diff --git a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
index cd7f0e719ad0c..9dbcab630d5a5 100644
--- a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
+++ b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
@@ -698,6 +698,11 @@ AArch64TargetLowering::AArch64TargetLowering(const TargetMachine &TM,
     setOperationAction(ISD::ABS, MVT::i64, Custom);
   }
 
+  setOperationAction(ISD::ABDS, MVT::i32, Custom);
+  setOperationAction(ISD::ABDS, MVT::i64, Custom);
+  setOperationAction(ISD::ABDU, MVT::i32, Custom);
+  setOperationAction(ISD::ABDU, MVT::i64, Custom);
+
   setOperationAction(ISD::SDIVREM, MVT::i32, Expand);
   setOperationAction(ISD::SDIVREM, MVT::i64, Expand);
   for (MVT VT : MVT::fixedlen_vector_valuetypes()) {
@@ -3712,7 +3717,8 @@ static SDValue emitStrictFPComparison(SDValue LHS, SDValue RHS, const SDLoc &DL,
 }
 
 static SDValue emitComparison(SDValue LHS, SDValue RHS, ISD::CondCode CC,
-                              const SDLoc &DL, SelectionDAG &DAG) {
+                              const SDLoc &DL, SelectionDAG &DAG,
+                              bool MIOrPLSupported = false) {
   EVT VT = LHS.getValueType();
   const bool FullFP16 = DAG.getSubtarget<AArch64Subtarget>().hasFullFP16();
 
@@ -3755,6 +3761,33 @@ static SDValue emitComparison(SDValue LHS, SDValue RHS, ISD::CondCode CC,
     } else if (LHS.getOpcode() == AArch64ISD::ANDS) {
       // Use result of ANDS
       return LHS.getValue(1);
+    } else if (MIOrPLSupported) {
+      // For MIOrPLSupported, optimize SUB/ADD operations with zero comparison
+      if (LHS.getOpcode() == ISD::SUB && CC == ISD::SETLT) {
+        // SUB(x, y) < 0 -> SUBS(x, y)
+        return DAG
+            .getNode(AArch64ISD::SUBS, DL, DAG.getVTList(VT, FlagsVT),
+                     LHS.getOperand(0), LHS.getOperand(1))
+            .getValue(1);
+      } else if (LHS.getOpcode() == ISD::ADD && CC == ISD::SETGE) {
+        // ADD(x, y) >= 0 -> ADDS(x, y)
+        return DAG
+            .getNode(AArch64ISD::ADDS, DL, DAG.getVTList(VT, FlagsVT),
+                     LHS.getOperand(0), LHS.getOperand(1))
+            .getValue(1);
+      } else if (LHS.getOpcode() == ISD::ADD && CC == ISD::SETLT) {
+        // ADD(x, y) < 0 -> SUBS(x, y)
+        return DAG
+            .getNode(AArch64ISD::ADDS, DL, DAG.getVTList(VT, FlagsVT),
+                     LHS.getOperand(0), LHS.getOperand(1))
+            .getValue(1);
+      } else if (LHS.getOpcode() == ISD::SUB && CC == ISD::SETGE) {
+        // SUB(x, y) >= 0 -> ADDS(x, y)
+        return DAG
+            .getNode(AArch64ISD::SUBS, DL, DAG.getVTList(VT, FlagsVT),
+                     LHS.getOperand(0), LHS.getOperand(1))
+            .getValue(1);
+      }
     }
   }
 
@@ -3819,7 +3852,8 @@ static SDValue emitConditionalComparison(SDValue LHS, SDValue RHS,
                                          ISD::CondCode CC, SDValue CCOp,
                                          AArch64CC::CondCode Predicate,
                                          AArch64CC::CondCode OutCC,
-                                         const SDLoc &DL, SelectionDAG &DAG) {
+                                         const SDLoc &DL, SelectionDAG &DAG,
+                                         bool MIOrPLSupported = false) {
   unsigned Opcode = 0;
   const bool FullFP16 = DAG.getSubtarget<AArch64Subtarget>().hasFullFP16();
 
@@ -3846,6 +3880,30 @@ static SDValue emitConditionalComparison(SDValue LHS, SDValue RHS,
     // we combine a (CCMP (sub 0, op1), op2) into a CCMN instruction ?
     Opcode = AArch64ISD::CCMN;
     LHS = LHS.getOperand(1);
+  } else if (isNullConstant(RHS) && !isUnsignedIntSetCC(CC) &&
+             MIOrPLSupported) {
+    // For MIOrPLSupported, optimize SUB/ADD operations with zero comparison
+    if (LHS.getOpcode() == ISD::SUB && CC == ISD::SETLT) {
+      // SUB(x, y) < 0 -> CCMP(x, y) with appropriate condition
+      Opcode = AArch64ISD::CCMP;
+      RHS = LHS.getOperand(1);
+      LHS = LHS.getOperand(0);
+    } else if (LHS.getOpcode() == ISD::ADD && CC == ISD::SETGE) {
+      // ADD(x, y) >= 0 -> CCMP(x, y) with appropriate condition
+      Opcode = AArch64ISD::CCMN;
+      RHS = LHS.getOperand(1);
+      LHS = LHS.getOperand(0);
+    } else if (LHS.getOpcode() == ISD::ADD && CC == ISD::SETLT) {
+      // ADD(x, y) < 0 -> CCMP(x, -y) with appropriate condition
+      Opcode = AArch64ISD::CCMN;
+      RHS = LHS.getOperand(1);
+      LHS = LHS.getOperand(0);
+    } else if (LHS.getOpcode() == ISD::SUB && CC == ISD::SETGE) {
+      // SUB(x, y) >= 0 -> CCMP(-x, y) with appropriate condition
+      Opcode = AArch64ISD::CCMP;
+      RHS = LHS.getOperand(1);
+      LHS = LHS.getOperand(0);
+    }
   }
   if (Opcode == 0)
     Opcode = AArch64ISD::CCMP;
@@ -3972,7 +4030,7 @@ static SDValue emitConjunctionRec(SelectionDAG &DAG, SDValue Val,
       return emitComparison(LHS, RHS, CC, DL, DAG);
     // Otherwise produce a ccmp.
     return emitConditionalComparison(LHS, RHS, CC, CCOp, Predicate, OutCC, DL,
-                                     DAG);
+                                     DAG, true);
   }
   assert(Val->hasOneUse() && "Valid conjunction/disjunction tree");
 
@@ -4251,7 +4309,7 @@ static SDValue getAArch64Cmp(SDValue LHS, SDValue RHS, ISD::CondCode CC,
   }
 
   if (!Cmp) {
-    Cmp = emitComparison(LHS, RHS, CC, DL, DAG);
+    Cmp = emitComparison(LHS, RHS, CC, DL, DAG, true);
     AArch64CC = changeIntCCToAArch64CC(CC, RHS);
   }
   AArch64cc = getCondCode(DAG, AArch64CC);
@@ -7371,13 +7429,100 @@ SDValue AArch64TargetLowering::LowerABS(SDValue Op, SelectionDAG &DAG) const {
     return LowerToPredicatedOp(Op, DAG, AArch64ISD::ABS_MERGE_PASSTHRU);
 
   SDLoc DL(Op);
-  SDValue Neg = DAG.getNegative(Op.getOperand(0), DL, VT);
 
-  // Generate SUBS & CSEL.
-  SDValue Cmp = DAG.getNode(AArch64ISD::SUBS, DL, DAG.getVTList(VT, FlagsVT),
-                            Op.getOperand(0), DAG.getConstant(0, DL, VT));
+  // Generate CMP & CSEL.
+  SDValue Cmp = emitComparison(Op.getOperand(0), DAG.getConstant(0, DL, VT),
+                               ISD::SETGE, DL, DAG, true);
+  SDValue Neg = DAG.getNegative(Op.getOperand(0), DL, VT);
   return DAG.getNode(AArch64ISD::CSEL, DL, VT, Op.getOperand(0), Neg,
-                     getCondCode(DAG, AArch64CC::PL), Cmp.getValue(1));
+                     getCondCode(DAG, AArch64CC::PL), Cmp);
+}
+
+// Generate SUBS and CNEG for absolute difference.
+SDValue AArch64TargetLowering::LowerABD(SDValue Op, SelectionDAG &DAG) const {
+  MVT VT = Op.getSimpleValueType();
+
+  bool IsSigned = Op.getOpcode() == ISD::ABDS;
+  if (VT.isVector()) {
+    if (IsSigned)
+      return LowerToPredicatedOp(Op, DAG, AArch64ISD::ABDS_PRED);
+    else
+      return LowerToPredicatedOp(Op, DAG, AArch64ISD::ABDU_PRED);
+  }
+
+  SDValue LHS = Op.getOperand(0);
+  SDValue RHS = Op.getOperand(1);
+  SDLoc DL(Op);
+
+  if (!isa<ConstantSDNode>(RHS) || !isLegalCmpImmed(RHS->getAsAPIntVal())) {
+    SDValue TheLHS = isCMN(LHS, IsSigned ? ISD::SETGE : ISD::SETUGE, DAG)
+                         ? LHS.getOperand(1)
+                         : LHS;
+    SDValue TheRHS = isCMN(RHS, IsSigned ? ISD::SETGE : ISD::SETUGE, DAG)
+                         ? RHS.getOperand(1)
+                         : RHS;
+    if (getCmpOperandFoldingProfit(TheLHS) >
+        getCmpOperandFoldingProfit(TheRHS)) {
+      std::swap(LHS, RHS);
+    }
+  }
+
+  // If the subtract doesn't overflow then just use abs(sub())
+  bool IsNonNegative = DAG.SignBitIsZero(LHS) && DAG.SignBitIsZero(RHS);
+
+  if (DAG.willNotOverflowSub(IsSigned || IsNonNegative, LHS, RHS))
+    return DAG.getNode(ISD::ABS, DL, VT,
+                       DAG.getNode(ISD::SUB, DL, VT, LHS, RHS));
+
+  if (DAG.willNotOverflowSub(IsSigned || IsNonNegative, RHS, LHS))
+    return DAG.getNode(ISD::ABS, DL, VT,
+                       DAG.getNode(ISD::SUB, DL, VT, RHS, LHS));
+
+  unsigned Opcode = AArch64ISD::SUBS;
+  // Check if RHS is a subtraction against 0: (0 - X)
+  if (RHS.getOpcode() == ISD::SUB) {
+    SDValue SubLHS = RHS.getOperand(0);
+    SDValue SubRHS = RHS.getOperand(1);
+
+    // Check if it's 0 - X
+    if (isNullConstant(SubLHS)) {
+      bool CanUseAdd = false;
+      if (IsSigned) {
+        // For UCMP: only if X is known to never be INT_MIN (to avoid overflow)
+        if (RHS->getFlags().hasNoSignedWrap() || !DAG.computeKnownBits(SubRHS)
+                                                      .getSignedMinValue()
+                                                      .isMinSignedValue()) {
+          CanUseAdd = true;
+        }
+      } else {
+        // For UCMP: only if X is known to never be zero
+        if (DAG.isKnownNeverZero(SubRHS)) {
+          CanUseAdd = true;
+        }
+      }
+
+      if (CanUseAdd) {
+        Opcode = AArch64ISD::ADDS;
+        RHS = SubRHS; // Replace RHS with X, so we do LHS + X instead of
+                      // LHS - (0 - X)
+      }
+    }
+  }
+
+  // Generate SUBS and CSEL for absolute difference (like LowerABS)
+  // Compute a - b with flags
+  SDValue Cmp = DAG.getNode(Opcode, DL, DAG.getVTList(VT, FlagsVT), LHS, RHS);
+
+  // Compute b - a (negative of a - b)
+  SDValue Neg = DAG.getNegative(Cmp.getValue(0), DL, VT);
+
+  // For unsigned: use HS (a >= b) to select a-b, otherwise b-a
+  // For signed: use GE (a >= b) to select a-b, otherwise b-a
+  AArch64CC::CondCode CC = IsSigned ? AArch64CC::GE : AArch64CC::HS;
+
+  // CSEL: if a > b, select a-b, otherwise b-a
+  return DAG.getNode(AArch64ISD::CSEL, DL, VT, Cmp.getValue(0), Neg,
+                     getCondCode(DAG, CC), Cmp.getValue(1));
 }
 
 static SDValue LowerBRCOND(SDValue Op, SelectionDAG &DAG) {
@@ -7832,9 +7977,8 @@ SDValue AArch64TargetLowering::LowerOperation(SDValue Op,
   case ISD::ABS:
     return LowerABS(Op, DAG);
   case ISD::ABDS:
-    return LowerToPredicatedOp(Op, DAG, AArch64ISD::ABDS_PRED);
   case ISD::ABDU:
-    return LowerToPredicatedOp(Op, DAG, AArch64ISD::ABDU_PRED);
+    return LowerABD(Op, DAG);
   case ISD::AVGFLOORS:
     return LowerAVG(Op, DAG, AArch64ISD::HADDS_PRED);
   case ISD::AVGFLOORU:
@@ -25815,29 +25959,6 @@ static SDValue performCSELCombine(SDNode *N,
     }
   }
 
-  // CSEL a, b, cc, SUBS(SUB(x,y), 0) -> CSEL a, b, cc, SUBS(x,y) if cc doesn't
-  // use overflow flags, to avoid the comparison with zero. In case of success,
-  // this also replaces the original SUB(x,y) with the newly created SUBS(x,y).
-  // NOTE: Perhaps in the future use performFlagSettingCombine to replace SUB
-  // nodes with their SUBS equivalent as is already done for other flag-setting
-  // operators, in which case doing the replacement here becomes redundant.
-  if (Cond.getOpcode() == AArch64ISD::SUBS && Cond->hasNUsesOfValue(1, 1) &&
-      isNullConstant(Cond.getOperand(1))) {
-    SDValue Sub = Cond.getOperand(0);
-    AArch64CC::CondCode CC =
-        static_cast<AArch64CC::CondCode>(N->getConstantOperandVal(2));
-    if (Sub.getOpcode() == ISD::SUB &&
-        (CC == AArch64CC::EQ || CC == AArch64CC::NE || CC == AArch64CC::MI ||
-         CC == AArch64CC::PL)) {
-      SDLoc DL(N);
-      SDValue Subs = DAG.getNode(AArch64ISD::SUBS, DL, Cond->getVTList(),
-                                 Sub.getOperand(0), Sub.getOperand(1));
-      DCI.CombineTo(Sub.getNode(), Subs);
-      DCI.CombineTo(Cond.getNode(), Subs, Subs.getValue(1));
-      return SDValue(N, 0);
-    }
-  }
-
   // CSEL (LASTB P, Z), X, NE(ANY P) -> CLASTB P, X, Z
   if (SDValue CondLast = foldCSELofLASTB(N, DAG))
     return CondLast;
diff --git a/llvm/lib/Target/AArch64/AArch64ISelLowering.h b/llvm/lib/Target/AArch64/AArch64ISelLowering.h
index ff073d3eafb1f..4de8985a44ff8 100644
--- a/llvm/lib/Target/AArch64/AArch64ISelLowering.h
+++ b/llvm/lib/Target/AArch64/AArch64ISelLowering.h
@@ -604,6 +604,7 @@ class AArch64TargetLowering : public TargetLowering {
   SDValue LowerSTORE(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerStore128(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerABS(SDValue Op, SelectionDAG &DAG) const;
+  SDValue LowerABD(SDValue Op, SelectionDAG &DAG) const;
 
   SDValue LowerMGATHER(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerMSCATTER(SDValue Op, SelectionDAG &DAG) const;
diff --git a/llvm/lib/Target/AArch64/AArch64InstrInfo.cpp b/llvm/lib/Target/AArch64/AArch64InstrInfo.cpp
index 5a51c812732e6..6f469d73761ae 100644
--- a/llvm/lib/Target/AArch64/AArch64InstrInfo.cpp
+++ b/llvm/lib/Target/AArch64/AArch64InstrInfo.cpp
@@ -1730,12 +1730,20 @@ static unsigned sForm(MachineInstr &Instr) {
 
   case AArch64::ADDSWrr:
   case AArch64::ADDSWri:
+  case AArch64::ADDSWrx:
   case AArch64::ADDSXrr:
   case AArch64::ADDSXri:
+  case AArch64::ADDSXrx:
   case AArch64::SUBSWrr:
   case AArch64::SUBSWri:
+  case AArch64::SUBSWrx:
   case AArch64::SUBSXrr:
   case AArch64::SUBSXri:
+  case AArch64::SUBSXrx:
+  case AArch64::ADCSWr:
+  case AArch64::ADCSXr:
+  case AArch64::SBCSWr:
+  case AArch64::SBCSXr:
     return Instr.getOpcode();
 
   case AArch64::ADDWrr:
@@ -1746,6 +1754,10 @@ static unsigned sForm(MachineInstr &Instr) {
     return AArch64::ADDSXrr;
   case AArch64::ADDXri:
     return AArch64::ADDSXri;
+  case AArch64::ADDWrx:
+    return AArch64::ADDSWrx;
+  case AArch64::ADDXrx:
+    return AArch64::ADDSXrx;
   case AArch64::ADCWr:
     return AArch64::ADCSWr;
   case AArch64::ADCXr:
@@ -1758,6 +1770,10 @@ static unsigned sForm(MachineInstr &Instr) {
     return AArch64::SUBSXrr;
   case AArch64::SUBXri:
     return AArch64::SUBSXri;
+  case AArch64::SUBWrx:
+    return AArch64::SUBSWrx;
+  case AArch64::SUBXrx:
+    return AArch64::SUBSXrx;
   case AArch64::SBCWr:
     return AArch64::SBCSWr;
   case AArch64::SBCXr:
diff --git a/llvm/test/CodeGen/AArch64/abds-neg.ll b/llvm/test/CodeGen/AArch64/abds-neg.ll
index 37319642f5b34..11bfcb92593cb 100644
--- a/llvm/test/CodeGen/AArch64/abds-neg.ll
+++ b/llvm/test/CodeGen/AArch64/abds-neg.ll
@@ -72,9 +72,8 @@ define i16 @abd_ext_i16(i16 %a, i16 %b) nounwind {
 define i16 @abd_ext_i16_i32(i16 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i16_i32:
 ; CHECK:       // %bb.0:
-; CHECK-NEXT:    sxth w8, w0
-; CHECK-NEXT:    subs w8, w8, w1
-; CHECK-NEXT:    cneg w0, w8, gt
+; CHECK-NEXT:    subs w8, w1, w0, sxth
+; CHECK-NEXT:    cneg w0, w8, ge
 ; CHECK-NEXT:    ret
   %aext = sext i16 %a to i64
   %bext = sext i32 %b to i64
@@ -105,7 +104,7 @@ define i32 @abd_ext_i32(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i32:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, gt
+; CHECK-NEXT:    cneg w0, w8, ge
 ; CHECK-NEXT:    ret
   %aext = sext i32 %a to i64
   %bext = sext i32 %b to i64
@@ -120,7 +119,7 @@ define i32 @abd_ext_i32_i16(i32 %a, i16 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i32_i16:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1, sxth
-; CHECK-NEXT:    cneg w0, w8, gt
+; CHECK-NEXT:    cneg w0, w8, ge
 ; CHECK-NEXT:    ret
   %aext = sext i32 %a to i64
   %bext = sext i16 %b to i64
@@ -135,7 +134,7 @@ define i32 @abd_ext_i32_undef(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i32_undef:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, gt
+; CHECK-NEXT:    cneg w0, w8, ge
 ; CHECK-NEXT:    ret
   %aext = sext i32 %a to i64
   %bext = sext i32 %b to i64
@@ -150,7 +149,7 @@ define i64 @abd_ext_i64(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i64:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, gt
+; CHECK-NEXT:    cneg x0, x8, ge
 ; CHECK-NEXT:    ret
   %aext = sext i64 %a to i128
   %bext = sext i64 %b to i128
@@ -165,7 +164,7 @@ define i64 @abd_ext_i64_undef(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i64_undef:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, gt
+; CHECK-NEXT:    cneg x0, x8, ge
 ; CHECK-NEXT:    ret
   %aext = sext i64 %a to i128
   %bext = sext i64 %b to i128
@@ -225,12 +224,9 @@ define i128 @abd_ext_i128_undef(i128 %a, i128 %b) nounwind {
 define i8 @abd_minmax_i8(i8 %a, i8 %b) nounwind {
 ; CHECK-LABEL: abd_minmax_i8:
 ; CHECK:       // %bb.0:
-; CHECK-NEXT:    sxtb w8, w1
-; CHECK-NEXT:    sxtb w9, w0
-; CHECK-NEXT:    cmp w9, w8
-; CHECK-NEXT:    csel w10, w9, w8, lt
-; CHECK-NEXT:    csel w8, w9, w8, gt
-; CHECK-NEXT:    sub w0, w10, w8
+; CHECK-NEXT:    sxtb w8, w0
+; CHECK-NEXT:    subs w8, w8, w1, sxtb
+; CHECK-NEXT:    cneg w0, w8, pl
 ; CHECK-NEXT:    ret
   %min = call i8 @llvm.smin.i8(i8 %a, i8 %b)
   %max = call i8 @llvm.smax.i8(i8 %a, i8 %b)
@@ -241,12 +237,9 @@ define i8 @abd_minmax_i8(i8 %a, i8 %b) nounwind {
 define i16 @abd_minmax_i16(i16 %a, i16 %b) nounwind {
 ; CHECK-LABEL: abd_minmax_i16:
 ; CHECK:       // %bb.0:
-; CHECK-NEXT:    sxth w8, w1
-; CHECK-NEXT:    sxth w9, w0
-; CHECK-NEXT:    cmp w9, w8
-; CHECK-NEXT:    csel w10, w9, w8, lt
-; CHECK-NEXT:    csel w8, w9, w8, gt
-; CHECK-NEXT:    sub w0, w10, w8
+; CHECK-NEXT:    sxth w8, w0
+; CHECK-NEXT:    subs w8, w8, w1, sxth
+; CHECK-NEXT:    cneg w0, w8, pl
 ; CHECK-NEXT:    ret
   %min = call i16 @llvm.smin.i16(i16 %a, i16 %b)
   %max = call i16 @llvm.smax.i16(i16 %a, i16 %b)
@@ -257,10 +250,8 @@ define i16 @abd_minmax_i16(i16 %a, i16 %b) nounwind {
 define i32 @abd_minmax_i32(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_minmax_i32:
 ; CHECK:       // %bb.0:
-; CHECK-NEXT:    cmp w0, w1
-; CHECK-NEXT:    csel w8, w0, w1, lt
-; CHECK-NEXT:    csel w9, w0, w1, gt
-; CHECK-NEXT:    sub w0, w8, w9
+; CHECK-NEXT:    subs w8, w0, w1
+; CHECK-NEXT:    cneg w0, w8, ge
 ; CHECK-NEXT:    ret
   %min = call i32 @llvm.smin.i32(i32 %a, i32 %b)
   %max = call i32 @llvm.smax.i32(i32 %a, i32 %b)
@@ -271,10 +262,8 @@ define i32 @abd_minmax_i32(i32 %a, i32 %b) nounwind {
 define i64 @abd_minmax_i64(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_minmax_i64:
 ; CHECK:       // %bb.0:
-; CHECK-NEXT:    cmp x0, x1
-; CHECK-NEXT:    csel x8, x0, x1, lt
-; CHECK-NEXT:    csel x9, x0, x1, gt
-; CHECK-NEXT:    sub x0, x8, x9
+; CHECK-NEXT:    subs x8, x0, x1
+; CHECK-NEXT:    cneg x0, x8, ge
 ; CHECK-NEXT:    ret
   %min = call i64 @llvm.smin.i64(i64 %a, i64 %b)
   %max = call i64 @llvm.smax.i64(i64 %a, i64 %b)
@@ -445,7 +434,7 @@ define i32 @abd_subnsw_i32(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_subnsw_i32:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, pl
+; CHECK-NEXT:    cneg w0, w8, ge
 ; CHECK-NEXT:    ret
   %sub = sub nsw i32 %a, %b
   %abs = call i32 @llvm.abs.i32(i32 %sub, i1 false)
@@ -457,7 +446,7 @@ define i32 @abd_subnsw_i32_undef(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_subnsw_i32_undef:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, pl
+; CHECK-NEXT:    cneg w0, w8, ge
 ; CHECK-NEXT:    ret
   %sub = sub nsw i32 %a, %b
   %abs = call i32 @llvm.abs.i32(i32 %sub, i1 true)
@@ -469,7 +458,7 @@ define i64 @abd_subnsw_i64(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_subnsw_i64:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, pl
+; CHECK-NEXT:    cneg x0, x8, ge
 ; CHECK-NEXT:    ret
   %sub = sub nsw i64 %a, %b
   %abs = call i64 @llvm.abs.i64(i64 %sub, i1 false)
@@ -481,7 +470,7 @@ define i64 @abd_subnsw_i64_undef(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_subnsw_i64_undef:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, pl
+; CHECK-NEXT:    cneg x0, x8, ge
 ; CHECK-NEXT:    ret
   %sub = sub nsw i64 %a, %b
   %abs = call i64 @llvm.abs.i64(i64 %sub, i1 true)
diff --git a/llvm/test/CodeGen/AArch64/abds.ll b/llvm/test/CodeGen/AArch64/abds.ll
index 1ef1c1c68c7bb..42618ac0cf68d 100644
--- a/llvm/test/CodeGen/AArch64/abds.ll
+++ b/llvm/test/CodeGen/AArch64/abds.ll
@@ -68,9 +68,8 @@ define i16 @abd_ext_i16(i16 %a, i16 %b) nounwind {
 define i16 @abd_ext_i16_i32(i16 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i16_i32:
 ; CHECK:       // %bb.0:
-; CHECK-NEXT:    sxth w8, w0
-; CHECK-NEXT:    subs w8, w8, w1
-; CHECK-NEXT:    cneg w0, w8, le
+; CHECK-NEXT:    subs w8, w1, w0, sxth
+; CHECK-NEXT:    cneg w0, w8, lt
 ; CHECK-NEXT:    ret
   %aext = sext i16 %a to i64
   %bext = sext i32 %b to i64
@@ -99,7 +98,7 @@ define i32 @abd_ext_i32(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i32:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, le
+; CHECK-NEXT:    cneg w0, w8, lt
 ; CHECK-NEXT:    ret
   %aext = sext i32 %a to i64
   %bext = sext i32 %b to i64
@@ -113,7 +112,7 @@ define i32 @abd_ext_i32_i16(i32 %a, i16 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i32_i16:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1, sxth
-; CHECK-NEXT:    cneg w0, w8, le
+; CHECK-NEXT:    cneg w0, w8, lt
 ; CHECK-NEXT:    ret
   %aext = sext i32 %a to i64
   %bext = sext i16 %b to i64
@@ -127,7 +126,7 @@ define i32 @abd_ext_i32_undef(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i32_undef:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, le
+; CHECK-NEXT:    cneg w0, w8, lt
 ; CHECK-NEXT:    ret
   %aext = sext i32 %a to i64
   %bext = sext i32 %b to i64
@@ -141,7 +140,7 @@ define i64 @abd_ext_i64(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i64:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, le
+; CHECK-NEXT:    cneg x0, x8, lt
 ; CHECK-NEXT:    ret
   %aext = sext i64 %a to i128
   %bext = sext i64 %b to i128
@@ -155,7 +154,7 @@ define i64 @abd_ext_i64_undef(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i64_undef:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, le
+; CHECK-NEXT:    cneg x0, x8, lt
 ; CHECK-NEXT:    ret
   %aext = sext i64 %a to i128
   %bext = sext i64 %b to i128
@@ -235,7 +234,7 @@ define i32 @abd_minmax_i32(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_minmax_i32:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, le
+; CHECK-NEXT:    cneg w0, w8, lt
 ; CHECK-NEXT:    ret
   %min = call i32 @llvm.smin.i32(i32 %a, i32 %b)
   %max = call i32 @llvm.smax.i32(i32 %a, i32 %b)
@@ -247,7 +246,7 @@ define i64 @abd_minmax_i64(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_minmax_i64:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, le
+; CHECK-NEXT:    cneg x0, x8, lt
 ; CHECK-NEXT:    ret
   %min = call i64 @llvm.smin.i64(i64 %a, i64 %b)
   %max = call i64 @llvm.smax.i64(i64 %a, i64 %b)
@@ -307,7 +306,7 @@ define i32 @abd_cmp_i32(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_cmp_i32:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, le
+; CHECK-NEXT:    cneg w0, w8, lt
 ; CHECK-NEXT:    ret
   %cmp = icmp slt i32 %a, %b
   %ab = sub i32 %a, %b
@@ -320,7 +319,7 @@ define i64 @abd_cmp_i64(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_cmp_i64:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, le
+; CHECK-NEXT:    cneg x0, x8, lt
 ; CHECK-NEXT:    ret
   %cmp = icmp sge i64 %a, %b
   %ab = sub i64 %a, %b
@@ -406,7 +405,7 @@ define i32 @abd_subnsw_i32(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_subnsw_i32:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, mi
+; CHECK-NEXT:    cneg w0, w8, lt
 ; CHECK-NEXT:    ret
   %sub = sub nsw i32 %a, %b
   %abs = call i32 @llvm.abs.i32(i32 %sub, i1 false)
@@ -417,7 +416,7 @@ define i32 @abd_subnsw_i32_undef(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_subnsw_i32_undef:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, mi
+; CHECK-NEXT:    cneg w0, w8, lt
 ; CHECK-NEXT:    ret
   %sub = sub nsw i32 %a, %b
   %abs = call i32 @llvm.abs.i32(i32 %sub, i1 true)
@@ -428,7 +427,7 @@ define i64 @abd_subnsw_i64(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_subnsw_i64:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, mi
+; CHECK-NEXT:    cneg x0, x8, lt
 ; CHECK-NEXT:    ret
   %sub = sub nsw i64 %a, %b
   %abs = call i64 @llvm.abs.i64(i64 %sub, i1 false)
@@ -439,7 +438,7 @@ define i64 @abd_subnsw_i64_undef(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_subnsw_i64_undef:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, mi
+; CHECK-NEXT:    cneg x0, x8, lt
 ; CHECK-NEXT:    ret
   %sub = sub nsw i64 %a, %b
   %abs = call i64 @llvm.abs.i64(i64 %sub, i1 true)
@@ -546,7 +545,7 @@ define i32 @abd_select_i32(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_select_i32:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, le
+; CHECK-NEXT:    cneg w0, w8, lt
 ; CHECK-NEXT:    ret
   %cmp = icmp sgt i32 %a, %b
   %ab = select i1 %cmp, i32 %a, i32 %b
@@ -559,7 +558,7 @@ define i64 @abd_select_i64(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_select_i64:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, le
+; CHECK-NEXT:    cneg x0, x8, lt
 ; CHECK-NEXT:    ret
   %cmp = icmp sge i64 %a, %b
   %ab = select i1 %cmp, i64 %a, i64 %b
diff --git a/llvm/test/CodeGen/AArch64/abdu-neg.ll b/llvm/test/CodeGen/AArch64/abdu-neg.ll
index 79fc12ea76f63..8da576121d881 100644
--- a/llvm/test/CodeGen/AArch64/abdu-neg.ll
+++ b/llvm/test/CodeGen/AArch64/abdu-neg.ll
