diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index a1a2cef..3ecd1c6 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -7564,6 +7564,19 @@
   return CGF.Builder.CreateMaskedLoad(Ops[0], Align, MaskVec, Ops[1]);
 }
 
+static Value *EmitX86MaskLogic(CodeGenFunction &CGF, Instruction::BinaryOps Opc,
+                              unsigned NumElts, SmallVectorImpl<Value *> &Ops,
+                              bool InvertLHS = false) {
+  Value *LHS = getMaskVecValue(CGF, Ops[0], NumElts);
+  Value *RHS = getMaskVecValue(CGF, Ops[1], NumElts);
+
+  if (InvertLHS)
+    LHS = CGF.Builder.CreateNot(LHS);
+
+  return CGF.Builder.CreateBitCast(CGF.Builder.CreateBinOp(Opc, LHS, RHS),
+                                  CGF.Builder.getIntNTy(std::max(NumElts, 8U)));
+}
+
 static Value *EmitX86SubVectorBroadcast(CodeGenFunction &CGF,
                                         SmallVectorImpl<Value *> &Ops,
                                         llvm::Type *DstTy,
@@ -8217,6 +8230,22 @@
     return EmitX86MaskedCompare(*this, CC, false, Ops);
   }
 
+  case X86::BI__builtin_ia32_kandhi:
+    return EmitX86MaskLogic(*this, Instruction::And, 16, Ops);
+  case X86::BI__builtin_ia32_kandnhi:
+    return EmitX86MaskLogic(*this, Instruction::And, 16, Ops, true);
+  case X86::BI__builtin_ia32_korhi:
+    return EmitX86MaskLogic(*this, Instruction::Or, 16, Ops);
+  case X86::BI__builtin_ia32_kxnorhi:
+    return EmitX86MaskLogic(*this, Instruction::Xor, 16, Ops, true);
+  case X86::BI__builtin_ia32_kxorhi:
+    return EmitX86MaskLogic(*this, Instruction::Xor, 16, Ops);
+  case X86::BI__builtin_ia32_knothi: {
+    Ops[0] = getMaskVecValue(*this, Ops[0], 16);
+    return Builder.CreateBitCast(Builder.CreateNot(Ops[0]),
+                                 Builder.getInt16Ty());
+  }
+
   case X86::BI__builtin_ia32_vplzcntd_128_mask:
   case X86::BI__builtin_ia32_vplzcntd_256_mask:
   case X86::BI__builtin_ia32_vplzcntd_512_mask: