diff --git a/include/llvm/Target/TargetLowering.h b/include/llvm/Target/TargetLowering.h
index aadfca964aa..d332c2d366b 100644
--- a/include/llvm/Target/TargetLowering.h
+++ b/include/llvm/Target/TargetLowering.h
@@ -2017,6 +2017,15 @@ public:
   ///
   virtual SDValue PerformDAGCombine(SDNode *N, DAGCombinerInfo &DCI) const;
 
+  /// Return true if it is profitable to move a following shift through this
+  //  node, adjusting any immediate operands as necessary to preserve semantics.
+  //  This transformation may not be desirable if it disrupts a particularly
+  //  auspicious target-specific tree (e.g. bitfield extractionon in AArch64).
+  //  By default, it returns true.
+  virtual bool isDesirableToCommuteWithShift(const SDNode *N /*Op*/) const {
+    return true;
+  }
+
   /// Return true if the target has native support for the specified value type
   /// and it is 'desirable' to use the type for the given node type. e.g. On x86
   /// i16 is legal, but undesirable since i16 instruction encodings are longer
diff --git a/lib/CodeGen/SelectionDAG/DAGCombiner.cpp b/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
index 014e97d7b2a..290f2a1ea27 100644
--- a/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
+++ b/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
@@ -3867,6 +3867,9 @@ SDValue DAGCombiner::visitShiftByConstant(SDNode *N, ConstantSDNode *Amt) {
       return SDValue();
   }
 
+  if (!TLI.isDesirableToCommuteWithShift(LHS))
+    return SDValue();
+
   // Fold the constants, shifting the binop RHS by the shift amount.
   SDValue NewRHS = DAG.getNode(N->getOpcode(), SDLoc(LHS->getOperand(1)),
                                N->getValueType(0),
diff --git a/lib/Target/ARM64/ARM64ISelLowering.cpp b/lib/Target/ARM64/ARM64ISelLowering.cpp
index ccd80175faf..6dd588c3705 100644
--- a/lib/Target/ARM64/ARM64ISelLowering.cpp
+++ b/lib/Target/ARM64/ARM64ISelLowering.cpp
@@ -5931,6 +5931,21 @@ ARM64TargetLowering::getScratchRegisters(CallingConv::ID) const {
   return ScratchRegs;
 }
 
+bool ARM64TargetLowering::isDesirableToCommuteWithShift(const SDNode *N) const {
+  EVT VT = N->getValueType(0);
+    // If N is unsigned bit extraction: ((x >> C) & mask), then do not combine
+    // it with shift to let it be lowered to UBFX.
+  if (N->getOpcode() == ISD::AND && (VT == MVT::i32 || VT == MVT::i64) &&
+      isa<ConstantSDNode>(N->getOperand(1))) {
+    uint64_t TruncMask = N->getConstantOperandVal(1);
+    if (isMask_64(TruncMask) &&
+      N->getOperand(0).getOpcode() == ISD::SRL &&
+      isa<ConstantSDNode>(N->getOperand(0)->getOperand(1)))
+      return false;
+  }
+  return true;
+}
+
 bool ARM64TargetLowering::shouldConvertConstantLoadToIntImm(const APInt &Imm,
                                                             Type *Ty) const {
   assert(Ty->isIntegerTy());
diff --git a/lib/Target/ARM64/ARM64ISelLowering.h b/lib/Target/ARM64/ARM64ISelLowering.h
index ddaf347bf1a..96a32ea85c0 100644
--- a/lib/Target/ARM64/ARM64ISelLowering.h
+++ b/lib/Target/ARM64/ARM64ISelLowering.h
@@ -284,6 +284,9 @@ public:
 
   const MCPhysReg *getScratchRegisters(CallingConv::ID CC) const override;
 
+  /// \brief Returns false if N is a bit extraction pattern of (X >> C) & Mask.
+  bool isDesirableToCommuteWithShift(const SDNode *N) const override;
+
   /// \brief Returns true if it is beneficial to convert a load of a constant
   /// to just the constant itself.
   bool shouldConvertConstantLoadToIntImm(const APInt &Imm,
diff --git a/test/CodeGen/ARM64/bitfield-extract.ll b/test/CodeGen/ARM64/bitfield-extract.ll
index 6de563c1eb6..3ea6d938e9d 100644
--- a/test/CodeGen/ARM64/bitfield-extract.ll
+++ b/test/CodeGen/ARM64/bitfield-extract.ll
@@ -501,6 +501,19 @@ end:
   ret i80 %conv3
 }
 
+; Check if we can still catch UBFX when "AND" is used by SHL.
+; CHECK-LABEL: fct21:
+; CHECK: ubfx
+@arr = external global [8 x [64 x i64]]
+define i64 @fct21(i64 %x) {
+entry:
+  %shr = lshr i64 %x, 4
+  %and = and i64 %shr, 15
+  %arrayidx = getelementptr inbounds [8 x [64 x i64]]* @arr, i64 0, i64 0, i64 %and
+  %0 = load i64* %arrayidx, align 8
+  ret i64 %0
+}
+
 define i16 @test_ignored_rightbits(i32 %dst, i32 %in) {
 ; CHECK-LABEL: test_ignored_rightbits: