Compute and preserve alignment more faithfully in IR-generation.

Introduce an Address type to bundle a pointer value with an
alignment.  Introduce APIs on CGBuilderTy to work with Address
values.  Change core APIs on CGF/CGM to traffic in Address where
appropriate.  Require alignments to be non-zero.  Update a ton
of code to compute and propagate alignment information.

As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment
helper function to CGF and made use of it in a number of places in
the expression emitter.

The end result is that we should now be significantly more correct
when performing operations on objects that are locally known to
be under-aligned.  Since alignment is not reliably tracked in the
type system, there are inherent limits to this, but at least we
are no longer confused by standard operations like derived-to-base
conversions and array-to-pointer decay.  I've also fixed a large
number of bugs where we were applying the complete-object alignment
to a pointer instead of the non-virtual alignment, although most of
these were hidden by the very conservative approach we took with
member alignment.

Also, because IRGen now reliably asserts on zero alignments, we
should no longer be subject to an absurd but frustrating recurring
bug where an incomplete type would report a zero alignment and then
we'd naively do a alignmentAtOffset on it and emit code using an
alignment equal to the largest power-of-two factor of the offset.

We should also now be emitting much more aggressive alignment
attributes in the presence of over-alignment.  In particular,
field access now uses alignmentAtOffset instead of min.

Several times in this patch, I had to change the existing
code-generation pattern in order to more effectively use
the Address APIs.  For the most part, this seems to be a strict
improvement, like doing pointer arithmetic with GEPs instead of
ptrtoint.  That said, I've tried very hard to not change semantics,
but it is likely that I've failed in a few places, for which I
apologize.

ABIArgInfo now always carries the assumed alignment of indirect and
indirect byval arguments.  In order to cut down on what was already
a dauntingly large patch, I changed the code to never set align
attributes in the IR on non-byval indirect arguments.  That is,
we still generate code which assumes that indirect arguments have
the given alignment, but we don't express this information to the
backend except where it's semantically required (i.e. on byvals).
This is likely a minor regression for those targets that did provide
this information, but it'll be trivial to add it back in a later
patch.

I partially punted on applying this work to CGBuiltin.  Please
do not add more uses of the CreateDefaultAligned{Load,Store}
APIs; they will be going away eventually.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@246985 91177308-0d34-0410-b5e6-96231b3b80d8

64 files changed, 1112 insertions, 894 deletions

diff --git a/test/CodeGen/aarch64-varargs.c b/test/CodeGen/aarch64-varargs.c
index 434337173c..08f39600c8 100644
--- a/test/CodeGen/aarch64-varargs.c
+++ b/test/CodeGen/aarch64-varargs.c
@@ -23,21 +23,19 @@ int simple_int(void) {
 
 // CHECK: [[VAARG_IN_REG]]
 // CHECK: [[REG_TOP:%[a-z_0-9]+]] = load i8*, i8** getelementptr inbounds (%struct.__va_list, %struct.__va_list* @the_list, i32 0, i32 1)
-// CHECK: [[REG_ADDR:%[a-z_0-9]+]] = getelementptr i8, i8* [[REG_TOP]], i32 [[GR_OFFS]]
-// CHECK-BE: [[REG_ADDR_VAL:%[0-9]+]] = ptrtoint i8* [[REG_ADDR]] to i64
-// CHECK-BE: [[REG_ADDR_VAL_ALIGNED:%[a-z_0-9]*]] = add i64 [[REG_ADDR_VAL]], 4
-// CHECK-BE: [[REG_ADDR:%[0-9]+]] = inttoptr i64 [[REG_ADDR_VAL_ALIGNED]] to i8*
-// CHECK: [[FROMREG_ADDR:%[a-z_0-9]+]] = bitcast i8* [[REG_ADDR]] to i32*
+// CHECK: [[REG_ADDR:%[a-z_0-9]+]] = getelementptr inbounds i8, i8* [[REG_TOP]], i32 [[GR_OFFS]]
+// CHECK-BE: [[REG_ADDR_ALIGNED:%[0-9]+]] = getelementptr inbounds i8, i8* [[REG_ADDR]], i64 4
+// CHECK-BE: [[FROMREG_ADDR:%[a-z_0-9]+]] = bitcast i8* [[REG_ADDR_ALIGNED]] to i32*
+// CHECK-LE: [[FROMREG_ADDR:%[a-z_0-9]+]] = bitcast i8* [[REG_ADDR]] to i32*
 // CHECK: br label %[[VAARG_END:[a-z._0-9]+]]
 
 // CHECK: [[VAARG_ON_STACK]]
 // CHECK: [[STACK:%[a-z_0-9]+]] = load i8*, i8** getelementptr inbounds (%struct.__va_list, %struct.__va_list* @the_list, i32 0, i32 0)
-// CHECK: [[NEW_STACK:%[a-z_0-9]+]] = getelementptr i8, i8* [[STACK]], i32 8
+// CHECK: [[NEW_STACK:%[a-z_0-9]+]] = getelementptr inbounds i8, i8* [[STACK]], i64 8
 // CHECK: store i8* [[NEW_STACK]], i8** getelementptr inbounds (%struct.__va_list, %struct.__va_list* @the_list, i32 0, i32 0)
-// CHECK-BE: [[STACK_VAL:%[0-9]+]] = ptrtoint i8* [[STACK]] to i64
-// CHECK-BE: [[STACK_VAL_ALIGNED:%[a-z_0-9]*]] = add i64 [[STACK_VAL]], 4
-// CHECK-BE: [[STACK:%[0-9]+]] = inttoptr i64 [[STACK_VAL_ALIGNED]] to i8*
-// CHECK: [[FROMSTACK_ADDR:%[a-z_0-9]+]] = bitcast i8* [[STACK]] to i32*
+// CHECK-BE: [[STACK_ALIGNED:%[a-z_0-9]*]] = getelementptr inbounds i8, i8* [[STACK]], i64 4
+// CHECK-BE: [[FROMSTACK_ADDR:%[a-z_0-9]+]] = bitcast i8* [[STACK_ALIGNED]] to i32*
+// CHECK-LE: [[FROMSTACK_ADDR:%[a-z_0-9]+]] = bitcast i8* [[STACK]] to i32*
 // CHECK: br label %[[VAARG_END]]
 
 // CHECK: [[VAARG_END]]
@@ -63,7 +61,7 @@ __int128 aligned_int(void) {
 
 // CHECK: [[VAARG_IN_REG]]
 // CHECK: [[REG_TOP:%[a-z_0-9]+]] = load i8*, i8** getelementptr inbounds (%struct.__va_list, %struct.__va_list* @the_list, i32 0, i32 1)
-// CHECK: [[REG_ADDR:%[a-z_0-9]+]] = getelementptr i8, i8* [[REG_TOP]], i32 [[ALIGNED_REGOFFS]]
+// CHECK: [[REG_ADDR:%[a-z_0-9]+]] = getelementptr inbounds i8, i8* [[REG_TOP]], i32 [[ALIGNED_REGOFFS]]
 // CHECK: [[FROMREG_ADDR:%[a-z_0-9]+]] = bitcast i8* [[REG_ADDR]] to i128*
 // CHECK: br label %[[VAARG_END:[a-z._0-9]+]]
 
@@ -73,7 +71,7 @@ __int128 aligned_int(void) {
 // CHECK: [[ALIGN_STACK:%[a-z_0-9]+]] = add i64 [[STACKINT]], 15
 // CHECK: [[ALIGNED_STACK_INT:%[a-z_0-9]+]] = and i64 [[ALIGN_STACK]], -16
 // CHECK: [[ALIGNED_STACK_PTR:%[a-z_0-9]+]] = inttoptr i64 [[ALIGNED_STACK_INT]] to i8*
-// CHECK: [[NEW_STACK:%[a-z_0-9]+]] = getelementptr i8, i8* [[ALIGNED_STACK_PTR]], i32 16
+// CHECK: [[NEW_STACK:%[a-z_0-9]+]] = getelementptr inbounds i8, i8* [[ALIGNED_STACK_PTR]], i64 16
 // CHECK: store i8* [[NEW_STACK]], i8** getelementptr inbounds (%struct.__va_list, %struct.__va_list* @the_list, i32 0, i32 0)
 // CHECK: [[FROMSTACK_ADDR:%[a-z_0-9]+]] = bitcast i8* [[ALIGNED_STACK_PTR]] to i128*
 // CHECK: br label %[[VAARG_END]]
@@ -104,14 +102,14 @@ struct bigstruct simple_indirect(void) {
 
 // CHECK: [[VAARG_IN_REG]]
 // CHECK: [[REG_TOP:%[a-z_0-9]+]] = load i8*, i8** getelementptr inbounds (%struct.__va_list, %struct.__va_list* @the_list, i32 0, i32 1)
-// CHECK: [[REG_ADDR:%[a-z_0-9]+]] = getelementptr i8, i8* [[REG_TOP]], i32 [[GR_OFFS]]
+// CHECK: [[REG_ADDR:%[a-z_0-9]+]] = getelementptr inbounds i8, i8* [[REG_TOP]], i32 [[GR_OFFS]]
 // CHECK: [[FROMREG_ADDR:%[a-z_0-9]+]] = bitcast i8* [[REG_ADDR]] to %struct.bigstruct**
 // CHECK: br label %[[VAARG_END:[a-z._0-9]+]]
 
 // CHECK: [[VAARG_ON_STACK]]
 // CHECK: [[STACK:%[a-z_0-9]+]] = load i8*, i8** getelementptr inbounds (%struct.__va_list, %struct.__va_list* @the_list, i32 0, i32 0)
 // CHECK-NOT: and i64
-// CHECK: [[NEW_STACK:%[a-z_0-9]+]] = getelementptr i8, i8* [[STACK]], i32 8
+// CHECK: [[NEW_STACK:%[a-z_0-9]+]] = getelementptr inbounds i8, i8* [[STACK]], i64 8
 // CHECK: store i8* [[NEW_STACK]], i8** getelementptr inbounds (%struct.__va_list, %struct.__va_list* @the_list, i32 0, i32 0)
 // CHECK: [[FROMSTACK_ADDR:%[a-z_0-9]+]] = bitcast i8* [[STACK]] to %struct.bigstruct**
 // CHECK: br label %[[VAARG_END]]
@@ -141,13 +139,13 @@ struct aligned_bigstruct simple_aligned_indirect(void) {
 
 // CHECK: [[VAARG_IN_REG]]
 // CHECK: [[REG_TOP:%[a-z_0-9]+]] = load i8*, i8** getelementptr inbounds (%struct.__va_list, %struct.__va_list* @the_list, i32 0, i32 1)
-// CHECK: [[REG_ADDR:%[a-z_0-9]+]] = getelementptr i8, i8* [[REG_TOP]], i32 [[GR_OFFS]]
+// CHECK: [[REG_ADDR:%[a-z_0-9]+]] = getelementptr inbounds i8, i8* [[REG_TOP]], i32 [[GR_OFFS]]
 // CHECK: [[FROMREG_ADDR:%[a-z_0-9]+]] = bitcast i8* [[REG_ADDR]] to %struct.aligned_bigstruct**
 // CHECK: br label %[[VAARG_END:[a-z._0-9]+]]
 
 // CHECK: [[VAARG_ON_STACK]]
 // CHECK: [[STACK:%[a-z_0-9]+]] = load i8*, i8** getelementptr inbounds (%struct.__va_list, %struct.__va_list* @the_list, i32 0, i32 0)
-// CHECK: [[NEW_STACK:%[a-z_0-9]+]] = getelementptr i8, i8* [[STACK]], i32 8
+// CHECK: [[NEW_STACK:%[a-z_0-9]+]] = getelementptr inbounds i8, i8* [[STACK]], i64 8
 // CHECK: store i8* [[NEW_STACK]], i8** getelementptr inbounds (%struct.__va_list, %struct.__va_list* @the_list, i32 0, i32 0)
 // CHECK: [[FROMSTACK_ADDR:%[a-z_0-9]+]] = bitcast i8* [[STACK]] to %struct.aligned_bigstruct**
 // CHECK: br label %[[VAARG_END]]
@@ -172,16 +170,15 @@ double simple_double(void) {
 
 // CHECK: [[VAARG_IN_REG]]
 // CHECK: [[REG_TOP:%[a-z_0-9]+]] = load i8*, i8** getelementptr inbounds (%struct.__va_list, %struct.__va_list* @the_list, i32 0, i32 2)
-// CHECK: [[REG_ADDR:%[a-z_0-9]+]] = getelementptr i8, i8* [[REG_TOP]], i32 [[VR_OFFS]]
-// CHECK-BE: [[REG_ADDR_VAL:%[0-9]+]] = ptrtoint i8* [[REG_ADDR]] to i64
-// CHECK-BE: [[REG_ADDR_VAL_ALIGNED:%[a-z_0-9]*]] = add i64 [[REG_ADDR_VAL]], 8
-// CHECK-BE: [[REG_ADDR:%[0-9]+]] = inttoptr i64 [[REG_ADDR_VAL_ALIGNED]] to i8*
-// CHECK: [[FROMREG_ADDR:%[a-z_0-9]+]] = bitcast i8* [[REG_ADDR]] to double*
+// CHECK: [[REG_ADDR:%[a-z_0-9]+]] = getelementptr inbounds i8, i8* [[REG_TOP]], i32 [[VR_OFFS]]
+// CHECK-BE: [[REG_ADDR_ALIGNED:%[a-z_0-9]*]] = getelementptr inbounds i8, i8* [[REG_ADDR]], i64 8
+// CHECK-BE: [[FROMREG_ADDR:%[a-z_0-9]+]] = bitcast i8* [[REG_ADDR_ALIGNED]] to double*
+// CHECK-LE: [[FROMREG_ADDR:%[a-z_0-9]+]] = bitcast i8* [[REG_ADDR]] to double*
 // CHECK: br label %[[VAARG_END:[a-z._0-9]+]]
 
 // CHECK: [[VAARG_ON_STACK]]
 // CHECK: [[STACK:%[a-z_0-9]+]] = load i8*, i8** getelementptr inbounds (%struct.__va_list, %struct.__va_list* @the_list, i32 0, i32 0)
-// CHECK: [[NEW_STACK:%[a-z_0-9]+]] = getelementptr i8, i8* [[STACK]], i32 8
+// CHECK: [[NEW_STACK:%[a-z_0-9]+]] = getelementptr inbounds i8, i8* [[STACK]], i64 8
 // CHECK: store i8* [[NEW_STACK]], i8** getelementptr inbounds (%struct.__va_list, %struct.__va_list* @the_list, i32 0, i32 0)
 // CHECK: [[FROMSTACK_ADDR:%[a-z_0-9]+]] = bitcast i8* [[STACK]] to double*
 // CHECK: br label %[[VAARG_END]]
@@ -211,17 +208,17 @@ struct hfa simple_hfa(void) {
 
 // CHECK: [[VAARG_IN_REG]]
 // CHECK: [[REG_TOP:%[a-z_0-9]+]] = load i8*, i8** getelementptr inbounds (%struct.__va_list, %struct.__va_list* @the_list, i32 0, i32 2)
-// CHECK: [[FIRST_REG:%[a-z_0-9]+]] = getelementptr i8, i8* [[REG_TOP]], i32 [[VR_OFFS]]
-// CHECK-LE: [[EL_ADDR:%[a-z_0-9]+]] = getelementptr i8, i8* [[FIRST_REG]], i32 0
-// CHECK-BE: [[EL_ADDR:%[a-z_0-9]+]] = getelementptr i8, i8* [[FIRST_REG]], i32 12
+// CHECK: [[FIRST_REG:%[a-z_0-9]+]] = getelementptr inbounds i8, i8* [[REG_TOP]], i32 [[VR_OFFS]]
+// CHECK-LE: [[EL_ADDR:%[a-z_0-9]+]] = getelementptr inbounds i8, i8* [[FIRST_REG]], i64 0
+// CHECK-BE: [[EL_ADDR:%[a-z_0-9]+]] = getelementptr inbounds i8, i8* [[FIRST_REG]], i64 12
 // CHECK: [[EL_TYPED:%[a-z_0-9]+]] = bitcast i8* [[EL_ADDR]] to float*
-// CHECK: [[EL_TMPADDR:%[a-z_0-9]+]] = getelementptr inbounds [2 x float], [2 x float]* %[[TMP_HFA:[a-z_.0-9]+]], i32 0, i32 0
+// CHECK: [[EL_TMPADDR:%[a-z_0-9]+]] = getelementptr inbounds [2 x float], [2 x float]* %[[TMP_HFA:[a-z_.0-9]+]], i64 0, i64 0
 // CHECK: [[EL:%[a-z_0-9]+]] = load float, float* [[EL_TYPED]]
 // CHECK: store float [[EL]], float* [[EL_TMPADDR]]
-// CHECK-LE: [[EL_ADDR:%[a-z_0-9]+]] = getelementptr i8, i8* [[FIRST_REG]], i32 16
-// CHECK-BE: [[EL_ADDR:%[a-z_0-9]+]] = getelementptr i8, i8* [[FIRST_REG]], i32 28
+// CHECK-LE: [[EL_ADDR:%[a-z_0-9]+]] = getelementptr inbounds i8, i8* [[FIRST_REG]], i64 16
+// CHECK-BE: [[EL_ADDR:%[a-z_0-9]+]] = getelementptr inbounds i8, i8* [[FIRST_REG]], i64 28
 // CHECK: [[EL_TYPED:%[a-z_0-9]+]] = bitcast i8* [[EL_ADDR]] to float*
-// CHECK: [[EL_TMPADDR:%[a-z_0-9]+]] = getelementptr inbounds [2 x float], [2 x float]* %[[TMP_HFA]], i32 0, i32 1
+// CHECK: [[EL_TMPADDR:%[a-z_0-9]+]] = getelementptr inbounds [2 x float], [2 x float]* %[[TMP_HFA]], i64 0, i64 1
 // CHECK: [[EL:%[a-z_0-9]+]] = load float, float* [[EL_TYPED]]
 // CHECK: store float [[EL]], float* [[EL_TMPADDR]]
 // CHECK: [[FROMREG_ADDR:%[a-z_0-9]+]] = bitcast [2 x float]* %[[TMP_HFA]] to %struct.hfa*
@@ -229,7 +226,7 @@ struct hfa simple_hfa(void) {
 
 // CHECK: [[VAARG_ON_STACK]]
 // CHECK: [[STACK:%[a-z_0-9]+]] = load i8*, i8** getelementptr inbounds (%struct.__va_list, %struct.__va_list* @the_list, i32 0, i32 0)
-// CHECK: [[NEW_STACK:%[a-z_0-9]+]] = getelementptr i8, i8* [[STACK]], i32 8
+// CHECK: [[NEW_STACK:%[a-z_0-9]+]] = getelementptr inbounds i8, i8* [[STACK]], i64 8
 // CHECK: store i8* [[NEW_STACK]], i8** getelementptr inbounds (%struct.__va_list, %struct.__va_list* @the_list, i32 0, i32 0)
 // CHECK: [[FROMSTACK_ADDR:%[a-z_0-9]+]] = bitcast i8* [[STACK]] to %struct.hfa*
 // CHECK: br label %[[VAARG_END]]
diff --git a/test/CodeGen/arm-abi-vector.c b/test/CodeGen/arm-abi-vector.c
index 88bf593ed2..9920332e48 100644
--- a/test/CodeGen/arm-abi-vector.c
+++ b/test/CodeGen/arm-abi-vector.c
@@ -14,18 +14,20 @@ typedef __attribute__(( ext_vector_type(5) ))  short __short5;
 // Passing legal vector types as varargs.
 double varargs_vec_2i(int fixed, ...) {
 // CHECK: varargs_vec_2i
-// CHECK: alloca <2 x i32>, align 8
-// CHECK: [[ALIGN:%.*]] = and i32 [[VAR:%.*]], -8
+// CHECK: [[VAR:%.*]] = alloca <2 x i32>, align 8
+// CHECK: [[ALIGN:%.*]] = and i32 {{%.*}}, -8
 // CHECK: [[AP_ALIGN:%.*]] = inttoptr i32 [[ALIGN]] to i8*
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_ALIGN]], i32 8
-// CHECK: bitcast i8* [[AP_ALIGN]] to <2 x i32>*
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_ALIGN]], i32 8
+// CHECK: [[AP_CAST:%.*]] = bitcast i8* [[AP_ALIGN]] to <2 x i32>*
+// CHECK: [[VEC:%.*]] = load <2 x i32>, <2 x i32>* [[AP_CAST]], align 8
+// CHECK: store <2 x i32> [[VEC]], <2 x i32>* [[VAR]], align 8
 // APCS-GNU: varargs_vec_2i
-// APCS-GNU: alloca <2 x i32>, align 8
-// APCS-GNU: [[VAR_ALIGN:%.*]] = alloca <2 x i32>
-// APCS-GNU: [[AP_NEXT:%.*]] = getelementptr i8, i8* {{%.*}}, i32 8
-// APCS-GNU: bitcast <2 x i32>* [[VAR_ALIGN]] to i8*
-// APCS-GNU: call void @llvm.memcpy
-// APCS-GNU: load <2 x i32>, <2 x i32>* [[VAR_ALIGN]]
+// APCS-GNU: [[VAR:%.*]] = alloca <2 x i32>, align 8
+// APCS-GNU: [[AP:%.*]] = load i8*,
+// APCS-GNU: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP]], i32 8
+// APCS-GNU: [[AP_CAST:%.*]] = bitcast i8* [[AP]] to <2 x i32>*
+// APCS-GNU: [[VEC:%.*]] = load <2 x i32>, <2 x i32>* [[AP_CAST]], align 4
+// APCS-GNU: store <2 x i32> [[VEC]], <2 x i32>* [[VAR]], align 8
   va_list ap;
   double sum = fixed;
   va_start(ap, fixed);
@@ -46,11 +48,11 @@ double test_2i(__int2 *in) {
 double varargs_vec_3c(int fixed, ...) {
 // CHECK: varargs_vec_3c
 // CHECK: alloca <3 x i8>, align 4
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP:%.*]], i32 4
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP:%.*]], i32 4
 // CHECK: bitcast i8* [[AP]] to <3 x i8>*
 // APCS-GNU: varargs_vec_3c
 // APCS-GNU: alloca <3 x i8>, align 4
-// APCS-GNU: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP:%.*]], i32 4
+// APCS-GNU: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP:%.*]], i32 4
 // APCS-GNU: bitcast i8* [[AP]] to <3 x i8>*
   va_list ap;
   double sum = fixed;
@@ -71,18 +73,20 @@ double test_3c(__char3 *in) {
 
 double varargs_vec_5c(int fixed, ...) {
 // CHECK: varargs_vec_5c
-// CHECK: alloca <5 x i8>, align 8
+// CHECK: [[VAR:%.*]] = alloca <5 x i8>, align 8
 // CHECK: [[ALIGN:%.*]] = and i32 {{%.*}}, -8
 // CHECK: [[AP_ALIGN:%.*]] = inttoptr i32 [[ALIGN]] to i8*
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_ALIGN]], i32 8
-// CHECK: bitcast i8* [[AP_ALIGN]] to <5 x i8>*
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_ALIGN]], i32 8
+// CHECK: [[AP_CAST:%.*]] = bitcast i8* [[AP_ALIGN]] to <5 x i8>*
+// CHECK: [[VEC:%.*]] = load <5 x i8>, <5 x i8>* [[AP_CAST]], align 8
+// CHECK: store <5 x i8> [[VEC]], <5 x i8>* [[VAR]], align 8
 // APCS-GNU: varargs_vec_5c
-// APCS-GNU: alloca <5 x i8>, align 8
-// APCS-GNU: [[VAR_ALIGN:%.*]] = alloca <5 x i8>
-// APCS-GNU: [[AP_NEXT:%.*]] = getelementptr i8, i8* {{%.*}}, i32 8
-// APCS-GNU: bitcast <5 x i8>* [[VAR_ALIGN]] to i8*
-// APCS-GNU: call void @llvm.memcpy
-// APCS-GNU: load <5 x i8>, <5 x i8>* [[VAR_ALIGN]]
+// APCS-GNU: [[VAR:%.*]] = alloca <5 x i8>, align 8
+// APCS-GNU: [[AP:%.*]] = load i8*,
+// APCS-GNU: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP]], i32 8
+// APCS-GNU: [[AP_CAST:%.*]] = bitcast i8* [[AP]] to <5 x i8>*
+// APCS-GNU: [[VEC:%.*]] = load <5 x i8>, <5 x i8>* [[AP_CAST]], align 4
+// APCS-GNU: store <5 x i8> [[VEC]], <5 x i8>* [[VAR]], align 8
   va_list ap;
   double sum = fixed;
   va_start(ap, fixed);
@@ -102,21 +106,20 @@ double test_5c(__char5 *in) {
 
 double varargs_vec_9c(int fixed, ...) {
 // CHECK: varargs_vec_9c
-// CHECK: alloca <9 x i8>, align 16
-// CHECK: [[VAR_ALIGN:%.*]] = alloca <9 x i8>
+// CHECK: [[VAR:%.*]] = alloca <9 x i8>, align 16
 // CHECK: [[ALIGN:%.*]] = and i32 {{%.*}}, -8
 // CHECK: [[AP_ALIGN:%.*]] = inttoptr i32 [[ALIGN]] to i8*
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_ALIGN]], i32 16
-// CHECK: bitcast <9 x i8>* [[VAR_ALIGN]] to i8*
-// CHECK: call void @llvm.memcpy
-// CHECK: load <9 x i8>, <9 x i8>* [[VAR_ALIGN]]
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_ALIGN]], i32 16
+// CHECK: [[AP_CAST:%.*]] = bitcast i8* [[AP_ALIGN]] to <9 x i8>*
+// CHECK: [[T0:%.*]] = load <9 x i8>, <9 x i8>* [[AP_CAST]], align 8
+// CHECK: store <9 x i8> [[T0]], <9 x i8>* [[VAR]], align 16
 // APCS-GNU: varargs_vec_9c
-// APCS-GNU: alloca <9 x i8>, align 16
-// APCS-GNU: [[VAR_ALIGN:%.*]] = alloca <9 x i8>
-// APCS-GNU: [[AP_NEXT:%.*]] = getelementptr i8, i8* {{%.*}}, i32 16
-// APCS-GNU: bitcast <9 x i8>* [[VAR_ALIGN]] to i8*
-// APCS-GNU: call void @llvm.memcpy
-// APCS-GNU: load <9 x i8>, <9 x i8>* [[VAR_ALIGN]]
+// APCS-GNU: [[VAR:%.*]] = alloca <9 x i8>, align 16
+// APCS-GNU: [[AP:%.*]] = load i8*,
+// APCS-GNU: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP]], i32 16
+// APCS-GNU: [[AP_CAST:%.*]] = bitcast i8* [[AP]] to <9 x i8>*
+// APCS-GNU: [[VEC:%.*]] = load <9 x i8>, <9 x i8>* [[AP_CAST]], align 4
+// APCS-GNU: store <9 x i8> [[VEC]], <9 x i8>* [[VAR]], align 16
   va_list ap;
   double sum = fixed;
   va_start(ap, fixed);
@@ -136,15 +139,13 @@ double test_9c(__char9 *in) {
 
 double varargs_vec_19c(int fixed, ...) {
 // CHECK: varargs_vec_19c
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP:%.*]], i32 4
-// CHECK: [[VAR:%.*]] = bitcast i8* [[AP]] to i8**
-// CHECK: [[VAR2:%.*]] = load i8*, i8** [[VAR]]
-// CHECK: bitcast i8* [[VAR2]] to <19 x i8>*
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP:%.*]], i32 4
+// CHECK: [[VAR:%.*]] = bitcast i8* [[AP]] to <19 x i8>**
+// CHECK: [[VAR2:%.*]] = load <19 x i8>*, <19 x i8>** [[VAR]]
 // APCS-GNU: varargs_vec_19c
-// APCS-GNU: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP:%.*]], i32 4
-// APCS-GNU: [[VAR:%.*]] = bitcast i8* [[AP]] to i8**
-// APCS-GNU: [[VAR2:%.*]] = load i8*, i8** [[VAR]]
-// APCS-GNU: bitcast i8* [[VAR2]] to <19 x i8>*
+// APCS-GNU: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP:%.*]], i32 4
+// APCS-GNU: [[VAR:%.*]] = bitcast i8* [[AP]] to <19 x i8>**
+// APCS-GNU: [[VAR2:%.*]] = load <19 x i8>*, <19 x i8>** [[VAR]]
   va_list ap;
   double sum = fixed;
   va_start(ap, fixed);
@@ -167,15 +168,14 @@ double varargs_vec_3s(int fixed, ...) {
 // CHECK: alloca <3 x i16>, align 8
 // CHECK: [[ALIGN:%.*]] = and i32 {{%.*}}, -8
 // CHECK: [[AP_ALIGN:%.*]] = inttoptr i32 [[ALIGN]] to i8*
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_ALIGN]], i32 8
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_ALIGN]], i32 8
 // CHECK: bitcast i8* [[AP_ALIGN]] to <3 x i16>*
 // APCS-GNU: varargs_vec_3s
-// APCS-GNU: alloca <3 x i16>, align 8
-// APCS-GNU: [[VAR_ALIGN:%.*]] = alloca <3 x i16>
-// APCS-GNU: [[AP_NEXT:%.*]] = getelementptr i8, i8* {{%.*}}, i32 8
-// APCS-GNU: bitcast <3 x i16>* [[VAR_ALIGN]] to i8*
-// APCS-GNU: call void @llvm.memcpy
-// APCS-GNU: load <3 x i16>, <3 x i16>* [[VAR_ALIGN]]
+// APCS-GNU: [[VAR:%.*]] = alloca <3 x i16>, align 8
+// APCS-GNU: [[AP:%.*]] = load i8*,
+// APCS-GNU: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP]], i32 8
+// APCS-GNU: [[AP_CAST:%.*]] = bitcast i8* [[AP]] to <3 x i16>*
+// APCS-GNU: [[VEC:%.*]] = load <3 x i16>, <3 x i16>* [[AP_CAST]], align 4
   va_list ap;
   double sum = fixed;
   va_start(ap, fixed);
@@ -195,21 +195,19 @@ double test_3s(__short3 *in) {
 
 double varargs_vec_5s(int fixed, ...) {
 // CHECK: varargs_vec_5s
-// CHECK: alloca <5 x i16>, align 16
-// CHECK: [[VAR_ALIGN:%.*]] = alloca <5 x i16>
+// CHECK: [[VAR_ALIGN:%.*]] = alloca <5 x i16>, align 16
 // CHECK: [[ALIGN:%.*]] = and i32 {{%.*}}, -8
 // CHECK: [[AP_ALIGN:%.*]] = inttoptr i32 [[ALIGN]] to i8*
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_ALIGN]], i32 16
-// CHECK: bitcast <5 x i16>* [[VAR_ALIGN]] to i8*
-// CHECK: call void @llvm.memcpy
-// CHECK: load <5 x i16>, <5 x i16>* [[VAR_ALIGN]]
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_ALIGN]], i32 16
+// CHECK: [[AP_CAST:%.*]] = bitcast i8* [[AP_ALIGN]] to <5 x i16>*
+// CHECK: [[VEC:%.*]] = load <5 x i16>, <5 x i16>* [[AP_CAST]], align 8
+// CHECK: store <5 x i16> [[VEC]], <5 x i16>* [[VAR_ALIGN]], align 16
 // APCS-GNU: varargs_vec_5s
-// APCS-GNU: alloca <5 x i16>, align 16
-// APCS-GNU: [[VAR_ALIGN:%.*]] = alloca <5 x i16>
-// APCS-GNU: [[AP_NEXT:%.*]] = getelementptr i8, i8* {{%.*}}, i32 16
-// APCS-GNU: bitcast <5 x i16>* [[VAR_ALIGN]] to i8*
-// APCS-GNU: call void @llvm.memcpy
-// APCS-GNU: load <5 x i16>, <5 x i16>* [[VAR_ALIGN]]
+// APCS-GNU: [[VAR:%.*]] = alloca <5 x i16>, align 16
+// APCS-GNU: [[AP:%.*]] = load i8*,
+// APCS-GNU: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP]], i32 16
+// APCS-GNU: [[AP_CAST:%.*]] = bitcast i8* [[AP]] to <5 x i16>*
+// APCS-GNU: [[VEC:%.*]] = load <5 x i16>, <5 x i16>* [[AP_CAST]], align 4
   va_list ap;
   double sum = fixed;
   va_start(ap, fixed);
@@ -238,11 +236,11 @@ double varargs_struct(int fixed, ...) {
 // CHECK: varargs_struct
 // CHECK: [[ALIGN:%.*]] = and i32 {{%.*}}, -8
 // CHECK: [[AP_ALIGN:%.*]] = inttoptr i32 [[ALIGN]] to i8*
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_ALIGN]], i32 16
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_ALIGN]], i32 16
 // CHECK: bitcast i8* [[AP_ALIGN]] to %struct.StructWithVec*
 // APCS-GNU: varargs_struct
 // APCS-GNU: [[VAR_ALIGN:%.*]] = alloca %struct.StructWithVec
-// APCS-GNU: [[AP_NEXT:%.*]] = getelementptr i8, i8* {{%.*}}, i32 16
+// APCS-GNU: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* {{%.*}}, i32 16
 // APCS-GNU: bitcast %struct.StructWithVec* [[VAR_ALIGN]] to i8*
 // APCS-GNU: call void @llvm.memcpy
   va_list ap;
diff --git a/test/CodeGen/arm-arguments.c b/test/CodeGen/arm-arguments.c
index b671626939..ec3e1734b0 100644
--- a/test/CodeGen/arm-arguments.c
+++ b/test/CodeGen/arm-arguments.c
@@ -159,13 +159,13 @@ struct s30 f30() {}
 struct s31 { char x; };
 void f31(struct s31 s) { }
 // AAPCS: @f31([1 x i32] %s.coerce)
-// AAPCS: %s = alloca %struct.s31, align 4
-// AAPCS: alloca [1 x i32]
-// AAPCS: store [1 x i32] %s.coerce, [1 x i32]*
+// AAPCS: %s = alloca %struct.s31, align 1
+// AAPCS: [[TEMP:%.*]] = alloca [1 x i32], align 4
+// AAPCS: store [1 x i32] %s.coerce, [1 x i32]* [[TEMP]], align 4
 // APCS-GNU: @f31([1 x i32] %s.coerce)
-// APCS-GNU: %s = alloca %struct.s31, align 4
-// APCS-GNU: alloca [1 x i32]
-// APCS-GNU: store [1 x i32] %s.coerce, [1 x i32]*
+// APCS-GNU: %s = alloca %struct.s31, align 1
+// APCS-GNU: [[TEMP:%.*]] = alloca [1 x i32], align 4
+// APCS-GNU: store [1 x i32] %s.coerce, [1 x i32]* [[TEMP]], align 4
 
 // PR13562
 struct s32 { double x; };
diff --git a/test/CodeGen/arm64-abi-vector.c b/test/CodeGen/arm64-abi-vector.c
index ebf7f51126..29aeadb66d 100644
--- a/test/CodeGen/arm64-abi-vector.c
+++ b/test/CodeGen/arm64-abi-vector.c
@@ -16,7 +16,7 @@ typedef __attribute__(( ext_vector_type(3) ))  double __double3;
 double varargs_vec_3c(int fixed, ...) {
 // CHECK: varargs_vec_3c
 // CHECK: alloca <3 x i8>, align 4
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_CUR:%.*]], i32 8
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_CUR:%.*]], i64 8
 // CHECK: bitcast i8* [[AP_CUR]] to <3 x i8>*
   va_list ap;
   double sum = fixed;
@@ -36,7 +36,7 @@ double test_3c(__char3 *in) {
 double varargs_vec_4c(int fixed, ...) {
 // CHECK: varargs_vec_4c
 // CHECK: alloca <4 x i8>, align 4
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_CUR:%.*]], i32 8
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_CUR:%.*]], i64 8
 // CHECK: bitcast i8* [[AP_CUR]] to <4 x i8>*
   va_list ap;
   double sum = fixed;
@@ -56,7 +56,7 @@ double test_4c(__char4 *in) {
 double varargs_vec_5c(int fixed, ...) {
 // CHECK: varargs_vec_5c
 // CHECK: alloca <5 x i8>, align 8
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_CUR:%.*]], i32 8
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_CUR:%.*]], i64 8
 // CHECK: bitcast i8* [[AP_CUR]] to <5 x i8>*
   va_list ap;
   double sum = fixed;
@@ -78,7 +78,7 @@ double varargs_vec_9c(int fixed, ...) {
 // CHECK: alloca <9 x i8>, align 16
 // CHECK: [[ALIGN:%.*]] = and i64 {{%.*}}, -16
 // CHECK: [[AP_ALIGN:%.*]] = inttoptr i64 [[ALIGN]] to i8*
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_ALIGN]], i32 16
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_ALIGN]], i64 16
 // CHECK: bitcast i8* [[AP_ALIGN]] to <9 x i8>*
   va_list ap;
   double sum = fixed;
@@ -97,10 +97,9 @@ double test_9c(__char9 *in) {
 
 double varargs_vec_19c(int fixed, ...) {
 // CHECK: varargs_vec_19c
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_CUR:%.*]], i32 8
-// CHECK: [[VAR:%.*]] = bitcast i8* [[AP_CUR]] to i8**
-// CHECK: [[VAR2:%.*]] = load i8*, i8** [[VAR]]
-// CHECK: bitcast i8* [[VAR2]] to <19 x i8>*
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_CUR:%.*]], i64 8
+// CHECK: [[VAR:%.*]] = bitcast i8* [[AP_CUR]] to <19 x i8>**
+// CHECK: [[VAR2:%.*]] = load <19 x i8>*, <19 x i8>** [[VAR]]
   va_list ap;
   double sum = fixed;
   va_start(ap, fixed);
@@ -119,7 +118,7 @@ double test_19c(__char19 *in) {
 double varargs_vec_3s(int fixed, ...) {
 // CHECK: varargs_vec_3s
 // CHECK: alloca <3 x i16>, align 8
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_CUR:%.*]], i32 8
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_CUR:%.*]], i64 8
 // CHECK: bitcast i8* [[AP_CUR]] to <3 x i16>*
   va_list ap;
   double sum = fixed;
@@ -141,7 +140,7 @@ double varargs_vec_5s(int fixed, ...) {
 // CHECK: alloca <5 x i16>, align 16
 // CHECK: [[ALIGN:%.*]] = and i64 {{%.*}}, -16
 // CHECK: [[AP_ALIGN:%.*]] = inttoptr i64 [[ALIGN]] to i8*
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_ALIGN]], i32 16
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_ALIGN]], i64 16
 // CHECK: bitcast i8* [[AP_ALIGN]] to <5 x i16>*
   va_list ap;
   double sum = fixed;
@@ -163,7 +162,7 @@ double varargs_vec_3i(int fixed, ...) {
 // CHECK: alloca <3 x i32>, align 16
 // CHECK: [[ALIGN:%.*]] = and i64 {{%.*}}, -16
 // CHECK: [[AP_ALIGN:%.*]] = inttoptr i64 [[ALIGN]] to i8*
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_ALIGN]], i32 16
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_ALIGN]], i64 16
 // CHECK: bitcast i8* [[AP_ALIGN]] to <3 x i32>*
   va_list ap;
   double sum = fixed;
@@ -183,10 +182,9 @@ double test_3i(__int3 *in) {
 double varargs_vec_5i(int fixed, ...) {
 // CHECK: varargs_vec_5i
 // CHECK: alloca <5 x i32>, align 16
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_CUR:%.*]], i32 8
-// CHECK: [[VAR:%.*]] = bitcast i8* [[AP_CUR]] to i8**
-// CHECK: [[VAR2:%.*]] = load i8*, i8** [[VAR]]
-// CHECK: bitcast i8* [[VAR2]] to <5 x i32>*
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_CUR:%.*]], i64 8
+// CHECK: [[VAR:%.*]] = bitcast i8* [[AP_CUR]] to <5 x i32>**
+// CHECK: [[VAR2:%.*]] = load <5 x i32>*, <5 x i32>** [[VAR]]
   va_list ap;
   double sum = fixed;
   va_start(ap, fixed);
@@ -205,10 +203,9 @@ double test_5i(__int5 *in) {
 double varargs_vec_3d(int fixed, ...) {
 // CHECK: varargs_vec_3d
 // CHECK: alloca <3 x double>, align 16
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_CUR:%.*]], i32 8
-// CHECK: [[VAR:%.*]] = bitcast i8* [[AP_CUR]] to i8**
-// CHECK: [[VAR2:%.*]] = load i8*, i8** [[VAR]]
-// CHECK: bitcast i8* [[VAR2]] to <3 x double>*
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_CUR:%.*]], i64 8
+// CHECK: [[VAR:%.*]] = bitcast i8* [[AP_CUR]] to <3 x double>**
+// CHECK: [[VAR2:%.*]] = load <3 x double>*, <3 x double>** [[VAR]]
   va_list ap;
   double sum = fixed;
   va_start(ap, fixed);
@@ -230,52 +227,49 @@ double varargs_vec(int fixed, ...) {
   double sum = fixed;
   va_start(ap, fixed);
   __char3 c3 = va_arg(ap, __char3);
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_CUR:%.*]], i32 8
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_CUR:%.*]], i64 8
 // CHECK: bitcast i8* [[AP_CUR]] to <3 x i8>*
   sum = sum + c3.x + c3.y;
   __char5 c5 = va_arg(ap, __char5);
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_CUR:%.*]], i32 8
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_CUR:%.*]], i64 8
 // CHECK: bitcast i8* [[AP_CUR]] to <5 x i8>*
   sum = sum + c5.x + c5.y;
   __char9 c9 = va_arg(ap, __char9);
 // CHECK: [[ALIGN:%.*]] = and i64 {{%.*}}, -16
 // CHECK: [[AP_ALIGN:%.*]] = inttoptr i64 [[ALIGN]] to i8*
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_ALIGN]], i32 16
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_ALIGN]], i64 16
 // CHECK: bitcast i8* [[AP_ALIGN]] to <9 x i8>*
   sum = sum + c9.x + c9.y;
   __char19 c19 = va_arg(ap, __char19);
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_CUR:%.*]], i32 8
-// CHECK: [[VAR:%.*]] = bitcast i8* [[AP_CUR]] to i8**
-// CHECK: [[VAR2:%.*]] = load i8*, i8** [[VAR]]
-// CHECK: bitcast i8* [[VAR2]] to <19 x i8>*
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_CUR:%.*]], i64 8
+// CHECK: [[VAR:%.*]] = bitcast i8* [[AP_CUR]] to <19 x i8>**
+// CHECK: [[VAR2:%.*]] = load <19 x i8>*, <19 x i8>** [[VAR]]
   sum = sum + c19.x + c19.y;
   __short3 s3 = va_arg(ap, __short3);
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_CUR:%.*]], i32 8
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_CUR:%.*]], i64 8
 // CHECK: bitcast i8* [[AP_CUR]] to <3 x i16>*
   sum = sum + s3.x + s3.y;
   __short5 s5 = va_arg(ap, __short5);
 // CHECK: [[ALIGN:%.*]] = and i64 {{%.*}}, -16
 // CHECK: [[AP_ALIGN:%.*]] = inttoptr i64 [[ALIGN]] to i8*
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_ALIGN]], i32 16
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_ALIGN]], i64 16
 // CHECK: bitcast i8* [[AP_ALIGN]] to <5 x i16>*
   sum = sum + s5.x + s5.y;
   __int3 i3 = va_arg(ap, __int3);
 // CHECK: [[ALIGN:%.*]] = and i64 {{%.*}}, -16
 // CHECK: [[AP_ALIGN:%.*]] = inttoptr i64 [[ALIGN]] to i8*
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_ALIGN]], i32 16
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_ALIGN]], i64 16
 // CHECK: bitcast i8* [[AP_ALIGN]] to <3 x i32>*
   sum = sum + i3.x + i3.y;
   __int5 i5 = va_arg(ap, __int5);
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_CUR:%.*]], i32 8
-// CHECK: [[VAR:%.*]] = bitcast i8* [[AP_CUR]] to i8**
-// CHECK: [[VAR2:%.*]] = load i8*, i8** [[VAR]]
-// CHECK: bitcast i8* [[VAR2]] to <5 x i32>*
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_CUR:%.*]], i64 8
+// CHECK: [[VAR:%.*]] = bitcast i8* [[AP_CUR]] to <5 x i32>**
+// CHECK: [[VAR2:%.*]] = load <5 x i32>*, <5 x i32>** [[VAR]]
   sum = sum + i5.x + i5.y;
   __double3 d3 = va_arg(ap, __double3);
-// CHECK: [[AP_NEXT:%.*]] = getelementptr i8, i8* [[AP_CUR:%.*]], i32 8
-// CHECK: [[VAR:%.*]] = bitcast i8* [[AP_CUR]] to i8**
-// CHECK: [[VAR2:%.*]] = load i8*, i8** [[VAR]]
-// CHECK: bitcast i8* [[VAR2]] to <3 x double>*
+// CHECK: [[AP_NEXT:%.*]] = getelementptr inbounds i8, i8* [[AP_CUR:%.*]], i64 8
+// CHECK: [[VAR:%.*]] = bitcast i8* [[AP_CUR]] to <3 x double>**
+// CHECK: [[VAR2:%.*]] = load <3 x double>*, <3 x double>** [[VAR]]
   sum = sum + d3.x + d3.y;
   va_end(ap);
   return sum;
diff --git a/test/CodeGen/arm64-arguments.c b/test/CodeGen/arm64-arguments.c
index 8b551c4af4..93a1a19895 100644
--- a/test/CodeGen/arm64-arguments.c
+++ b/test/CodeGen/arm64-arguments.c
@@ -117,7 +117,7 @@ struct s30 f30() {}
 struct s31 { char x; };
 void f31(struct s31 s) { }
 // CHECK: define void @f31(i64 %s.coerce)
-// CHECK: %s = alloca %struct.s31, align 8
+// CHECK: %s = alloca %struct.s31, align 1
 // CHECK: trunc i64 %s.coerce to i8
 // CHECK: store i8 %{{.*}},
 
@@ -273,10 +273,10 @@ typedef struct s38 s38_no_align;
 __attribute__ ((noinline))
 int f38(int i, s38_no_align s1, s38_no_align s2) {
 // CHECK: define i32 @f38(i32 %i, i64 %s1.coerce, i64 %s2.coerce)
-// CHECK: %s1 = alloca %struct.s38, align 8
-// CHECK: %s2 = alloca %struct.s38, align 8
-// CHECK: store i64 %s1.coerce, i64* %{{.*}}, align 8
-// CHECK: store i64 %s2.coerce, i64* %{{.*}}, align 8
+// CHECK: %s1 = alloca %struct.s38, align 4
+// CHECK: %s2 = alloca %struct.s38, align 4
+// CHECK: store i64 %s1.coerce, i64* %{{.*}}, align 4
+// CHECK: store i64 %s2.coerce, i64* %{{.*}}, align 4
 // CHECK: getelementptr inbounds %struct.s38, %struct.s38* %s1, i32 0, i32 0
 // CHECK: getelementptr inbounds %struct.s38, %struct.s38* %s2, i32 0, i32 0
 // CHECK: getelementptr inbounds %struct.s38, %struct.s38* %s1, i32 0, i32 1
@@ -297,10 +297,10 @@ __attribute__ ((noinline))
 int f38_stack(int i, int i2, int i3, int i4, int i5, int i6, int i7, int i8,
               int i9, s38_no_align s1, s38_no_align s2) {
 // CHECK: define i32 @f38_stack(i32 %i, i32 %i2, i32 %i3, i32 %i4, i32 %i5, i32 %i6, i32 %i7, i32 %i8, i32 %i9, i64 %s1.coerce, i64 %s2.coerce)
-// CHECK: %s1 = alloca %struct.s38, align 8
-// CHECK: %s2 = alloca %struct.s38, align 8
-// CHECK: store i64 %s1.coerce, i64* %{{.*}}, align 8
-// CHECK: store i64 %s2.coerce, i64* %{{.*}}, align 8
+// CHECK: %s1 = alloca %struct.s38, align 4
+// CHECK: %s2 = alloca %struct.s38, align 4
+// CHECK: store i64 %s1.coerce, i64* %{{.*}}, align 4
+// CHECK: store i64 %s2.coerce, i64* %{{.*}}, align 4
 // CHECK: getelementptr inbounds %struct.s38, %struct.s38* %s1, i32 0, i32 0
 // CHECK: getelementptr inbounds %struct.s38, %struct.s38* %s2, i32 0, i32 0
 // CHECK: getelementptr inbounds %struct.s38, %struct.s38* %s1, i32 0, i32 1
@@ -381,10 +381,10 @@ typedef struct s40 s40_no_align;
 __attribute__ ((noinline))
 int f40(int i, s40_no_align s1, s40_no_align s2) {
 // CHECK: define i32 @f40(i32 %i, [2 x i64] %s1.coerce, [2 x i64] %s2.coerce)
-// CHECK: %s1 = alloca %struct.s40, align 8
-// CHECK: %s2 = alloca %struct.s40, align 8
-// CHECK: store [2 x i64] %s1.coerce, [2 x i64]* %{{.*}}, align 8
-// CHECK: store [2 x i64] %s2.coerce, [2 x i64]* %{{.*}}, align 8
+// CHECK: %s1 = alloca %struct.s40, align 4
+// CHECK: %s2 = alloca %struct.s40, align 4
+// CHECK: store [2 x i64] %s1.coerce, [2 x i64]* %{{.*}}, align 4
+// CHECK: store [2 x i64] %s2.coerce, [2 x i64]* %{{.*}}, align 4
 // CHECK: getelementptr inbounds %struct.s40, %struct.s40* %s1, i32 0, i32 0
 // CHECK: getelementptr inbounds %struct.s40, %struct.s40* %s2, i32 0, i32 0
 // CHECK: getelementptr inbounds %struct.s40, %struct.s40* %s1, i32 0, i32 1
@@ -405,10 +405,10 @@ __attribute__ ((noinline))
 int f40_stack(int i, int i2, int i3, int i4, int i5, int i6, int i7, int i8,
               int i9, s40_no_align s1, s40_no_align s2) {
 // CHECK: define i32 @f40_stack(i32 %i, i32 %i2, i32 %i3, i32 %i4, i32 %i5, i32 %i6, i32 %i7, i32 %i8, i32 %i9, [2 x i64] %s1.coerce, [2 x i64] %s2.coerce)
-// CHECK: %s1 = alloca %struct.s40, align 8
-// CHECK: %s2 = alloca %struct.s40, align 8
-// CHECK: store [2 x i64] %s1.coerce, [2 x i64]* %{{.*}}, align 8
-// CHECK: store [2 x i64] %s2.coerce, [2 x i64]* %{{.*}}, align 8
+// CHECK: %s1 = alloca %struct.s40, align 4
+// CHECK: %s2 = alloca %struct.s40, align 4
+// CHECK: store [2 x i64] %s1.coerce, [2 x i64]* %{{.*}}, align 4
+// CHECK: store [2 x i64] %s2.coerce, [2 x i64]* %{{.*}}, align 4
 // CHECK: getelementptr inbounds %struct.s40, %struct.s40* %s1, i32 0, i32 0
 // CHECK: getelementptr inbounds %struct.s40, %struct.s40* %s2, i32 0, i32 0
 // CHECK: getelementptr inbounds %struct.s40, %struct.s40* %s1, i32 0, i32 1
@@ -629,7 +629,7 @@ float test_hfa(int n, ...) {
 // CHECK: [[CURLIST:%.*]] = load i8*, i8** [[THELIST]]
 
   // HFA is not indirect, so occupies its full 16 bytes on the stack.
-// CHECK: [[NEXTLIST:%.*]] = getelementptr i8, i8* [[CURLIST]], i32 16
+// CHECK: [[NEXTLIST:%.*]] = getelementptr inbounds i8, i8* [[CURLIST]], i64 16
 // CHECK: store i8* [[NEXTLIST]], i8** [[THELIST]]
 
