diff --git a/lib/CodeGen/CGBuiltin.cpp b/lib/CodeGen/CGBuiltin.cpp
index c6ac3cc9445783d64d4f4713f673e4c1093f747c..97145348b3534f05ae8ecc21bd9969fe2631f4f4 100644
--- a/lib/CodeGen/CGBuiltin.cpp
+++ b/lib/CodeGen/CGBuiltin.cpp
@@ -2493,7 +2493,7 @@ static Value *EmitCommonNeonSISDBuiltinExpr(CodeGenFunction &CGF,
   Function *F = CGF.LookupNeonLLVMIntrinsic(Int, Modifier, ArgTy, E);
 
   int j = 0;
-  ConstantInt *C0 = ConstantInt::get(CGF.Int32Ty, 0);
+  ConstantInt *C0 = ConstantInt::get(CGF.SizeTy, 0);
   for (Function::const_arg_iterator ai = F->arg_begin(), ae = F->arg_end();
        ai != ae; ++ai, ++j) {
     llvm::Type *ArgTy = ai->getType();
@@ -2733,7 +2733,7 @@ Value *CodeGenFunction::EmitCommonNeonBuiltinExpr(
     Ops[0] = Builder.CreateBitCast(Ops[0], Ty);
     LoadInst *Ld = Builder.CreateLoad(Ops[0]);
     Ld->setAlignment(cast<ConstantInt>(Align)->getZExtValue());
-    llvm::Constant *CI = ConstantInt::get(Int32Ty, 0);
+    llvm::Constant *CI = ConstantInt::get(SizeTy, 0);
     Ops[0] = Builder.CreateInsertElement(V, Ld, CI);
     return EmitNeonSplat(Ops[0], CI);
   }
@@ -3703,7 +3703,7 @@ Value *CodeGenFunction::vectorWrapScalar16(Value *Op) {
   llvm::Type *VTy = llvm::VectorType::get(Int16Ty, 4);
   Op = Builder.CreateBitCast(Op, Int16Ty);
   Value *V = UndefValue::get(VTy);
-  llvm::Constant *CI = ConstantInt::get(Int32Ty, 0);
+  llvm::Constant *CI = ConstantInt::get(SizeTy, 0);
   Op = Builder.CreateInsertElement(V, Op, CI);
   return Op;
 }
@@ -3712,7 +3712,7 @@ Value *CodeGenFunction::vectorWrapScalar8(Value *Op) {
   llvm::Type *VTy = llvm::VectorType::get(Int8Ty, 8);
   Op = Builder.CreateBitCast(Op, Int8Ty);
   Value *V = UndefValue::get(VTy);
-  llvm::Constant *CI = ConstantInt::get(Int32Ty, 0);
+  llvm::Constant *CI = ConstantInt::get(SizeTy, 0);
   Op = Builder.CreateInsertElement(V, Op, CI);
   return Op;
 }
@@ -3729,7 +3729,7 @@ emitVectorWrappedScalar8Intrinsic(unsigned Int, SmallVectorImpl<Value*> &Ops,
   Ops[1] = vectorWrapScalar8(Ops[1]);
   llvm::Type *VTy = llvm::VectorType::get(Int8Ty, 8);
   Value *V = EmitNeonCall(CGM.getIntrinsic(Int, VTy), Ops, Name);
-  Constant *CI = ConstantInt::get(Int32Ty, 0);
+  Constant *CI = ConstantInt::get(SizeTy, 0);
   return Builder.CreateExtractElement(V, CI, "lane0");
 }
 
@@ -3745,7 +3745,7 @@ emitVectorWrappedScalar16Intrinsic(unsigned Int, SmallVectorImpl<Value*> &Ops,
   Ops[1] = vectorWrapScalar16(Ops[1]);
   llvm::Type *VTy = llvm::VectorType::get(Int16Ty, 4);
   Value *V = EmitNeonCall(CGM.getIntrinsic(Int, VTy), Ops, Name);
-  Constant *CI = ConstantInt::get(Int32Ty, 0);
+  Constant *CI = ConstantInt::get(SizeTy, 0);
   return Builder.CreateExtractElement(V, CI, "lane0");
 }
 
@@ -3950,8 +3950,8 @@ Value *CodeGenFunction::EmitAArch64BuiltinExpr(unsigned BuiltinID,
     Value *Vec = EmitScalarExpr(E->getArg(0));
     // The vector is v2f64, so make sure it's bitcast to that.
     Vec = Builder.CreateBitCast(Vec, Ty, "v2i64");
-    llvm::Value *Idx0 = llvm::ConstantInt::get(Int32Ty, 0);
-    llvm::Value *Idx1 = llvm::ConstantInt::get(Int32Ty, 1);
+    llvm::Value *Idx0 = llvm::ConstantInt::get(SizeTy, 0);
+    llvm::Value *Idx1 = llvm::ConstantInt::get(SizeTy, 1);
     Value *Op0 = Builder.CreateExtractElement(Vec, Idx0, "lane0");
     Value *Op1 = Builder.CreateExtractElement(Vec, Idx1, "lane1");
     // Pairwise addition of a v2f64 into a scalar f64.
@@ -3963,8 +3963,8 @@ Value *CodeGenFunction::EmitAArch64BuiltinExpr(unsigned BuiltinID,
     Value *Vec = EmitScalarExpr(E->getArg(0));
     // The vector is v2f64, so make sure it's bitcast to that.
     Vec = Builder.CreateBitCast(Vec, Ty, "v2f64");
-    llvm::Value *Idx0 = llvm::ConstantInt::get(Int32Ty, 0);
-    llvm::Value *Idx1 = llvm::ConstantInt::get(Int32Ty, 1);
+    llvm::Value *Idx0 = llvm::ConstantInt::get(SizeTy, 0);
+    llvm::Value *Idx1 = llvm::ConstantInt::get(SizeTy, 1);
     Value *Op0 = Builder.CreateExtractElement(Vec, Idx0, "lane0");
     Value *Op1 = Builder.CreateExtractElement(Vec, Idx1, "lane1");
     // Pairwise addition of a v2f64 into a scalar f64.
@@ -3976,8 +3976,8 @@ Value *CodeGenFunction::EmitAArch64BuiltinExpr(unsigned BuiltinID,
     Value *Vec = EmitScalarExpr(E->getArg(0));
     // The vector is v2f32, so make sure it's bitcast to that.
     Vec = Builder.CreateBitCast(Vec, Ty, "v2f32");
-    llvm::Value *Idx0 = llvm::ConstantInt::get(Int32Ty, 0);
-    llvm::Value *Idx1 = llvm::ConstantInt::get(Int32Ty, 1);
+    llvm::Value *Idx0 = llvm::ConstantInt::get(SizeTy, 0);
+    llvm::Value *Idx1 = llvm::ConstantInt::get(SizeTy, 1);
     Value *Op0 = Builder.CreateExtractElement(Vec, Idx0, "lane0");
     Value *Op1 = Builder.CreateExtractElement(Vec, Idx1, "lane1");
     // Pairwise addition of a v2f32 into a scalar f32.
@@ -4228,7 +4228,7 @@ Value *CodeGenFunction::EmitAArch64BuiltinExpr(unsigned BuiltinID,
     llvm::Type *VTy = llvm::VectorType::get(Int32Ty, 4);
     Ops[1] = EmitNeonCall(CGM.getIntrinsic(Intrinsic::aarch64_neon_sqdmull, VTy),
                           ProductOps, "vqdmlXl");
-    Constant *CI = ConstantInt::get(Int32Ty, 0);
+    Constant *CI = ConstantInt::get(SizeTy, 0);
     Ops[1] = Builder.CreateExtractElement(Ops[1], CI, "lane0");
 
     unsigned AccumInt = BuiltinID == NEON::BI__builtin_neon_vqdmlalh_s16
@@ -4325,7 +4325,7 @@ Value *CodeGenFunction::EmitAArch64BuiltinExpr(unsigned BuiltinID,
     llvm::Type *VTy = llvm::VectorType::get(Int32Ty, 4);
     Ops[1] = EmitNeonCall(CGM.getIntrinsic(Intrinsic::aarch64_neon_sqdmull, VTy),
                           ProductOps, "vqdmlXl");
-    Constant *CI = ConstantInt::get(Int32Ty, 0);
+    Constant *CI = ConstantInt::get(SizeTy, 0);
     Ops[1] = Builder.CreateExtractElement(Ops[1], CI, "lane0");
     Ops.pop_back();
 
@@ -5232,7 +5232,7 @@ Value *CodeGenFunction::EmitAArch64BuiltinExpr(unsigned BuiltinID,
     Ty = llvm::PointerType::getUnqual(VTy->getElementType());
     Ops[0] = Builder.CreateBitCast(Ops[0], Ty);
     Ops[0] = Builder.CreateLoad(Ops[0]);
-    llvm::Constant *CI = ConstantInt::get(Int32Ty, 0);
+    llvm::Constant *CI = ConstantInt::get(SizeTy, 0);
     Ops[0] = Builder.CreateInsertElement(V, Ops[0], CI);
     return EmitNeonSplat(Ops[0], CI);
   }
@@ -5611,7 +5611,7 @@ Value *CodeGenFunction::EmitX86BuiltinExpr(unsigned BuiltinID,
 
     // extract (0, 1)
     unsigned Index = BuiltinID == X86::BI__builtin_ia32_storelps ? 0 : 1;
-    llvm::Value *Idx = llvm::ConstantInt::get(Int32Ty, Index);
+    llvm::Value *Idx = llvm::ConstantInt::get(SizeTy, Index);
     Ops[1] = Builder.CreateExtractElement(Ops[1], Idx, "extract");
 
     // cast pointer to i64 & store
diff --git a/lib/CodeGen/CGExpr.cpp b/lib/CodeGen/CGExpr.cpp
index 966edc46f2e5e6bd80e7646b9aaad6d0232096f0..e43e61c94636c19a9c37a9a3ae1d70f639028d84 100644
--- a/lib/CodeGen/CGExpr.cpp
+++ b/lib/CodeGen/CGExpr.cpp
@@ -1331,7 +1331,7 @@ RValue CodeGenFunction::EmitLoadOfExtVectorElementLValue(LValue LV) {
   const VectorType *ExprVT = LV.getType()->getAs<VectorType>();
   if (!ExprVT) {
     unsigned InIdx = getAccessedFieldNo(0, Elts);
-    llvm::Value *Elt = llvm::ConstantInt::get(Int32Ty, InIdx);
+    llvm::Value *Elt = llvm::ConstantInt::get(SizeTy, InIdx);
     return RValue::get(Builder.CreateExtractElement(Vec, Elt));
   }
 
@@ -1590,7 +1590,7 @@ void CodeGenFunction::EmitStoreThroughExtVectorComponentLValue(RValue Src,
   } else {
     // If the Src is a scalar (not a vector) it must be updating one element.
     unsigned InIdx = getAccessedFieldNo(0, Elts);
-    llvm::Value *Elt = llvm::ConstantInt::get(Int32Ty, InIdx);
+    llvm::Value *Elt = llvm::ConstantInt::get(SizeTy, InIdx);
     Vec = Builder.CreateInsertElement(Vec, SrcVal, Elt);
   }
 
@@ -2307,7 +2307,6 @@ LValue CodeGenFunction::EmitArraySubscriptExpr(const ArraySubscriptExpr *E,
     // Emit the vector as an lvalue to get its address.
     LValue LHS = EmitLValue(E->getBase());
     assert(LHS.isSimple() && "Can only subscript lvalue vectors here!");
-    Idx = Builder.CreateIntCast(Idx, Int32Ty, IdxSigned, "vidx");
     return LValue::MakeVectorElt(LHS.getAddress(), Idx,
                                  E->getBase()->getType(), LHS.getAlignment());
   }
diff --git a/lib/CodeGen/CGExprScalar.cpp b/lib/CodeGen/CGExprScalar.cpp
index 1ee3ce84ce2a0a59a8ea0eda6264951fde3855b3..4dfa16961580ff4ceefd386d969df3159f608ed1 100644
--- a/lib/CodeGen/CGExprScalar.cpp
+++ b/lib/CodeGen/CGExprScalar.cpp
@@ -940,9 +940,8 @@ Value *ScalarExprEmitter::VisitShuffleVectorExpr(ShuffleVectorExpr *E) {
                                                   MTy->getNumElements());
     Value* NewV = llvm::UndefValue::get(RTy);
     for (unsigned i = 0, e = MTy->getNumElements(); i != e; ++i) {
-      Value *IIndx = Builder.getInt32(i);
+      Value *IIndx = llvm::ConstantInt::get(CGF.SizeTy, i);
       Value *Indx = Builder.CreateExtractElement(Mask, IIndx, "shuf_idx");
-      Indx = Builder.CreateZExt(Indx, CGF.Int32Ty, "idx_zext");
 
       Value *VExt = Builder.CreateExtractElement(LHS, Indx, "shuf_elt");
       NewV = Builder.CreateInsertElement(NewV, VExt, IIndx, "shuf_ins");
@@ -1073,8 +1072,6 @@ Value *ScalarExprEmitter::VisitArraySubscriptExpr(ArraySubscriptExpr *E) {
   if (CGF.SanOpts->ArrayBounds)
     CGF.EmitBoundsCheck(E, E->getBase(), Idx, IdxTy, /*Accessed*/true);
 
