isNegated, bit isSub,string asm,
                               SDPatternOperator node> {
+  def Hrrr : BaseThreeOperandFPData<isNegated, isSub, FPR16, asm,
+            [(set FPR16:$Rd,
+                  (node (f16 FPR16:$Rn), (f16 FPR16:$Rm), (f16 FPR16:$Ra)))]> {
+    let Inst{23-22} = 0b11; // 16-bit size flag
+    let Predicates = [HasFullFP16];
+  }
+
   def Srrr : BaseThreeOperandFPData<isNegated, isSub, FPR32, asm,
             [(set FPR32:$Rd,
                   (node (f32 FPR32:$Rn), (f32 FPR32:$Rm), (f32 FPR32:$Ra)))]> {
-    let Inst{22} = 0; // 32-bit size flag
+    let Inst{23-22} = 0b00; // 32-bit size flag
   }
 
   def Drrr : BaseThreeOperandFPData<isNegated, isSub, FPR64, asm,
             [(set FPR64:$Rd,
                   (node (f64 FPR64:$Rn), (f64 FPR64:$Rm), (f64 FPR64:$Ra)))]> {
-    let Inst{22} = 1; // 64-bit size flag
+    let Inst{23-22} = 0b01; // 64-bit size flag
   }
 }
 
@@ -3867,7 +4116,7 @@ class BaseOneOperandFPComparison<bit signalAllNans,
     : I<(outs), (ins regtype:$Rn), asm, "\t$Rn, #0.0", "", pat>,
       Sched<[WriteFCmp]> {
   bits<5> Rn;
-  let Inst{31-23} = 0b000111100;
+  let Inst{31-24} = 0b00011110;
   let Inst{21}    = 1;
 
   let Inst{15-10} = 0b001000;
@@ -3886,7 +4135,7 @@ class BaseTwoOperandFPComparison<bit signalAllNans, RegisterClass regtype,
       Sched<[WriteFCmp]> {
   bits<5> Rm;
   bits<5> Rn;
-  let Inst{31-23} = 0b000111100;
+  let Inst{31-24} = 0b00011110;
   let Inst{21}    = 1;
   let Inst{20-16} = Rm;
   let Inst{15-10} = 0b001000;
@@ -3898,24 +4147,36 @@ class BaseTwoOperandFPComparison<bit signalAllNans, RegisterClass regtype,
 multiclass FPComparison<bit signalAllNans, string asm,
                         SDPatternOperator OpNode = null_frag> {
   let Defs = [NZCV] in {
+  def Hrr : BaseTwoOperandFPComparison<signalAllNans, FPR16, asm,
+      [(OpNode FPR16:$Rn, (f16 FPR16:$Rm)), (implicit NZCV)]> {
+    let Inst{23-22} = 0b11;
+    let Predicates = [HasFullFP16];
+  }
+
+  def Hri : BaseOneOperandFPComparison<signalAllNans, FPR16, asm,
+      [(OpNode (f16 FPR16:$Rn), fpimm0), (implicit NZCV)]> {
+    let Inst{23-22} = 0b11;
+    let Predicates = [HasFullFP16];
+  }
+
   def Srr : BaseTwoOperandFPComparison<signalAllNans, FPR32, asm,
       [(OpNode FPR32:$Rn, (f32 FPR32:$Rm)), (implicit NZCV)]> {
-    let Inst{22} = 0;
+    let Inst{23-22} = 0b00;
   }
 
   def Sri : BaseOneOperandFPComparison<signalAllNans, FPR32, asm,
       [(OpNode (f32 FPR32:$Rn), fpimm0), (implicit NZCV)]> {
-    let Inst{22} = 0;
+    let Inst{23-22} = 0b00;
   }
 
   def Drr : BaseTwoOperandFPComparison<signalAllNans, FPR64, asm,
       [(OpNode FPR64:$Rn, (f64 FPR64:$Rm)), (implicit NZCV)]> {
-    let Inst{22} = 1;
+    let Inst{23-22} = 0b01;
   }
 
   def Dri : BaseOneOperandFPComparison<signalAllNans, FPR64, asm,
       [(OpNode (f64 FPR64:$Rn), fpimm0), (implicit NZCV)]> {
-    let Inst{22} = 1;
+    let Inst{23-22} = 0b01;
   }
   } // Defs = [NZCV]
 }
@@ -3925,17 +4186,20 @@ multiclass FPComparison<bit signalAllNans, string asm,
 //---
 
 let mayLoad = 0, mayStore = 0, hasSideEffects = 0 in
-class BaseFPCondComparison<bit signalAllNans,
-                              RegisterClass regtype, string asm>
-    : I<(outs), (ins regtype:$Rn, regtype:$Rm, imm0_15:$nzcv, ccode:$cond),
-         asm, "\t$Rn, $Rm, $nzcv, $cond", "", []>,
+class BaseFPCondComparison<bit signalAllNans, RegisterClass regtype,
+                           string mnemonic, list<dag> pat>
+    : I<(outs), (ins regtype:$Rn, regtype:$Rm, imm32_0_15:$nzcv, ccode:$cond),
+         mnemonic, "\t$Rn, $Rm, $nzcv, $cond", "", pat>,
       Sched<[WriteFCmp]> {
+  let Uses = [NZCV];
+  let Defs = [NZCV];
+
   bits<5> Rn;
   bits<5> Rm;
   bits<4> nzcv;
   bits<4> cond;
 