 // CHECK: bitcast i8* [[CURLIST]] to %struct.HFA*
@@ -656,12 +656,11 @@ float test_toobig_hfa(int n, ...) {
 
   // TooBigHFA is not actually an HFA, so gets passed indirectly. Only 8 bytes
   // of stack consumed.
-// CHECK: [[NEXTLIST:%.*]] = getelementptr i8, i8* [[CURLIST]], i32 8
+// CHECK: [[NEXTLIST:%.*]] = getelementptr inbounds i8, i8* [[CURLIST]], i64 8
 // CHECK: store i8* [[NEXTLIST]], i8** [[THELIST]]
 
-// CHECK: [[HFAPTRPTR:%.*]] = bitcast i8* [[CURLIST]] to i8**
-// CHECK: [[HFAPTR:%.*]] = load i8*, i8** [[HFAPTRPTR]]
-// CHECK: bitcast i8* [[HFAPTR]] to %struct.TooBigHFA*
+// CHECK: [[HFAPTRPTR:%.*]] = bitcast i8* [[CURLIST]] to %struct.TooBigHFA**
+// CHECK: [[HFAPTR:%.*]] = load %struct.TooBigHFA*, %struct.TooBigHFA** [[HFAPTRPTR]]
   __builtin_va_list thelist;
   __builtin_va_start(thelist, n);
   struct TooBigHFA h = __builtin_va_arg(thelist, struct TooBigHFA);
@@ -679,12 +678,12 @@ int32x4_t test_hva(int n, ...) {
 
   // HVA is not indirect, so occupies its full 16 bytes on the stack. but it
   // must be properly aligned.
-// CHECK: [[ALIGN0:%.*]] = getelementptr i8, i8* [[CURLIST]], i32 15
-// CHECK: [[ALIGN1:%.*]] = ptrtoint i8* [[ALIGN0]] to i64
+// CHECK: [[ALIGN0:%.*]] = ptrtoint i8* [[CURLIST]] to i64
+// CHECK: [[ALIGN1:%.*]] = add i64 [[ALIGN0]], 15
 // CHECK: [[ALIGN2:%.*]] = and i64 [[ALIGN1]], -16
 // CHECK: [[ALIGNED_LIST:%.*]] = inttoptr i64 [[ALIGN2]] to i8*
 
-// CHECK: [[NEXTLIST:%.*]] = getelementptr i8, i8* [[ALIGNED_LIST]], i32 32
+// CHECK: [[NEXTLIST:%.*]] = getelementptr inbounds i8, i8* [[ALIGNED_LIST]], i64 32
 // CHECK: store i8* [[NEXTLIST]], i8** [[THELIST]]
 
 // CHECK: bitcast i8* [[ALIGNED_LIST]] to %struct.HVA*
@@ -705,12 +704,11 @@ int32x4_t test_toobig_hva(int n, ...) {
 
   // TooBigHVA is not actually an HVA, so gets passed indirectly. Only 8 bytes
   // of stack consumed.
-// CHECK: [[NEXTLIST:%.*]] = getelementptr i8, i8* [[CURLIST]], i32 8
+// CHECK: [[NEXTLIST:%.*]] = getelementptr inbounds i8, i8* [[CURLIST]], i64 8
 // CHECK: store i8* [[NEXTLIST]], i8** [[THELIST]]
 
-// CHECK: [[HVAPTRPTR:%.*]] = bitcast i8* [[CURLIST]] to i8**
-// CHECK: [[HVAPTR:%.*]] = load i8*, i8** [[HVAPTRPTR]]
-// CHECK: bitcast i8* [[HVAPTR]] to %struct.TooBigHVA*
+// CHECK: [[HVAPTRPTR:%.*]] = bitcast i8* [[CURLIST]] to %struct.TooBigHVA**
+// CHECK: [[HVAPTR:%.*]] = load %struct.TooBigHVA*, %struct.TooBigHVA** [[HVAPTRPTR]]
   __builtin_va_list thelist;
   __builtin_va_start(thelist, n);
   struct TooBigHVA h = __builtin_va_arg(thelist, struct TooBigHVA);
diff --git a/test/CodeGen/arm64-be-hfa-vararg.c b/test/CodeGen/arm64-be-hfa-vararg.c
index 537aab52b3..c22572459b 100644
--- a/test/CodeGen/arm64-be-hfa-vararg.c
+++ b/test/CodeGen/arm64-be-hfa-vararg.c
@@ -4,7 +4,15 @@
 
 // A single member HFA must be aligned just like a non-HFA register argument.
 double callee(int a, ...) {
-// CHECK: = add i64 %{{.*}}, 8
+// CHECK: [[REGPP:%.*]] = getelementptr inbounds %struct.__va_list, %struct.__va_list* [[VA:%.*]], i32 0, i32 2
+// CHECK: [[REGP:%.*]] = load i8*, i8** [[REGPP]], align 8
+// CHECK: [[OFFSET0:%.*]] = getelementptr inbounds i8, i8* [[REGP]], i32 {{.*}}
+// CHECK: [[OFFSET1:%.*]] = getelementptr inbounds i8, i8* [[OFFSET0]], i64 8
+
+// CHECK: [[MEMPP:%.*]] = getelementptr inbounds %struct.__va_list, %struct.__va_list* [[VA:%.*]], i32 0, i32 0
+// CHECK: [[MEMP:%.*]] = load i8*, i8** [[MEMPP]], align 8
+// CHECK: [[NEXTP:%.*]] = getelementptr inbounds i8, i8* [[MEMP]], i64 8
+// CHECK: store i8* [[NEXTP]], i8** [[MEMPP]], align 8
   va_list vl;
   va_start(vl, a);
   double result = va_arg(vl, struct { double a; }).a;
diff --git a/test/CodeGen/atomic-arm64.c b/test/CodeGen/atomic-arm64.c
index 98f27aba4f..e871536866 100644
--- a/test/CodeGen/atomic-arm64.c
+++ b/test/CodeGen/atomic-arm64.c
@@ -21,7 +21,7 @@ extern _Atomic(void*) a_pointer;
 extern _Atomic(pointer_pair_t) a_pointer_pair;
 extern _Atomic(pointer_quad_t) a_pointer_quad;
 
-// CHECK:    define void @test0()
+// CHECK-LABEL:define void @test0()
 // CHECK:      [[TEMP:%.*]] = alloca i8, align 1
 // CHECK-NEXT: store i8 1, i8* [[TEMP]]
 // CHECK-NEXT: [[T0:%.*]] = load i8, i8* [[TEMP]], align 1
@@ -30,7 +30,7 @@ void test0() {
   __c11_atomic_store(&a_bool, 1, memory_order_seq_cst);
 }
 
-// CHECK:    define void @test1()
+// CHECK-LABEL:define void @test1()
 // CHECK:      [[TEMP:%.*]] = alloca float, align 4
 // CHECK-NEXT: store float 3.000000e+00, float* [[TEMP]]
 // CHECK-NEXT: [[T0:%.*]] = bitcast float* [[TEMP]] to i32*
@@ -40,7 +40,7 @@ void test1() {
   __c11_atomic_store(&a_float, 3, memory_order_seq_cst);
 }
 
-// CHECK:    define void @test2()
+// CHECK-LABEL:define void @test2()
 // CHECK:      [[TEMP:%.*]] = alloca i8*, align 8
 // CHECK-NEXT: store i8* @a_bool, i8** [[TEMP]]
 // CHECK-NEXT: [[T0:%.*]] = bitcast i8** [[TEMP]] to i64*
@@ -50,18 +50,18 @@ void test2() {
   __c11_atomic_store(&a_pointer, &a_bool, memory_order_seq_cst);
 }
 
-// CHECK:    define void @test3(
+// CHECK-LABEL:define void @test3(
 // CHECK:      [[PAIR:%.*]] = alloca [[PAIR_T:%.*]], align 8
 // CHECK-NEXT: [[TEMP:%.*]] = alloca [[PAIR_T]], align 8
 // CHECK:      llvm.memcpy
 // CHECK-NEXT: [[T0:%.*]] = bitcast [[PAIR_T]]* [[TEMP]] to i128*
-// CHECK-NEXT: [[T1:%.*]] = load i128, i128* [[T0]], align 16
+// CHECK-NEXT: [[T1:%.*]] = load i128, i128* [[T0]], align 8
 // CHECK-NEXT: store atomic i128 [[T1]], i128* bitcast ([[PAIR_T]]* @a_pointer_pair to i128*) seq_cst, align 16
 void test3(pointer_pair_t pair) {
   __c11_atomic_store(&a_pointer_pair, pair, memory_order_seq_cst);
 }
 
-// CHECK:    define void @test4([[QUAD_T:%.*]]*
+// CHECK-LABEL:define void @test4(
 // CHECK:      [[TEMP:%.*]] = alloca [[QUAD_T:%.*]], align 8
 // CHECK-NEXT: [[T0:%.*]] = bitcast [[QUAD_T]]* [[TEMP]] to i8*
 // CHECK-NEXT: [[T1:%.*]] = bitcast [[QUAD_T]]* {{%.*}} to i8*
diff --git a/test/CodeGen/block-byref-aggr.c b/test/CodeGen/block-byref-aggr.c
index 910f6da3cc..7d146a2d47 100644
--- a/test/CodeGen/block-byref-aggr.c
+++ b/test/CodeGen/block-byref-aggr.c
@@ -16,7 +16,7 @@ void test0() {
 // CHECK:      [[A:%.*]] = alloca [[BYREF:%.*]], align 8
 // CHECK-NEXT: [[TEMP:%.*]] = alloca [[AGG]], align 4
 // CHECK:      [[RESULT:%.*]] = call i32 @makeAgg()
-// CHECK-NEXT: [[T0:%.*]] = getelementptr [[AGG]], [[AGG]]* [[TEMP]], i32 0, i32 0
+// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[AGG]], [[AGG]]* [[TEMP]], i32 0, i32 0
 // CHECK-NEXT: store i32 [[RESULT]], i32* [[T0]]
 //   Check that we properly assign into the forwarding pointer.
 // CHECK-NEXT: [[A_FORWARDING:%.*]] = getelementptr inbounds [[BYREF]], [[BYREF]]* [[A]], i32 0, i32 1
@@ -42,7 +42,7 @@ void test1() {
 // CHECK-NEXT: [[B:%.*]] = alloca [[B_BYREF:%.*]], align 8
 // CHECK-NEXT: [[TEMP:%.*]] = alloca [[AGG]], align 4
 // CHECK:      [[RESULT:%.*]] = call i32 @makeAgg()
-// CHECK-NEXT: [[T0:%.*]] = getelementptr [[AGG]], [[AGG]]* [[TEMP]], i32 0, i32 0
+// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[AGG]], [[AGG]]* [[TEMP]], i32 0, i32 0
 // CHECK-NEXT: store i32 [[RESULT]], i32* [[T0]]
 //   Check that we properly assign into the forwarding pointer, first for b:
 // CHECK-NEXT: [[B_FORWARDING:%.*]] = getelementptr inbounds [[B_BYREF]], [[B_BYREF]]* [[B]], i32 0, i32 1
diff --git a/test/CodeGen/c11atomics-ios.c b/test/CodeGen/c11atomics-ios.c
index a869982b17..138db696bf 100644
--- a/test/CodeGen/c11atomics-ios.c
+++ b/test/CodeGen/c11atomics-ios.c
@@ -103,21 +103,21 @@ void testStruct(_Atomic(S) *fp) {
 
 // CHECK-NEXT: [[P:%.*]] = load [[S]]*, [[S]]** [[FP]]
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[S]], [[S]]* [[P]], i32 0, i32 0
-// CHECK-NEXT: store i16 1, i16* [[T0]], align 2
+// CHECK-NEXT: store i16 1, i16* [[T0]], align 8
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[S]], [[S]]* [[P]], i32 0, i32 1
 // CHECK-NEXT: store i16 2, i16* [[T0]], align 2
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[S]], [[S]]* [[P]], i32 0, i32 2
-// CHECK-NEXT: store i16 3, i16* [[T0]], align 2
+// CHECK-NEXT: store i16 3, i16* [[T0]], align 4
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[S]], [[S]]* [[P]], i32 0, i32 3
 // CHECK-NEXT: store i16 4, i16* [[T0]], align 2
   __c11_atomic_init(fp, (S){1,2,3,4});
 
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[S]], [[S]]* [[X]], i32 0, i32 0
-// CHECK-NEXT: store i16 1, i16* [[T0]], align 2
+// CHECK-NEXT: store i16 1, i16* [[T0]], align 8
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[S]], [[S]]* [[X]], i32 0, i32 1
 // CHECK-NEXT: store i16 2, i16* [[T0]], align 2
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[S]], [[S]]* [[X]], i32 0, i32 2
-// CHECK-NEXT: store i16 3, i16* [[T0]], align 2
+// CHECK-NEXT: store i16 3, i16* [[T0]], align 4
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[S]], [[S]]* [[X]], i32 0, i32 3
 // CHECK-NEXT: store i16 4, i16* [[T0]], align 2
   _Atomic(S) x = (S){1,2,3,4};
@@ -157,29 +157,29 @@ void testPromotedStruct(_Atomic(PS) *fp) {
 // CHECK-NEXT: call void @llvm.memset.p0i8.i64(i8* [[T0]], i8 0, i64 8, i32 8, i1 false)
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[APS]], [[APS]]* [[P]], i32 0, i32 0
 // CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds [[PS]], [[PS]]* [[T0]], i32 0, i32 0
-// CHECK-NEXT: store i16 1, i16* [[T1]], align 2
+// CHECK-NEXT: store i16 1, i16* [[T1]], align 8
 // CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds [[PS]], [[PS]]* [[T0]], i32 0, i32 1
 // CHECK-NEXT: store i16 2, i16* [[T1]], align 2
 // CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds [[PS]], [[PS]]* [[T0]], i32 0, i32 2
-// CHECK-NEXT: store i16 3, i16* [[T1]], align 2
+// CHECK-NEXT: store i16 3, i16* [[T1]], align 4
   __c11_atomic_init(fp, (PS){1,2,3});
 
 // CHECK-NEXT: [[T0:%.*]] = bitcast [[APS]]* [[X]] to i8*
 // CHECK-NEXT: call void @llvm.memset.p0i8.i32(i8* [[T0]], i8 0, i32 8, i32 8, i1 false)
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[APS]], [[APS]]* [[X]], i32 0, i32 0
 // CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds [[PS]], [[PS]]* [[T0]], i32 0, i32 0
-// CHECK-NEXT: store i16 1, i16* [[T1]], align 2
+// CHECK-NEXT: store i16 1, i16* [[T1]], align 8
 // CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds [[PS]], [[PS]]* [[T0]], i32 0, i32 1
 // CHECK-NEXT: store i16 2, i16* [[T1]], align 2
 // CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds [[PS]], [[PS]]* [[T0]], i32 0, i32 2
-// CHECK-NEXT: store i16 3, i16* [[T1]], align 2
+// CHECK-NEXT: store i16 3, i16* [[T1]], align 4
   _Atomic(PS) x = (PS){1,2,3};
 
 // CHECK-NEXT: [[T0:%.*]] = load [[APS]]*, [[APS]]** [[FP]]
 // CHECK-NEXT: [[T1:%.*]] = bitcast [[APS]]* [[T0]] to i64*
 // CHECK-NEXT: [[T2:%.*]] = load atomic i64, i64* [[T1]] seq_cst, align 8
 // CHECK-NEXT: [[T3:%.*]] = bitcast [[APS]]* [[TMP0]] to i64*
-// CHECK-NEXT: store i64 [[T2]], i64* [[T3]], align 2
+// CHECK-NEXT: store i64 [[T2]], i64* [[T3]], align 8
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[APS]], [[APS]]* [[TMP0]], i32 0, i32 0
 // CHECK-NEXT: [[T1:%.*]] = bitcast [[PS]]* [[F]] to i8*
 // CHECK-NEXT: [[T2:%.*]] = bitcast [[PS]]* [[T0]] to i8*
diff --git a/test/CodeGen/c11atomics.c b/test/CodeGen/c11atomics.c
index d1e4478d7e..c6eaca6f2a 100644
--- a/test/CodeGen/c11atomics.c
+++ b/test/CodeGen/c11atomics.c
@@ -254,21 +254,21 @@ void testStruct(_Atomic(S) *fp) {
 
 // CHECK-NEXT: [[P:%.*]] = load [[S]]*, [[S]]** [[FP]]
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[S]], [[S]]* [[P]], i32 0, i32 0
-// CHECK-NEXT: store i16 1, i16* [[T0]], align 2
+// CHECK-NEXT: store i16 1, i16* [[T0]], align 8
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[S]], [[S]]* [[P]], i32 0, i32 1
 // CHECK-NEXT: store i16 2, i16* [[T0]], align 2
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[S]], [[S]]* [[P]], i32 0, i32 2
-// CHECK-NEXT: store i16 3, i16* [[T0]], align 2
+// CHECK-NEXT: store i16 3, i16* [[T0]], align 4
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[S]], [[S]]* [[P]], i32 0, i32 3
 // CHECK-NEXT: store i16 4, i16* [[T0]], align 2
   __c11_atomic_init(fp, (S){1,2,3,4});
 
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[S]], [[S]]* [[X]], i32 0, i32 0
-// CHECK-NEXT: store i16 1, i16* [[T0]], align 2
+// CHECK-NEXT: store i16 1, i16* [[T0]], align 8
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[S]], [[S]]* [[X]], i32 0, i32 1
 // CHECK-NEXT: store i16 2, i16* [[T0]], align 2
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[S]], [[S]]* [[X]], i32 0, i32 2
-// CHECK-NEXT: store i16 3, i16* [[T0]], align 2
+// CHECK-NEXT: store i16 3, i16* [[T0]], align 4
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[S]], [[S]]* [[X]], i32 0, i32 3
 // CHECK-NEXT: store i16 4, i16* [[T0]], align 2
   _Atomic(S) x = (S){1,2,3,4};
@@ -310,22 +310,22 @@ void testPromotedStruct(_Atomic(PS) *fp) {
 // CHECK-NEXT: call void @llvm.memset.p0i8.i64(i8* [[T0]], i8 0, i64 8, i32 8, i1 false)
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[APS]], [[APS]]* [[P]], i32 0, i32 0
 // CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds [[PS]], [[PS]]* [[T0]], i32 0, i32 0
-// CHECK-NEXT: store i16 1, i16* [[T1]], align 2
+// CHECK-NEXT: store i16 1, i16* [[T1]], align 8
 // CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds [[PS]], [[PS]]* [[T0]], i32 0, i32 1
 // CHECK-NEXT: store i16 2, i16* [[T1]], align 2
 // CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds [[PS]], [[PS]]* [[T0]], i32 0, i32 2
-// CHECK-NEXT: store i16 3, i16* [[T1]], align 2
+// CHECK-NEXT: store i16 3, i16* [[T1]], align 4
   __c11_atomic_init(fp, (PS){1,2,3});
 
 // CHECK-NEXT: [[T0:%.*]] = bitcast [[APS]]* [[X]] to i8*
 // CHECK-NEXT: call void @llvm.memset.p0i8.i32(i8* [[T0]], i8 0, i32 8, i32 8, i1 false)
 // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [[APS]], [[APS]]* [[X]], i32 0, i32 0
 // CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds [[PS]], [[PS]]* [[T0]], i32 0, i32 0
-// CHECK-NEXT: store i16 1, i16* [[T1]], align 2
+// CHECK-NEXT: store i16 1, i16* [[T1]], align 8
 // CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds [[PS]], [[PS]]* [[T0]], i32 0, i32 1
 // CHECK-NEXT: store i16 2, i16* [[T1]], align 2
 // CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds [[PS]], [[PS]]* [[T0]], i32 0, i32 2
-// CHECK-NEXT: store i16 3, i16* [[T1]], align 2
+// CHECK-NEXT: store i16 3, i16* [[T1]], align 4
   _Atomic(PS) x = (PS){1,2,3};
 
 // CHECK-NEXT: [[T0:%.*]] = load [[APS]]*, [[APS]]** [[FP]]
diff --git a/test/CodeGen/catch-undef-behavior.c b/test/CodeGen/catch-undef-behavior.c
index 013a4a4cce..739cf89f5b 100644
--- a/test/CodeGen/catch-undef-behavior.c
+++ b/test/CodeGen/catch-undef-behavior.c
@@ -269,7 +269,7 @@ int long_double_int_overflow(long double ld) {
   // CHECK-COMMON: %[[INBOUNDS:.*]] = and i1 %[[GE]], %[[LE]]
   // CHECK-COMMON-NEXT: br i1 %[[INBOUNDS]]
 
-  // CHECK-UBSAN: store x86_fp80 %[[F]], x86_fp80* %[[ALLOCA:.*]], !nosanitize
+  // CHECK-UBSAN: store x86_fp80 %[[F]], x86_fp80* %[[ALLOCA:.*]], align 16, !nosanitize
   // CHECK-UBSAN: %[[ARG:.*]] = ptrtoint x86_fp80* %[[ALLOCA]] to i64
   // CHECK-UBSAN: call void @__ubsan_handle_float_cast_overflow(i8* bitcast ({{.*}} @[[LINE_1300]] to i8*), i64 %[[ARG]]
 
diff --git a/test/CodeGen/exprs.c b/test/CodeGen/exprs.c
index 59afa802b1..f46b5748f2 100644
--- a/test/CodeGen/exprs.c
+++ b/test/CodeGen/exprs.c
@@ -127,9 +127,10 @@ int f11(long X) {
   return A[X];
 
 // CHECK: [[Xaddr:%[^ ]+]] = alloca i64, align 8
-// CHECK: load {{.*}}, {{.*}}* [[Xaddr]]
-// CHECK-NEXT: getelementptr inbounds [100 x i32], [100 x i32]* %A, i32 0, 
-// CHECK-NEXT: load i32, i32*
+// CHECK: [[A:%.*]] = alloca [100 x i32], align
+// CHECK: [[X:%.*]] = load {{.*}}, {{.*}}* [[Xaddr]]
+// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [100 x i32], [100 x i32]* [[A]], i64 0, i64 [[X]]
+// CHECK-NEXT: load i32, i32* [[T0]], align 4
 }
 
 int f12() {
diff --git a/test/CodeGen/ext-vector-member-alignment.c b/test/CodeGen/ext-vector-member-alignment.c
index 5f044b8a28..686051e193 100644
--- a/test/CodeGen/ext-vector-member-alignment.c
+++ b/test/CodeGen/ext-vector-member-alignment.c
@@ -14,14 +14,12 @@ void func(struct struct1* p, float *a, float *b, float c) {
   *a = p->position.y;
   *b = p->position[0];
   p->position[2] = c;
-  // FIXME: We should be able to come up with a more aggressive alignment
-  // estimate.
   // CHECK: @func
-  // CHECK: load <4 x float>, <4 x float>* {{%.*}}, align 1
-  // CHECK: store <4 x float> {{%.*}}, <4 x float>* {{%.*}}, align 1
-  // CHECK: load <4 x float>, <4 x float>* {{%.*}}, align 1
-  // CHECK: load <4 x float>, <4 x float>* {{%.*}}, align 1
-  // CHECK: load <4 x float>, <4 x float>* {{%.*}}, align 1
-  // CHECK: store <4 x float> {{%.*}}, <4 x float>* {{%.*}}, align 1
+  // CHECK: load <4 x float>, <4 x float>* {{%.*}}, align 4
+  // CHECK: store <4 x float> {{%.*}}, <4 x float>* {{%.*}}, align 4
+  // CHECK: load <4 x float>, <4 x float>* {{%.*}}, align 4
+  // CHECK: load <4 x float>, <4 x float>* {{%.*}}, align 4
+  // CHECK: load <4 x float>, <4 x float>* {{%.*}}, align 4
+  // CHECK: store <4 x float> {{%.*}}, <4 x float>* {{%.*}}, align 4
   // CHECK: ret void
 }
diff --git a/test/CodeGen/mips-varargs.c b/test/CodeGen/mips-varargs.c
index 891769c711..6608017062 100644
--- a/test/CodeGen/mips-varargs.c
+++ b/test/CodeGen/mips-varargs.c
@@ -1,9 +1,9 @@
-// RUN: %clang_cc1 -triple mips-unknown-linux -o - -O1 -emit-llvm %s | FileCheck %s -check-prefix=ALL -check-prefix=O32
-// RUN: %clang_cc1 -triple mipsel-unknown-linux -o - -O1 -emit-llvm %s | FileCheck %s -check-prefix=ALL -check-prefix=O32
-// RUN: %clang_cc1 -triple mips64-unknown-linux -o - -O1 -emit-llvm  -target-abi n32 %s | FileCheck %s -check-prefix=ALL -check-prefix=N32 -check-prefix=NEW
-// RUN: %clang_cc1 -triple mips64-unknown-linux -o - -O1 -emit-llvm  -target-abi n32 %s | FileCheck %s -check-prefix=ALL -check-prefix=N32 -check-prefix=NEW
-// RUN: %clang_cc1 -triple mips64-unknown-linux -o - -O1 -emit-llvm %s | FileCheck %s -check-prefix=ALL -check-prefix=N64 -check-prefix=NEW
-// RUN: %clang_cc1 -triple mips64el-unknown-linux -o - -O1 -emit-llvm %s | FileCheck %s -check-prefix=ALL -check-prefix=N64 -check-prefix=NEW
+// RUN: %clang_cc1 -triple mips-unknown-linux -o - -emit-llvm %s | FileCheck %s -check-prefix=ALL -check-prefix=O32
+// RUN: %clang_cc1 -triple mipsel-unknown-linux -o - -emit-llvm %s | FileCheck %s -check-prefix=ALL -check-prefix=O32
+// RUN: %clang_cc1 -triple mips64-unknown-linux -o - -emit-llvm  -target-abi n32 %s | FileCheck %s -check-prefix=ALL -check-prefix=N32 -check-prefix=NEW
+// RUN: %clang_cc1 -triple mips64-unknown-linux -o - -emit-llvm  -target-abi n32 %s | FileCheck %s -check-prefix=ALL -check-prefix=N32 -check-prefix=NEW
+// RUN: %clang_cc1 -triple mips64-unknown-linux -o - -emit-llvm %s | FileCheck %s -check-prefix=ALL -check-prefix=N64 -check-prefix=NEW
+// RUN: %clang_cc1 -triple mips64el-unknown-linux -o - -emit-llvm %s | FileCheck %s -check-prefix=ALL -check-prefix=N64 -check-prefix=NEW
 
 #include <stdarg.h>
 
@@ -24,75 +24,31 @@ int test_i32(char *fmt, ...) {
 // O32:   %va = alloca i8*, align [[PTRALIGN:4]]
 // N32:   %va = alloca i8*, align [[PTRALIGN:4]]
 // N64:   %va = alloca i8*, align [[PTRALIGN:8]]
+// ALL:   [[V:%.*]] = alloca i32, align 4
+// NEW:   [[PROMOTION_TEMP:%.*]] = alloca i32, align 4
 //
-// ALL:   [[VA1:%.+]] = bitcast i8** %va to i8*
-// ALL:   call void @llvm.va_start(i8* [[VA1]])
-//
-// O32:   [[TMP0:%.+]] = bitcast i8** %va to i32**
-// O32:   [[AP_CUR:%.+]] = load i32*, i32** [[TMP0]], align [[PTRALIGN]]
-// NEW:   [[TMP0:%.+]] = bitcast i8** %va to i64**
-// NEW:   [[AP_CUR:%.+]] = load i64*, i64** [[TMP0]], align [[PTRALIGN]]
+// ALL:   [[VA:%.+]] = bitcast i8** %va to i8*
+// ALL:   call void @llvm.va_start(i8* [[VA]])
+// ALL:   [[AP_CUR:%.+]] = load i8*, i8** %va, align [[PTRALIGN]]
+// O32:   [[AP_NEXT:%.+]] = getelementptr inbounds i8, i8* [[AP_CUR]], [[INTPTR_T:i32]] [[CHUNKSIZE:4]]
+// NEW:   [[AP_NEXT:%.+]] = getelementptr inbounds i8, i8* [[AP_CUR]], [[INTPTR_T:i32|i64]] [[CHUNKSIZE:8]]
 //
-// O32:   [[AP_NEXT:%.+]] = getelementptr i32, i32* [[AP_CUR]], i32 1
-// NEW:   [[AP_NEXT:%.+]] = getelementptr i64, i64* [[AP_CUR]], {{i32|i64}} 1
+// ALL:   store i8* [[AP_NEXT]], i8** %va, align [[PTRALIGN]]
 //
-// O32:   store i32* [[AP_NEXT]], i32** [[TMP0]], align [[PTRALIGN]]
-// NEW:   store i64* [[AP_NEXT]], i64** [[TMP0]], align [[PTRALIGN]]
+// O32:   [[AP_CAST:%.+]] = bitcast i8* [[AP_CUR]] to [[CHUNK_T:i32]]*
+// O32:   [[ARG:%.+]] = load i32, i32* [[AP_CAST]], align [[CHUNKALIGN:4]]
 //
-// O32:   [[ARG1:%.+]] = load i32, i32* [[AP_CUR]], align 4
-// NEW:   [[TMP2:%.+]] = load i64, i64* [[AP_CUR]], align 8
-// NEW:   [[ARG1:%.+]] = trunc i64 [[TMP2]] to i32
+// N32:   [[AP_CAST:%.+]] = bitcast i8* [[AP_CUR]] to [[CHUNK_T:i64]]*
+// N32:   [[TMP:%.+]] = load i64, i64* [[AP_CAST]], align [[CHUNKALIGN:8]]
+// N64:   [[AP_CAST:%.+]] = bitcast i8* [[AP_CUR]] to [[CHUNK_T:i64]]*
+// N64:   [[TMP:%.+]] = load i64, i64* [[AP_CAST]], align [[CHUNKALIGN:8]]
+// NEW:   [[TMP2:%.+]] = trunc i64 [[TMP]] to i32
+// NEW:   store i32 [[TMP2]], i32* [[PROMOTION_TEMP]], align 4
+// NEW:   [[ARG:%.+]] = load i32, i32* [[PROMOTION_TEMP]], align 4
+// ALL:   store i32 [[ARG]], i32* [[V]], align 4
 //
-// ALL:   call void @llvm.va_end(i8* [[VA1]])
-// ALL:   ret i32 [[ARG1]]
-// ALL: }
-
-int test_i32_2args(char *fmt, ...) {
-  va_list va;
-
-  va_start(va, fmt);
-  int v1 = va_arg(va, int);
-  int v2 = va_arg(va, int);
-  va_end(va);
-
-  return v1 + v2;
-}
-
-// ALL-LABEL: define i32 @test_i32_2args(i8*{{.*}} %fmt, ...)
-//
-// ALL:   %va = alloca i8*, align [[PTRALIGN]]
 // ALL:   [[VA1:%.+]] = bitcast i8** %va to i8*
-// ALL:   call void @llvm.va_start(i8* [[VA1]])
-//
-// O32:   [[TMP0:%.+]] = bitcast i8** %va to i32**
-// O32:   [[AP_CUR:%.+]] = load i32*, i32** [[TMP0]], align [[PTRALIGN]]
-// NEW:   [[TMP0:%.+]] = bitcast i8** %va to i64**
-// NEW:   [[AP_CUR:%.+]] = load i64*, i64** [[TMP0]], align [[PTRALIGN]]
-//
-// O32:   [[AP_NEXT1:%.+]] = getelementptr i32, i32* [[AP_CUR]], i32 1
-// NEW:   [[AP_NEXT1:%.+]] = getelementptr i64, i64* [[AP_CUR]], [[INTPTR_T:i32|i64]] 1
-//
-// O32:   store i32* [[AP_NEXT1]], i32** [[TMP0]], align [[PTRALIGN]]
-// FIXME: N32 optimised this store out. Why only for this ABI?
-// N64:   store i64* [[AP_NEXT1]], i64** [[TMP0]], align [[PTRALIGN]]
-//
-// O32:   [[ARG1:%.+]] = load i32, i32* [[AP_CUR]], align 4
-// NEW:   [[TMP3:%.+]] = load i64, i64* [[AP_CUR]], align 8
-// NEW:   [[ARG1:%.+]] = trunc i64 [[TMP3]] to i32
-//
-// O32:   [[AP_NEXT2:%.+]] = getelementptr i32, i32* [[AP_CUR]], i32 2
-// NEW:   [[AP_NEXT2:%.+]] = getelementptr i64, i64* [[AP_CUR]], [[INTPTR_T]] 2
-//
-// O32:   store i32* [[AP_NEXT2]], i32** [[TMP0]], align [[PTRALIGN]]
-// NEW:   store i64* [[AP_NEXT2]], i64** [[TMP0]], align [[PTRALIGN]]
-//
-// O32:   [[ARG2:%.+]] = load i32, i32* [[AP_NEXT1]], align 4
-// NEW:   [[TMP4:%.+]] = load i64, i64* [[AP_NEXT1]], align 8
-// NEW:   [[ARG2:%.+]] = trunc i64 [[TMP4]] to i32
-//
 // ALL:   call void @llvm.va_end(i8* [[VA1]])
-// ALL:   [[ADD:%.+]] = add nsw i32 [[ARG2]], [[ARG1]]
-// ALL:   ret i32 [[ADD]]
 // ALL: }
 
 long long test_i64(char *fmt, ...) {
@@ -108,32 +64,25 @@ long long test_i64(char *fmt, ...) {
 // ALL-LABEL: define i64 @test_i64(i8*{{.*}} %fmt, ...)
 //
 // ALL:   %va = alloca i8*, align [[PTRALIGN]]
-// ALL:   [[VA1:%.+]] = bitcast i8** %va to i8*
-// ALL:   call void @llvm.va_start(i8* [[VA1]])
-//
-// O32:   [[TMP0:%.+]] = bitcast i8** %va to i32*
-// O32:   [[AP_CUR:%.+]] = load [[INTPTR_T:i32]], i32* [[TMP0]], align [[PTRALIGN]]
-// NEW:   [[TMP0:%.+]] = bitcast i8** %va to i64**
-// NEW:   [[AP_CUR:%.+]] = load i64*, i64** [[TMP0]], align [[PTRALIGN]]
+// ALL:   [[VA:%.+]] = bitcast i8** %va to i8*
+// ALL:   call void @llvm.va_start(i8* [[VA]])
+// ALL:   [[AP_CUR:%.+]] = load i8*, i8** %va, align [[PTRALIGN]]
 //
 // i64 is 8-byte aligned, while this is within O32's stack alignment there's no
 // guarantee that the offset is still 8-byte aligned after earlier reads.
-// O32:   [[PTR1:%.+]] = add i32 [[AP_CUR]], 7
-// O32:   [[PTR2:%.+]] = and i32 [[PTR1]], -8
-// O32:   [[PTR3:%.+]] = inttoptr [[INTPTR_T]] [[PTR2]] to i64*
-// O32:   [[PTR4:%.+]] = inttoptr [[INTPTR_T]] [[PTR2]] to i8*
-//
-// O32:   [[AP_NEXT:%.+]] = getelementptr i8, i8* [[PTR4]], [[INTPTR_T]] 8
-// NEW:   [[AP_NEXT:%.+]] = getelementptr i64, i64* [[AP_CUR]], [[INTPTR_T:i32|i64]] 1
+// O32:   [[TMP1:%.+]] = ptrtoint i8* [[AP_CUR]] to i32
+// O32:   [[TMP2:%.+]] = add i32 [[TMP1]], 7
+// O32:   [[TMP3:%.+]] = and i32 [[TMP2]], -8
+// O32:   [[AP_CUR:%.+]] = inttoptr i32 [[TMP3]] to i8*
 //
-// O32:   store i8* [[AP_NEXT]], i8** %va, align [[PTRALIGN]]
-// NEW:   store i64* [[AP_NEXT]], i64** [[TMP0]], align [[PTRALIGN]]
+// ALL:   [[AP_NEXT:%.+]] = getelementptr inbounds i8, i8* [[AP_CUR]], [[INTPTR_T]] 8
+// ALL:   store i8* [[AP_NEXT]], i8** %va, align [[PTRALIGN]]
 //
-// O32:   [[ARG1:%.+]] = load i64, i64* [[PTR3]], align 8
-// NEW:   [[ARG1:%.+]] = load i64, i64* [[AP_CUR]], align 8
+// ALL:   [[AP_CAST:%.*]] = bitcast i8* [[AP_CUR]] to i64*
+// ALL:   [[ARG:%.+]] = load i64, i64* [[AP_CAST]], align 8
 //
+// ALL:   [[VA1:%.+]] = bitcast i8** %va to i8*
 // ALL:   call void @llvm.va_end(i8* [[VA1]])
-// ALL:   ret i64 [[ARG1]]
 // ALL: }
 
 char *test_ptr(char *fmt, ...) {
@@ -148,41 +97,30 @@ char *test_ptr(char *fmt, ...) {
 
 // ALL-LABEL: define i8* @test_ptr(i8*{{.*}} %fmt, ...)
 //
-// O32:   %va = alloca i8*, align [[PTRALIGN:4]]
-// N32:   %va = alloca i8*, align [[PTRALIGN:4]]
-// N64:   %va = alloca i8*, align [[PTRALIGN:8]]
-//
-// ALL:   [[VA1:%.+]] = bitcast i8** %va to i8*
-// ALL:   call void @llvm.va_start(i8* [[VA1]])
+// ALL:   %va = alloca i8*, align [[PTRALIGN]]
+// ALL:   [[V:%.*]] = alloca i8*, align [[PTRALIGN]]
+// N32:   [[AP_CAST:%.+]] = alloca i8*, align 4
+// ALL:   [[VA:%.+]] = bitcast i8** %va to i8*
+// ALL:   call void @llvm.va_start(i8* [[VA]])
+// ALL:   [[AP_CUR:%.+]] = load i8*, i8** %va, align [[PTRALIGN]]
+// ALL:   [[AP_NEXT:%.+]] = getelementptr inbounds i8, i8* [[AP_CUR]], [[INTPTR_T]] [[CHUNKSIZE]]
+// ALL:   store i8* [[AP_NEXT]], i8** %va, align [[PTRALIGN]]
 //
-// O32:   [[TMP0:%.+]] = bitcast i8** %va to i8***
-// O32:   [[AP_CUR:%.+]] = load i8**, i8*** [[TMP0]], align [[PTRALIGN]]
-// N32 differs because the vararg is not a N32 pointer. It's been promoted to 64-bit.
-// N32:   [[TMP0:%.+]] = bitcast i8** %va to i64**
-// N32:   [[AP_CUR:%.+]] = load i64*, i64** [[TMP0]], align [[PTRALIGN]]
-// N64:   [[TMP0:%.+]] = bitcast i8** %va to i8***
-// N64:   [[AP_CUR:%.+]] = load i8**, i8*** [[TMP0]], align [[PTRALIGN]]
-//
-// O32:   [[AP_NEXT:%.+]] = getelementptr i8*, i8** [[AP_CUR]], i32 1
-// N32 differs because the vararg is not a N32 pointer. It's been promoted to 64-bit.
-// N32:   [[AP_NEXT:%.+]] = getelementptr i64, i64* [[AP_CUR]], {{i32|i64}} 1
-// N64:   [[AP_NEXT:%.+]] = getelementptr i8*, i8** [[AP_CUR]], {{i32|i64}} 1
-//
-// O32:   store i8** [[AP_NEXT]], i8*** [[TMP0]], align [[PTRALIGN]]
-// N32 differs because the vararg is not a N32 pointer. It's been promoted to 64-bit.
-// N32:   store i64* [[AP_NEXT]], i64** [[TMP0]], align [[PTRALIGN]]
-// N64:   store i8** [[AP_NEXT]], i8*** [[TMP0]], align [[PTRALIGN]]
-//
-// O32:   [[ARG1:%.+]] = load i8*, i8** [[AP_CUR]], align 4
-// N32 differs because the vararg is not a N32 pointer. It's been promoted to
-// 64-bit so we must truncate the excess and bitcast to a N32 pointer.
-// N32:   [[TMP2:%.+]] = load i64, i64* [[AP_CUR]], align 8
+// When the chunk size matches the pointer size, this is easy.
+// O32:   [[AP_CAST:%.+]] = bitcast i8* [[AP_CUR]] to i8**
+// N64:   [[AP_CAST:%.+]] = bitcast i8* [[AP_CUR]] to i8**
+// Otherwise we need a promotion temporary.
+// N32:   [[TMP1:%.+]] = bitcast i8* [[AP_CUR]] to i64*
+// N32:   [[TMP2:%.+]] = load i64, i64* [[TMP1]], align 8
 // N32:   [[TMP3:%.+]] = trunc i64 [[TMP2]] to i32
-// N32:   [[ARG1:%.+]] = inttoptr i32 [[TMP3]] to i8*
-// N64:   [[ARG1:%.+]] = load i8*, i8** [[AP_CUR]], align 8
+// N32:   [[PTR:%.+]] = inttoptr i32 [[TMP3]] to i8*
+// N32:   store i8* [[PTR]], i8** [[AP_CAST]], align 4
+//
+// ALL:   [[ARG:%.+]] = load i8*, i8** [[AP_CAST]], align [[PTRALIGN]]
+// ALL:   store i8* [[ARG]], i8** [[V]], align [[PTRALIGN]]
 //
+// ALL:   [[VA:%.+]] = bitcast i8** %va to i8*
 // ALL:   call void @llvm.va_end(i8* [[VA1]])
-// ALL:   ret i8* [[ARG1]]
 // ALL: }
 
 int test_v4i32(char *fmt, ...) {
@@ -198,31 +136,33 @@ int test_v4i32(char *fmt, ...) {
 // ALL-LABEL: define i32 @test_v4i32(i8*{{.*}} %fmt, ...)
 //
 // ALL:   %va = alloca i8*, align [[PTRALIGN]]
+// ALL:   [[V]] = alloca <4 x i32>, align 16
 // ALL:   [[VA1:%.+]] = bitcast i8** %va to i8*
 // ALL:   call void @llvm.va_start(i8* [[VA1]])
-//
-// O32:   [[TMP0:%.+]] = bitcast i8** %va to i32*
-// N32:   [[TMP0:%.+]] = bitcast i8** %va to i32*
-// N64:   [[TMP0:%.+]] = bitcast i8** %va to i64*
-//
-// O32:   [[PTR0:%.+]] = load [[INTPTR_T:i32]], i32* [[TMP0]], align [[PTRALIGN]]
-// N32:   [[PTR0:%.+]] = load [[INTPTR_T:i32]], i32* [[TMP0]], align [[PTRALIGN]]
-// N64:   [[PTR0:%.+]] = load [[INTPTR_T:i64]], i64* [[TMP0]], align [[PTRALIGN]]
+// ALL:   [[AP_CUR:%.+]] = load i8*, i8** %va, align [[PTRALIGN]]
 //
 // Vectors are 16-byte aligned, however the O32 ABI has a maximum alignment of
 // 8-bytes since the base of the stack is 8-byte aligned.
-// O32:   [[PTR1:%.+]] = add i32 [[PTR0]], 7
-// O32:   [[PTR2:%.+]] = and i32 [[PTR1]], -8
+// O32:   [[TMP1:%.+]] = ptrtoint i8* [[AP_CUR]] to i32
+// O32:   [[TMP2:%.+]] = add i32 [[TMP1]], 7
+// O32:   [[TMP3:%.+]] = and i32 [[TMP2]], -8
+// O32:   [[AP_CUR:%.+]] = inttoptr i32 [[TMP3]] to i8*
 //
-// NEW:   [[PTR1:%.+]] = add [[INTPTR_T]] [[PTR0]], 15
-// NEW:   [[PTR2:%.+]] = and [[INTPTR_T]] [[PTR1]], -16
+// NEW:   [[TMP1:%.+]] = ptrtoint i8* [[AP_CUR]] to [[INTPTR_T]]
+// NEW:   [[TMP2:%.+]] = add [[INTPTR_T]] [[TMP1]], 15
+// NEW:   [[TMP3:%.+]] = and [[INTPTR_T]] [[TMP2]], -16
+// NEW:   [[AP_CUR:%.+]] = inttoptr [[INTPTR_T]] [[TMP3]] to i8*
 //
-// ALL:   [[PTR3:%.+]] = inttoptr [[INTPTR_T]] [[PTR2]] to <4 x i32>*
-// ALL:   [[PTR4:%.+]] = inttoptr [[INTPTR_T]] [[PTR2]] to i8*
-// ALL:   [[AP_NEXT:%.+]] = getelementptr i8, i8* [[PTR4]], [[INTPTR_T]] 16
+// ALL:   [[AP_NEXT:%.+]] = getelementptr inbounds i8, i8* [[AP_CUR]], [[INTPTR_T]] 16
 // ALL:   store i8* [[AP_NEXT]], i8** %va, align [[PTRALIGN]]
-// ALL:   [[PTR5:%.+]] = load <4 x i32>, <4 x i32>* [[PTR3]], align 16
+//
+// ALL:   [[AP_CAST:%.+]] = bitcast i8* [[AP_CUR]] to <4 x i32>*
+// O32:   [[ARG:%.+]] = load <4 x i32>, <4 x i32>* [[AP_CAST]], align 8
+// N64:   [[ARG:%.+]] = load <4 x i32>, <4 x i32>* [[AP_CAST]], align 16
+// ALL:   store <4 x i32> [[ARG]], <4 x i32>* [[V]], align 16
+//
+// ALL:   [[VA1:%.+]] = bitcast i8** %va to i8*
 // ALL:   call void @llvm.va_end(i8* [[VA1]])
-// ALL:   [[VECEXT:%.+]] = extractelement <4 x i32> [[PTR5]], i32 0
+// ALL:   [[VECEXT:%.+]] = extractelement <4 x i32> {{.*}}, i32 0
 // ALL:   ret i32 [[VECEXT]]
 // ALL: }
diff --git a/test/CodeGen/object-size.c b/test/CodeGen/object-size.c
index 367318d6cc..e69764087e 100644
--- a/test/CodeGen/object-size.c
+++ b/test/CodeGen/object-size.c
@@ -15,7 +15,7 @@ int gi, gj;
 
 // CHECK-LABEL: define void @test1
 void test1() {
-  // CHECK:     = call i8* @__strcpy_chk(i8* getelementptr inbounds ([63 x i8], [63 x i8]* @gbuf, i32 0, i64 4), i8* getelementptr inbounds ([9 x i8], [9 x i8]* @.str, i32 0, i32 0), i64 59)
+  // CHECK:     = call i8* @__strcpy_chk(i8* getelementptr inbounds ([63 x i8], [63 x i8]* @gbuf, i64 0, i64 4), i8* getelementptr inbounds ([9 x i8], [9 x i8]* @.str, i32 0, i32 0), i64 59)
   strcpy(&gbuf[4], "Hi there");
 }
 
@@ -33,7 +33,7 @@ void test3() {
 
 // CHECK-LABEL: define void @test4
 void test4() {
-  // CHECK:     = call i8* @__strcpy_chk(i8* getelementptr inbounds ([63 x i8], [63 x i8]* @gbuf, i32 0, i64 -1), i8* getelementptr inbounds ([9 x i8], [9 x i8]* @.str, i32 0, i32 0), i64 0)
+  // CHECK:     = call i8* @__strcpy_chk(i8* getelementptr inbounds ([63 x i8], [63 x i8]* @gbuf, i64 0, i64 -1), i8* getelementptr inbounds ([9 x i8], [9 x i8]* @.str, i32 0, i32 0), i64 0)
   strcpy((char*)(void*)&gbuf[-1], "Hi there");
 }
 
diff --git a/test/CodeGen/packed-arrays.c b/test/CodeGen/packed-arrays.c
index 993d88e277..bb742c6f31 100644
--- a/test/CodeGen/packed-arrays.c
+++ b/test/CodeGen/packed-arrays.c
@@ -64,10 +64,12 @@ int f0_b(struct s0 *a) {
   return *(a->x + 1);
 }
 