@@ -72,9 +72,8 @@ define i16 @abd_ext_i16(i16 %a, i16 %b) nounwind {
 define i16 @abd_ext_i16_i32(i16 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i16_i32:
 ; CHECK:       // %bb.0:
-; CHECK-NEXT:    and w8, w0, #0xffff
-; CHECK-NEXT:    subs w8, w8, w1
-; CHECK-NEXT:    cneg w0, w8, hi
+; CHECK-NEXT:    subs w8, w1, w0, uxth
+; CHECK-NEXT:    cneg w0, w8, hs
 ; CHECK-NEXT:    ret
   %aext = zext i16 %a to i64
   %bext = zext i32 %b to i64
@@ -105,7 +104,7 @@ define i32 @abd_ext_i32(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i32:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, hi
+; CHECK-NEXT:    cneg w0, w8, hs
 ; CHECK-NEXT:    ret
   %aext = zext i32 %a to i64
   %bext = zext i32 %b to i64
@@ -120,7 +119,7 @@ define i32 @abd_ext_i32_i16(i32 %a, i16 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i32_i16:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1, uxth
-; CHECK-NEXT:    cneg w0, w8, hi
+; CHECK-NEXT:    cneg w0, w8, hs
 ; CHECK-NEXT:    ret
   %aext = zext i32 %a to i64
   %bext = zext i16 %b to i64
@@ -135,7 +134,7 @@ define i32 @abd_ext_i32_undef(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i32_undef:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, hi
+; CHECK-NEXT:    cneg w0, w8, hs
 ; CHECK-NEXT:    ret
   %aext = zext i32 %a to i64
   %bext = zext i32 %b to i64
@@ -150,7 +149,7 @@ define i64 @abd_ext_i64(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i64:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, hi
+; CHECK-NEXT:    cneg x0, x8, hs
 ; CHECK-NEXT:    ret
   %aext = zext i64 %a to i128
   %bext = zext i64 %b to i128
@@ -165,7 +164,7 @@ define i64 @abd_ext_i64_undef(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i64_undef:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, hi
+; CHECK-NEXT:    cneg x0, x8, hs
 ; CHECK-NEXT:    ret
   %aext = zext i64 %a to i128
   %bext = zext i64 %b to i128
@@ -229,12 +228,9 @@ define i128 @abd_ext_i128_undef(i128 %a, i128 %b) nounwind {
 define i8 @abd_minmax_i8(i8 %a, i8 %b) nounwind {
 ; CHECK-LABEL: abd_minmax_i8:
 ; CHECK:       // %bb.0:
-; CHECK-NEXT:    and w8, w1, #0xff
-; CHECK-NEXT:    and w9, w0, #0xff
-; CHECK-NEXT:    cmp w9, w8
-; CHECK-NEXT:    csel w10, w9, w8, lo
-; CHECK-NEXT:    csel w8, w9, w8, hi
-; CHECK-NEXT:    sub w0, w10, w8
+; CHECK-NEXT:    and w8, w0, #0xff
+; CHECK-NEXT:    subs w8, w8, w1, uxtb
+; CHECK-NEXT:    cneg w0, w8, pl
 ; CHECK-NEXT:    ret
   %min = call i8 @llvm.umin.i8(i8 %a, i8 %b)
   %max = call i8 @llvm.umax.i8(i8 %a, i8 %b)
@@ -245,12 +241,9 @@ define i8 @abd_minmax_i8(i8 %a, i8 %b) nounwind {
 define i16 @abd_minmax_i16(i16 %a, i16 %b) nounwind {
 ; CHECK-LABEL: abd_minmax_i16:
 ; CHECK:       // %bb.0:
-; CHECK-NEXT:    and w8, w1, #0xffff
-; CHECK-NEXT:    and w9, w0, #0xffff
-; CHECK-NEXT:    cmp w9, w8
-; CHECK-NEXT:    csel w10, w9, w8, lo
-; CHECK-NEXT:    csel w8, w9, w8, hi
-; CHECK-NEXT:    sub w0, w10, w8
+; CHECK-NEXT:    and w8, w0, #0xffff
+; CHECK-NEXT:    subs w8, w8, w1, uxth
+; CHECK-NEXT:    cneg w0, w8, pl
 ; CHECK-NEXT:    ret
   %min = call i16 @llvm.umin.i16(i16 %a, i16 %b)
   %max = call i16 @llvm.umax.i16(i16 %a, i16 %b)
@@ -261,10 +254,8 @@ define i16 @abd_minmax_i16(i16 %a, i16 %b) nounwind {
 define i32 @abd_minmax_i32(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_minmax_i32:
 ; CHECK:       // %bb.0:
-; CHECK-NEXT:    cmp w0, w1
-; CHECK-NEXT:    csel w8, w0, w1, lo
-; CHECK-NEXT:    csel w9, w0, w1, hi
-; CHECK-NEXT:    sub w0, w8, w9
+; CHECK-NEXT:    subs w8, w0, w1
+; CHECK-NEXT:    cneg w0, w8, hs
 ; CHECK-NEXT:    ret
   %min = call i32 @llvm.umin.i32(i32 %a, i32 %b)
   %max = call i32 @llvm.umax.i32(i32 %a, i32 %b)
@@ -275,10 +266,8 @@ define i32 @abd_minmax_i32(i32 %a, i32 %b) nounwind {
 define i64 @abd_minmax_i64(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_minmax_i64:
 ; CHECK:       // %bb.0:
-; CHECK-NEXT:    cmp x0, x1
-; CHECK-NEXT:    csel x8, x0, x1, lo
-; CHECK-NEXT:    csel x9, x0, x1, hi
-; CHECK-NEXT:    sub x0, x8, x9
+; CHECK-NEXT:    subs x8, x0, x1
+; CHECK-NEXT:    cneg x0, x8, hs
 ; CHECK-NEXT:    ret
   %min = call i64 @llvm.umin.i64(i64 %a, i64 %b)
   %max = call i64 @llvm.umax.i64(i64 %a, i64 %b)
diff --git a/llvm/test/CodeGen/AArch64/abdu.ll b/llvm/test/CodeGen/AArch64/abdu.ll
index 6db7693fb3a1c..e7e38b6cd5d99 100644
--- a/llvm/test/CodeGen/AArch64/abdu.ll
+++ b/llvm/test/CodeGen/AArch64/abdu.ll
@@ -68,9 +68,8 @@ define i16 @abd_ext_i16(i16 %a, i16 %b) nounwind {
 define i16 @abd_ext_i16_i32(i16 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i16_i32:
 ; CHECK:       // %bb.0:
-; CHECK-NEXT:    and w8, w0, #0xffff
-; CHECK-NEXT:    subs w8, w8, w1
-; CHECK-NEXT:    cneg w0, w8, ls
+; CHECK-NEXT:    subs w8, w1, w0, uxth
+; CHECK-NEXT:    cneg w0, w8, lo
 ; CHECK-NEXT:    ret
   %aext = zext i16 %a to i64
   %bext = zext i32 %b to i64
@@ -99,7 +98,7 @@ define i32 @abd_ext_i32(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i32:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, ls
+; CHECK-NEXT:    cneg w0, w8, lo
 ; CHECK-NEXT:    ret
   %aext = zext i32 %a to i64
   %bext = zext i32 %b to i64
@@ -113,7 +112,7 @@ define i32 @abd_ext_i32_i16(i32 %a, i16 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i32_i16:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1, uxth
-; CHECK-NEXT:    cneg w0, w8, ls
+; CHECK-NEXT:    cneg w0, w8, lo
 ; CHECK-NEXT:    ret
   %aext = zext i32 %a to i64
   %bext = zext i16 %b to i64
@@ -127,7 +126,7 @@ define i32 @abd_ext_i32_undef(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i32_undef:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, ls
+; CHECK-NEXT:    cneg w0, w8, lo
 ; CHECK-NEXT:    ret
   %aext = zext i32 %a to i64
   %bext = zext i32 %b to i64
@@ -141,7 +140,7 @@ define i64 @abd_ext_i64(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i64:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, ls
+; CHECK-NEXT:    cneg x0, x8, lo
 ; CHECK-NEXT:    ret
   %aext = zext i64 %a to i128
   %bext = zext i64 %b to i128
@@ -155,7 +154,7 @@ define i64 @abd_ext_i64_undef(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_ext_i64_undef:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, ls
+; CHECK-NEXT:    cneg x0, x8, lo
 ; CHECK-NEXT:    ret
   %aext = zext i64 %a to i128
   %bext = zext i64 %b to i128
@@ -239,7 +238,7 @@ define i32 @abd_minmax_i32(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_minmax_i32:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, ls
+; CHECK-NEXT:    cneg w0, w8, lo
 ; CHECK-NEXT:    ret
   %min = call i32 @llvm.umin.i32(i32 %a, i32 %b)
   %max = call i32 @llvm.umax.i32(i32 %a, i32 %b)
@@ -251,7 +250,7 @@ define i64 @abd_minmax_i64(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_minmax_i64:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, ls
+; CHECK-NEXT:    cneg x0, x8, lo
 ; CHECK-NEXT:    ret
   %min = call i64 @llvm.umin.i64(i64 %a, i64 %b)
   %max = call i64 @llvm.umax.i64(i64 %a, i64 %b)
@@ -313,7 +312,7 @@ define i32 @abd_cmp_i32(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_cmp_i32:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, ls
+; CHECK-NEXT:    cneg w0, w8, lo
 ; CHECK-NEXT:    ret
   %cmp = icmp ult i32 %a, %b
   %ab = sub i32 %a, %b
@@ -326,7 +325,7 @@ define i64 @abd_cmp_i64(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_cmp_i64:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, ls
+; CHECK-NEXT:    cneg x0, x8, lo
 ; CHECK-NEXT:    ret
   %cmp = icmp uge i64 %a, %b
   %ab = sub i64 %a, %b
@@ -411,7 +410,7 @@ define i32 @abd_select_i32(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: abd_select_i32:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, ls
+; CHECK-NEXT:    cneg w0, w8, lo
 ; CHECK-NEXT:    ret
   %cmp = icmp ugt i32 %a, %b
   %ab = select i1 %cmp, i32 %a, i32 %b
@@ -424,7 +423,7 @@ define i64 @abd_select_i64(i64 %a, i64 %b) nounwind {
 ; CHECK-LABEL: abd_select_i64:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x8, x0, x1
-; CHECK-NEXT:    cneg x0, x8, ls
+; CHECK-NEXT:    cneg x0, x8, lo
 ; CHECK-NEXT:    ret
   %cmp = icmp uge i64 %a, %b
   %ab = select i1 %cmp, i64 %a, i64 %b
diff --git a/llvm/test/CodeGen/AArch64/abs.ll b/llvm/test/CodeGen/AArch64/abs.ll
index 0f56d25a47b2a..92d8ba242e0de 100644
--- a/llvm/test/CodeGen/AArch64/abs.ll
+++ b/llvm/test/CodeGen/AArch64/abs.ll
@@ -388,3 +388,33 @@ entry:
   ret <3 x i32> %res
 }
 declare <3 x i32> @llvm.abs.v3i32(<3 x i32>, i1)
+
+define i32 @combine_subs_multiple_sub_uses(i32 %a, i32 %b) {
+; CHECK-LABEL: combine_subs_multiple_sub_uses:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    subs w8, w0, w1
+; CHECK-NEXT:    csel w9, w0, w1, ne
+; CHECK-NEXT:    add w0, w9, w8
+; CHECK-NEXT:    ret
+  %sub = sub i32 %a, %b
+  %cc = icmp ne i32 %sub, 0
+  %sel = select i1 %cc, i32 %a, i32 %b
+  %add = add i32 %sel, %sub
+  ret i32 %add
+}
+
+define i32 @do_not_combine_subs_multiple_flag_uses(i32 %a, i32 %b, i32 %c, i32 %d) {
+; CHECK-LABEL: do_not_combine_subs_multiple_flag_uses:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    cmp w0, w1
+; CHECK-NEXT:    csel w8, w0, w1, ne
+; CHECK-NEXT:    csel w9, w2, w3, ne
+; CHECK-NEXT:    add w0, w8, w9
+; CHECK-NEXT:    ret
+  %sub = sub i32 %a, %b
+  %cc = icmp ne i32 %sub, 0
+  %sel = select i1 %cc, i32 %a, i32 %b
+  %other = select i1 %cc, i32 %c, i32 %d
+  %add = add i32 %sel, %other
+  ret i32 %add
+}
diff --git a/llvm/test/CodeGen/AArch64/alias_mask.ll b/llvm/test/CodeGen/AArch64/alias_mask.ll
index 9b9c020016bab..48c66ad5bac1c 100644
--- a/llvm/test/CodeGen/AArch64/alias_mask.ll
+++ b/llvm/test/CodeGen/AArch64/alias_mask.ll
@@ -393,70 +393,70 @@ entry:
 define <32 x i1> @whilewr_32_expand3(ptr %a, ptr %b) {
 ; CHECK-LABEL: whilewr_32_expand3:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    subs x9, x1, x0
+; CHECK-NEXT:    subs x10, x1, x0
 ; CHECK-NEXT:    index z0.d, #0, #1
-; CHECK-NEXT:    add x10, x9, #3
-; CHECK-NEXT:    sub x11, x9, #61
-; CHECK-NEXT:    csel x10, x10, x9, mi
-; CHECK-NEXT:    subs x9, x9, #64
-; CHECK-NEXT:    csel x9, x11, x9, mi
-; CHECK-NEXT:    asr x10, x10, #2
-; CHECK-NEXT:    asr x9, x9, #2
-; CHECK-NEXT:    mov z1.d, z0.d
+; CHECK-NEXT:    add x9, x10, #3
+; CHECK-NEXT:    sub x12, x10, #61
+; CHECK-NEXT:    csel x9, x9, x10, mi
+; CHECK-NEXT:    asr x11, x9, #2
 ; CHECK-NEXT:    mov z2.d, z0.d
 ; CHECK-NEXT:    mov z3.d, z0.d
 ; CHECK-NEXT:    mov z4.d, z0.d
+; CHECK-NEXT:    cmp x11, #1
+; CHECK-NEXT:    dup v1.2d, x11
 ; CHECK-NEXT:    mov z5.d, z0.d
+; CHECK-NEXT:    cset w9, lt
+; CHECK-NEXT:    subs x10, x10, #64
 ; CHECK-NEXT:    mov z6.d, z0.d
-; CHECK-NEXT:    dup v7.2d, x10
-; CHECK-NEXT:    dup v16.2d, x9
-; CHECK-NEXT:    add z1.d, z1.d, #12 // =0xc
-; CHECK-NEXT:    add z2.d, z2.d, #10 // =0xa
-; CHECK-NEXT:    cmp x9, #1
-; CHECK-NEXT:    add z3.d, z3.d, #8 // =0x8
-; CHECK-NEXT:    add z4.d, z4.d, #6 // =0x6
-; CHECK-NEXT:    add z5.d, z5.d, #4 // =0x4
-; CHECK-NEXT:    add z6.d, z6.d, #2 // =0x2
-; CHECK-NEXT:    cmhi v17.2d, v7.2d, v0.2d
+; CHECK-NEXT:    csel x10, x12, x10, mi
+; CHECK-NEXT:    mov z7.d, z0.d
+; CHECK-NEXT:    add z2.d, z2.d, #12 // =0xc
+; CHECK-NEXT:    asr x10, x10, #2
+; CHECK-NEXT:    add z3.d, z3.d, #10 // =0xa
+; CHECK-NEXT:    add z4.d, z4.d, #8 // =0x8
+; CHECK-NEXT:    add z5.d, z5.d, #6 // =0x6
+; CHECK-NEXT:    add z6.d, z6.d, #4 // =0x4
+; CHECK-NEXT:    cmhi v17.2d, v1.2d, v0.2d
+; CHECK-NEXT:    dup v16.2d, x10
+; CHECK-NEXT:    add z7.d, z7.d, #2 // =0x2
+; CHECK-NEXT:    cmhi v19.2d, v1.2d, v2.2d
+; CHECK-NEXT:    cmhi v20.2d, v1.2d, v3.2d
+; CHECK-NEXT:    cmhi v21.2d, v1.2d, v4.2d
+; CHECK-NEXT:    cmp x10, #1
+; CHECK-NEXT:    cmhi v22.2d, v1.2d, v5.2d
+; CHECK-NEXT:    cset w10, lt
 ; CHECK-NEXT:    cmhi v18.2d, v16.2d, v0.2d
 ; CHECK-NEXT:    add z0.d, z0.d, #14 // =0xe
-; CHECK-NEXT:    cmhi v19.2d, v7.2d, v1.2d
-; CHECK-NEXT:    cmhi v20.2d, v7.2d, v2.2d
-; CHECK-NEXT:    cmhi v21.2d, v7.2d, v3.2d
-; CHECK-NEXT:    cmhi v22.2d, v7.2d, v4.2d
-; CHECK-NEXT:    cmhi v23.2d, v7.2d, v5.2d
-; CHECK-NEXT:    cmhi v24.2d, v7.2d, v6.2d
-; CHECK-NEXT:    cmhi v1.2d, v16.2d, v1.2d
 ; CHECK-NEXT:    cmhi v2.2d, v16.2d, v2.2d
 ; CHECK-NEXT:    cmhi v3.2d, v16.2d, v3.2d
 ; CHECK-NEXT:    cmhi v4.2d, v16.2d, v4.2d
-; CHECK-NEXT:    cmhi v7.2d, v7.2d, v0.2d
 ; CHECK-NEXT:    cmhi v5.2d, v16.2d, v5.2d
-; CHECK-NEXT:    cmhi v6.2d, v16.2d, v6.2d
-; CHECK-NEXT:    cset w9, lt
-; CHECK-NEXT:    cmhi v0.2d, v16.2d, v0.2d
-; CHECK-NEXT:    uzp1 v16.4s, v21.4s, v20.4s
-; CHECK-NEXT:    cmp x10, #1
-; CHECK-NEXT:    uzp1 v20.4s, v23.4s, v22.4s
-; CHECK-NEXT:    uzp1 v17.4s, v17.4s, v24.4s
-; CHECK-NEXT:    cset w10, lt
-; CHECK-NEXT:    uzp1 v2.4s, v3.4s, v2.4s
-; CHECK-NEXT:    uzp1 v3.4s, v19.4s, v7.4s
-; CHECK-NEXT:    uzp1 v4.4s, v5.4s, v4.4s
-; CHECK-NEXT:    uzp1 v5.4s, v18.4s, v6.4s
-; CHECK-NEXT:    uzp1 v0.4s, v1.4s, v0.4s
-; CHECK-NEXT:    uzp1 v1.8h, v17.8h, v20.8h
-; CHECK-NEXT:    uzp1 v3.8h, v16.8h, v3.8h
+; CHECK-NEXT:    cmhi v23.2d, v16.2d, v6.2d
+; CHECK-NEXT:    cmhi v24.2d, v16.2d, v7.2d
+; CHECK-NEXT:    cmhi v6.2d, v1.2d, v6.2d
+; CHECK-NEXT:    cmhi v16.2d, v16.2d, v0.2d
+; CHECK-NEXT:    cmhi v7.2d, v1.2d, v7.2d
+; CHECK-NEXT:    cmhi v0.2d, v1.2d, v0.2d
+; CHECK-NEXT:    uzp1 v1.4s, v21.4s, v20.4s
+; CHECK-NEXT:    uzp1 v3.4s, v4.4s, v3.4s
+; CHECK-NEXT:    uzp1 v4.4s, v23.4s, v5.4s
+; CHECK-NEXT:    uzp1 v5.4s, v18.4s, v24.4s
+; CHECK-NEXT:    uzp1 v6.4s, v6.4s, v22.4s
+; CHECK-NEXT:    uzp1 v2.4s, v2.4s, v16.4s
+; CHECK-NEXT:    uzp1 v7.4s, v17.4s, v7.4s
+; CHECK-NEXT:    uzp1 v0.4s, v19.4s, v0.4s
 ; CHECK-NEXT:    uzp1 v4.8h, v5.8h, v4.8h
-; CHECK-NEXT:    uzp1 v0.8h, v2.8h, v0.8h
+; CHECK-NEXT:    uzp1 v2.8h, v3.8h, v2.8h
+; CHECK-NEXT:    uzp1 v3.8h, v7.8h, v6.8h
+; CHECK-NEXT:    uzp1 v0.8h, v1.8h, v0.8h
+; CHECK-NEXT:    uzp1 v1.16b, v4.16b, v2.16b
+; CHECK-NEXT:    uzp1 v0.16b, v3.16b, v0.16b
 ; CHECK-NEXT:    dup v2.16b, w10
-; CHECK-NEXT:    uzp1 v1.16b, v1.16b, v3.16b
 ; CHECK-NEXT:    dup v3.16b, w9
 ; CHECK-NEXT:    adrp x9, .LCPI14_0
-; CHECK-NEXT:    uzp1 v0.16b, v4.16b, v0.16b
 ; CHECK-NEXT:    orr v1.16b, v1.16b, v2.16b
-; CHECK-NEXT:    ldr q2, [x9, :lo12:.LCPI14_0]
 ; CHECK-NEXT:    orr v0.16b, v0.16b, v3.16b
+; CHECK-NEXT:    ldr q2, [x9, :lo12:.LCPI14_0]