-  let Inst{31-23} = 0b000111100;
+  let Inst{31-24} = 0b00011110;
   let Inst{21}    = 1;
   let Inst{20-16} = Rm;
   let Inst{15-12} = cond;
@@ -3945,16 +4209,24 @@ class BaseFPCondComparison<bit signalAllNans,
   let Inst{3-0}   = nzcv;
 }
 
-multiclass FPCondComparison<bit signalAllNans, string asm> {
-  let Defs = [NZCV], Uses = [NZCV] in {
-  def Srr : BaseFPCondComparison<signalAllNans, FPR32, asm> {
-    let Inst{22} = 0;
+multiclass FPCondComparison<bit signalAllNans, string mnemonic,
+                            SDPatternOperator OpNode = null_frag> {
+  def Hrr : BaseFPCondComparison<signalAllNans, FPR16, mnemonic, []> {
+    let Inst{23-22} = 0b11;
+    let Predicates = [HasFullFP16];
   }
 
-  def Drr : BaseFPCondComparison<signalAllNans, FPR64, asm> {
-    let Inst{22} = 1;
+  def Srr : BaseFPCondComparison<signalAllNans, FPR32, mnemonic,
+      [(set NZCV, (OpNode (f32 FPR32:$Rn), (f32 FPR32:$Rm), (i32 imm:$nzcv),
+                          (i32 imm:$cond), NZCV))]> {
+    let Inst{23-22} = 0b00;
+  }
+
+  def Drr : BaseFPCondComparison<signalAllNans, FPR64, mnemonic,
+      [(set NZCV, (OpNode (f64 FPR64:$Rn), (f64 FPR64:$Rm), (i32 imm:$nzcv),
+                          (i32 imm:$cond), NZCV))]> {
+    let Inst{23-22} = 0b01;
   }
-  } // Defs = [NZCV], Uses = [NZCV]
 }
 
 //---
@@ -3973,7 +4245,7 @@ class BaseFPCondSelect<RegisterClass regtype, ValueType vt, string asm>
   bits<5> Rm;
   bits<4> cond;
 
-  let Inst{31-23} = 0b000111100;
+  let Inst{31-24} = 0b00011110;
   let Inst{21}    = 1;
   let Inst{20-16} = Rm;
   let Inst{15-12} = cond;
@@ -3984,12 +4256,17 @@ class BaseFPCondSelect<RegisterClass regtype, ValueType vt, string asm>
 
 multiclass FPCondSelect<string asm> {
   let Uses = [NZCV] in {
+  def Hrrr : BaseFPCondSelect<FPR16, f16, asm> {
+    let Inst{23-22} = 0b11;
+    let Predicates = [HasFullFP16];
+  }
+
   def Srrr : BaseFPCondSelect<FPR32, f32, asm> {
-    let Inst{22} = 0;
+    let Inst{23-22} = 0b00;
   }
 
   def Drrr : BaseFPCondSelect<FPR64, f64, asm> {
-    let Inst{22} = 1;
+    let Inst{23-22} = 0b01;
   }
   } // Uses = [NZCV]
 }
@@ -4004,7 +4281,7 @@ class BaseFPMoveImmediate<RegisterClass regtype, Operand fpimmtype, string asm>
     Sched<[WriteFImm]> {
   bits<5> Rd;
   bits<8> imm;
-  let Inst{31-23} = 0b000111100;
+  let Inst{31-24} = 0b00011110;
   let Inst{21}    = 1;
   let Inst{20-13} = imm;
   let Inst{12-5}  = 0b10000000;
@@ -4012,12 +4289,17 @@ class BaseFPMoveImmediate<RegisterClass regtype, Operand fpimmtype, string asm>
 }
 
 multiclass FPMoveImmediate<string asm> {
+  def Hi : BaseFPMoveImmediate<FPR16, fpimm16, asm> {
+    let Inst{23-22} = 0b11;
+    let Predicates = [HasFullFP16];
+  }
+
   def Si : BaseFPMoveImmediate<FPR32, fpimm32, asm> {
-    let Inst{22} = 0;
+    let Inst{23-22} = 0b00;
   }
 
   def Di : BaseFPMoveImmediate<FPR64, fpimm64, asm> {
-    let Inst{22} = 1;
+    let Inst{23-22} = 0b01;
   }
 }
 } // end of 'let Predicates = [HasFPARMv8]'
@@ -4385,7 +4667,7 @@ class BaseSIMDVectorLShiftLongBySize<bit Q, bits<2> size,
 }
 
 multiclass SIMDVectorLShiftLongBySizeBHS {
-  let neverHasSideEffects = 1 in {
+  let hasSideEffects = 0 in {
   def v8i8  : BaseSIMDVectorLShiftLongBySize<0, 0b00, V64,
                                              "shll", ".8h",  ".8b", "8">;
   def v16i8 : BaseSIMDVectorLShiftLongBySize<1, 0b00, V128,
@@ -4723,17 +5005,17 @@ multiclass SIMDFPCmpTwoVector<bit U, bit S, bits<5> opc,
                                      asm, ".2d", "0.0",
                                      v2i64, v2f64, OpNode>;
 
-  def : InstAlias<asm # " $Vd.2s, $Vn.2s, #0",
+  def : InstAlias<asm # "\t$Vd.2s, $Vn.2s, #0",
                   (!cast<Instruction>(NAME # v2i32rz) V64:$Vd, V64:$Vn), 0>;
-  def : InstAlias<asm # " $Vd.4s, $Vn.4s, #0",
+  def : InstAlias<asm # "\t$Vd.4s, $Vn.4s, #0",
                   (!cast<Instruction>(NAME # v4i32rz) V128:$Vd, V128:$Vn), 0>;
-  def : InstAlias<asm # " $Vd.2d, $Vn.2d, #0",
+  def : InstAlias<asm # "\t$Vd.2d, $Vn.2d, #0",
                   (!cast<Instruction>(NAME # v2i64rz) V128:$Vd, V128:$Vn), 0>;
-  def : InstAlias<asm # ".2s $Vd, $Vn, #0",
+  def : InstAlias<asm # ".2s\t$Vd, $Vn, #0",
                   (!cast<Instruction>(NAME # v2i32rz) V64:$Vd, V64:$Vn), 0>;
-  def : InstAlias<asm # ".4s $Vd, $Vn, #0",
+  def : InstAlias<asm # ".4s\t$Vd, $Vn, #0",
                   (!cast<Instruction>(NAME # v4i32rz) V128:$Vd, V128:$Vn), 0>;
-  def : InstAlias<asm # ".2d $Vd, $Vn, #0",
+  def : InstAlias<asm # ".2d\t$Vd, $Vn, #0",
                   (!cast<Instruction>(NAME # v2i64rz) V128:$Vd, V128:$Vn), 0>;
 }
 
@@ -5262,6 +5544,10 @@ multiclass SIMDZipVector<bits<3>opc, string asm,
   def v2i64  : BaseSIMDZipVector<0b111, opc, V128,
       asm, ".2d", OpNode, v2i64>;
 
+  def : Pat<(v4f16 (OpNode V64:$Rn, V64:$Rm)),
+        (!cast<Instruction>(NAME#"v4i16") V64:$Rn, V64:$Rm)>;
+  def : Pat<(v8f16 (OpNode V128:$Rn, V128:$Rm)),
+        (!cast<Instruction>(NAME#"v8i16") V128:$Rn, V128:$Rm)>;
   def : Pat<(v2f32 (OpNode V64:$Rn, V64:$Rm)),
         (!cast<Instruction>(NAME#"v2i32") V64:$Rn, V64:$Rm)>;
   def : Pat<(v4f32 (OpNode V128:$Rn, V128:$Rm)),
@@ -5296,6 +5582,27 @@ class BaseSIMDThreeScalar<bit U, bits<2> size, bits<5> opcode,
   let Inst{4-0}   = Rd;
 }
 
+let mayStore = 0, mayLoad = 0, hasSideEffects = 0 in
+class BaseSIMDThreeScalarTied<bit U, bits<2> size, bit R, bits<5> opcode,
+                        dag oops, dag iops, string asm,
+            list<dag> pattern>
+  : I<oops, iops, asm, "\t$Rd, $Rn, $Rm", "$Rd = $dst", pattern>,
+    Sched<[WriteV]> {
+  bits<5> Rd;
+  bits<5> Rn;
+  bits<5> Rm;
+  let Inst{31-30} = 0b01;
+  let Inst{29}    = U;
+  let Inst{28-24} = 0b11110;
+  let Inst{23-22} = size;
+  let Inst{21}    = R;
+  let Inst{20-16} = Rm;
+  let Inst{15-11} = opcode;
+  let Inst{10}    = 1;
+  let Inst{9-5}   = Rn;
+  let Inst{4-0}   = Rd;
+}
+
 multiclass SIMDThreeScalarD<bit U, bits<5> opc, string asm,
                             SDPatternOperator OpNode> {
   def v1i64  : BaseSIMDThreeScalar<U, 0b11, opc, FPR64, asm,
@@ -5323,6 +5630,16 @@ multiclass SIMDThreeScalarHS<bit U, bits<5> opc, string asm,
   def v1i16  : BaseSIMDThreeScalar<U, 0b01, opc, FPR16, asm, []>;
 }
 
+multiclass SIMDThreeScalarHSTied<bit U, bit R, bits<5> opc, string asm,
+                                 SDPatternOperator OpNode = null_frag> {
+  def v1i32: BaseSIMDThreeScalarTied<U, 0b10, R, opc, (outs FPR32:$dst),
+                                     (ins FPR32:$Rd, FPR32:$Rn, FPR32:$Rm), 
+                                     asm, []>;
+  def v1i16: BaseSIMDThreeScalarTied<U, 0b01, R, opc, (outs FPR16:$dst),
+                                     (ins FPR16:$Rd, FPR16:$Rn, FPR16:$Rm), 
+                                     asm, []>;
+}
+
 multiclass SIMDThreeScalarSD<bit U, bit S, bits<5> opc, string asm,
                              SDPatternOperator OpNode = null_frag> {
   let mayLoad = 0, mayStore = 0, hasSideEffects = 0 in {
@@ -5481,14 +5798,14 @@ multiclass SIMDCmpTwoScalarD<bit U, bits<5> opc, string asm,
             (!cast<Instruction>(NAME # v1i64rz) FPR64:$Rn)>;
 }
 