+// Note that 'y' still causes struct s1 to be four-byte aligned.
+
 // Note that we are incompatible with GCC on this example.
 // 
 // CHECK-LABEL: define i32 @f1_a
-// CHECK:   load i32, i32* %{{.*}}, align 1
+// CHECK:   load i32, i32* %{{.*}}, align 4
 // CHECK: }
 // CHECK-LABEL: define i32 @f1_b
 // CHECK:   load i32, i32* %{{.*}}, align 4
@@ -79,7 +81,7 @@ int f0_b(struct s0 *a) {
 // CHECK:   load i32, i32* %{{.*}}, align 4
 // CHECK: }
 // CHECK-LABEL: define i32 @f1_d
-// CHECK:   load i32, i32* %{{.*}}, align 1
+// CHECK:   load i32, i32* %{{.*}}, align 4
 // CHECK: }
 int f1_a(struct s1 *a) {
   return a->x[1];
diff --git a/test/CodeGen/packed-structure.c b/test/CodeGen/packed-structure.c
index 8de31d6a81..7d1183dc5c 100644
--- a/test/CodeGen/packed-structure.c
+++ b/test/CodeGen/packed-structure.c
@@ -25,7 +25,7 @@ int s0_load_x(struct s0 *a) { return a->x; }
 // with align 1 (in 2363.1 at least).
 //
 // CHECK-FUNCTIONS-LABEL: define i32 @s0_load_y
-// CHECK-FUNCTIONS: [[s0_load_y:%.*]] = load i32, i32* {{.*}}, align 1
+// CHECK-FUNCTIONS: [[s0_load_y:%.*]] = load i32, i32* {{.*}}, align 4
 // CHECK-FUNCTIONS: ret i32 [[s0_load_y]]
 int s0_load_y(struct s0 *a) { return a->y; }
 // CHECK-FUNCTIONS-LABEL: define void @s0_copy
@@ -95,6 +95,6 @@ int s3_1 = __alignof(((struct s3*) 0)->anInt);
 // CHECK-FUNCTIONS-LABEL: define i32 @test3(
 int test3(struct s3 *ptr) {
   // CHECK-FUNCTIONS:      [[PTR:%.*]] = getelementptr inbounds {{%.*}}, {{%.*}}* {{%.*}}, i32 0, i32 1
-  // CHECK-FUNCTIONS-NEXT: load i32, i32* [[PTR]], align 1
+  // CHECK-FUNCTIONS-NEXT: load i32, i32* [[PTR]], align 2
   return ptr->anInt;
 }
diff --git a/test/CodeGen/ppc-varargs-struct.c b/test/CodeGen/ppc-varargs-struct.c
index 1c983c0e43..a6ba13799d 100644
--- a/test/CodeGen/ppc-varargs-struct.c
+++ b/test/CodeGen/ppc-varargs-struct.c
@@ -19,89 +19,69 @@ void testva (int n, ...)
 // CHECK: bitcast %struct.x* %t to i8*
 // CHECK: bitcast %struct.x* %{{[0-9]+}} to i8*
 // CHECK: call void @llvm.memcpy
-// CHECK-PPC:  [[ARRAYDECAY:%[a-z0-9]+]] = getelementptr inbounds [1 x %struct.__va_list_tag], [1 x %struct.__va_list_tag]* %ap, i32 0, i32 0
-// CHECK-PPC-NEXT:  [[GPRPTR:%[a-z0-9]+]] = bitcast %struct.__va_list_tag* [[ARRAYDECAY]] to i8*
-// CHECK-PPC-NEXT:  [[ZERO:%[0-9]+]] = ptrtoint i8* [[GPRPTR]] to i32
-// CHECK-PPC-NEXT:  [[ONE:%[0-9]+]] = add i32 [[ZERO]], 1
-// CHECK-PPC-NEXT:  [[TWO:%[0-9]+]] = inttoptr i32 [[ONE]] to i8*
-// CHECK-PPC-NEXT:  [[THREE:%[0-9]+]] = add i32 [[ONE]], 3
-// CHECK-PPC-NEXT:  [[FOUR:%[0-9]+]] = inttoptr i32 [[THREE]] to i8**
-// CHECK-PPC-NEXT:  [[FIVE:%[0-9]+]] = add i32 [[THREE]], 4
-// CHECK-PPC-NEXT:  [[SIX:%[0-9]+]] = inttoptr i32 [[FIVE]] to i8**
-// CHECK-PPC-NEXT:  [[GPR:%[a-z0-9]+]] = load i8, i8* [[GPRPTR]]
-// CHECK-PPC-NEXT:  [[FPR:%[a-z0-9]+]] = load i8, i8* [[TWO]] 
-// CHECK-PPC-NEXT:  [[OVERFLOW_AREA:%[a-z_0-9]+]] = load i8*, i8** [[FOUR]]
-// CHECK-PPC-NEXT:  [[SEVEN:%[0-9]+]] = ptrtoint i8* [[OVERFLOW_AREA]] to i32
-// CHECK-PPC-NEXT:  [[REGSAVE_AREA:%[a-z_0-9]+]] = load i8*, i8** [[SIX]]
-// CHECK-PPC-NEXT:  [[EIGHT:%[0-9]+]] = ptrtoint i8* [[REGSAVE_AREA]] to i32
-// CHECK-PPC-NEXT:  [[COND:%[a-z0-9]+]] = icmp ult i8 [[GPR]], 8
-// CHECK-PPC-NEXT:  [[NINE:%[0-9]+]] = mul i8 [[GPR]], 4
-// CHECK-PPC-NEXT:  [[TEN:%[0-9]+]] = sext i8 [[NINE]] to i32
-// CHECK-PPC-NEXT:  [[ELEVEN:%[0-9]+]] = add i32 [[EIGHT]], [[TEN]]
-// CHECK-PPC-NEXT:  br i1 [[COND]], label [[USING_REGS:%[a-z_0-9]+]], label [[USING_OVERFLOW:%[a-z_0-9]+]]
+
+// CHECK-PPC:  [[ARRAYDECAY:%.+]] = getelementptr inbounds [1 x %struct.__va_list_tag], [1 x %struct.__va_list_tag]* %ap, i32 0, i32 0
+// CHECK-PPC-NEXT:  [[GPRPTR:%.+]] = getelementptr inbounds %struct.__va_list_tag, %struct.__va_list_tag* [[ARRAYDECAY]], i32 0, i32 0
+// CHECK-PPC-NEXT:  [[GPR:%.+]] = load i8, i8* [[GPRPTR]], align 4
+// CHECK-PPC-NEXT:  [[COND:%.+]] = icmp ult i8 [[GPR]], 8
+// CHECK-PPC-NEXT:  br i1 [[COND]], label %[[USING_REGS:[a-z_0-9]+]], label %[[USING_OVERFLOW:[a-z_0-9]+]]
 //
-// CHECK-PPC1:[[USING_REGS]]
-// CHECK-PPC:  [[TWELVE:%[0-9]+]] = inttoptr i32 [[ELEVEN]] to %struct.x*
-// CHECK-PPC-NEXT:  [[THIRTEEN:%[0-9]+]] = add i8 [[GPR]], 1
-// CHECK-PPC-NEXT:  store i8 [[THIRTEEN]], i8* [[GPRPTR]]
-// CHECK-PPC-NEXT:  br label [[CONT:%[a-z0-9]+]]
+// CHECK-PPC:[[USING_REGS]]
+// CHECK-PPC-NEXT:  [[REGSAVE_AREA_P:%.+]] = getelementptr inbounds %struct.__va_list_tag, %struct.__va_list_tag* [[ARRAYDECAY]], i32 0, i32 4
+// CHECK-PPC-NEXT:  [[REGSAVE_AREA:%.+]] = load i8*, i8** [[REGSAVE_AREA_P]], align 4
+// CHECK-PPC-NEXT:  [[OFFSET:%.+]] = mul i8 [[GPR]], 4
+// CHECK-PPC-NEXT:  [[RAW_REGADDR:%.+]] = getelementptr inbounds i8, i8* [[REGSAVE_AREA]], i8 [[OFFSET]]
+// CHECK-PPC-NEXT:  [[REGADDR:%.+]] = bitcast i8* [[RAW_REGADDR]] to %struct.x**
+// CHECK-PPC-NEXT:  [[USED_GPR:%[0-9]+]] = add i8 [[GPR]], 1
+// CHECK-PPC-NEXT:  store i8 [[USED_GPR]], i8* [[GPRPTR]], align 4
+// CHECK-PPC-NEXT:  br label %[[CONT:[a-z0-9]+]]
 //
-// CHECK-PPC1:[[USING_OVERFLOW]]
-// CHECK-PPC:  [[FOURTEEN:%[0-9]+]] = inttoptr i32 [[SEVEN]] to %struct.x*
-// CHECK-PPC-NEXT:  [[FIFTEEN:%[0-9]+]] = add i32 [[SEVEN]], 4
-// CHECK-PPC-NEXT:  [[SIXTEEN:%[0-9]+]] = inttoptr i32 [[FIFTEEN]] to i8*
-// CHECK-PPC-NEXT:  store i8* [[SIXTEEN]], i8** [[FOUR]]
-// CHECK-PPC-NEXT:  br label [[CONT]]
+// CHECK-PPC:[[USING_OVERFLOW]]
+// CHECK-PPC-NEXT:  [[OVERFLOW_AREA_P:%[0-9]+]] = getelementptr inbounds %struct.__va_list_tag, %struct.__va_list_tag* [[ARRAYDECAY]], i32 0, i32 3
+// CHECK-PPC-NEXT:  [[OVERFLOW_AREA:%.+]] = load i8*, i8** [[OVERFLOW_AREA_P]], align 4
+// CHECK-PPC-NEXT:  [[MEMADDR:%.+]] = bitcast i8* [[OVERFLOW_AREA]] to %struct.x**
+// CHECK-PPC-NEXT:  [[NEW_OVERFLOW_AREA:%[0-9]+]] = getelementptr inbounds i8, i8* [[OVERFLOW_AREA]], i32 4
+// CHECK-PPC-NEXT:  store i8* [[NEW_OVERFLOW_AREA]], i8** [[OVERFLOW_AREA_P]]
+// CHECK-PPC-NEXT:  br label %[[CONT]]
 //
-// CHECK-PPC1:[[CONT]]
-// CHECK-PPC:  [[VAARG_ADDR:%[a-z.0-9]+]] = phi %struct.x* [ [[TWELVE]], [[USING_REGS]] ], [ [[FOURTEEN]], [[USING_OVERFLOW]] ]
-// CHECK-PPC-NEXT:  [[AGGRPTR:%[a-z0-9]+]] = bitcast %struct.x* [[VAARG_ADDR]] to i8**
-// CHECK-PPC-NEXT:  [[AGGR:%[a-z0-9]+]] = load i8*, i8** [[AGGRPTR]]
-// CHECK-PPC-NEXT:  [[SEVENTEEN:%[0-9]+]] = bitcast %struct.x* %t to i8*
-// CHECK-PPC-NEXT:  call void @llvm.memcpy.p0i8.p0i8.i32(i8* [[SEVENTEEN]], i8* [[AGGR]], i32 16, i32 8, i1 false)
+// CHECK-PPC:[[CONT]]
+// CHECK-PPC-NEXT:  [[VAARG_ADDR:%[a-z.0-9]+]] = phi %struct.x** [ [[REGADDR]], %[[USING_REGS]] ], [ [[MEMADDR]], %[[USING_OVERFLOW]] ]
+// CHECK-PPC-NEXT:  [[AGGR:%[a-z0-9]+]] = load %struct.x*, %struct.x** [[VAARG_ADDR]]
+// CHECK-PPC-NEXT:  [[DEST:%[0-9]+]] = bitcast %struct.x* %t to i8*
+// CHECK-PPC-NEXT:  [[SRC:%.+]] = bitcast %struct.x* [[AGGR]] to i8*
+// CHECK-PPC-NEXT:  call void @llvm.memcpy.p0i8.p0i8.i32(i8* [[DEST]], i8* [[SRC]], i32 16, i32 8, i1 false)
 
   int v = va_arg (ap, int);
-// CHECK: ptrtoint i8* %{{[a-z.0-9]*}} to i64
-// CHECK: add i64 %{{[0-9]+}}, 4
-// CHECK: inttoptr i64 %{{[0-9]+}} to i8*
+  
+// CHECK: getelementptr inbounds i8, i8* %{{[a-z.0-9]*}}, i64 4
 // CHECK: bitcast i8* %{{[0-9]+}} to i32*
-// CHECK-PPC:  [[ARRAYDECAY1:%[a-z0-9]+]] = getelementptr inbounds [1 x %struct.__va_list_tag], [1 x %struct.__va_list_tag]* %ap, i32 0, i32 0
-// CHECK-PPC-NEXT:  [[GPRPTR1:%[a-z0-9]+]] = bitcast %struct.__va_list_tag* [[ARRAYDECAY1]] to i8*
-// CHECK-PPC-NEXT:  [[EIGHTEEN:%[0-9]+]] = ptrtoint i8* [[GPRPTR1]] to i32
-// CHECK-PPC-NEXT:  [[NINETEEN:%[0-9]+]] = add i32 [[EIGHTEEN]], 1
-// CHECK-PPC-NEXT:  [[TWENTY:%[0-9]+]] = inttoptr i32 [[NINETEEN]] to i8*
-// CHECK-PPC-NEXT:  [[TWENTYONE:%[0-9]+]] = add i32 [[NINETEEN]], 3
-// CHECK-PPC-NEXT:  [[TWENTYTWO:%[0-9]+]] = inttoptr i32 [[TWENTYONE]] to i8**
-// CHECK-PPC-NEXT:  [[TWENTYTHREE:%[0-9]+]] = add i32 [[TWENTYONE]], 4
-// CHECK-PPC-NEXT:  [[TWENTYFOUR:%[0-9]+]] = inttoptr i32 [[TWENTYTHREE]] to i8**
-// CHECK-PPC-NEXT:  [[GPR1:%[a-z0-9]+]] = load i8, i8* [[GPRPTR1]]
-// CHECK-PPC-NEXT:  [[FPR1:%[a-z0-9]+]] = load i8, i8* [[TWENTY]]
-// CHECK-PPC-NEXT:  [[OVERFLOW_AREA1:%[a-z_0-9]+]] = load i8*, i8** [[TWENTYTWO]]
-// CHECK-PPC-NEXT:  [[TWENTYFIVE:%[0-9]+]] = ptrtoint i8* [[OVERFLOW_AREA1]] to i32
-// CHECK-PPC-NEXT:  [[REGSAVE_AREA1:%[a-z_0-9]+]] = load i8*, i8** [[TWENTYFOUR]]
-// CHECK-PPC-NEXT:  [[TWENTYSIX:%[0-9]+]] = ptrtoint i8* [[REGSAVE_AREA1]] to i32
-// CHECK-PPC-NEXT:  [[COND1:%[a-z0-9]+]] = icmp ult i8 [[GPR1]], 8
-// CHECK-PPC-NEXT:  [[TWENTYSEVEN:%[0-9]+]] = mul i8 [[GPR1]], 4
-// CHECK-PPC-NEXT:  [[TWENTYEIGHT:%[0-9]+]] = sext i8 [[TWENTYSEVEN]] to i32
-// CHECK-PPC-NEXT:  [[TWENTYNINE:%[0-9]+]] = add i32 [[TWENTYSIX]], [[TWENTYEIGHT]]
-// CHECK-PPC-NEXT:  br i1 [[COND1]], label [[USING_REGS1:%[.a-z_0-9]+]], label [[USING_OVERFLOW1:%[.a-z_0-9]+]]
+// CHECK-PPC:       [[ARRAYDECAY:%[a-z0-9]+]] = getelementptr inbounds [1 x %struct.__va_list_tag], [1 x %struct.__va_list_tag]* %ap, i32 0, i32 0
+// CHECK-PPC-NEXT:  [[GPRPTR:%.+]] = getelementptr inbounds %struct.__va_list_tag, %struct.__va_list_tag* [[ARRAYDECAY]], i32 0, i32 0
+// CHECK-PPC-NEXT:  [[GPR:%.+]] = load i8, i8* [[GPRPTR]], align 4
+// CHECK-PPC-NEXT:  [[COND:%.+]] = icmp ult i8 [[GPR]], 8
+// CHECK-PPC-NEXT:  br i1 [[COND]], label %[[USING_REGS:.+]], label %[[USING_OVERFLOW:.+]]{{$}}
 //
-// CHECK-PPC1:[[USING_REGS1]]:
-// CHECK-PPC:  [[THIRTY:%[0-9]+]] = inttoptr i32 [[TWENTYNINE]] to i32*
-// CHECK-PPC-NEXT:  [[THIRTYONE:%[0-9]+]] = add i8 [[GPR1]], 1
-// CHECK-PPC-NEXT:  store i8 [[THIRTYONE]], i8* [[GPRPTR1]]
-// CHECK-PPC-NEXT:  br label [[CONT1:%[a-z0-9]+]]
+// CHECK-PPC:[[USING_REGS]]
+// CHECK-PPC-NEXT:  [[REGSAVE_AREA_P:%.+]] = getelementptr inbounds %struct.__va_list_tag, %struct.__va_list_tag* [[ARRAYDECAY]], i32 0, i32 4
+// CHECK-PPC-NEXT:  [[REGSAVE_AREA:%.+]] = load i8*, i8** [[REGSAVE_AREA_P]], align 4
+// CHECK-PPC-NEXT:  [[OFFSET:%.+]] = mul i8 [[GPR]], 4
+// CHECK-PPC-NEXT:  [[RAW_REGADDR:%.+]] = getelementptr inbounds i8, i8* [[REGSAVE_AREA]], i8 [[OFFSET]]
+// CHECK-PPC-NEXT:  [[REGADDR:%.+]] = bitcast i8* [[RAW_REGADDR]] to i32*
+// CHECK-PPC-NEXT:  [[USED_GPR:%[0-9]+]] = add i8 [[GPR]], 1
+// CHECK-PPC-NEXT:  store i8 [[USED_GPR]], i8* [[GPRPTR]], align 4
+// CHECK-PPC-NEXT:  br label %[[CONT:[a-z0-9]+]]
 //
-// CHECK-PPC1:[[USING_OVERFLOW1]]:
-// CHECK-PPC:  [[THIRTYTWO:%[0-9]+]] = inttoptr i32 [[TWENTYFIVE]] to i32*
-// CHECK-PPC-NEXT:  [[THIRTYTHREE:%[0-9]+]] = add i32 [[TWENTYFIVE]], 4
-// CHECK-PPC-NEXT:  [[THIRTYFOUR:%[0-9]+]] = inttoptr i32 [[THIRTYTHREE]] to i8*
-// CHECK-PPC-NEXT:  store i8* [[THIRTYFOUR]], i8** [[TWENTYTWO]]
-// CHECK-PPC-NEXT:  br label [[CONT1]]
+// CHECK-PPC:[[USING_OVERFLOW]]
+// CHECK-PPC-NEXT:  [[OVERFLOW_AREA_P:%[0-9]+]] = getelementptr inbounds %struct.__va_list_tag, %struct.__va_list_tag* [[ARRAYDECAY]], i32 0, i32 3
+// CHECK-PPC-NEXT:  [[OVERFLOW_AREA:%.+]] = load i8*, i8** [[OVERFLOW_AREA_P]], align 4
+// CHECK-PPC-NEXT:  [[MEMADDR:%.+]] = bitcast i8* [[OVERFLOW_AREA]] to i32*
+// CHECK-PPC-NEXT:  [[NEW_OVERFLOW_AREA:%[0-9]+]] = getelementptr inbounds i8, i8* [[OVERFLOW_AREA]], i32 4
+// CHECK-PPC-NEXT:  store i8* [[NEW_OVERFLOW_AREA]], i8** [[OVERFLOW_AREA_P]]
+// CHECK-PPC-NEXT:  br label %[[CONT]]
 //
-// CHECK-PPC1:[[CONT1]]:
-// CHECK-PPC:  [[VAARG_ADDR1:%[a-z.0-9]+]] = phi i32* [ [[THIRTY]], [[USING_REGS1]] ], [ [[THIRTYTWO]], [[USING_OVERFLOW1]] ]
-// CHECK-PPC-NEXT:  [[THIRTYFIVE:%[0-9]+]] = load i32, i32* [[VAARG_ADDR1]]
+// CHECK-PPC:[[CONT]]
+// CHECK-PPC-NEXT:  [[VAARG_ADDR:%[a-z.0-9]+]] = phi i32* [ [[REGADDR]], %[[USING_REGS]] ], [ [[MEMADDR]], %[[USING_OVERFLOW]] ]
+// CHECK-PPC-NEXT:  [[THIRTYFIVE:%[0-9]+]] = load i32, i32* [[VAARG_ADDR]]
 // CHECK-PPC-NEXT:  store i32 [[THIRTYFIVE]], i32* %v, align 4
 
 #ifdef __powerpc64__
diff --git a/test/CodeGen/ppc64-align-struct.c b/test/CodeGen/ppc64-align-struct.c
index 8c4437a38d..6a04d0cd84 100644
--- a/test/CodeGen/ppc64-align-struct.c
+++ b/test/CodeGen/ppc64-align-struct.c
@@ -41,18 +41,22 @@ void test6 (int x, struct test6 y)
 }
 
 // This case requires run-time realignment of the incoming struct
-// CHECK: define void @test7(i32 signext %x, %struct.test7* byval align 16)
+// CHECK-LABEL: define void @test7(i32 signext %x, %struct.test7* byval align 16)
 // CHECK: %y = alloca %struct.test7, align 32
 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64
 void test7 (int x, struct test7 y)
 {
 }
 
-// CHECK: define void @test1va(%struct.test1* noalias sret %agg.result, i32 signext %x, ...)
+// CHECK-LABEL: define void @test1va(%struct.test1* noalias sret %agg.result, i32 signext %x, ...)
+// CHECK: %y = alloca %struct.test1, align 4
 // CHECK: %[[CUR:[^ ]+]] = load i8*, i8** %ap
-// CHECK: %[[NEXT:[^ ]+]] = getelementptr i8, i8* %[[CUR]], i64 8
+// CHECK: %[[NEXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[CUR]], i64 8
 // CHECK: store i8* %[[NEXT]], i8** %ap
-// CHECK: bitcast i8* %[[CUR]] to %struct.test1*
+// CHECK: [[T0:%.*]] = bitcast i8* %[[CUR]] to %struct.test1*
+// CHECK: [[DEST:%.*]] = bitcast %struct.test1* %y to i8*
+// CHECK: [[SRC:%.*]] = bitcast %struct.test1* [[T0]] to i8*
+// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[DEST]], i8* [[SRC]], i64 8, i32 4, i1 false)
 struct test1 test1va (int x, ...)
 {
   struct test1 y;
@@ -63,15 +67,19 @@ struct test1 test1va (int x, ...)
   return y;
 }
 
-// CHECK: define void @test2va(%struct.test2* noalias sret %agg.result, i32 signext %x, ...)
+// CHECK-LABEL: define void @test2va(%struct.test2* noalias sret %agg.result, i32 signext %x, ...)
+// CHECK: %y = alloca %struct.test2, align 16
 // CHECK: %[[CUR:[^ ]+]] = load i8*, i8** %ap
 // CHECK: %[[TMP0:[^ ]+]] = ptrtoint i8* %[[CUR]] to i64
 // CHECK: %[[TMP1:[^ ]+]] = add i64 %[[TMP0]], 15
 // CHECK: %[[TMP2:[^ ]+]] = and i64 %[[TMP1]], -16
 // CHECK: %[[ALIGN:[^ ]+]] = inttoptr i64 %[[TMP2]] to i8*
-// CHECK: %[[NEXT:[^ ]+]] = getelementptr i8, i8* %[[ALIGN]], i64 16
+// CHECK: %[[NEXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[ALIGN]], i64 16
 // CHECK: store i8* %[[NEXT]], i8** %ap
-// CHECK: bitcast i8* %[[ALIGN]] to %struct.test2*
+// CHECK: [[T0:%.*]] = bitcast i8* %[[ALIGN]] to %struct.test2*
+// CHECK: [[DEST:%.*]] = bitcast %struct.test2* %y to i8*
+// CHECK: [[SRC:%.*]] = bitcast %struct.test2* [[T0]] to i8*
+// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[DEST]], i8* [[SRC]], i64 16, i32 16, i1 false)
 struct test2 test2va (int x, ...)
 {
   struct test2 y;
@@ -82,15 +90,19 @@ struct test2 test2va (int x, ...)
   return y;
 }
 
-// CHECK: define void @test3va(%struct.test3* noalias sret %agg.result, i32 signext %x, ...)
+// CHECK-LABEL: define void @test3va(%struct.test3* noalias sret %agg.result, i32 signext %x, ...)
+// CHECK: %y = alloca %struct.test3, align 32
 // CHECK: %[[CUR:[^ ]+]] = load i8*, i8** %ap
 // CHECK: %[[TMP0:[^ ]+]] = ptrtoint i8* %[[CUR]] to i64
 // CHECK: %[[TMP1:[^ ]+]] = add i64 %[[TMP0]], 15
 // CHECK: %[[TMP2:[^ ]+]] = and i64 %[[TMP1]], -16
 // CHECK: %[[ALIGN:[^ ]+]] = inttoptr i64 %[[TMP2]] to i8*
-// CHECK: %[[NEXT:[^ ]+]] = getelementptr i8, i8* %[[ALIGN]], i64 32
+// CHECK: %[[NEXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[ALIGN]], i64 32
 // CHECK: store i8* %[[NEXT]], i8** %ap
-// CHECK: bitcast i8* %[[ALIGN]] to %struct.test3*
+// CHECK: [[T0:%.*]] = bitcast i8* %[[ALIGN]] to %struct.test3*
+// CHECK: [[DEST:%.*]] = bitcast %struct.test3* %y to i8*
+// CHECK: [[SRC:%.*]] = bitcast %struct.test3* [[T0]] to i8*
+// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[DEST]], i8* [[SRC]], i64 32, i32 16, i1 false)
 struct test3 test3va (int x, ...)
 {
   struct test3 y;
@@ -101,11 +113,15 @@ struct test3 test3va (int x, ...)
   return y;
 }
 
-// CHECK: define void @test4va(%struct.test4* noalias sret %agg.result, i32 signext %x, ...)
+// CHECK-LABEL: define void @test4va(%struct.test4* noalias sret %agg.result, i32 signext %x, ...)
+// CHECK: %y = alloca %struct.test4, align 4
 // CHECK: %[[CUR:[^ ]+]] = load i8*, i8** %ap
-// CHECK: %[[NEXT:[^ ]+]] = getelementptr i8, i8* %[[CUR]], i64 16
+// CHECK: %[[NEXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[CUR]], i64 16
 // CHECK: store i8* %[[NEXT]], i8** %ap
-// CHECK: bitcast i8* %[[CUR]] to %struct.test4*
+// CHECK: [[T0:%.*]] = bitcast i8* %[[CUR]] to %struct.test4*
+// CHECK: [[DEST:%.*]] = bitcast %struct.test4* %y to i8*
+// CHECK: [[SRC:%.*]] = bitcast %struct.test4* [[T0]] to i8*
+// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[DEST]], i8* [[SRC]], i64 12, i32 4, i1 false)
 struct test4 test4va (int x, ...)
 {
   struct test4 y;
@@ -116,11 +132,15 @@ struct test4 test4va (int x, ...)
   return y;
 }
 
-// CHECK: define void @testva_longdouble(%struct.test_longdouble* noalias sret %agg.result, i32 signext %x, ...)
+// CHECK-LABEL: define void @testva_longdouble(%struct.test_longdouble* noalias sret %agg.result, i32 signext %x, ...)
+// CHECK: %y = alloca %struct.test_longdouble, align 16
 // CHECK: %[[CUR:[^ ]+]] = load i8*, i8** %ap
-// CHECK: %[[NEXT:[^ ]+]] = getelementptr i8, i8* %[[CUR]], i64 16
+// CHECK: %[[NEXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[CUR]], i64 16
 // CHECK: store i8* %[[NEXT]], i8** %ap
-// CHECK: bitcast i8* %[[CUR]] to %struct.test_longdouble*
+// CHECK: [[T0:%.*]] = bitcast i8* %[[CUR]] to %struct.test_longdouble*
+// CHECK: [[DEST:%.*]] = bitcast %struct.test_longdouble* %y to i8*
+// CHECK: [[SRC:%.*]] = bitcast %struct.test_longdouble* [[T0]] to i8*
+// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[DEST]], i8* [[SRC]], i64 16, i32 8, i1 false)
 struct test_longdouble { long double x; };
 struct test_longdouble testva_longdouble (int x, ...)
 {
@@ -132,15 +152,19 @@ struct test_longdouble testva_longdouble (int x, ...)
   return y;
 }
 