 ; CHECK-NEXT:    shl v1.16b, v1.16b, #7
 ; CHECK-NEXT:    shl v0.16b, v0.16b, #7
 ; CHECK-NEXT:    cmlt v1.16b, v1.16b, #0
@@ -469,8 +469,8 @@ define <32 x i1> @whilewr_32_expand3(ptr %a, ptr %b) {
 ; CHECK-NEXT:    zip1 v0.16b, v0.16b, v3.16b
 ; CHECK-NEXT:    addv h1, v1.8h
 ; CHECK-NEXT:    addv h0, v0.8h
-; CHECK-NEXT:    str h1, [x8]
-; CHECK-NEXT:    str h0, [x8, #2]
+; CHECK-NEXT:    str h1, [x8, #2]
+; CHECK-NEXT:    str h0, [x8]
 ; CHECK-NEXT:    ret
 entry:
   %0 = call <32 x i1> @llvm.loop.dependence.war.mask.v32i1(ptr %a, ptr %b, i64 4)
@@ -586,70 +586,70 @@ entry:
 define <32 x i1> @whilewr_64_expand4(ptr %a, ptr %b) {
 ; CHECK-LABEL: whilewr_64_expand4:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    subs x9, x1, x0
+; CHECK-NEXT:    subs x10, x1, x0
 ; CHECK-NEXT:    index z0.d, #0, #1
-; CHECK-NEXT:    add x10, x9, #7
-; CHECK-NEXT:    sub x11, x9, #121
-; CHECK-NEXT:    csel x10, x10, x9, mi
-; CHECK-NEXT:    subs x9, x9, #128
-; CHECK-NEXT:    csel x9, x11, x9, mi
-; CHECK-NEXT:    asr x10, x10, #3
-; CHECK-NEXT:    asr x9, x9, #3
-; CHECK-NEXT:    mov z1.d, z0.d
+; CHECK-NEXT:    add x9, x10, #7
+; CHECK-NEXT:    sub x12, x10, #121
+; CHECK-NEXT:    csel x9, x9, x10, mi
+; CHECK-NEXT:    asr x11, x9, #3
 ; CHECK-NEXT:    mov z2.d, z0.d
 ; CHECK-NEXT:    mov z3.d, z0.d
 ; CHECK-NEXT:    mov z4.d, z0.d
+; CHECK-NEXT:    cmp x11, #1
+; CHECK-NEXT:    dup v1.2d, x11
 ; CHECK-NEXT:    mov z5.d, z0.d
+; CHECK-NEXT:    cset w9, lt
+; CHECK-NEXT:    subs x10, x10, #128
 ; CHECK-NEXT:    mov z6.d, z0.d
-; CHECK-NEXT:    dup v7.2d, x10
-; CHECK-NEXT:    dup v16.2d, x9
-; CHECK-NEXT:    add z1.d, z1.d, #12 // =0xc
-; CHECK-NEXT:    add z2.d, z2.d, #10 // =0xa
-; CHECK-NEXT:    cmp x9, #1
-; CHECK-NEXT:    add z3.d, z3.d, #8 // =0x8
-; CHECK-NEXT:    add z4.d, z4.d, #6 // =0x6
-; CHECK-NEXT:    add z5.d, z5.d, #4 // =0x4
-; CHECK-NEXT:    add z6.d, z6.d, #2 // =0x2
-; CHECK-NEXT:    cmhi v17.2d, v7.2d, v0.2d
+; CHECK-NEXT:    csel x10, x12, x10, mi
+; CHECK-NEXT:    mov z7.d, z0.d
+; CHECK-NEXT:    add z2.d, z2.d, #12 // =0xc
+; CHECK-NEXT:    asr x10, x10, #3
+; CHECK-NEXT:    add z3.d, z3.d, #10 // =0xa
+; CHECK-NEXT:    add z4.d, z4.d, #8 // =0x8
+; CHECK-NEXT:    add z5.d, z5.d, #6 // =0x6
+; CHECK-NEXT:    add z6.d, z6.d, #4 // =0x4
+; CHECK-NEXT:    cmhi v17.2d, v1.2d, v0.2d
+; CHECK-NEXT:    dup v16.2d, x10
+; CHECK-NEXT:    add z7.d, z7.d, #2 // =0x2
+; CHECK-NEXT:    cmhi v19.2d, v1.2d, v2.2d
+; CHECK-NEXT:    cmhi v20.2d, v1.2d, v3.2d
+; CHECK-NEXT:    cmhi v21.2d, v1.2d, v4.2d
+; CHECK-NEXT:    cmp x10, #1
+; CHECK-NEXT:    cmhi v22.2d, v1.2d, v5.2d
+; CHECK-NEXT:    cset w10, lt
 ; CHECK-NEXT:    cmhi v18.2d, v16.2d, v0.2d
 ; CHECK-NEXT:    add z0.d, z0.d, #14 // =0xe
-; CHECK-NEXT:    cmhi v19.2d, v7.2d, v1.2d
-; CHECK-NEXT:    cmhi v20.2d, v7.2d, v2.2d
-; CHECK-NEXT:    cmhi v21.2d, v7.2d, v3.2d
-; CHECK-NEXT:    cmhi v22.2d, v7.2d, v4.2d
-; CHECK-NEXT:    cmhi v23.2d, v7.2d, v5.2d
-; CHECK-NEXT:    cmhi v24.2d, v7.2d, v6.2d
-; CHECK-NEXT:    cmhi v1.2d, v16.2d, v1.2d
 ; CHECK-NEXT:    cmhi v2.2d, v16.2d, v2.2d
 ; CHECK-NEXT:    cmhi v3.2d, v16.2d, v3.2d
 ; CHECK-NEXT:    cmhi v4.2d, v16.2d, v4.2d
-; CHECK-NEXT:    cmhi v7.2d, v7.2d, v0.2d
 ; CHECK-NEXT:    cmhi v5.2d, v16.2d, v5.2d
-; CHECK-NEXT:    cmhi v6.2d, v16.2d, v6.2d
-; CHECK-NEXT:    cset w9, lt
-; CHECK-NEXT:    cmhi v0.2d, v16.2d, v0.2d
-; CHECK-NEXT:    uzp1 v16.4s, v21.4s, v20.4s
-; CHECK-NEXT:    cmp x10, #1
-; CHECK-NEXT:    uzp1 v20.4s, v23.4s, v22.4s
-; CHECK-NEXT:    uzp1 v17.4s, v17.4s, v24.4s
-; CHECK-NEXT:    cset w10, lt
-; CHECK-NEXT:    uzp1 v2.4s, v3.4s, v2.4s
-; CHECK-NEXT:    uzp1 v3.4s, v19.4s, v7.4s
-; CHECK-NEXT:    uzp1 v4.4s, v5.4s, v4.4s
-; CHECK-NEXT:    uzp1 v5.4s, v18.4s, v6.4s
-; CHECK-NEXT:    uzp1 v0.4s, v1.4s, v0.4s
-; CHECK-NEXT:    uzp1 v1.8h, v17.8h, v20.8h
-; CHECK-NEXT:    uzp1 v3.8h, v16.8h, v3.8h
+; CHECK-NEXT:    cmhi v23.2d, v16.2d, v6.2d
+; CHECK-NEXT:    cmhi v24.2d, v16.2d, v7.2d
+; CHECK-NEXT:    cmhi v6.2d, v1.2d, v6.2d
+; CHECK-NEXT:    cmhi v16.2d, v16.2d, v0.2d
+; CHECK-NEXT:    cmhi v7.2d, v1.2d, v7.2d
+; CHECK-NEXT:    cmhi v0.2d, v1.2d, v0.2d
+; CHECK-NEXT:    uzp1 v1.4s, v21.4s, v20.4s
+; CHECK-NEXT:    uzp1 v3.4s, v4.4s, v3.4s
+; CHECK-NEXT:    uzp1 v4.4s, v23.4s, v5.4s
+; CHECK-NEXT:    uzp1 v5.4s, v18.4s, v24.4s
+; CHECK-NEXT:    uzp1 v6.4s, v6.4s, v22.4s
+; CHECK-NEXT:    uzp1 v2.4s, v2.4s, v16.4s
+; CHECK-NEXT:    uzp1 v7.4s, v17.4s, v7.4s
+; CHECK-NEXT:    uzp1 v0.4s, v19.4s, v0.4s
 ; CHECK-NEXT:    uzp1 v4.8h, v5.8h, v4.8h
-; CHECK-NEXT:    uzp1 v0.8h, v2.8h, v0.8h
+; CHECK-NEXT:    uzp1 v2.8h, v3.8h, v2.8h
+; CHECK-NEXT:    uzp1 v3.8h, v7.8h, v6.8h
+; CHECK-NEXT:    uzp1 v0.8h, v1.8h, v0.8h
+; CHECK-NEXT:    uzp1 v1.16b, v4.16b, v2.16b
+; CHECK-NEXT:    uzp1 v0.16b, v3.16b, v0.16b
 ; CHECK-NEXT:    dup v2.16b, w10
-; CHECK-NEXT:    uzp1 v1.16b, v1.16b, v3.16b
 ; CHECK-NEXT:    dup v3.16b, w9
 ; CHECK-NEXT:    adrp x9, .LCPI18_0
-; CHECK-NEXT:    uzp1 v0.16b, v4.16b, v0.16b
 ; CHECK-NEXT:    orr v1.16b, v1.16b, v2.16b
-; CHECK-NEXT:    ldr q2, [x9, :lo12:.LCPI18_0]
 ; CHECK-NEXT:    orr v0.16b, v0.16b, v3.16b
+; CHECK-NEXT:    ldr q2, [x9, :lo12:.LCPI18_0]
 ; CHECK-NEXT:    shl v1.16b, v1.16b, #7
 ; CHECK-NEXT:    shl v0.16b, v0.16b, #7
 ; CHECK-NEXT:    cmlt v1.16b, v1.16b, #0
@@ -662,8 +662,8 @@ define <32 x i1> @whilewr_64_expand4(ptr %a, ptr %b) {
 ; CHECK-NEXT:    zip1 v0.16b, v0.16b, v3.16b
 ; CHECK-NEXT:    addv h1, v1.8h
 ; CHECK-NEXT:    addv h0, v0.8h
-; CHECK-NEXT:    str h1, [x8]
-; CHECK-NEXT:    str h0, [x8, #2]
+; CHECK-NEXT:    str h1, [x8, #2]
+; CHECK-NEXT:    str h0, [x8]
 ; CHECK-NEXT:    ret
 entry:
   %0 = call <32 x i1> @llvm.loop.dependence.war.mask.v32i1(ptr %a, ptr %b, i64 8)
@@ -790,11 +790,10 @@ entry:
 define <1 x i1> @whilewr_8_scalarize(ptr %a, ptr %b) {
 ; CHECK-LABEL: whilewr_8_scalarize:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    subs x8, x1, x0
+; CHECK-NEXT:    sub x8, x1, x0
 ; CHECK-NEXT:    cmp x8, #0
-; CHECK-NEXT:    cset w8, gt
-; CHECK-NEXT:    cmp x1, x0
-; CHECK-NEXT:    csinc w0, w8, wzr, ne
+; CHECK-NEXT:    ccmp x8, #0, #4, le
+; CHECK-NEXT:    cset w0, eq
 ; CHECK-NEXT:    ret
 entry:
   %0 = call <1 x i1> @llvm.loop.dependence.war.mask.v1i1(ptr %a, ptr %b, i64 1)
@@ -804,11 +803,10 @@ entry:
 define <1 x i1> @whilewr_16_scalarize(ptr %a, ptr %b) {
 ; CHECK-LABEL: whilewr_16_scalarize:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    subs x8, x1, x0
+; CHECK-NEXT:    sub x8, x1, x0
 ; CHECK-NEXT:    cmp x8, #1
-; CHECK-NEXT:    cset w8, gt
-; CHECK-NEXT:    cmp x1, x0
-; CHECK-NEXT:    csinc w0, w8, wzr, ne
+; CHECK-NEXT:    ccmp x8, #0, #4, le
+; CHECK-NEXT:    cset w0, eq
 ; CHECK-NEXT:    ret
 entry:
   %0 = call <1 x i1> @llvm.loop.dependence.war.mask.v1i1(ptr %a, ptr %b, i64 2)
@@ -818,11 +816,10 @@ entry:
 define <1 x i1> @whilewr_32_scalarize(ptr %a, ptr %b) {
 ; CHECK-LABEL: whilewr_32_scalarize:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    subs x8, x1, x0
+; CHECK-NEXT:    sub x8, x1, x0
 ; CHECK-NEXT:    cmp x8, #3
-; CHECK-NEXT:    cset w8, gt