-multiclass SIMDCmpTwoScalarSD<bit U, bit S, bits<5> opc, string asm,
+multiclass SIMDFPCmpTwoScalar<bit U, bit S, bits<5> opc, string asm,
                               SDPatternOperator OpNode> {
   def v1i64rz  : BaseSIMDCmpTwoScalar<U, {S,1}, opc, FPR64, asm, "0.0">;
   def v1i32rz  : BaseSIMDCmpTwoScalar<U, {S,0}, opc, FPR32, asm, "0.0">;
 
-  def : InstAlias<asm # " $Rd, $Rn, #0",
+  def : InstAlias<asm # "\t$Rd, $Rn, #0",
                   (!cast<Instruction>(NAME # v1i64rz) FPR64:$Rd, FPR64:$Rn), 0>;
-  def : InstAlias<asm # " $Rd, $Rn, #0",
+  def : InstAlias<asm # "\t$Rd, $Rn, #0",
                   (!cast<Instruction>(NAME # v1i32rz) FPR32:$Rd, FPR32:$Rn), 0>;
 
   def : Pat<(v1i64 (OpNode (v1f64 FPR64:$Rn))),
@@ -5504,7 +5821,7 @@ multiclass SIMDTwoScalarD<bit U, bits<5> opc, string asm,
             (!cast<Instruction>(NAME # "v1i64") FPR64:$Rn)>;
 }
 
-multiclass SIMDTwoScalarSD<bit U, bit S, bits<5> opc, string asm> {
+multiclass SIMDFPTwoScalar<bit U, bit S, bits<5> opc, string asm> {
   def v1i64       : BaseSIMDTwoScalar<U, {S,1}, opc, FPR64, FPR64, asm,[]>;
   def v1i32       : BaseSIMDTwoScalar<U, {S,0}, opc, FPR32, FPR32, asm,[]>;
 }
@@ -5587,7 +5904,7 @@ multiclass SIMDPairwiseScalarD<bit U, bits<5> opc, string asm> {
                                       asm, ".2d">;
 }
 
-multiclass SIMDPairwiseScalarSD<bit U, bit S, bits<5> opc, string asm> {
+multiclass SIMDFPPairwiseScalar<bit U, bit S, bits<5> opc, string asm> {
   def v2i32p : BaseSIMDPairwiseScalar<U, {S,0}, opc, FPR32Op, V64,
                                       asm, ".2s">;
   def v2i64p : BaseSIMDPairwiseScalar<U, {S,1}, opc, FPR64Op, V128,
@@ -5844,7 +6161,7 @@ class SIMDInsMainMovAlias<string size, Instruction inst,
 class SIMDInsElementMovAlias<string size, Instruction inst,
                              Operand idxtype>
     : InstAlias<"mov" # "{\t$dst" # size # "$idx, $src" # size # "$idx2" #
-                      # "|" # size #" $dst$idx, $src$idx2}",
+                      # "|" # size #"\t$dst$idx, $src$idx2}",
                 (inst V128:$dst, idxtype:$idx, V128:$src, idxtype:$idx2)>;
 
 
@@ -5883,7 +6200,7 @@ multiclass SIMDIns {
     let Inst{20-18} = idx;
     let Inst{17-16} = 0b10;
     let Inst{14-12} = idx2;
-    let Inst{11} = 0;
+    let Inst{11} = {?};
   }
   def vi32lane : SIMDInsFromElement<".s", v4i32, i32, VectorIndexS> {
     bits<2> idx;
@@ -5891,7 +6208,7 @@ multiclass SIMDIns {
     let Inst{20-19} = idx;
     let Inst{18-16} = 0b100;
     let Inst{14-13} = idx2;
-    let Inst{12-11} = 0;
+    let Inst{12-11} = {?,?};
   }
   def vi64lane : SIMDInsFromElement<".d", v2i64, i64, VectorIndexD> {
     bits<1> idx;
@@ -5899,7 +6216,7 @@ multiclass SIMDIns {
     let Inst{20} = idx;
     let Inst{19-16} = 0b1000;
     let Inst{14} = idx2;
-    let Inst{13-11} = 0;
+    let Inst{13-11} = {?,?,?};
   }
 
   // For all forms of the INS instruction, the "mov" mnemonic is the
@@ -6357,8 +6674,8 @@ class BaseSIMDIndexedTied<bit Q, bit U, bit Scalar, bits<2> size, bits<4> opc,
   let Inst{4-0}   = Rd;
 }
 