-// CHECK: define void @testva_vector(%struct.test_vector* noalias sret %agg.result, i32 signext %x, ...)
+// CHECK-LABEL: define void @testva_vector(%struct.test_vector* noalias sret %agg.result, i32 signext %x, ...)
+// CHECK: %y = alloca %struct.test_vector, align 16
 // CHECK: %[[CUR:[^ ]+]] = load i8*, i8** %ap
 // CHECK: %[[TMP0:[^ ]+]] = ptrtoint i8* %[[CUR]] to i64
 // CHECK: %[[TMP1:[^ ]+]] = add i64 %[[TMP0]], 15
 // CHECK: %[[TMP2:[^ ]+]] = and i64 %[[TMP1]], -16
 // CHECK: %[[ALIGN:[^ ]+]] = inttoptr i64 %[[TMP2]] to i8*
-// CHECK: %[[NEXT:[^ ]+]] = getelementptr i8, i8* %[[ALIGN]], i64 16
+// CHECK: %[[NEXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[ALIGN]], i64 16
 // CHECK: store i8* %[[NEXT]], i8** %ap
-// CHECK: bitcast i8* %[[ALIGN]] to %struct.test_vector*
+// CHECK: [[T0:%.*]] = bitcast i8* %[[ALIGN]] to %struct.test_vector*
+// CHECK: [[DEST:%.*]] = bitcast %struct.test_vector* %y to i8*
+// CHECK: [[SRC:%.*]] = bitcast %struct.test_vector* [[T0]] to i8*
+// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[DEST]], i8* [[SRC]], i64 16, i32 16, i1 false)
 struct test_vector { vector int x; };
 struct test_vector testva_vector (int x, ...)
 {
diff --git a/test/CodeGen/ppc64-complex-parms.c b/test/CodeGen/ppc64-complex-parms.c
index f5583a0742..3f2a0c2142 100644
--- a/test/CodeGen/ppc64-complex-parms.c
+++ b/test/CodeGen/ppc64-complex-parms.c
@@ -62,10 +62,10 @@ void bar_float(void) {
 // CHECK: %[[VAR3:[A-Za-z0-9.]+]] = getelementptr inbounds { float, float }, { float, float }* %[[VAR1]], i32 0, i32 1
 // CHECK: store float 2.000000e+00, float* %[[VAR2]]
 // CHECK: store float -2.500000e+00, float* %[[VAR3]]
-// CHECK: %[[VAR4:[A-Za-z0-9.]+]] = getelementptr { float, float }, { float, float }* %[[VAR1]], i32 0, i32 0
-// CHECK: %[[VAR5:[A-Za-z0-9.]+]] = load float, float* %[[VAR4]], align 1
-// CHECK: %[[VAR6:[A-Za-z0-9.]+]] = getelementptr { float, float }, { float, float }* %[[VAR1]], i32 0, i32 1
-// CHECK: %[[VAR7:[A-Za-z0-9.]+]] = load float, float* %[[VAR6]], align 1
+// CHECK: %[[VAR4:[A-Za-z0-9.]+]] = getelementptr inbounds { float, float }, { float, float }* %[[VAR1]], i32 0, i32 0
+// CHECK: %[[VAR5:[A-Za-z0-9.]+]] = load float, float* %[[VAR4]], align 4
+// CHECK: %[[VAR6:[A-Za-z0-9.]+]] = getelementptr inbounds { float, float }, { float, float }* %[[VAR1]], i32 0, i32 1
+// CHECK: %[[VAR7:[A-Za-z0-9.]+]] = load float, float* %[[VAR6]], align 4
 // CHECK: %{{[A-Za-z0-9.]+}} = call float @foo_float(float %[[VAR5]], float %[[VAR7]])
 
 void bar_double(void) {
@@ -78,10 +78,10 @@ void bar_double(void) {
 // CHECK: %[[VAR13:[A-Za-z0-9.]+]] = getelementptr inbounds { double, double }, { double, double }* %[[VAR11]], i32 0, i32 1
 // CHECK: store double 2.000000e+00, double* %[[VAR12]]
 // CHECK: store double -2.500000e+00, double* %[[VAR13]]
-// CHECK: %[[VAR14:[A-Za-z0-9.]+]] = getelementptr { double, double }, { double, double }* %[[VAR11]], i32 0, i32 0
-// CHECK: %[[VAR15:[A-Za-z0-9.]+]] = load double, double* %[[VAR14]], align 1
-// CHECK: %[[VAR16:[A-Za-z0-9.]+]] = getelementptr { double, double }, { double, double }* %[[VAR11]], i32 0, i32 1
-// CHECK: %[[VAR17:[A-Za-z0-9.]+]] = load double, double* %[[VAR16]], align 1
+// CHECK: %[[VAR14:[A-Za-z0-9.]+]] = getelementptr inbounds { double, double }, { double, double }* %[[VAR11]], i32 0, i32 0
+// CHECK: %[[VAR15:[A-Za-z0-9.]+]] = load double, double* %[[VAR14]], align 8
+// CHECK: %[[VAR16:[A-Za-z0-9.]+]] = getelementptr inbounds { double, double }, { double, double }* %[[VAR11]], i32 0, i32 1
+// CHECK: %[[VAR17:[A-Za-z0-9.]+]] = load double, double* %[[VAR16]], align 8
 // CHECK: %{{[A-Za-z0-9.]+}} = call double @foo_double(double %[[VAR15]], double %[[VAR17]])
 
 void bar_long_double(void) {
@@ -94,10 +94,10 @@ void bar_long_double(void) {
 // CHECK: %[[VAR23:[A-Za-z0-9.]+]] = getelementptr inbounds { ppc_fp128, ppc_fp128 }, { ppc_fp128, ppc_fp128 }* %[[VAR21]], i32 0, i32 1
 // CHECK: store ppc_fp128 0xM40000000000000000000000000000000, ppc_fp128* %[[VAR22]]
 // CHECK: store ppc_fp128 0xMC0040000000000000000000000000000, ppc_fp128* %[[VAR23]]
-// CHECK: %[[VAR24:[A-Za-z0-9.]+]] = getelementptr { ppc_fp128, ppc_fp128 }, { ppc_fp128, ppc_fp128 }* %[[VAR21]], i32 0, i32 0
-// CHECK: %[[VAR25:[A-Za-z0-9.]+]] = load ppc_fp128, ppc_fp128* %[[VAR24]], align 1
-// CHECK: %[[VAR26:[A-Za-z0-9.]+]] = getelementptr { ppc_fp128, ppc_fp128 }, { ppc_fp128, ppc_fp128 }* %[[VAR21]], i32 0, i32 1
-// CHECK: %[[VAR27:[A-Za-z0-9.]+]] = load ppc_fp128, ppc_fp128* %[[VAR26]], align 1
+// CHECK: %[[VAR24:[A-Za-z0-9.]+]] = getelementptr inbounds { ppc_fp128, ppc_fp128 }, { ppc_fp128, ppc_fp128 }* %[[VAR21]], i32 0, i32 0
+// CHECK: %[[VAR25:[A-Za-z0-9.]+]] = load ppc_fp128, ppc_fp128* %[[VAR24]], align 16
+// CHECK: %[[VAR26:[A-Za-z0-9.]+]] = getelementptr inbounds { ppc_fp128, ppc_fp128 }, { ppc_fp128, ppc_fp128 }* %[[VAR21]], i32 0, i32 1
+// CHECK: %[[VAR27:[A-Za-z0-9.]+]] = load ppc_fp128, ppc_fp128* %[[VAR26]], align 16
 // CHECK: %{{[A-Za-z0-9.]+}} = call ppc_fp128 @foo_long_double(ppc_fp128 %[[VAR25]], ppc_fp128 %[[VAR27]])
 
 void bar_int(void) {
@@ -110,10 +110,10 @@ void bar_int(void) {
 // CHECK: %[[VAR33:[A-Za-z0-9.]+]] = getelementptr inbounds { i32, i32 }, { i32, i32 }* %[[VAR31]], i32 0, i32 1
 // CHECK: store i32 2, i32* %[[VAR32]]
 // CHECK: store i32 -3, i32* %[[VAR33]]
-// CHECK: %[[VAR34:[A-Za-z0-9.]+]] = getelementptr { i32, i32 }, { i32, i32 }* %[[VAR31]], i32 0, i32 0
-// CHECK: %[[VAR35:[A-Za-z0-9.]+]] = load i32, i32* %[[VAR34]], align 1
-// CHECK: %[[VAR36:[A-Za-z0-9.]+]] = getelementptr { i32, i32 }, { i32, i32 }* %[[VAR31]], i32 0, i32 1
-// CHECK: %[[VAR37:[A-Za-z0-9.]+]] = load i32, i32* %[[VAR36]], align 1
+// CHECK: %[[VAR34:[A-Za-z0-9.]+]] = getelementptr inbounds { i32, i32 }, { i32, i32 }* %[[VAR31]], i32 0, i32 0
+// CHECK: %[[VAR35:[A-Za-z0-9.]+]] = load i32, i32* %[[VAR34]], align 4
+// CHECK: %[[VAR36:[A-Za-z0-9.]+]] = getelementptr inbounds { i32, i32 }, { i32, i32 }* %[[VAR31]], i32 0, i32 1
+// CHECK: %[[VAR37:[A-Za-z0-9.]+]] = load i32, i32* %[[VAR36]], align 4
 // CHECK: %{{[A-Za-z0-9.]+}} = call signext i32 @foo_int(i32 %[[VAR35]], i32 %[[VAR37]])
 
 void bar_short(void) {
@@ -126,10 +126,10 @@ void bar_short(void) {
 // CHECK: %[[VAR43:[A-Za-z0-9.]+]] = getelementptr inbounds { i16, i16 }, { i16, i16 }* %[[VAR41]], i32 0, i32 1
 // CHECK: store i16 2, i16* %[[VAR42]]
 // CHECK: store i16 -3, i16* %[[VAR43]]
-// CHECK: %[[VAR44:[A-Za-z0-9.]+]] = getelementptr { i16, i16 }, { i16, i16 }* %[[VAR41]], i32 0, i32 0
-// CHECK: %[[VAR45:[A-Za-z0-9.]+]] = load i16, i16* %[[VAR44]], align 1
-// CHECK: %[[VAR46:[A-Za-z0-9.]+]] = getelementptr { i16, i16 }, { i16, i16 }* %[[VAR41]], i32 0, i32 1
-// CHECK: %[[VAR47:[A-Za-z0-9.]+]] = load i16, i16* %[[VAR46]], align 1
+// CHECK: %[[VAR44:[A-Za-z0-9.]+]] = getelementptr inbounds { i16, i16 }, { i16, i16 }* %[[VAR41]], i32 0, i32 0
+// CHECK: %[[VAR45:[A-Za-z0-9.]+]] = load i16, i16* %[[VAR44]], align 2
+// CHECK: %[[VAR46:[A-Za-z0-9.]+]] = getelementptr inbounds { i16, i16 }, { i16, i16 }* %[[VAR41]], i32 0, i32 1
+// CHECK: %[[VAR47:[A-Za-z0-9.]+]] = load i16, i16* %[[VAR46]], align 2
 // CHECK: %{{[A-Za-z0-9.]+}} = call signext i16 @foo_short(i16 %[[VAR45]], i16 %[[VAR47]])
 
 void bar_char(void) {
@@ -142,9 +142,9 @@ void bar_char(void) {
 // CHECK: %[[VAR53:[A-Za-z0-9.]+]] = getelementptr inbounds { i8, i8 }, { i8, i8 }* %[[VAR51]], i32 0, i32 1
 // CHECK: store i8 2, i8* %[[VAR52]]
 // CHECK: store i8 -3, i8* %[[VAR53]]
-// CHECK: %[[VAR54:[A-Za-z0-9.]+]] = getelementptr { i8, i8 }, { i8, i8 }* %[[VAR51]], i32 0, i32 0
+// CHECK: %[[VAR54:[A-Za-z0-9.]+]] = getelementptr inbounds { i8, i8 }, { i8, i8 }* %[[VAR51]], i32 0, i32 0
 // CHECK: %[[VAR55:[A-Za-z0-9.]+]] = load i8, i8* %[[VAR54]], align 1
-// CHECK: %[[VAR56:[A-Za-z0-9.]+]] = getelementptr { i8, i8 }, { i8, i8 }* %[[VAR51]], i32 0, i32 1
+// CHECK: %[[VAR56:[A-Za-z0-9.]+]] = getelementptr inbounds { i8, i8 }, { i8, i8 }* %[[VAR51]], i32 0, i32 1
 // CHECK: %[[VAR57:[A-Za-z0-9.]+]] = load i8, i8* %[[VAR56]], align 1
 // CHECK: %{{[A-Za-z0-9.]+}} = call signext i8 @foo_char(i8 %[[VAR55]], i8 %[[VAR57]])
 
@@ -158,10 +158,10 @@ void bar_long(void) {
 // CHECK: %[[VAR63:[A-Za-z0-9.]+]] = getelementptr inbounds { i64, i64 }, { i64, i64 }* %[[VAR61]], i32 0, i32 1
 // CHECK: store i64 2, i64* %[[VAR62]]
 // CHECK: store i64 -3, i64* %[[VAR63]]
-// CHECK: %[[VAR64:[A-Za-z0-9.]+]] = getelementptr { i64, i64 }, { i64, i64 }* %[[VAR61]], i32 0, i32 0
-// CHECK: %[[VAR65:[A-Za-z0-9.]+]] = load i64, i64* %[[VAR64]], align 1
-// CHECK: %[[VAR66:[A-Za-z0-9.]+]] = getelementptr { i64, i64 }, { i64, i64 }* %[[VAR61]], i32 0, i32 1
-// CHECK: %[[VAR67:[A-Za-z0-9.]+]] = load i64, i64* %[[VAR66]], align 1
+// CHECK: %[[VAR64:[A-Za-z0-9.]+]] = getelementptr inbounds { i64, i64 }, { i64, i64 }* %[[VAR61]], i32 0, i32 0
+// CHECK: %[[VAR65:[A-Za-z0-9.]+]] = load i64, i64* %[[VAR64]], align 8
+// CHECK: %[[VAR66:[A-Za-z0-9.]+]] = getelementptr inbounds { i64, i64 }, { i64, i64 }* %[[VAR61]], i32 0, i32 1
+// CHECK: %[[VAR67:[A-Za-z0-9.]+]] = load i64, i64* %[[VAR66]], align 8
 // CHECK: %{{[A-Za-z0-9.]+}} = call i64 @foo_long(i64 %[[VAR65]], i64 %[[VAR67]])
 
 void bar_long_long(void) {
@@ -174,10 +174,10 @@ void bar_long_long(void) {
 // CHECK: %[[VAR73:[A-Za-z0-9.]+]] = getelementptr inbounds { i64, i64 }, { i64, i64 }* %[[VAR71]], i32 0, i32 1
 // CHECK: store i64 2, i64* %[[VAR72]]
 // CHECK: store i64 -3, i64* %[[VAR73]]
-// CHECK: %[[VAR74:[A-Za-z0-9.]+]] = getelementptr { i64, i64 }, { i64, i64 }* %[[VAR71]], i32 0, i32 0
-// CHECK: %[[VAR75:[A-Za-z0-9.]+]] = load i64, i64* %[[VAR74]], align 1
-// CHECK: %[[VAR76:[A-Za-z0-9.]+]] = getelementptr { i64, i64 }, { i64, i64 }* %[[VAR71]], i32 0, i32 1
-// CHECK: %[[VAR77:[A-Za-z0-9.]+]] = load i64, i64* %[[VAR76]], align 1
+// CHECK: %[[VAR74:[A-Za-z0-9.]+]] = getelementptr inbounds { i64, i64 }, { i64, i64 }* %[[VAR71]], i32 0, i32 0
+// CHECK: %[[VAR75:[A-Za-z0-9.]+]] = load i64, i64* %[[VAR74]], align 8
+// CHECK: %[[VAR76:[A-Za-z0-9.]+]] = getelementptr inbounds { i64, i64 }, { i64, i64 }* %[[VAR71]], i32 0, i32 1
+// CHECK: %[[VAR77:[A-Za-z0-9.]+]] = load i64, i64* %[[VAR76]], align 8
 // CHECK: %{{[A-Za-z0-9.]+}} = call i64 @foo_long_long(i64 %[[VAR75]], i64 %[[VAR77]])
 
 // CHECK: attributes [[NUW]] = { nounwind{{.*}} }
diff --git a/test/CodeGen/ppc64-struct-onefloat.c b/test/CodeGen/ppc64-struct-onefloat.c
index 534e5116f9..efc6fe9d8d 100644
--- a/test/CodeGen/ppc64-struct-onefloat.c
+++ b/test/CodeGen/ppc64-struct-onefloat.c
@@ -13,15 +13,15 @@ void bar(Sf a, Sd b, SSf d, SSd e) {}
 // CHECK:  %b = alloca %struct.s2, align 8
 // CHECK:  %d = alloca %struct.s4, align 4
 // CHECK:  %e = alloca %struct.s5, align 8
-// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr %struct.s1, %struct.s1* %a, i32 0, i32 0
+// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr inbounds %struct.s1, %struct.s1* %a, i32 0, i32 0
 // CHECK:  store float %a.coerce, float* %{{[a-zA-Z0-9.]+}}, align 4
-// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr %struct.s2, %struct.s2* %b, i32 0, i32 0
+// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr inbounds %struct.s2, %struct.s2* %b, i32 0, i32 0
 // CHECK:  store double %b.coerce, double* %{{[a-zA-Z0-9.]+}}, align 8
-// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr %struct.s4, %struct.s4* %d, i32 0, i32 0
-// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr %struct.s1, %struct.s1* %{{[a-zA-Z0-9.]+}}, i32 0, i32 0
+// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr inbounds %struct.s4, %struct.s4* %d, i32 0, i32 0
+// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr inbounds %struct.s1, %struct.s1* %{{[a-zA-Z0-9.]+}}, i32 0, i32 0
 // CHECK:  store float %d.coerce, float* %{{[a-zA-Z0-9.]+}}, align 4
-// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr %struct.s5, %struct.s5* %e, i32 0, i32 0
-// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr %struct.s2, %struct.s2* %{{[a-zA-Z0-9.]+}}, i32 0, i32 0
+// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr inbounds %struct.s5, %struct.s5* %e, i32 0, i32 0
+// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr inbounds %struct.s2, %struct.s2* %{{[a-zA-Z0-9.]+}}, i32 0, i32 0
 // CHECK:  store double %e.coerce, double* %{{[a-zA-Z0-9.]+}}, align 8
 // CHECK:  ret void
 
@@ -35,15 +35,15 @@ void foo(void)
 }
 
 // CHECK-LABEL: define void @foo
-// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr %struct.s1, %struct.s1* %p1, i32 0, i32 0
+// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr inbounds %struct.s1, %struct.s1* %p1, i32 0, i32 0
 // CHECK:  %{{[0-9]+}} = load float, float* %{{[a-zA-Z0-9.]+}}, align 4
-// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr %struct.s2, %struct.s2* %p2, i32 0, i32 0
+// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr inbounds %struct.s2, %struct.s2* %p2, i32 0, i32 0
 // CHECK:  %{{[0-9]+}} = load double, double* %{{[a-zA-Z0-9.]+}}, align 8
-// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr %struct.s4, %struct.s4* %p4, i32 0, i32 0
-// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr %struct.s1, %struct.s1* %{{[a-zA-Z0-9.]+}}, i32 0, i32 0
+// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr inbounds %struct.s4, %struct.s4* %p4, i32 0, i32 0
+// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr inbounds %struct.s1, %struct.s1* %{{[a-zA-Z0-9.]+}}, i32 0, i32 0
 // CHECK:  %{{[0-9]+}} = load float, float* %{{[a-zA-Z0-9.]+}}, align 4
-// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr %struct.s5, %struct.s5* %p5, i32 0, i32 0
-// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr %struct.s2, %struct.s2* %{{[a-zA-Z0-9.]+}}, i32 0, i32 0
+// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr inbounds %struct.s5, %struct.s5* %p5, i32 0, i32 0
+// CHECK:  %{{[a-zA-Z0-9.]+}} = getelementptr inbounds %struct.s2, %struct.s2* %{{[a-zA-Z0-9.]+}}, i32 0, i32 0
 // CHECK:  %{{[0-9]+}} = load double, double* %{{[a-zA-Z0-9.]+}}, align 8
 // CHECK:  call void @bar(float inreg %{{[0-9]+}}, double inreg %{{[0-9]+}}, float inreg %{{[0-9]+}}, double inreg %{{[0-9]+}})
 // CHECK:  ret void
diff --git a/test/CodeGen/ppc64-varargs-complex.c b/test/CodeGen/ppc64-varargs-complex.c
index f790629345..58206801ac 100644
--- a/test/CodeGen/ppc64-varargs-complex.c
+++ b/test/CodeGen/ppc64-varargs-complex.c
@@ -9,15 +9,14 @@ void testva (int n, ...)
 
   _Complex int i   = va_arg(ap, _Complex int);
   // CHECK: %[[VAR40:[A-Za-z0-9.]+]] = load i8*, i8** %[[VAR100:[A-Za-z0-9.]+]]
-  // CHECK-NEXT: %[[VAR41:[A-Za-z0-9.]+]] = getelementptr i8, i8* %[[VAR40]], i64 16
+  // CHECK-NEXT: %[[VAR41:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR40]], i64 16
   // CHECK-NEXT: store i8* %[[VAR41]], i8** %[[VAR100]]
-  // CHECK-NEXT: %[[VAR1:[A-Za-z0-9.]+]] = ptrtoint i8* %[[VAR40]] to i64
-  // CHECK-NEXT: %[[VAR2:[A-Za-z0-9.]+]] = add i64 %[[VAR1]], 4
-  // CHECK-NEXT: %[[VAR3:[A-Za-z0-9.]+]] = add i64 %[[VAR1]], 12
-  // CHECK-NEXT: %[[VAR4:[A-Za-z0-9.]+]] = inttoptr i64 %[[VAR2]] to i32*
-  // CHECK-NEXT: %[[VAR5:[A-Za-z0-9.]+]] = inttoptr i64 %[[VAR3]] to i32*
-  // CHECK-NEXT: %[[VAR6:[A-Za-z0-9.]+]] = load i32, i32* %[[VAR4]]
-  // CHECK-NEXT: %[[VAR7:[A-Za-z0-9.]+]] = load i32, i32* %[[VAR5]]
+  // CHECK-NEXT: %[[VAR1:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR40]], i64 4
+  // CHECK-NEXT: %[[VAR2:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR40]], i64 12
+  // CHECK-NEXT: %[[VAR4:[A-Za-z0-9.]+]] = bitcast i8* %[[VAR1]] to i32*
+  // CHECK-NEXT: %[[VAR5:[A-Za-z0-9.]+]] = bitcast i8* %[[VAR2]] to i32*
+  // CHECK-NEXT: %[[VAR6:[A-Za-z0-9.]+]] = load i32, i32* %[[VAR4]], align 4
+  // CHECK-NEXT: %[[VAR7:[A-Za-z0-9.]+]] = load i32, i32* %[[VAR5]], align 4
   // CHECK-NEXT: %[[VAR8:[A-Za-z0-9.]+]] = getelementptr inbounds { i32, i32 }, { i32, i32 }* %[[VAR0:[A-Za-z0-9.]+]], i32 0, i32 0
   // CHECK-NEXT: %[[VAR9:[A-Za-z0-9.]+]] = getelementptr inbounds { i32, i32 }, { i32, i32 }* %[[VAR0]], i32 0, i32 1
   // CHECK-NEXT: store i32 %[[VAR6]], i32* %[[VAR8]]
@@ -25,15 +24,14 @@ void testva (int n, ...)
 
   _Complex short s = va_arg(ap, _Complex short);
   // CHECK: %[[VAR50:[A-Za-z0-9.]+]] = load i8*, i8** %[[VAR100:[A-Za-z0-9.]+]]
-  // CHECK-NEXT: %[[VAR51:[A-Za-z0-9.]+]] = getelementptr i8, i8* %[[VAR50]], i64 16
+  // CHECK-NEXT: %[[VAR51:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR50]], i64 16
   // CHECK-NEXT: store i8* %[[VAR51]], i8** %[[VAR100]]
-  // CHECK: %[[VAR11:[A-Za-z0-9.]+]] = ptrtoint i8* %{{[A-Za-z0-9.]+}} to i64
-  // CHECK-NEXT: %[[VAR12:[A-Za-z0-9.]+]] = add i64 %[[VAR11]], 6
-  // CHECK-NEXT: %[[VAR13:[A-Za-z0-9.]+]] = add i64 %[[VAR11]], 14
-  // CHECK-NEXT: %[[VAR14:[A-Za-z0-9.]+]] = inttoptr i64 %[[VAR12]] to i16*
-  // CHECK-NEXT: %[[VAR15:[A-Za-z0-9.]+]] = inttoptr i64 %[[VAR13]] to i16*
-  // CHECK-NEXT: %[[VAR16:[A-Za-z0-9.]+]] = load i16, i16* %[[VAR14]]
-  // CHECK-NEXT: %[[VAR17:[A-Za-z0-9.]+]] = load i16, i16* %[[VAR15]]
+  // CHECK-NEXT: %[[VAR12:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR50]], i64 6
+  // CHECK-NEXT: %[[VAR13:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR50]], i64 14
+  // CHECK-NEXT: %[[VAR14:[A-Za-z0-9.]+]] = bitcast i8* %[[VAR12]] to i16*
+  // CHECK-NEXT: %[[VAR15:[A-Za-z0-9.]+]] = bitcast i8* %[[VAR13]] to i16*
+  // CHECK-NEXT: %[[VAR16:[A-Za-z0-9.]+]] = load i16, i16* %[[VAR14]], align 2
+  // CHECK-NEXT: %[[VAR17:[A-Za-z0-9.]+]] = load i16, i16* %[[VAR15]], align 2
   // CHECK-NEXT: %[[VAR18:[A-Za-z0-9.]+]] = getelementptr inbounds { i16, i16 }, { i16, i16 }* %[[VAR10:[A-Za-z0-9.]+]], i32 0, i32 0
   // CHECK-NEXT: %[[VAR19:[A-Za-z0-9.]+]] = getelementptr inbounds { i16, i16 }, { i16, i16 }* %[[VAR10]], i32 0, i32 1
   // CHECK-NEXT: store i16 %[[VAR16]], i16* %[[VAR18]]
@@ -41,15 +39,12 @@ void testva (int n, ...)
 