-; CHECK-NEXT:    cmp x1, x0
-; CHECK-NEXT:    csinc w0, w8, wzr, ne
+; CHECK-NEXT:    ccmp x8, #0, #4, le
+; CHECK-NEXT:    cset w0, eq
 ; CHECK-NEXT:    ret
 entry:
   %0 = call <1 x i1> @llvm.loop.dependence.war.mask.v1i1(ptr %a, ptr %b, i64 4)
@@ -832,11 +829,10 @@ entry:
 define <1 x i1> @whilewr_64_scalarize(ptr %a, ptr %b) {
 ; CHECK-LABEL: whilewr_64_scalarize:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    subs x8, x1, x0
+; CHECK-NEXT:    sub x8, x1, x0
 ; CHECK-NEXT:    cmp x8, #7
-; CHECK-NEXT:    cset w8, gt
-; CHECK-NEXT:    cmp x1, x0
-; CHECK-NEXT:    csinc w0, w8, wzr, ne
+; CHECK-NEXT:    ccmp x8, #0, #4, le
+; CHECK-NEXT:    cset w0, eq
 ; CHECK-NEXT:    ret
 entry:
   %0 = call <1 x i1> @llvm.loop.dependence.war.mask.v1i1(ptr %a, ptr %b, i64 8)
@@ -846,11 +842,10 @@ entry:
 define <1 x i1> @whilerw_8_scalarize(ptr %a, ptr %b) {
 ; CHECK-LABEL: whilerw_8_scalarize:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    subs x8, x1, x0
+; CHECK-NEXT:    sub x8, x1, x0
 ; CHECK-NEXT:    cmp x8, #0
-; CHECK-NEXT:    cset w8, gt
-; CHECK-NEXT:    cmp x1, x0
-; CHECK-NEXT:    csinc w0, w8, wzr, ne
+; CHECK-NEXT:    ccmp x8, #0, #4, le
+; CHECK-NEXT:    cset w0, eq
 ; CHECK-NEXT:    ret
 entry:
   %0 = call <1 x i1> @llvm.loop.dependence.raw.mask.v1i1(ptr %a, ptr %b, i64 1)
@@ -860,11 +855,10 @@ entry:
 define <1 x i1> @whilerw_16_scalarize(ptr %a, ptr %b) {
 ; CHECK-LABEL: whilerw_16_scalarize:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    subs x8, x1, x0
+; CHECK-NEXT:    sub x8, x1, x0
 ; CHECK-NEXT:    cmp x8, #1
-; CHECK-NEXT:    cset w8, gt
-; CHECK-NEXT:    cmp x1, x0
-; CHECK-NEXT:    csinc w0, w8, wzr, ne
+; CHECK-NEXT:    ccmp x8, #0, #4, le
+; CHECK-NEXT:    cset w0, eq
 ; CHECK-NEXT:    ret
 entry:
   %0 = call <1 x i1> @llvm.loop.dependence.raw.mask.v1i1(ptr %a, ptr %b, i64 2)
@@ -874,11 +868,10 @@ entry:
 define <1 x i1> @whilerw_32_scalarize(ptr %a, ptr %b) {
 ; CHECK-LABEL: whilerw_32_scalarize:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    subs x8, x1, x0
+; CHECK-NEXT:    sub x8, x1, x0
 ; CHECK-NEXT:    cmp x8, #3
-; CHECK-NEXT:    cset w8, gt
-; CHECK-NEXT:    cmp x1, x0
-; CHECK-NEXT:    csinc w0, w8, wzr, ne
+; CHECK-NEXT:    ccmp x8, #0, #4, le
+; CHECK-NEXT:    cset w0, eq
 ; CHECK-NEXT:    ret
 entry:
   %0 = call <1 x i1> @llvm.loop.dependence.raw.mask.v1i1(ptr %a, ptr %b, i64 4)
@@ -888,11 +881,10 @@ entry:
 define <1 x i1> @whilerw_64_scalarize(ptr %a, ptr %b) {
 ; CHECK-LABEL: whilerw_64_scalarize:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    subs x8, x1, x0
+; CHECK-NEXT:    sub x8, x1, x0
 ; CHECK-NEXT:    cmp x8, #7
-; CHECK-NEXT:    cset w8, gt
-; CHECK-NEXT:    cmp x1, x0
-; CHECK-NEXT:    csinc w0, w8, wzr, ne
+; CHECK-NEXT:    ccmp x8, #0, #4, le
+; CHECK-NEXT:    cset w0, eq
 ; CHECK-NEXT:    ret
 entry:
   %0 = call <1 x i1> @llvm.loop.dependence.raw.mask.v1i1(ptr %a, ptr %b, i64 8)
diff --git a/llvm/test/CodeGen/AArch64/arm64-csel.ll b/llvm/test/CodeGen/AArch64/arm64-csel.ll
index 69fad57a683ac..584e61b734527 100644
--- a/llvm/test/CodeGen/AArch64/arm64-csel.ll
+++ b/llvm/test/CodeGen/AArch64/arm64-csel.ll
@@ -65,7 +65,7 @@ define i32@foo5(i32 %a, i32 %b) nounwind ssp {
 ; CHECK-LABEL: foo5:
 ; CHECK:       // %bb.0: // %entry
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w0, w8, mi
+; CHECK-NEXT:    cneg w0, w8, lt
 ; CHECK-NEXT:    ret
 entry:
   %sub = sub nsw i32 %a, %b
@@ -98,7 +98,7 @@ define i32 @foo7(i32 %a, i32 %b) nounwind {
 ; CHECK-LABEL: foo7:
 ; CHECK:       // %bb.0: // %entry
 ; CHECK-NEXT:    subs w8, w0, w1
-; CHECK-NEXT:    cneg w9, w8, mi
+; CHECK-NEXT:    cneg w9, w8, lt
 ; CHECK-NEXT:    cmn w8, #1
 ; CHECK-NEXT:    csel w8, w9, w0, lt
 ; CHECK-NEXT:    csel w0, w8, w9, gt
diff --git a/llvm/test/CodeGen/AArch64/csel-subs-dag-combine.ll b/llvm/test/CodeGen/AArch64/csel-subs-dag-combine.ll
deleted file mode 100644
index 5036be9c45e69..0000000000000
--- a/llvm/test/CodeGen/AArch64/csel-subs-dag-combine.ll
+++ /dev/null
@@ -1,112 +0,0 @@
-; RUN: llc -debug-only=isel -o /dev/null < %s 2>&1 | FileCheck %s
-
-; REQUIRES: asserts
-
-; These tests ensure that we don't combine
-;   CSEL a, b, cc, SUBS(SUB(x,y), 0) -> CSEL a, b, cc, SUBS(x,y)
-; if the flags set by SUBS(SUB(x,y), 0) have more than one use.
-;
-; This restriction exists because combining SUBS(SUB(x,y), 0) -> SUBS(x,y) is
-; only valid if there are no users of the overflow flags (C/V) generated by the
-; SUBS. Currently, we only check the flags used by the CSEL, and therefore we
-; conservatively reject cases where the SUBS's flags have other uses.
-
-target triple = "aarch64-unknown-linux-gnu"
-
-; CHECK-LABEL: Legalized selection DAG: %bb.0 'combine_subs:'
-; CHECK-NEXT: SelectionDAG has 13 nodes:
-; CHECK-NEXT:   t0: ch,glue = EntryToken
-; CHECK-NEXT:   t2: i32,ch = CopyFromReg t0, Register:i32 %0
-; CHECK-NEXT:   t4: i32,ch = CopyFromReg t0, Register:i32 %1
-; CHECK-NEXT:         t5: i32 = sub t2, t4
-; CHECK-NEXT:       t14: i32,i32 = AArch64ISD::SUBS t5, Constant:i32<0>
-; CHECK-NEXT:     t16: i32 = AArch64ISD::CSEL t2, t4, Constant:i32<1>, t14:1
-; CHECK-NEXT:   t11: ch,glue = CopyToReg t0, Register:i32 $w0, t16
-; CHECK-NEXT:   t12: ch = AArch64ISD::RET_GLUE t11, Register:i32 $w0, t11:1
-
-; CHECK-LABEL: Optimized legalized selection DAG: %bb.0 'combine_subs:'
-; CHECK-NEXT: SelectionDAG has 11 nodes:
-; CHECK-NEXT:   t0: ch,glue = EntryToken
-; CHECK-NEXT:   t2: i32,ch = CopyFromReg t0, Register:i32 %0
-; CHECK-NEXT:   t4: i32,ch = CopyFromReg t0, Register:i32 %1
-; CHECK-NEXT:       t18: i32,i32 = AArch64ISD::SUBS t2, t4
-; CHECK-NEXT:     t16: i32 = AArch64ISD::CSEL t2, t4, Constant:i32<1>, t18:1
-; CHECK-NEXT:   t11: ch,glue = CopyToReg t0, Register:i32 $w0, t16
-; CHECK-NEXT:   t12: ch = AArch64ISD::RET_GLUE t11, Register:i32 $w0, t11:1
-
-define i32 @combine_subs(i32 %a, i32 %b) {
-  %sub = sub i32 %a, %b
-  %cc = icmp ne i32 %sub, 0
-  %sel = select i1 %cc, i32 %a, i32 %b
-  ret i32 %sel
-}
-
-; CHECK-LABEL: Legalized selection DAG: %bb.0 'combine_subs_multiple_sub_uses:'
-; CHECK-NEXT: SelectionDAG has 14 nodes:
-; CHECK-NEXT:   t0: ch,glue = EntryToken
-; CHECK-NEXT:   t2: i32,ch = CopyFromReg t0, Register:i32 %0
-; CHECK-NEXT:   t4: i32,ch = CopyFromReg t0, Register:i32 %1
-; CHECK-NEXT:   t5: i32 = sub t2, t4
-; CHECK-NEXT:         t15: i32,i32 = AArch64ISD::SUBS t5, Constant:i32<0>
-; CHECK-NEXT:       t17: i32 = AArch64ISD::CSEL t2, t4, Constant:i32<1>, t15:1
-; CHECK-NEXT:     t10: i32 = add t17, t5
-; CHECK-NEXT:   t12: ch,glue = CopyToReg t0, Register:i32 $w0, t10
-; CHECK-NEXT:   t13: ch = AArch64ISD::RET_GLUE t12, Register:i32 $w0, t12:1
-
-; CHECK-LABEL: Optimized legalized selection DAG: %bb.0 'combine_subs_multiple_sub_uses:'
-; CHECK-NEXT: SelectionDAG has 12 nodes:
-; CHECK-NEXT:   t0: ch,glue = EntryToken
-; CHECK-NEXT:   t2: i32,ch = CopyFromReg t0, Register:i32 %0
-; CHECK-NEXT:   t4: i32,ch = CopyFromReg t0, Register:i32 %1
-; CHECK-NEXT:       t17: i32 = AArch64ISD::CSEL t2, t4, Constant:i32<1>, t19:1
-; CHECK-NEXT:     t10: i32 = add t17, t19
-; CHECK-NEXT:   t12: ch,glue = CopyToReg t0, Register:i32 $w0, t10
-; CHECK-NEXT:   t19: i32,i32 = AArch64ISD::SUBS t2, t4
-; CHECK-NEXT:   t13: ch = AArch64ISD::RET_GLUE t12, Register:i32 $w0, t12:1
-
-define i32 @combine_subs_multiple_sub_uses(i32 %a, i32 %b) {
-  %sub = sub i32 %a, %b
-  %cc = icmp ne i32 %sub, 0
-  %sel = select i1 %cc, i32 %a, i32 %b
-  %add = add i32 %sel, %sub
-  ret i32 %add
-}
-
-; CHECK-LABEL: Legalized selection DAG: %bb.0 'do_not_combine_subs_multiple_flag_uses:'
-; CHECK-NEXT: SelectionDAG has 19 nodes:
-; CHECK-NEXT:   t0: ch,glue = EntryToken
-; CHECK-NEXT:   t2: i32,ch = CopyFromReg t0, Register:i32 %0
-; CHECK-NEXT:   t4: i32,ch = CopyFromReg t0, Register:i32 %1
-; CHECK-NEXT:       t24: i32 = AArch64ISD::CSEL t2, t4, Constant:i32<1>, t21:1
-; CHECK-NEXT:         t6: i32,ch = CopyFromReg t0, Register:i32 %2
-; CHECK-NEXT:         t8: i32,ch = CopyFromReg t0, Register:i32 %3
-; CHECK-NEXT:       t23: i32 = AArch64ISD::CSEL t6, t8, Constant:i32<1>, t21:1
-; CHECK-NEXT:     t15: i32 = add t24, t23
-; CHECK-NEXT:   t17: ch,glue = CopyToReg t0, Register:i32 $w0, t15
-; CHECK-NEXT:     t9: i32 = sub t2, t4
-; CHECK-NEXT:   t21: i32,i32 = AArch64ISD::SUBS t9, Constant:i32<0>
-; CHECK-NEXT:   t18: ch = AArch64ISD::RET_GLUE t17, Register:i32 $w0, t17:1
-
-; CHECK-LABEL: Optimized legalized selection DAG: %bb.0 'do_not_combine_subs_multiple_flag_uses:'
-; CHECK-NEXT: SelectionDAG has 19 nodes:
-; CHECK-NEXT:   t0: ch,glue = EntryToken
-; CHECK-NEXT:   t2: i32,ch = CopyFromReg t0, Register:i32 %0
-; CHECK-NEXT:   t4: i32,ch = CopyFromReg t0, Register:i32 %1
-; CHECK-NEXT:       t24: i32 = AArch64ISD::CSEL t2, t4, Constant:i32<1>, t21:1
-; CHECK-NEXT:         t6: i32,ch = CopyFromReg t0, Register:i32 %2
-; CHECK-NEXT:         t8: i32,ch = CopyFromReg t0, Register:i32 %3
-; CHECK-NEXT:       t23: i32 = AArch64ISD::CSEL t6, t8, Constant:i32<1>, t21:1
-; CHECK-NEXT:     t15: i32 = add t24, t23
-; CHECK-NEXT:   t17: ch,glue = CopyToReg t0, Register:i32 $w0, t15
-; CHECK-NEXT:     t9: i32 = sub t2, t4
-; CHECK-NEXT:   t21: i32,i32 = AArch64ISD::SUBS t9, Constant:i32<0>
-; CHECK-NEXT:   t18: ch = AArch64ISD::RET_GLUE t17, Register:i32 $w0, t17:1
-
-define i32 @do_not_combine_subs_multiple_flag_uses(i32 %a, i32 %b, i32 %c, i32 %d) {
-  %sub = sub i32 %a, %b
-  %cc = icmp ne i32 %sub, 0
-  %sel = select i1 %cc, i32 %a, i32 %b
-  %other = select i1 %cc, i32 %c, i32 %d
-  %add = add i32 %sel, %other
-  ret i32 %add
-}
diff --git a/llvm/test/CodeGen/AArch64/midpoint-int.ll b/llvm/test/CodeGen/AArch64/midpoint-int.ll
index 79bba5363188b..02bbca72b5a97 100644
--- a/llvm/test/CodeGen/AArch64/midpoint-int.ll
+++ b/llvm/test/CodeGen/AArch64/midpoint-int.ll
@@ -15,7 +15,7 @@ define i32 @scalar_i32_signed_reg_reg(i32 %a1, i32 %a2) nounwind {
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w9, w0, w1
 ; CHECK-NEXT:    mov w8, #-1 // =0xffffffff
-; CHECK-NEXT:    cneg w9, w9, le
+; CHECK-NEXT:    cneg w9, w9, lt
 ; CHECK-NEXT:    cneg w8, w8, le
 ; CHECK-NEXT:    lsr w9, w9, #1
 ; CHECK-NEXT:    madd w0, w9, w8, w0
@@ -36,7 +36,7 @@ define i32 @scalar_i32_unsigned_reg_reg(i32 %a1, i32 %a2) nounwind {
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs w9, w0, w1
 ; CHECK-NEXT:    mov w8, #-1 // =0xffffffff
-; CHECK-NEXT:    cneg w9, w9, ls
+; CHECK-NEXT:    cneg w9, w9, lo
 ; CHECK-NEXT:    cneg w8, w8, ls
 ; CHECK-NEXT:    lsr w9, w9, #1
 ; CHECK-NEXT:    madd w0, w9, w8, w0
@@ -60,7 +60,7 @@ define i32 @scalar_i32_signed_mem_reg(ptr %a1_addr, i32 %a2) nounwind {
 ; CHECK-NEXT:    ldr w9, [x0]
 ; CHECK-NEXT:    mov w8, #-1 // =0xffffffff
 ; CHECK-NEXT:    subs w10, w9, w1
-; CHECK-NEXT:    cneg w10, w10, le
+; CHECK-NEXT:    cneg w10, w10, lt
 ; CHECK-NEXT:    cneg w8, w8, le
 ; CHECK-NEXT:    lsr w10, w10, #1
 ; CHECK-NEXT:    madd w0, w10, w8, w9
@@ -83,7 +83,7 @@ define i32 @scalar_i32_signed_reg_mem(i32 %a1, ptr %a2_addr) nounwind {
 ; CHECK-NEXT:    ldr w9, [x1]
 ; CHECK-NEXT:    mov w8, #-1 // =0xffffffff
 ; CHECK-NEXT:    subs w9, w0, w9
-; CHECK-NEXT:    cneg w9, w9, le
+; CHECK-NEXT:    cneg w9, w9, lt
 ; CHECK-NEXT:    cneg w8, w8, le
 ; CHECK-NEXT:    lsr w9, w9, #1
 ; CHECK-NEXT:    madd w0, w9, w8, w0
@@ -107,7 +107,7 @@ define i32 @scalar_i32_signed_mem_mem(ptr %a1_addr, ptr %a2_addr) nounwind {
 ; CHECK-NEXT:    ldr w10, [x1]
 ; CHECK-NEXT:    mov w8, #-1 // =0xffffffff
 ; CHECK-NEXT:    subs w10, w9, w10
-; CHECK-NEXT:    cneg w10, w10, le
+; CHECK-NEXT:    cneg w10, w10, lt
 ; CHECK-NEXT:    cneg w8, w8, le
 ; CHECK-NEXT:    lsr w10, w10, #1
 ; CHECK-NEXT:    madd w0, w10, w8, w9
@@ -136,7 +136,7 @@ define i64 @scalar_i64_signed_reg_reg(i64 %a1, i64 %a2) nounwind {
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x9, x0, x1
 ; CHECK-NEXT:    mov x8, #-1 // =0xffffffffffffffff
-; CHECK-NEXT:    cneg x9, x9, le
+; CHECK-NEXT:    cneg x9, x9, lt
 ; CHECK-NEXT:    cneg x8, x8, le
 ; CHECK-NEXT:    lsr x9, x9, #1
 ; CHECK-NEXT:    madd x0, x9, x8, x0
@@ -157,7 +157,7 @@ define i64 @scalar_i64_unsigned_reg_reg(i64 %a1, i64 %a2) nounwind {
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    subs x9, x0, x1
 ; CHECK-NEXT:    mov x8, #-1 // =0xffffffffffffffff
-; CHECK-NEXT:    cneg x9, x9, ls
+; CHECK-NEXT:    cneg x9, x9, lo
 ; CHECK-NEXT:    cneg x8, x8, ls
 ; CHECK-NEXT:    lsr x9, x9, #1
 ; CHECK-NEXT:    madd x0, x9, x8, x0
@@ -181,7 +181,7 @@ define i64 @scalar_i64_signed_mem_reg(ptr %a1_addr, i64 %a2) nounwind {
 ; CHECK-NEXT:    ldr x9, [x0]
 ; CHECK-NEXT:    mov x8, #-1 // =0xffffffffffffffff
 ; CHECK-NEXT:    subs x10, x9, x1
-; CHECK-NEXT:    cneg x10, x10, le
+; CHECK-NEXT:    cneg x10, x10, lt
 ; CHECK-NEXT:    cneg x8, x8, le
 ; CHECK-NEXT:    lsr x10, x10, #1
 ; CHECK-NEXT:    madd x0, x10, x8, x9
@@ -204,7 +204,7 @@ define i64 @scalar_i64_signed_reg_mem(i64 %a1, ptr %a2_addr) nounwind {
 ; CHECK-NEXT:    ldr x9, [x1]
 ; CHECK-NEXT:    mov x8, #-1 // =0xffffffffffffffff
 ; CHECK-NEXT:    subs x9, x0, x9
-; CHECK-NEXT:    cneg x9, x9, le
+; CHECK-NEXT:    cneg x9, x9, lt
 ; CHECK-NEXT:    cneg x8, x8, le
 ; CHECK-NEXT:    lsr x9, x9, #1
 ; CHECK-NEXT:    madd x0, x9, x8, x0
@@ -228,7 +228,7 @@ define i64 @scalar_i64_signed_mem_mem(ptr %a1_addr, ptr %a2_addr) nounwind {
 ; CHECK-NEXT:    ldr x10, [x1]
 ; CHECK-NEXT:    mov x8, #-1 // =0xffffffffffffffff
 ; CHECK-NEXT:    subs x10, x9, x10
-; CHECK-NEXT:    cneg x10, x10, le
+; CHECK-NEXT:    cneg x10, x10, lt
 ; CHECK-NEXT:    cneg x8, x8, le
 ; CHECK-NEXT:    lsr x10, x10, #1
 ; CHECK-NEXT:    madd x0, x10, x8, x9
diff --git a/llvm/test/CodeGen/AArch64/pr72777.ll b/llvm/test/CodeGen/AArch64/pr72777.ll
index fa9f82f8c93c2..c8966bf9ce68b 100644
--- a/llvm/test/CodeGen/AArch64/pr72777.ll
+++ b/llvm/test/CodeGen/AArch64/pr72777.ll
@@ -5,9 +5,9 @@ define i64 @f(i64 %0, i64 %1) {
 ; CHECK-LABEL: f:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    orr x8, x1, #0x1
-; CHECK-NEXT:    add x9, x0, x0
 ; CHECK-NEXT:    mov x10, #-9223372036854775808 // =0x8000000000000000
-; CHECK-NEXT:    adds x8, x8, x9
+; CHECK-NEXT:    add x8, x8, x0
+; CHECK-NEXT:    adds x8, x8, x0
 ; CHECK-NEXT:    lsl x9, x8, #1
 ; CHECK-NEXT:    cinv x10, x10, pl
 ; CHECK-NEXT:    cmp x8, x9, asr #1