-multiclass SIMDFPIndexedSD<bit U, bits<4> opc, string asm,
-                           SDPatternOperator OpNode> {
+multiclass SIMDFPIndexed<bit U, bits<4> opc, string asm,
+                         SDPatternOperator OpNode> {
   def v2i32_indexed : BaseSIMDIndexed<0, U, 0, 0b10, opc,
                                       V64, V64,
                                       V128, VectorIndexS,
@@ -6420,7 +6737,7 @@ multiclass SIMDFPIndexedSD<bit U, bits<4> opc, string asm,
   }
 }
 
-multiclass SIMDFPIndexedSDTiedPatterns<string INST, SDPatternOperator OpNode> {
+multiclass SIMDFPIndexedTiedPatterns<string INST, SDPatternOperator OpNode> {
   // 2 variants for the .2s version: DUPLANE from 128-bit and DUP scalar.
   def : Pat<(v2f32 (OpNode (v2f32 V64:$Rd), (v2f32 V64:$Rn),
                            (AArch64duplane32 (v4f32 V128:$Rm),
@@ -6472,7 +6789,7 @@ multiclass SIMDFPIndexedSDTiedPatterns<string INST, SDPatternOperator OpNode> {
                 V128:$Rm, VectorIndexD:$idx)>;
 }
 
-multiclass SIMDFPIndexedSDTied<bit U, bits<4> opc, string asm> {
+multiclass SIMDFPIndexedTied<bit U, bits<4> opc, string asm> {
   def v2i32_indexed : BaseSIMDIndexedTied<0, U, 0, 0b10, opc, V64, V64,
                                           V128, VectorIndexS,
                                           asm, ".2s", ".2s", ".2s", ".s", []> {
@@ -8514,6 +8831,174 @@ multiclass SIMDLdSt4SingleAliases<string asm> {
 }
 } // end of 'let Predicates = [HasNEON]'
 
+//----------------------------------------------------------------------------
+// AdvSIMD v8.1 Rounding Double Multiply Add/Subtract
+//----------------------------------------------------------------------------
+
+let Predicates = [HasNEON, HasV8_1a] in {
+
+class BaseSIMDThreeSameVectorTiedR0<bit Q, bit U, bits<2> size, bits<5> opcode,
+                                    RegisterOperand regtype, string asm, 
+                                    string kind, list<dag> pattern>
+  : BaseSIMDThreeSameVectorTied<Q, U, size, opcode, regtype, asm, kind, 
+                                pattern> {
+  let Inst{21}=0;
+}
+multiclass SIMDThreeSameVectorSQRDMLxHTiedHS<bit U, bits<5> opc, string asm,
+                                             SDPatternOperator Accum> {
+  def v4i16 : BaseSIMDThreeSameVectorTiedR0<0, U, 0b01, opc, V64, asm, ".4h",
+    [(set (v4i16 V64:$dst),
+          (Accum (v4i16 V64:$Rd),
+                 (v4i16 (int_aarch64_neon_sqrdmulh (v4i16 V64:$Rn),
+                                                   (v4i16 V64:$Rm)))))]>;         
+  def v8i16 : BaseSIMDThreeSameVectorTiedR0<1, U, 0b01, opc, V128, asm, ".8h",
+    [(set (v8i16 V128:$dst),
+          (Accum (v8i16 V128:$Rd),
+                 (v8i16 (int_aarch64_neon_sqrdmulh (v8i16 V128:$Rn),
+                                                   (v8i16 V128:$Rm)))))]>;
+  def v2i32 : BaseSIMDThreeSameVectorTiedR0<0, U, 0b10, opc, V64, asm, ".2s",
+    [(set (v2i32 V64:$dst),
+          (Accum (v2i32 V64:$Rd),
+                 (v2i32 (int_aarch64_neon_sqrdmulh (v2i32 V64:$Rn),
+                                                   (v2i32 V64:$Rm)))))]>;
+  def v4i32 : BaseSIMDThreeSameVectorTiedR0<1, U, 0b10, opc, V128, asm, ".4s",
+    [(set (v4i32 V128:$dst),
+          (Accum (v4i32 V128:$Rd),
+                 (v4i32 (int_aarch64_neon_sqrdmulh (v4i32 V128:$Rn),
+                                                   (v4i32 V128:$Rm)))))]>;
+}
+
+multiclass SIMDIndexedSQRDMLxHSDTied<bit U, bits<4> opc, string asm,
+                                     SDPatternOperator Accum> {
+  def v4i16_indexed : BaseSIMDIndexedTied<0, U, 0, 0b01, opc,
+                                          V64, V64, V128_lo, VectorIndexH,
+                                          asm, ".4h", ".4h", ".4h", ".h",
+    [(set (v4i16 V64:$dst),
+          (Accum (v4i16 V64:$Rd),
+                 (v4i16 (int_aarch64_neon_sqrdmulh
+                          (v4i16 V64:$Rn),
+                          (v4i16 (AArch64duplane16 (v8i16 V128_lo:$Rm),
+                                                    VectorIndexH:$idx))))))]> {
+    bits<3> idx;
+    let Inst{11} = idx{2};
+    let Inst{21} = idx{1};
+    let Inst{20} = idx{0};
+  }
+
+  def v8i16_indexed : BaseSIMDIndexedTied<1, U, 0, 0b01, opc,
+                                          V128, V128, V128_lo, VectorIndexH,
+                                          asm, ".8h", ".8h", ".8h", ".h",
+    [(set (v8i16 V128:$dst),
+          (Accum (v8i16 V128:$Rd),
+                 (v8i16 (int_aarch64_neon_sqrdmulh
+                          (v8i16 V128:$Rn),
+                          (v8i16 (AArch64duplane16 (v8i16 V128_lo:$Rm),
+                                                   VectorIndexH:$idx))))))]> {
+    bits<3> idx;
+    let Inst{11} = idx{2};
+    let Inst{21} = idx{1};
+    let Inst{20} = idx{0};
+  }
+
+  def v2i32_indexed : BaseSIMDIndexedTied<0, U, 0, 0b10, opc,
+                                          V64, V64, V128, VectorIndexS,
+                                          asm, ".2s", ".2s", ".2s", ".s",
+    [(set (v2i32 V64:$dst),
+        (Accum (v2i32 V64:$Rd),
+               (v2i32 (int_aarch64_neon_sqrdmulh
+                        (v2i32 V64:$Rn),
+                        (v2i32 (AArch64duplane32 (v4i32 V128:$Rm),
+                                                 VectorIndexS:$idx))))))]> {
+    bits<2> idx;
+    let Inst{11} = idx{1};
+    let Inst{21} = idx{0};
+  }
+
+  // FIXME: it would be nice to use the scalar (v1i32) instruction here, but 
+  // an intermediate EXTRACT_SUBREG would be untyped.
+  // FIXME: direct EXTRACT_SUBREG from v2i32 to i32 is illegal, that's why we 
+  // got it lowered here as (i32 vector_extract (v4i32 insert_subvector(..)))
+  def : Pat<(i32 (Accum (i32 FPR32Op:$Rd),
+                       (i32 (vector_extract 
+                               (v4i32 (insert_subvector
+                                       (undef), 
+                                        (v2i32 (int_aarch64_neon_sqrdmulh 
+                                                 (v2i32 V64:$Rn),
+                                                 (v2i32 (AArch64duplane32 
+                                                          (v4i32 V128:$Rm),
+                                                          VectorIndexS:$idx)))),
+                                      (i32 0))),
+                               (i64 0))))),
+            (EXTRACT_SUBREG
+                (v2i32 (!cast<Instruction>(NAME # v2i32_indexed)
+                          (v2i32 (INSERT_SUBREG (v2i32 (IMPLICIT_DEF)), 
+                                                FPR32Op:$Rd, 
+                                                ssub)), 
+                          V64:$Rn,
+                          V128:$Rm, 
+                          VectorIndexS:$idx)),
+                ssub)>;
+
+  def v4i32_indexed : BaseSIMDIndexedTied<1, U, 0, 0b10, opc,
+                                          V128, V128, V128, VectorIndexS,
+                                          asm, ".4s", ".4s", ".4s", ".s",
+    [(set (v4i32 V128:$dst),
+          (Accum (v4i32 V128:$Rd),
+                 (v4i32 (int_aarch64_neon_sqrdmulh
+                          (v4i32 V128:$Rn),
+                          (v4i32 (AArch64duplane32 (v4i32 V128:$Rm),
+                                                   VectorIndexS:$idx))))))]> {
+    bits<2> idx;
+    let Inst{11} = idx{1};
+    let Inst{21} = idx{0};
+  }
+
+  // FIXME: it would be nice to use the scalar (v1i32) instruction here, but
+  // an intermediate EXTRACT_SUBREG would be untyped.
+  def : Pat<(i32 (Accum (i32 FPR32Op:$Rd),
+                        (i32 (vector_extract 
+                               (v4i32 (int_aarch64_neon_sqrdmulh 
+                                        (v4i32 V128:$Rn),
+                                        (v4i32 (AArch64duplane32 
+                                                 (v4i32 V128:$Rm),
+                                                 VectorIndexS:$idx)))),
+                               (i64 0))))),
+            (EXTRACT_SUBREG
+                (v4i32 (!cast<Instruction>(NAME # v4i32_indexed)
+                         (v4i32 (INSERT_SUBREG (v4i32 (IMPLICIT_DEF)), 
+                                               FPR32Op:$Rd, 
+                                               ssub)), 
+                         V128:$Rn,
+                         V128:$Rm, 
+                         VectorIndexS:$idx)),
+                ssub)>;
+
+  def i16_indexed : BaseSIMDIndexedTied<1, U, 1, 0b01, opc,
+                                        FPR16Op, FPR16Op, V128_lo,
+                                        VectorIndexH, asm, ".h", "", "", ".h", 
+                                        []> {
+    bits<3> idx;
+    let Inst{11} = idx{2};
+    let Inst{21} = idx{1};
+    let Inst{20} = idx{0};
+  }
+
+  def i32_indexed : BaseSIMDIndexedTied<1, U, 1, 0b10, opc,
+                                        FPR32Op, FPR32Op, V128, VectorIndexS,
+                                        asm, ".s", "", "", ".s",
+    [(set (i32 FPR32Op:$dst),
+          (Accum (i32 FPR32Op:$Rd),
+                 (i32 (int_aarch64_neon_sqrdmulh
+                        (i32 FPR32Op:$Rn),
+                        (i32 (vector_extract (v4i32 V128:$Rm),
+                                             VectorIndexS:$idx))))))]> {
+    bits<2> idx;
+    let Inst{11} = idx{1};
+    let Inst{21} = idx{0};
+  }
+}
+} // let Predicates = [HasNeon, HasV8_1a]
+
 //----------------------------------------------------------------------------
 // Crypto extensions
 //----------------------------------------------------------------------------
@@ -8610,6 +9095,178 @@ class SHAInstSS<bits<4> opc, string asm, Intrinsic OpNode>
                [(set (i32 FPR32:$Rd), (OpNode (i32 FPR32:$Rn)))]>;
 } // end of 'let Predicates = [HasCrypto]'
 
+//----------------------------------------------------------------------------
+// v8.1 atomic instructions extension:
+// * CAS
+// * CASP
+// * SWP
+// * LDOPregister<OP>, and aliases STOPregister<OP>
+
+// Instruction encodings:
+//
+//      31 30|29  24|23|22|21|20 16|15|14  10|9 5|4 0
+// CAS  SZ   |001000|1 |A |1 |Rs   |R |11111 |Rn |Rt
+// CASP  0|SZ|001000|0 |A |1 |Rs   |R |11111 |Rn |Rt
+// SWP  SZ   |111000|A |R |1 |Rs   |1 |OPC|00|Rn |Rt
+// LD   SZ   |111000|A |R |1 |Rs   |0 |OPC|00|Rn |Rt
+// ST   SZ   |111000|A |R |1 |Rs   |0 |OPC|00|Rn |11111
+
+// Instruction syntax:
+//
+// CAS{<order>}[<size>] <Ws>, <Wt>, [<Xn|SP>]
+// CAS{<order>} <Xs>, <Xt>, [<Xn|SP>]
+// CASP{<order>} <Ws>, <W(s+1)>, <Wt>, <W(t+1)>, [<Xn|SP>]
+// CASP{<order>} <Xs>, <X(s+1)>, <Xt>, <X(t+1)>, [<Xn|SP>]
+// SWP{<order>}[<size>] <Ws>, <Wt>, [<Xn|SP>]
+// SWP{<order>} <Xs>, <Xt>, [<Xn|SP>]
+// LD<OP>{<order>}[<size>] <Ws>, <Wt>, [<Xn|SP>]
+// LD<OP>{<order>} <Xs>, <Xt>, [<Xn|SP>]
+// ST<OP>{<order>}[<size>] <Ws>, [<Xn|SP>]
+// ST<OP>{<order>} <Xs>, [<Xn|SP>]
+
+let Predicates = [HasV8_1a], mayLoad = 1, mayStore = 1, hasSideEffects = 1 in
+class BaseCASEncoding<dag oops, dag iops, string asm, string operands,
+                      string cstr, list<dag> pattern>
+      : I<oops, iops, asm, operands, cstr, pattern> {
+  bits<2> Sz;
+  bit NP;
+  bit Acq;
+  bit Rel;
+  bits<5> Rs;
+  bits<5> Rn;
+  bits<5> Rt;
+  let Inst{31-30} = Sz;
+  let Inst{29-24} = 0b001000;
+  let Inst{23} = NP;
+  let Inst{22} = Acq;
+  let Inst{21} = 0b1;
+  let Inst{20-16} = Rs;
+  let Inst{15} = Rel;
+  let Inst{14-10} = 0b11111;
+  let Inst{9-5} = Rn;
+  let Inst{4-0} = Rt;
+}
+
+class BaseCAS<string order, string size, RegisterClass RC>
+      : BaseCASEncoding<(outs RC:$out),(ins RC:$Rs, RC:$Rt, GPR64sp:$Rn),
+                        "cas" # order # size, "\t$Rs, $Rt, [$Rn]",
+                        "$out = $Rs",[]> {
+  let NP = 1;
+}
+
+multiclass CompareAndSwap<bits<1> Acq, bits<1> Rel, string order> {
+  let Sz = 0b00, Acq = Acq, Rel = Rel in def b : BaseCAS<order, "b", GPR32>;
+  let Sz = 0b01, Acq = Acq, Rel = Rel in def h : BaseCAS<order, "h", GPR32>;
+  let Sz = 0b10, Acq = Acq, Rel = Rel in def s : BaseCAS<order, "", GPR32>;
+  let Sz = 0b11, Acq = Acq, Rel = Rel in def d : BaseCAS<order, "", GPR64>;
+}
+
+class BaseCASP<string order, string size, RegisterOperand RC>
+      : BaseCASEncoding<(outs RC:$out),(ins RC:$Rs, RC:$Rt, GPR64sp:$Rn),
+                        "casp" # order # size, "\t$Rs, $Rt, [$Rn]",
+                        "$out = $Rs",[]> {
+  let NP = 0;
+}
+
+multiclass CompareAndSwapPair<bits<1> Acq, bits<1> Rel, string order> {
+  let Sz = 0b00, Acq = Acq, Rel = Rel in 
+    def s : BaseCASP<order, "", WSeqPairClassOperand>;
+  let Sz = 0b01, Acq = Acq, Rel = Rel in 
+    def d : BaseCASP<order, "", XSeqPairClassOperand>;
+}
+
+let Predicates = [HasV8_1a] in
+class BaseSWP<string order, string size, RegisterClass RC>
+      : I<(outs RC:$Rt),(ins RC:$Rs, GPR64sp:$Rn), "swp" # order # size,
+          "\t$Rs, $Rt, [$Rn]","",[]> {
+  bits<2> Sz;
+  bit Acq;
+  bit Rel;
+  bits<5> Rs;
+  bits<3> opc = 0b000;
+  bits<5> Rn;
+  bits<5> Rt;