   _Complex char c  = va_arg(ap, _Complex char);
   // CHECK: %[[VAR60:[A-Za-z0-9.]+]] = load i8*, i8** %[[VAR100:[A-Za-z0-9.]+]]
-  // CHECK-NEXT: %[[VAR61:[A-Za-z0-9.]+]] = getelementptr i8, i8* %[[VAR60]], i64 16
+  // CHECK-NEXT: %[[VAR61:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR60]], i64 16
   // CHECK-NEXT: store i8* %[[VAR61]], i8** %[[VAR100]]
-  // CHECK: %[[VAR21:[A-Za-z0-9.]+]] = ptrtoint i8* %{{[A-Za-z0-9.]+}} to i64
-  // CHECK-NEXT: %[[VAR22:[A-Za-z0-9.]+]] = add i64 %[[VAR21]], 7
-  // CHECK-NEXT: %[[VAR23:[A-Za-z0-9.]+]] = add i64 %[[VAR21]], 15
-  // CHECK-NEXT: %[[VAR24:[A-Za-z0-9.]+]] = inttoptr i64 %[[VAR22]] to i8*
-  // CHECK-NEXT: %[[VAR25:[A-Za-z0-9.]+]] = inttoptr i64 %[[VAR23]] to i8*
-  // CHECK-NEXT: %[[VAR26:[A-Za-z0-9.]+]] = load i8, i8* %[[VAR24]]
-  // CHECK-NEXT: %[[VAR27:[A-Za-z0-9.]+]] = load i8, i8* %[[VAR25]]
+  // CHECK-NEXT: %[[VAR24:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR60]], i64 7
+  // CHECK-NEXT: %[[VAR25:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR60]], i64 15
+  // CHECK-NEXT: %[[VAR26:[A-Za-z0-9.]+]] = load i8, i8* %[[VAR24]], align 1
+  // CHECK-NEXT: %[[VAR27:[A-Za-z0-9.]+]] = load i8, i8* %[[VAR25]], align 1
   // CHECK-NEXT: %[[VAR28:[A-Za-z0-9.]+]] = getelementptr inbounds { i8, i8 }, { i8, i8 }* %[[VAR20:[A-Za-z0-9.]+]], i32 0, i32 0
   // CHECK-NEXT: %[[VAR29:[A-Za-z0-9.]+]] = getelementptr inbounds { i8, i8 }, { i8, i8 }* %[[VAR20]], i32 0, i32 1
   // CHECK-NEXT: store i8 %[[VAR26]], i8* %[[VAR28]]
@@ -57,15 +52,14 @@ void testva (int n, ...)
 
   _Complex float f = va_arg(ap, _Complex float);
   // CHECK: %[[VAR70:[A-Za-z0-9.]+]] = load i8*, i8** %[[VAR100:[A-Za-z0-9.]+]]
-  // CHECK-NEXT: %[[VAR71:[A-Za-z0-9.]+]] = getelementptr i8, i8* %[[VAR70]], i64 16
+  // CHECK-NEXT: %[[VAR71:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR70]], i64 16
   // CHECK-NEXT: store i8* %[[VAR71]], i8** %[[VAR100]]
-  // CHECK: %[[VAR31:[A-Za-z0-9.]+]] = ptrtoint i8* %{{[A-Za-z0-9.]+}} to i64
-  // CHECK-NEXT: %[[VAR32:[A-Za-z0-9.]+]] = add i64 %[[VAR31]], 4
-  // CHECK-NEXT: %[[VAR33:[A-Za-z0-9.]+]] = add i64 %[[VAR31]], 12
-  // CHECK-NEXT: %[[VAR34:[A-Za-z0-9.]+]] = inttoptr i64 %[[VAR32]] to float*
-  // CHECK-NEXT: %[[VAR35:[A-Za-z0-9.]+]] = inttoptr i64 %[[VAR33]] to float*
-  // CHECK-NEXT: %[[VAR36:[A-Za-z0-9.]+]] = load float, float* %[[VAR34]]
-  // CHECK-NEXT: %[[VAR37:[A-Za-z0-9.]+]] = load float, float* %[[VAR35]]
+  // CHECK-NEXT: %[[VAR32:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR70]], i64 4
+  // CHECK-NEXT: %[[VAR33:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR70]], i64 12
+  // CHECK-NEXT: %[[VAR34:[A-Za-z0-9.]+]] = bitcast i8* %[[VAR32]] to float*
+  // CHECK-NEXT: %[[VAR35:[A-Za-z0-9.]+]] = bitcast i8* %[[VAR33]] to float*
+  // CHECK-NEXT: %[[VAR36:[A-Za-z0-9.]+]] = load float, float* %[[VAR34]], align 4
+  // CHECK-NEXT: %[[VAR37:[A-Za-z0-9.]+]] = load float, float* %[[VAR35]], align 4
   // CHECK-NEXT: %[[VAR38:[A-Za-z0-9.]+]] = getelementptr inbounds { float, float }, { float, float }* %[[VAR30:[A-Za-z0-9.]+]], i32 0, i32 0
   // CHECK-NEXT: %[[VAR39:[A-Za-z0-9.]+]] = getelementptr inbounds { float, float }, { float, float }* %[[VAR30]], i32 0, i32 1
   // CHECK-NEXT: store float %[[VAR36]], float* %[[VAR38]]
diff --git a/test/CodeGen/ppc64le-varargs-complex.c b/test/CodeGen/ppc64le-varargs-complex.c
index 68dfa0b69f..399371b6e6 100644
--- a/test/CodeGen/ppc64le-varargs-complex.c
+++ b/test/CodeGen/ppc64le-varargs-complex.c
@@ -9,14 +9,13 @@ void testva (int n, ...)
 
   _Complex int i   = va_arg(ap, _Complex int);
   // CHECK: %[[VAR40:[A-Za-z0-9.]+]] = load i8*, i8** %[[VAR100:[A-Za-z0-9.]+]]
-  // CHECK-NEXT: %[[VAR41:[A-Za-z0-9.]+]] = getelementptr i8, i8* %[[VAR40]], i64 16
+  // CHECK-NEXT: %[[VAR41:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR40]], i64 16
   // CHECK-NEXT: store i8* %[[VAR41]], i8** %[[VAR100]]
-  // CHECK-NEXT: %[[VAR1:[A-Za-z0-9.]+]] = ptrtoint i8* %[[VAR40]] to i64
-  // CHECK-NEXT: %[[VAR3:[A-Za-z0-9.]+]] = add i64 %[[VAR1]], 8
-  // CHECK-NEXT: %[[VAR4:[A-Za-z0-9.]+]] = inttoptr i64 %[[VAR1]] to i32*
-  // CHECK-NEXT: %[[VAR5:[A-Za-z0-9.]+]] = inttoptr i64 %[[VAR3]] to i32*
-  // CHECK-NEXT: %[[VAR6:[A-Za-z0-9.]+]] = load i32, i32* %[[VAR4]]
-  // CHECK-NEXT: %[[VAR7:[A-Za-z0-9.]+]] = load i32, i32* %[[VAR5]]
+  // CHECK-NEXT: %[[VAR3:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR40]], i64 8
+  // CHECK-NEXT: %[[VAR4:[A-Za-z0-9.]+]] = bitcast i8* %[[VAR40]] to i32*
+  // CHECK-NEXT: %[[VAR5:[A-Za-z0-9.]+]] = bitcast i8* %[[VAR3]] to i32*
+  // CHECK-NEXT: %[[VAR6:[A-Za-z0-9.]+]] = load i32, i32* %[[VAR4]], align 8
+  // CHECK-NEXT: %[[VAR7:[A-Za-z0-9.]+]] = load i32, i32* %[[VAR5]], align 8
   // CHECK-NEXT: %[[VAR8:[A-Za-z0-9.]+]] = getelementptr inbounds { i32, i32 }, { i32, i32 }* %[[VAR0:[A-Za-z0-9.]+]], i32 0, i32 0
   // CHECK-NEXT: %[[VAR9:[A-Za-z0-9.]+]] = getelementptr inbounds { i32, i32 }, { i32, i32 }* %[[VAR0]], i32 0, i32 1
   // CHECK-NEXT: store i32 %[[VAR6]], i32* %[[VAR8]]
@@ -24,14 +23,13 @@ void testva (int n, ...)
 
   _Complex short s = va_arg(ap, _Complex short);
   // CHECK: %[[VAR50:[A-Za-z0-9.]+]] = load i8*, i8** %[[VAR100:[A-Za-z0-9.]+]]
-  // CHECK-NEXT: %[[VAR51:[A-Za-z0-9.]+]] = getelementptr i8, i8* %[[VAR50]], i64 16
+  // CHECK-NEXT: %[[VAR51:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR50]], i64 16
   // CHECK-NEXT: store i8* %[[VAR51]], i8** %[[VAR100]]
-  // CHECK: %[[VAR11:[A-Za-z0-9.]+]] = ptrtoint i8* %{{[A-Za-z0-9.]+}} to i64
-  // CHECK-NEXT: %[[VAR13:[A-Za-z0-9.]+]] = add i64 %[[VAR11]], 8
-  // CHECK-NEXT: %[[VAR14:[A-Za-z0-9.]+]] = inttoptr i64 %[[VAR11]] to i16*
-  // CHECK-NEXT: %[[VAR15:[A-Za-z0-9.]+]] = inttoptr i64 %[[VAR13]] to i16*
-  // CHECK-NEXT: %[[VAR16:[A-Za-z0-9.]+]] = load i16, i16* %[[VAR14]]
-  // CHECK-NEXT: %[[VAR17:[A-Za-z0-9.]+]] = load i16, i16* %[[VAR15]]
+  // CHECK-NEXT: %[[VAR13:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR50]], i64 8
+  // CHECK-NEXT: %[[VAR14:[A-Za-z0-9.]+]] = bitcast i8* %[[VAR50]] to i16*
+  // CHECK-NEXT: %[[VAR15:[A-Za-z0-9.]+]] = bitcast i8* %[[VAR13]] to i16*
+  // CHECK-NEXT: %[[VAR16:[A-Za-z0-9.]+]] = load i16, i16* %[[VAR14]], align 8
+  // CHECK-NEXT: %[[VAR17:[A-Za-z0-9.]+]] = load i16, i16* %[[VAR15]], align 8
   // CHECK-NEXT: %[[VAR18:[A-Za-z0-9.]+]] = getelementptr inbounds { i16, i16 }, { i16, i16 }* %[[VAR10:[A-Za-z0-9.]+]], i32 0, i32 0
   // CHECK-NEXT: %[[VAR19:[A-Za-z0-9.]+]] = getelementptr inbounds { i16, i16 }, { i16, i16 }* %[[VAR10]], i32 0, i32 1
   // CHECK-NEXT: store i16 %[[VAR16]], i16* %[[VAR18]]
@@ -39,14 +37,11 @@ void testva (int n, ...)
 
   _Complex char c  = va_arg(ap, _Complex char);
   // CHECK: %[[VAR60:[A-Za-z0-9.]+]] = load i8*, i8** %[[VAR100:[A-Za-z0-9.]+]]
-  // CHECK-NEXT: %[[VAR61:[A-Za-z0-9.]+]] = getelementptr i8, i8* %[[VAR60]], i64 16
+  // CHECK-NEXT: %[[VAR61:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR60]], i64 16
   // CHECK-NEXT: store i8* %[[VAR61]], i8** %[[VAR100]]
-  // CHECK: %[[VAR21:[A-Za-z0-9.]+]] = ptrtoint i8* %{{[A-Za-z0-9.]+}} to i64
-  // CHECK-NEXT: %[[VAR23:[A-Za-z0-9.]+]] = add i64 %[[VAR21]], 8
-  // CHECK-NEXT: %[[VAR24:[A-Za-z0-9.]+]] = inttoptr i64 %[[VAR21]] to i8*
-  // CHECK-NEXT: %[[VAR25:[A-Za-z0-9.]+]] = inttoptr i64 %[[VAR23]] to i8*
-  // CHECK-NEXT: %[[VAR26:[A-Za-z0-9.]+]] = load i8, i8* %[[VAR24]]
-  // CHECK-NEXT: %[[VAR27:[A-Za-z0-9.]+]] = load i8, i8* %[[VAR25]]
+  // CHECK-NEXT: %[[VAR25:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR60]], i64 8
+  // CHECK-NEXT: %[[VAR26:[A-Za-z0-9.]+]] = load i8, i8* %[[VAR60]], align 8
+  // CHECK-NEXT: %[[VAR27:[A-Za-z0-9.]+]] = load i8, i8* %[[VAR25]], align 8
   // CHECK-NEXT: %[[VAR28:[A-Za-z0-9.]+]] = getelementptr inbounds { i8, i8 }, { i8, i8 }* %[[VAR20:[A-Za-z0-9.]+]], i32 0, i32 0
   // CHECK-NEXT: %[[VAR29:[A-Za-z0-9.]+]] = getelementptr inbounds { i8, i8 }, { i8, i8 }* %[[VAR20]], i32 0, i32 1
   // CHECK-NEXT: store i8 %[[VAR26]], i8* %[[VAR28]]
@@ -54,14 +49,13 @@ void testva (int n, ...)
 
   _Complex float f = va_arg(ap, _Complex float);
   // CHECK: %[[VAR70:[A-Za-z0-9.]+]] = load i8*, i8** %[[VAR100:[A-Za-z0-9.]+]]
-  // CHECK-NEXT: %[[VAR71:[A-Za-z0-9.]+]] = getelementptr i8, i8* %[[VAR70]], i64 16
+  // CHECK-NEXT: %[[VAR71:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR70]], i64 16
   // CHECK-NEXT: store i8* %[[VAR71]], i8** %[[VAR100]]
-  // CHECK: %[[VAR31:[A-Za-z0-9.]+]] = ptrtoint i8* %{{[A-Za-z0-9.]+}} to i64
-  // CHECK-NEXT: %[[VAR33:[A-Za-z0-9.]+]] = add i64 %[[VAR31]], 8
-  // CHECK-NEXT: %[[VAR34:[A-Za-z0-9.]+]] = inttoptr i64 %[[VAR31]] to float*
-  // CHECK-NEXT: %[[VAR35:[A-Za-z0-9.]+]] = inttoptr i64 %[[VAR33]] to float*
-  // CHECK-NEXT: %[[VAR36:[A-Za-z0-9.]+]] = load float, float* %[[VAR34]]
-  // CHECK-NEXT: %[[VAR37:[A-Za-z0-9.]+]] = load float, float* %[[VAR35]]
+  // CHECK-NEXT: %[[VAR33:[A-Za-z0-9.]+]] = getelementptr inbounds i8, i8* %[[VAR70]], i64 8
+  // CHECK-NEXT: %[[VAR34:[A-Za-z0-9.]+]] = bitcast i8* %[[VAR70]] to float*
+  // CHECK-NEXT: %[[VAR35:[A-Za-z0-9.]+]] = bitcast i8* %[[VAR33]] to float*
+  // CHECK-NEXT: %[[VAR36:[A-Za-z0-9.]+]] = load float, float* %[[VAR34]], align 8
+  // CHECK-NEXT: %[[VAR37:[A-Za-z0-9.]+]] = load float, float* %[[VAR35]], align 8
   // CHECK-NEXT: %[[VAR38:[A-Za-z0-9.]+]] = getelementptr inbounds { float, float }, { float, float }* %[[VAR30:[A-Za-z0-9.]+]], i32 0, i32 0
   // CHECK-NEXT: %[[VAR39:[A-Za-z0-9.]+]] = getelementptr inbounds { float, float }, { float, float }* %[[VAR30]], i32 0, i32 1
   // CHECK-NEXT: store float %[[VAR36]], float* %[[VAR38]]
diff --git a/test/CodeGen/sparcv9-abi.c b/test/CodeGen/sparcv9-abi.c
index bf447198cd..5984fa558c 100644
--- a/test/CodeGen/sparcv9-abi.c
+++ b/test/CodeGen/sparcv9-abi.c
@@ -132,9 +132,9 @@ int f_variable(char *f, ...) {
   while ((c = *f++)) switch (c) {
 
 // CHECK: %[[CUR:[^ ]+]] = load i8*, i8** %ap
-// CHECK-DAG: %[[NXT:[^ ]+]] = getelementptr i8, i8* %[[CUR]], i32 8
+// CHECK-DAG: %[[NXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[CUR]], i64 8
 // CHECK-DAG: store i8* %[[NXT]], i8** %ap
-// CHECK-DAG: %[[EXT:[^ ]+]] = getelementptr i8, i8* %[[CUR]], i32 4
+// CHECK-DAG: %[[EXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[CUR]], i64 4
 // CHECK-DAG: %[[ADR:[^ ]+]] = bitcast i8* %[[EXT]] to i32*
 // CHECK-DAG: load i32, i32* %[[ADR]]
 // CHECK: br
@@ -143,7 +143,7 @@ int f_variable(char *f, ...) {
     break;
 
 // CHECK: %[[CUR:[^ ]+]] = load i8*, i8** %ap
-// CHECK-DAG: %[[NXT:[^ ]+]] = getelementptr i8, i8* %[[CUR]], i32 8
+// CHECK-DAG: %[[NXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[CUR]], i64 8
 // CHECK-DAG: store i8* %[[NXT]], i8** %ap
 // CHECK-DAG: %[[ADR:[^ ]+]] = bitcast i8* %[[CUR]] to i64*
 // CHECK-DAG: load i64, i64* %[[ADR]]
@@ -153,7 +153,7 @@ int f_variable(char *f, ...) {
     break;
 
 // CHECK: %[[CUR:[^ ]+]] = load i8*, i8** %ap
-// CHECK-DAG: %[[NXT:[^ ]+]] = getelementptr i8, i8* %[[CUR]], i32 8
+// CHECK-DAG: %[[NXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[CUR]], i64 8
 // CHECK-DAG: store i8* %[[NXT]], i8** %ap
 // CHECK-DAG: %[[ADR:[^ ]+]] = bitcast i8* %[[CUR]] to %struct.tiny*
 // CHECK: br
@@ -162,7 +162,7 @@ int f_variable(char *f, ...) {
     break;
 
 // CHECK: %[[CUR:[^ ]+]] = load i8*, i8** %ap
-// CHECK-DAG: %[[NXT:[^ ]+]] = getelementptr i8, i8* %[[CUR]], i32 16
+// CHECK-DAG: %[[NXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[CUR]], i64 16
 // CHECK-DAG: store i8* %[[NXT]], i8** %ap
 // CHECK-DAG: %[[ADR:[^ ]+]] = bitcast i8* %[[CUR]] to %struct.small*
 // CHECK: br
@@ -171,7 +171,7 @@ int f_variable(char *f, ...) {
     break;
 
 // CHECK: %[[CUR:[^ ]+]] = load i8*, i8** %ap
-// CHECK-DAG: %[[NXT:[^ ]+]] = getelementptr i8, i8* %[[CUR]], i32 8
+// CHECK-DAG: %[[NXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[CUR]], i64 8
 // CHECK-DAG: store i8* %[[NXT]], i8** %ap
 // CHECK-DAG: %[[IND:[^ ]+]] = bitcast i8* %[[CUR]] to %struct.medium**
 // CHECK-DAG: %[[ADR:[^ ]+]] = load %struct.medium*, %struct.medium** %[[IND]]
diff --git a/test/CodeGen/tbaa-class.cpp b/test/CodeGen/tbaa-class.cpp
index a8005d6057..f611ae5abb 100644
--- a/test/CodeGen/tbaa-class.cpp
+++ b/test/CodeGen/tbaa-class.cpp
@@ -51,10 +51,10 @@ public:
 };
 
 uint32_t g(uint32_t *s, StructA *A, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z1g
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32:!.*]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z1g
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32:!.*]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_A_f32:!.*]]
   *s = 1;
@@ -63,22 +63,22 @@ uint32_t g(uint32_t *s, StructA *A, uint64_t count) {
 }
 
 uint32_t g2(uint32_t *s, StructA *A, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z2g2
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// CHECK: store i16 4, i16* %{{.*}}, align 2, !tbaa [[TAG_i16:!.*]]
-// PATH: define i32 @{{.*}}(
+// CHECK: store i16 4, i16* %{{.*}}, align 4, !tbaa [[TAG_i16:!.*]]
+// PATH-LABEL: define i32 @_Z2g2
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: store i16 4, i16* %{{.*}}, align 2, !tbaa [[TAG_A_f16:!.*]]
+// PATH: store i16 4, i16* %{{.*}}, align 4, !tbaa [[TAG_A_f16:!.*]]
   *s = 1;
   A->f16 = 4;
   return *s;
 }
 
 uint32_t g3(StructA *A, StructB *B, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z2g3
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z2g3
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_A_f32]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_B_a_f32:!.*]]
   A->f32 = 1;
@@ -87,22 +87,22 @@ uint32_t g3(StructA *A, StructB *B, uint64_t count) {
 }
 
 uint32_t g4(StructA *A, StructB *B, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z2g4
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// CHECK: store i16 4, i16* %{{.*}}, align 2, !tbaa [[TAG_i16]]
-// PATH: define i32 @{{.*}}(
+// CHECK: store i16 4, i16* %{{.*}}, align 4, !tbaa [[TAG_i16]]
+// PATH-LABEL: define i32 @_Z2g4
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_A_f32]]
-// PATH: store i16 4, i16* %{{.*}}, align 2, !tbaa [[TAG_B_a_f16:!.*]]
+// PATH: store i16 4, i16* %{{.*}}, align 4, !tbaa [[TAG_B_a_f16:!.*]]
   A->f32 = 1;
   B->a.f16 = 4;
   return A->f32;
 }
 
 uint32_t g5(StructA *A, StructB *B, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z2g5
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z2g5
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_A_f32]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_B_f32:!.*]]
   A->f32 = 1;
@@ -111,10 +111,10 @@ uint32_t g5(StructA *A, StructB *B, uint64_t count) {
 }
 
 uint32_t g6(StructA *A, StructB *B, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z2g6
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z2g6
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_A_f32]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_B_a_f32_2:!.*]]
   A->f32 = 1;
@@ -123,10 +123,10 @@ uint32_t g6(StructA *A, StructB *B, uint64_t count) {
 }
 
 uint32_t g7(StructA *A, StructS *S, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z2g7
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z2g7
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_A_f32]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_S_f32:!.*]]
   A->f32 = 1;
@@ -135,22 +135,22 @@ uint32_t g7(StructA *A, StructS *S, uint64_t count) {
 }
 
 uint32_t g8(StructA *A, StructS *S, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z2g8
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// CHECK: store i16 4, i16* %{{.*}}, align 2, !tbaa [[TAG_i16]]
-// PATH: define i32 @{{.*}}(
+// CHECK: store i16 4, i16* %{{.*}}, align 4, !tbaa [[TAG_i16]]
+// PATH-LABEL: define i32 @_Z2g8
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_A_f32]]
-// PATH: store i16 4, i16* %{{.*}}, align 2, !tbaa [[TAG_S_f16:!.*]]
+// PATH: store i16 4, i16* %{{.*}}, align 4, !tbaa [[TAG_S_f16:!.*]]
   A->f32 = 1;
   S->f16 = 4;
   return A->f32;
 }
 
 uint32_t g9(StructS *S, StructS2 *S2, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z2g9
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z2g9
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_S_f32]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_S_f32:!.*]]
   S->f32 = 1;
@@ -159,10 +159,10 @@ uint32_t g9(StructS *S, StructS2 *S2, uint64_t count) {
 }
 
 uint32_t g10(StructS *S, StructS2 *S2, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z3g10
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z3g10
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_S_f32]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_S2_f32_2:!.*]]
   S->f32 = 1;
@@ -171,10 +171,10 @@ uint32_t g10(StructS *S, StructS2 *S2, uint64_t count) {
 }
 
 uint32_t g11(StructC *C, StructD *D, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z3g11
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z3g11
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_C_b_a_f32:!.*]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_D_b_a_f32:!.*]]
   C->b.a.f32 = 1;
@@ -183,11 +183,11 @@ uint32_t g11(StructC *C, StructD *D, uint64_t count) {
 }
 
 uint32_t g12(StructC *C, StructD *D, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z3g12
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // TODO: differentiate the two accesses.
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z3g12
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_B_a_f32]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_B_a_f32]]
   StructB *b1 = &(C->b);
diff --git a/test/CodeGen/tbaa.cpp b/test/CodeGen/tbaa.cpp
index 2bff5d0ba0..c43ca58bc3 100644
--- a/test/CodeGen/tbaa.cpp
+++ b/test/CodeGen/tbaa.cpp
@@ -45,10 +45,10 @@ typedef struct
 } StructS2;
 
 uint32_t g(uint32_t *s, StructA *A, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z1g
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32:!.*]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z1g
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32:!.*]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_A_f32:!.*]]
   *s = 1;
@@ -57,22 +57,22 @@ uint32_t g(uint32_t *s, StructA *A, uint64_t count) {
 }
 
 uint32_t g2(uint32_t *s, StructA *A, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z2g2
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// CHECK: store i16 4, i16* %{{.*}}, align 2, !tbaa [[TAG_i16:!.*]]
-// PATH: define i32 @{{.*}}(
+// CHECK: store i16 4, i16* %{{.*}}, align 4, !tbaa [[TAG_i16:!.*]]
+// PATH-LABEL: define i32 @_Z2g2
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: store i16 4, i16* %{{.*}}, align 2, !tbaa [[TAG_A_f16:!.*]]
+// PATH: store i16 4, i16* %{{.*}}, align 4, !tbaa [[TAG_A_f16:!.*]]
   *s = 1;
   A->f16 = 4;
   return *s;
 }
 
 uint32_t g3(StructA *A, StructB *B, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z2g3
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z2g3
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_A_f32]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_B_a_f32:!.*]]
   A->f32 = 1;
@@ -81,22 +81,22 @@ uint32_t g3(StructA *A, StructB *B, uint64_t count) {
 }
 
 uint32_t g4(StructA *A, StructB *B, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z2g4
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// CHECK: store i16 4, i16* %{{.*}}, align 2, !tbaa [[TAG_i16]]
-// PATH: define i32 @{{.*}}(
+// CHECK: store i16 4, i16* %{{.*}}, align 4, !tbaa [[TAG_i16]]
+// PATH-LABEL: define i32 @_Z2g4
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_A_f32]]
-// PATH: store i16 4, i16* %{{.*}}, align 2, !tbaa [[TAG_B_a_f16:!.*]]
+// PATH: store i16 4, i16* %{{.*}}, align 4, !tbaa [[TAG_B_a_f16:!.*]]
   A->f32 = 1;
   B->a.f16 = 4;
   return A->f32;
 }
 
 uint32_t g5(StructA *A, StructB *B, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z2g5
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z2g5
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_A_f32]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_B_f32:!.*]]
   A->f32 = 1;
@@ -105,10 +105,10 @@ uint32_t g5(StructA *A, StructB *B, uint64_t count) {
 }
 
 uint32_t g6(StructA *A, StructB *B, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z2g6
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z2g6
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_A_f32]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_B_a_f32_2:!.*]]
   A->f32 = 1;
@@ -117,10 +117,10 @@ uint32_t g6(StructA *A, StructB *B, uint64_t count) {
 }
 
 uint32_t g7(StructA *A, StructS *S, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z2g7
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z2g7
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_A_f32]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_S_f32:!.*]]
   A->f32 = 1;
@@ -129,22 +129,22 @@ uint32_t g7(StructA *A, StructS *S, uint64_t count) {
 }
 
 uint32_t g8(StructA *A, StructS *S, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z2g8
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// CHECK: store i16 4, i16* %{{.*}}, align 2, !tbaa [[TAG_i16]]
-// PATH: define i32 @{{.*}}(
+// CHECK: store i16 4, i16* %{{.*}}, align 4, !tbaa [[TAG_i16]]
+// PATH-LABEL: define i32 @_Z2g8
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_A_f32]]
-// PATH: store i16 4, i16* %{{.*}}, align 2, !tbaa [[TAG_S_f16:!.*]]
+// PATH: store i16 4, i16* %{{.*}}, align 4, !tbaa [[TAG_S_f16:!.*]]
   A->f32 = 1;
   S->f16 = 4;
   return A->f32;
 }
 
 uint32_t g9(StructS *S, StructS2 *S2, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z2g9
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z2g9
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_S_f32]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_S2_f32:!.*]]
   S->f32 = 1;
@@ -153,22 +153,22 @@ uint32_t g9(StructS *S, StructS2 *S2, uint64_t count) {
 }
 
 uint32_t g10(StructS *S, StructS2 *S2, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z3g10
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// CHECK: store i16 4, i16* %{{.*}}, align 2, !tbaa [[TAG_i16]]
-// PATH: define i32 @{{.*}}(
+// CHECK: store i16 4, i16* %{{.*}}, align 4, !tbaa [[TAG_i16]]
+// PATH-LABEL: define i32 @_Z3g10
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_S_f32]]
-// PATH: store i16 4, i16* %{{.*}}, align 2, !tbaa [[TAG_S2_f16:!.*]]
+// PATH: store i16 4, i16* %{{.*}}, align 4, !tbaa [[TAG_S2_f16:!.*]]
   S->f32 = 1;
   S2->f16 = 4;
   return S->f32;
 }
 
 uint32_t g11(StructC *C, StructD *D, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z3g11
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z3g11
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_C_b_a_f32:!.*]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_D_b_a_f32:!.*]]
   C->b.a.f32 = 1;
@@ -177,11 +177,11 @@ uint32_t g11(StructC *C, StructD *D, uint64_t count) {
 }
 
 uint32_t g12(StructC *C, StructD *D, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z3g12
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // TODO: differentiate the two accesses.
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z3g12
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_B_a_f32]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_B_a_f32]]
   StructB *b1 = &(C->b);
@@ -202,9 +202,9 @@ struct five {
 } ATTR;
 char g13(struct five *a, struct five *b) {
   return a->b;
-// CHECK: define signext i8 @{{.*}}(
+// CHECK-LABEL: define signext i8 @_Z3g13
 // CHECK: load i8, i8* %{{.*}}, align 1, !tbaa [[TAG_char:!.*]]
-// PATH: define signext i8 @{{.*}}(
+// PATH-LABEL: define signext i8 @_Z3g13
 // PATH: load i8, i8* %{{.*}}, align 1, !tbaa [[TAG_five_b:!.*]]
 }
 