-  bool IdxSigned = IdxTy->isSignedIntegerOrEnumerationType();
-  Idx = Builder.CreateIntCast(Idx, CGF.Int32Ty, IdxSigned, "vecidxcast");
   return Builder.CreateExtractElement(Base, Idx, "vecext");
 }
 
diff --git a/test/CodeGen/builtinshufflevector2.c b/test/CodeGen/builtinshufflevector2.c
index 04405b5cd159e880021c3676852b46ed1009a63c..8712c991fe18cef02ea667fe446f9e63f051daa1 100644
--- a/test/CodeGen/builtinshufflevector2.c
+++ b/test/CodeGen/builtinshufflevector2.c
@@ -6,23 +6,23 @@ typedef unsigned int uint4 __attribute__((ext_vector_type(4)));
 // CHECK-LABEL: define void @clang_shufflevector_v_v(
 void clang_shufflevector_v_v( float4* A, float4 x, uint4 mask ) {
 // CHECK: [[MASK:%.*]] = and <4 x i32> {{%.*}}, <i32 3, i32 3, i32 3, i32 3>
-// CHECK: [[I:%.*]] = extractelement <4 x i32> [[MASK]], i32 0
-// CHECK: [[E:%.*]] = extractelement <4 x float> [[X:%.*]], i32 [[I]]
+// CHECK: [[I:%.*]] = extractelement <4 x i32> [[MASK]], i{{[0-9]+}} 0
+// CHECK: [[E:%.*]] = extractelement <4 x float> [[X:%.*]], i{{[0-9]+}} [[I]]
 //
 // Here is where ToT Clang code generation makes a mistake.  
 // It uses [[I]] as the insertion index instead of 0.
 // Similarly on the remaining insertelement.
-// CHECK: [[V:%[a-zA-Z0-9._]+]] = insertelement <4 x float> undef, float [[E]], i32 0
+// CHECK: [[V:%[a-zA-Z0-9._]+]] = insertelement <4 x float> undef, float [[E]], i{{[0-9]+}} 0
 
-// CHECK: [[I:%.*]] = extractelement <4 x i32> [[MASK]], i32 1
-// CHECK: [[E:%.*]] = extractelement <4 x float> [[X]], i32 [[I]]
-// CHECK: [[V2:%.*]] = insertelement <4 x float> [[V]], float [[E]], i32 1
-// CHECK: [[I:%.*]] = extractelement <4 x i32> [[MASK]], i32 2
-// CHECK: [[E:%.*]] = extractelement <4 x float> [[X]], i32 [[I]]
-// CHECK: [[V3:%.*]] = insertelement <4 x float> [[V2]], float [[E]], i32 2
-// CHECK: [[I:%.*]] = extractelement <4 x i32> [[MASK]], i32 3
-// CHECK: [[E:%.*]] = extractelement <4 x float> [[X]], i32 [[I]]
-// CHECK: [[V4:%.*]] = insertelement <4 x float> [[V3]], float [[E]], i32 3
+// CHECK: [[I:%.*]] = extractelement <4 x i32> [[MASK]], i{{[0-9]+}} 1
+// CHECK: [[E:%.*]] = extractelement <4 x float> [[X]], i{{[0-9]+}} [[I]]
+// CHECK: [[V2:%.*]] = insertelement <4 x float> [[V]], float [[E]], i{{[0-9]+}} 1
+// CHECK: [[I:%.*]] = extractelement <4 x i32> [[MASK]], i{{[0-9]+}} 2
+// CHECK: [[E:%.*]] = extractelement <4 x float> [[X]], i{{[0-9]+}} [[I]]
+// CHECK: [[V3:%.*]] = insertelement <4 x float> [[V2]], float [[E]], i{{[0-9]+}} 2
+// CHECK: [[I:%.*]] = extractelement <4 x i32> [[MASK]], i{{[0-9]+}} 3
+// CHECK: [[E:%.*]] = extractelement <4 x float> [[X]], i{{[0-9]+}} [[I]]
+// CHECK: [[V4:%.*]] = insertelement <4 x float> [[V3]], float [[E]], i{{[0-9]+}} 3
 // CHECK: store <4 x float> [[V4]], <4 x float>* {{%.*}},
   *A = __builtin_shufflevector( x, mask );
 }