+  let Inst{31-30} = Sz;
+  let Inst{29-24} = 0b111000;
+  let Inst{23} = Acq;
+  let Inst{22} = Rel;
+  let Inst{21} = 0b1;
+  let Inst{20-16} = Rs;
+  let Inst{15} = 0b1;
+  let Inst{14-12} = opc;
+  let Inst{11-10} = 0b00;
+  let Inst{9-5} = Rn;
+  let Inst{4-0} = Rt;
+}
+
+multiclass Swap<bits<1> Acq, bits<1> Rel, string order> {
+  let Sz = 0b00, Acq = Acq, Rel = Rel in def b : BaseSWP<order, "b", GPR32>;
+  let Sz = 0b01, Acq = Acq, Rel = Rel in def h : BaseSWP<order, "h", GPR32>;
+  let Sz = 0b10, Acq = Acq, Rel = Rel in def s : BaseSWP<order, "", GPR32>;
+  let Sz = 0b11, Acq = Acq, Rel = Rel in def d : BaseSWP<order, "", GPR64>;
+}
+
+let Predicates = [HasV8_1a], mayLoad = 1, mayStore = 1, hasSideEffects = 1 in
+class BaseLDOPregister<string op, string order, string size, RegisterClass RC>
+      : I<(outs RC:$Rt),(ins RC:$Rs, GPR64sp:$Rn), "ld" # op # order # size,
+          "\t$Rs, $Rt, [$Rn]","",[]> {
+  bits<2> Sz;
+  bit Acq;
+  bit Rel;
+  bits<5> Rs;
+  bits<3> opc;
+  bits<5> Rn;
+  bits<5> Rt;
+  let Inst{31-30} = Sz;
+  let Inst{29-24} = 0b111000;
+  let Inst{23} = Acq;
+  let Inst{22} = Rel;
+  let Inst{21} = 0b1;
+  let Inst{20-16} = Rs;
+  let Inst{15} = 0b0;
+  let Inst{14-12} = opc;
+  let Inst{11-10} = 0b00;
+  let Inst{9-5} = Rn;
+  let Inst{4-0} = Rt;
+}
+
+multiclass LDOPregister<bits<3> opc, string op, bits<1> Acq, bits<1> Rel, 
+                        string order> {
+  let Sz = 0b00, Acq = Acq, Rel = Rel, opc = opc in 
+    def b : BaseLDOPregister<op, order, "b", GPR32>;
+  let Sz = 0b01, Acq = Acq, Rel = Rel, opc = opc in 
+    def h : BaseLDOPregister<op, order, "h", GPR32>;
+  let Sz = 0b10, Acq = Acq, Rel = Rel, opc = opc in 
+    def s : BaseLDOPregister<op, order, "", GPR32>;
+  let Sz = 0b11, Acq = Acq, Rel = Rel, opc = opc in 
+    def d : BaseLDOPregister<op, order, "", GPR64>;
+}
+
+let Predicates = [HasV8_1a] in
+class BaseSTOPregister<string asm, RegisterClass OP, Register Reg,
+                        Instruction inst> :
+      InstAlias<asm # "\t$Rs, [$Rn]", (inst Reg, OP:$Rs, GPR64sp:$Rn)>;
+
+multiclass STOPregister<string asm, string instr> {
+  def : BaseSTOPregister<asm # "lb", GPR32, WZR, 
+                    !cast<Instruction>(instr # "Lb")>;
+  def : BaseSTOPregister<asm # "lh", GPR32, WZR, 
+                    !cast<Instruction>(instr # "Lh")>;
+  def : BaseSTOPregister<asm # "l",  GPR32, WZR, 
+                    !cast<Instruction>(instr # "Ls")>;
+  def : BaseSTOPregister<asm # "l",  GPR64, XZR, 
+                    !cast<Instruction>(instr # "Ld")>;
+  def : BaseSTOPregister<asm # "b",  GPR32, WZR, 
+                    !cast<Instruction>(instr # "b")>;
+  def : BaseSTOPregister<asm # "h",  GPR32, WZR, 
+                    !cast<Instruction>(instr # "h")>;
+  def : BaseSTOPregister<asm,        GPR32, WZR, 
+                    !cast<Instruction>(instr # "s")>;
+  def : BaseSTOPregister<asm,        GPR64, XZR, 
+                    !cast<Instruction>(instr # "d")>;
+}
+
+//----------------------------------------------------------------------------
 // Allow the size specifier tokens to be upper case, not just lower.
 def : TokenAlias<".8B", ".8b">;
 def : TokenAlias<".4H", ".4h">;