@@ -215,9 +215,9 @@ struct six {
   char c;
 };
 char g14(struct six *a, struct six *b) {
-// CHECK: define signext i8 @{{.*}}(
+// CHECK-LABEL: define signext i8 @_Z3g14
 // CHECK: load i8, i8* %{{.*}}, align 1, !tbaa [[TAG_char]]
-// PATH: define signext i8 @{{.*}}(
+// PATH-LABEL: define signext i8 @_Z3g14
 // PATH: load i8, i8* %{{.*}}, align 1, !tbaa [[TAG_six_b:!.*]]
   return a->b;
 }
@@ -225,10 +225,10 @@ char g14(struct six *a, struct six *b) {
 // Types that differ only by name may alias.
 typedef StructS StructS3;
 uint32_t g15(StructS *S, StructS3 *S3, uint64_t count) {
-// CHECK: define i32 @{{.*}}(
+// CHECK-LABEL: define i32 @_Z3g15
 // CHECK: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
 // CHECK: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_i32]]
-// PATH: define i32 @{{.*}}(
+// PATH-LABEL: define i32 @_Z3g15
 // PATH: store i32 1, i32* %{{.*}}, align 4, !tbaa [[TAG_S_f32]]
 // PATH: store i32 4, i32* %{{.*}}, align 4, !tbaa [[TAG_S_f32]]
   S->f32 = 1;
diff --git a/test/CodeGen/vectorcall.c b/test/CodeGen/vectorcall.c
index 17927c7a3d..9ee35b1a02 100644
--- a/test/CodeGen/vectorcall.c
+++ b/test/CodeGen/vectorcall.c
@@ -32,13 +32,13 @@ void __vectorcall hfa1(int a, struct HFA4 b, int c) {}
 // registers.
 void __vectorcall hfa2(struct HFA4 a, struct HFA4 b, double c) {}
 // CHECK: define x86_vectorcallcc void @"\01hfa2@@72"(double %a.0, double %a.1, double %a.2, double %a.3, %struct.HFA4* inreg %b, double %c)
-// X64: define x86_vectorcallcc void @"\01hfa2@@72"(double %a.0, double %a.1, double %a.2, double %a.3, %struct.HFA4* align 8 %b, double %c)
+// X64: define x86_vectorcallcc void @"\01hfa2@@72"(double %a.0, double %a.1, double %a.2, double %a.3, %struct.HFA4* %b, double %c)
 
 // Ensure that we pass builtin types directly while counting them against the
 // SSE register usage.
 void __vectorcall hfa3(double a, double b, double c, double d, double e, struct HFA2 f) {}
 // CHECK: define x86_vectorcallcc void @"\01hfa3@@56"(double %a, double %b, double %c, double %d, double %e, %struct.HFA2* inreg %f)
-// X64: define x86_vectorcallcc void @"\01hfa3@@56"(double %a, double %b, double %c, double %d, double %e, %struct.HFA2* align 8 %f)
+// X64: define x86_vectorcallcc void @"\01hfa3@@56"(double %a, double %b, double %c, double %d, double %e, %struct.HFA2* %f)
 
 // Aggregates with more than four elements are not HFAs and are passed byval.
 // Because they are not classified as homogeneous, they don't get special
@@ -63,11 +63,11 @@ void __vectorcall hva1(int a, struct HVA4 b, int c) {}
 
 void __vectorcall hva2(struct HVA4 a, struct HVA4 b, v4f32 c) {}
 // CHECK: define x86_vectorcallcc void @"\01hva2@@144"(<4 x float> %a.0, <4 x float> %a.1, <4 x float> %a.2, <4 x float> %a.3, %struct.HVA4* inreg %b, <4 x float> %c)
-// X64: define x86_vectorcallcc void @"\01hva2@@144"(<4 x float> %a.0, <4 x float> %a.1, <4 x float> %a.2, <4 x float> %a.3, %struct.HVA4* align 16 %b, <4 x float> %c)
+// X64: define x86_vectorcallcc void @"\01hva2@@144"(<4 x float> %a.0, <4 x float> %a.1, <4 x float> %a.2, <4 x float> %a.3, %struct.HVA4* %b, <4 x float> %c)
 
 void __vectorcall hva3(v4f32 a, v4f32 b, v4f32 c, v4f32 d, v4f32 e, struct HVA2 f) {}
 // CHECK: define x86_vectorcallcc void @"\01hva3@@112"(<4 x float> %a, <4 x float> %b, <4 x float> %c, <4 x float> %d, <4 x float> %e, %struct.HVA2* inreg %f)
-// X64: define x86_vectorcallcc void @"\01hva3@@112"(<4 x float> %a, <4 x float> %b, <4 x float> %c, <4 x float> %d, <4 x float> %e, %struct.HVA2* align 16 %f)
+// X64: define x86_vectorcallcc void @"\01hva3@@112"(<4 x float> %a, <4 x float> %b, <4 x float> %c, <4 x float> %d, <4 x float> %e, %struct.HVA2* %f)
 
 typedef float __attribute__((ext_vector_type(3))) v3f32;
 struct OddSizeHVA { v3f32 x, y; };
diff --git a/test/CodeGen/xcore-abi.c b/test/CodeGen/xcore-abi.c
index 23fb441469..2bac78d92e 100644
--- a/test/CodeGen/xcore-abi.c
+++ b/test/CodeGen/xcore-abi.c
@@ -33,7 +33,7 @@ void testva (int n, ...) {
   f(v1);
   // CHECK: [[I:%[a-z0-9]+]] = load i8*, i8** [[AP]]
   // CHECK: [[P:%[a-z0-9]+]] = bitcast i8* [[I]] to i8**
-  // CHECK: [[IN:%[a-z0-9]+]] = getelementptr i8, i8* [[I]], i32 4
+  // CHECK: [[IN:%[a-z0-9]+]] = getelementptr inbounds i8, i8* [[I]], i32 4
   // CHECK: store i8* [[IN]], i8** [[AP]]
   // CHECK: [[V1:%[a-z0-9]+]] = load i8*, i8** [[P]]
   // CHECK: store i8* [[V1]], i8** [[V:%[a-z0-9]+]], align 4
@@ -43,7 +43,7 @@ void testva (int n, ...) {
   char v2 = va_arg (ap, char); // expected-warning{{second argument to 'va_arg' is of promotable type 'char'}}
   f(&v2);
   // CHECK: [[I:%[a-z0-9]+]] = load i8*, i8** [[AP]]
-  // CHECK: [[IN:%[a-z0-9]+]] = getelementptr i8, i8* [[I]], i32 4
+  // CHECK: [[IN:%[a-z0-9]+]] = getelementptr inbounds i8, i8* [[I]], i32 4
   // CHECK: store i8* [[IN]], i8** [[AP]]
   // CHECK: [[V1:%[a-z0-9]+]] = load i8, i8* [[I]]
   // CHECK: store i8 [[V1]], i8* [[V:%[a-z0-9]+]], align 1
@@ -53,7 +53,7 @@ void testva (int n, ...) {
   f(&v3);
   // CHECK: [[I:%[a-z0-9]+]] = load i8*, i8** [[AP]]
   // CHECK: [[P:%[a-z0-9]+]] = bitcast i8* [[I]] to i32*
-  // CHECK: [[IN:%[a-z0-9]+]] = getelementptr i8, i8* [[I]], i32 4
+  // CHECK: [[IN:%[a-z0-9]+]] = getelementptr inbounds i8, i8* [[I]], i32 4
   // CHECK: store i8* [[IN]], i8** [[AP]]
   // CHECK: [[V1:%[a-z0-9]+]] = load i32, i32* [[P]]
   // CHECK: store i32 [[V1]], i32* [[V:%[a-z0-9]+]], align 4
@@ -64,7 +64,7 @@ void testva (int n, ...) {
   f(&v4);
   // CHECK: [[I:%[a-z0-9]+]] = load i8*, i8** [[AP]]
   // CHECK: [[P:%[a-z0-9]+]] = bitcast i8* [[I]] to i64*
-  // CHECK: [[IN:%[a-z0-9]+]] = getelementptr i8, i8* [[I]], i32 8
+  // CHECK: [[IN:%[a-z0-9]+]] = getelementptr inbounds i8, i8* [[I]], i32 8
   // CHECK: store i8* [[IN]], i8** [[AP]]
   // CHECK: [[V1:%[a-z0-9]+]] = load i64, i64* [[P]]
   // CHECK: store i64 [[V1]], i64* [[V:%[a-z0-9]+]], align 4
@@ -76,7 +76,7 @@ void testva (int n, ...) {
   // CHECK: [[I:%[a-z0-9]+]] = load i8*, i8** [[AP]]
   // CHECK: [[I2:%[a-z0-9]+]] = bitcast i8* [[I]] to %struct.x**
   // CHECK: [[P:%[a-z0-9]+]] = load %struct.x*, %struct.x** [[I2]]
-  // CHECK: [[IN:%[a-z0-9]+]] = getelementptr i8, i8* [[I]], i32 4
+  // CHECK: [[IN:%[a-z0-9]+]] = getelementptr inbounds i8, i8* [[I]], i32 4
   // CHECK: store i8* [[IN]], i8** [[AP]]
   // CHECK: [[V1:%[a-z0-9]+]] = bitcast %struct.x* [[V:%[a-z0-9]+]] to i8*
   // CHECK: [[P1:%[a-z0-9]+]] = bitcast %struct.x* [[P]] to i8*
@@ -89,7 +89,7 @@ void testva (int n, ...) {
   // CHECK: [[I:%[a-z0-9]+]] = load i8*, i8** [[AP]]
   // CHECK: [[I2:%[a-z0-9]+]] = bitcast i8* [[I]] to [4 x i32]**
   // CHECK: [[P:%[a-z0-9]+]] = load [4 x i32]*, [4 x i32]** [[I2]]
-  // CHECK: [[IN:%[a-z0-9]+]] = getelementptr i8, i8* [[I]], i32 4
+  // CHECK: [[IN:%[a-z0-9]+]] = getelementptr inbounds i8, i8* [[I]], i32 4
   // CHECK: store i8* [[IN]], i8** [[AP]]
   // CHECK: [[V1:%[a-z0-9]+]] = bitcast [4 x i32]* [[V0:%[a-z0-9]+]] to i8*
   // CHECK: [[P1:%[a-z0-9]+]] = bitcast [4 x i32]* [[P]] to i8*
@@ -104,7 +104,7 @@ void testva (int n, ...) {
   f(&v7);
   // CHECK: [[I:%[a-z0-9]+]] = load i8*, i8** [[AP]]
   // CHECK: [[P:%[a-z0-9]+]] = bitcast i8* [[I]] to double*
-  // CHECK: [[IN:%[a-z0-9]+]] = getelementptr i8, i8* [[I]], i32 8
+  // CHECK: [[IN:%[a-z0-9]+]] = getelementptr inbounds i8, i8* [[I]], i32 8
   // CHECK: store i8* [[IN]], i8** [[AP]]
   // CHECK: [[V1:%[a-z0-9]+]] = load double, double* [[P]]
   // CHECK: store double [[V1]], double* [[V:%[a-z0-9]+]], align 4
diff --git a/test/CodeGenCXX/alignment.cpp b/test/CodeGenCXX/alignment.cpp
new file mode 100644
index 0000000000..62648c7317
--- /dev/null
+++ b/test/CodeGenCXX/alignment.cpp
@@ -0,0 +1,297 @@
+// RUN: %clang_cc1 %s -emit-llvm -o - -triple=x86_64-apple-darwin10 | FileCheck %s
+
+extern int int_source();
+extern void int_sink(int x);
+
+namespace test0 {
+  struct A {
+    int aField;
+    int bField;
+  };
+
+  struct B {
+    int onebit : 2;
+    int twobit : 6;
+    int intField;
+  };
+
+  struct __attribute__((packed, aligned(2))) C : A, B {
+  };
+
+  // These accesses should have alignment 4 because they're at offset 0
+  // in a reference with an assumed alignment of 4.
+  // CHECK-LABEL: @_ZN5test01aERNS_1BE
+  void a(B &b) {
+    // CHECK: [[CALL:%.*]] = call i32 @_Z10int_sourcev()
+    // CHECK: [[B_P:%.*]] = load [[B:%.*]]*, [[B]]**
+    // CHECK: [[FIELD_P:%.*]] = bitcast [[B]]* [[B_P]] to i8*
+    // CHECK: [[TRUNC:%.*]] = trunc i32 [[CALL]] to i8
+    // CHECK: [[OLD_VALUE:%.*]] = load i8, i8* [[FIELD_P]], align 4
+    // CHECK: [[T0:%.*]] = and i8 [[TRUNC]], 3
+    // CHECK: [[T1:%.*]] = and i8 [[OLD_VALUE]], -4
+    // CHECK: [[T2:%.*]] = or i8 [[T1]], [[T0]]
+    // CHECK: store i8 [[T2]], i8* [[FIELD_P]], align 4
+    b.onebit = int_source();
+    
+    // CHECK: [[B_P:%.*]] = load [[B]]*, [[B]]**
+    // CHECK: [[FIELD_P:%.*]] = bitcast [[B]]* [[B_P]] to i8*
+    // CHECK: [[VALUE:%.*]] = load i8, i8* [[FIELD_P]], align 4
+    // CHECK: [[T0:%.*]] = shl i8 [[VALUE]], 6
+    // CHECK: [[T1:%.*]] = ashr i8 [[T0]], 6
+    // CHECK: [[T2:%.*]] = sext i8 [[T1]] to i32
+    // CHECK: call void @_Z8int_sinki(i32 [[T2]])
+    int_sink(b.onebit);
+  }
+
+  // These accesses should have alignment 2 because they're at offset 8
+  // in a reference/pointer with an assumed alignment of 2.
+  // CHECK-LABEL: @_ZN5test01bERNS_1CE
+  void b(C &c) {
+    // CHECK: [[CALL:%.*]] = call i32 @_Z10int_sourcev()
+    // CHECK: [[C_P:%.*]] = load [[C:%.*]]*, [[C]]**
+    // CHECK: [[T0:%.*]] = bitcast [[C]]* [[C_P]] to i8*
+    // CHECK: [[T1:%.*]] = getelementptr inbounds i8, i8* [[T0]], i64 8
+    // CHECK: [[B_P:%.*]] = bitcast i8* [[T1]] to [[B]]*
+    // CHECK: [[FIELD_P:%.*]] = bitcast [[B]]* [[B_P]] to i8*
+    // CHECK: [[TRUNC:%.*]] = trunc i32 [[CALL]] to i8
+    // CHECK: [[OLD_VALUE:%.*]] = load i8, i8* [[FIELD_P]], align 2
+    // CHECK: [[T0:%.*]] = and i8 [[TRUNC]], 3
+    // CHECK: [[T1:%.*]] = and i8 [[OLD_VALUE]], -4
+    // CHECK: [[T2:%.*]] = or i8 [[T1]], [[T0]]
+    // CHECK: store i8 [[T2]], i8* [[FIELD_P]], align 2
+    c.onebit = int_source();
+    
+    // CHECK: [[C_P:%.*]] = load [[C]]*, [[C]]**
+    // CHECK: [[T0:%.*]] = bitcast [[C]]* [[C_P]] to i8*
+    // CHECK: [[T1:%.*]] = getelementptr inbounds i8, i8* [[T0]], i64 8
+    // CHECK: [[B_P:%.*]] = bitcast i8* [[T1]] to [[B]]*
+    // CHECK: [[FIELD_P:%.*]] = bitcast [[B]]* [[B_P]] to i8*
+    // CHECK: [[VALUE:%.*]] = load i8, i8* [[FIELD_P]], align 2
+    // CHECK: [[T0:%.*]] = shl i8 [[VALUE]], 6
+    // CHECK: [[T1:%.*]] = ashr i8 [[T0]], 6
+    // CHECK: [[T2:%.*]] = sext i8 [[T1]] to i32
+    // CHECK: call void @_Z8int_sinki(i32 [[T2]])
+    int_sink(c.onebit);
+  }
+
+  // CHECK-LABEL: @_ZN5test01cEPNS_1CE
+  void c(C *c) {
+    // CHECK: [[CALL:%.*]] = call i32 @_Z10int_sourcev()
+    // CHECK: [[C_P:%.*]] = load [[C]]*, [[C]]**
+    // CHECK: [[T0:%.*]] = bitcast [[C]]* [[C_P]] to i8*
+    // CHECK: [[T1:%.*]] = getelementptr inbounds i8, i8* [[T0]], i64 8
+    // CHECK: [[B_P:%.*]] = bitcast i8* [[T1]] to [[B]]*
+    // CHECK: [[FIELD_P:%.*]] = bitcast [[B]]* [[B_P]] to i8*
+    // CHECK: [[TRUNC:%.*]] = trunc i32 [[CALL]] to i8
+    // CHECK: [[OLD_VALUE:%.*]] = load i8, i8* [[FIELD_P]], align 2
+    // CHECK: [[T0:%.*]] = and i8 [[TRUNC]], 3
+    // CHECK: [[T1:%.*]] = and i8 [[OLD_VALUE]], -4
+    // CHECK: [[T2:%.*]] = or i8 [[T1]], [[T0]]
+    // CHECK: store i8 [[T2]], i8* [[FIELD_P]], align 2
+    c->onebit = int_source();
+
+    // CHECK: [[C_P:%.*]] = load [[C:%.*]]*, [[C]]**
+    // CHECK: [[T0:%.*]] = bitcast [[C]]* [[C_P]] to i8*
+    // CHECK: [[T1:%.*]] = getelementptr inbounds i8, i8* [[T0]], i64 8
+    // CHECK: [[B_P:%.*]] = bitcast i8* [[T1]] to [[B:%.*]]*
+    // CHECK: [[FIELD_P:%.*]] = bitcast [[B]]* [[B_P]] to i8*
+    // CHECK: [[VALUE:%.*]] = load i8, i8* [[FIELD_P]], align 2
+    // CHECK: [[T0:%.*]] = shl i8 [[VALUE]], 6
+    // CHECK: [[T1:%.*]] = ashr i8 [[T0]], 6
+    // CHECK: [[T2:%.*]] = sext i8 [[T1]] to i32
+    // CHECK: call void @_Z8int_sinki(i32 [[T2]])
+    int_sink(c->onebit);
+  }
+
+  // These accesses should have alignment 2 because they're at offset 8
+  // in an alignment-2 variable.
+  // CHECK-LABEL: @_ZN5test01dEv
+  void d() {
+    // CHECK: [[C_P:%.*]] = alloca [[C:%.*]], align 2
+    C c;
+
+    // CHECK: [[T0:%.*]] = bitcast [[C]]* [[C_P]] to i8*
+    // CHECK: [[T1:%.*]] = getelementptr inbounds i8, i8* [[T0]], i64 8
+    // CHECK: [[B_P:%.*]] = bitcast i8* [[T1]] to [[B]]*
+    // CHECK: [[FIELD_P:%.*]] = bitcast [[B]]* [[B_P]] to i8*
+    // CHECK: [[TRUNC:%.*]] = trunc i32 [[CALL]] to i8
+    // CHECK: [[OLD_VALUE:%.*]] = load i8, i8* [[FIELD_P]], align 2
+    // CHECK: [[T0:%.*]] = and i8 [[TRUNC]], 3
+    // CHECK: [[T1:%.*]] = and i8 [[OLD_VALUE]], -4
+    // CHECK: [[T2:%.*]] = or i8 [[T1]], [[T0]]
+    // CHECK: store i8 [[T2]], i8* [[FIELD_P]], align 2
+    c.onebit = int_source();
+
+    // CHECK: [[T0:%.*]] = bitcast [[C]]* [[C_P]] to i8*
+    // CHECK: [[T1:%.*]] = getelementptr inbounds i8, i8* [[T0]], i64 8
+    // CHECK: [[B_P:%.*]] = bitcast i8* [[T1]] to [[B:%.*]]*
+    // CHECK: [[FIELD_P:%.*]] = bitcast [[B]]* [[B_P]] to i8*
+    // CHECK: [[VALUE:%.*]] = load i8, i8* [[FIELD_P]], align 2
+    // CHECK: [[T0:%.*]] = shl i8 [[VALUE]], 6
+    // CHECK: [[T1:%.*]] = ashr i8 [[T0]], 6
+    // CHECK: [[T2:%.*]] = sext i8 [[T1]] to i32
+    // CHECK: call void @_Z8int_sinki(i32 [[T2]])
+    int_sink(c.onebit);
+  }
+
+  // These accesses should have alignment 8 because they're at offset 8
+  // in an alignment-16 variable.
+  // CHECK-LABEL: @_ZN5test01eEv
+  void e() {
+    // CHECK: [[C_P:%.*]] = alloca [[C:%.*]], align 16
+    __attribute__((aligned(16))) C c;
+
+    // CHECK: [[T0:%.*]] = bitcast [[C]]* [[C_P]] to i8*
+    // CHECK: [[T1:%.*]] = getelementptr inbounds i8, i8* [[T0]], i64 8
+    // CHECK: [[B_P:%.*]] = bitcast i8* [[T1]] to [[B]]*
+    // CHECK: [[FIELD_P:%.*]] = bitcast [[B]]* [[B_P]] to i8*
+    // CHECK: [[TRUNC:%.*]] = trunc i32 [[CALL]] to i8
+    // CHECK: [[OLD_VALUE:%.*]] = load i8, i8* [[FIELD_P]], align 8
+    // CHECK: [[T0:%.*]] = and i8 [[TRUNC]], 3
+    // CHECK: [[T1:%.*]] = and i8 [[OLD_VALUE]], -4
+    // CHECK: [[T2:%.*]] = or i8 [[T1]], [[T0]]
+    // CHECK: store i8 [[T2]], i8* [[FIELD_P]], align 8
+    c.onebit = int_source();
+
+    // CHECK: [[T0:%.*]] = bitcast [[C]]* [[C_P]] to i8*
+    // CHECK: [[T1:%.*]] = getelementptr inbounds i8, i8* [[T0]], i64 8
+    // CHECK: [[B_P:%.*]] = bitcast i8* [[T1]] to [[B:%.*]]*
+    // CHECK: [[FIELD_P:%.*]] = bitcast [[B]]* [[B_P]] to i8*
+    // CHECK: [[VALUE:%.*]] = load i8, i8* [[FIELD_P]], align 8
+    // CHECK: [[T0:%.*]] = shl i8 [[VALUE]], 6
+    // CHECK: [[T1:%.*]] = ashr i8 [[T0]], 6
+    // CHECK: [[T2:%.*]] = sext i8 [[T1]] to i32
+    // CHECK: call void @_Z8int_sinki(i32 [[T2]])
+    int_sink(c.onebit);
+  }
+}
+
+namespace test1 {
+  struct Array {
+    int elts[4];
+  };
+
+  struct A {
+    __attribute__((aligned(16))) Array aArray;
+  };
+
+  struct B : virtual A {
+    void *bPointer; // puts bArray at offset 16
+    Array bArray;
+  };
+
+  struct C : virtual A { // must be viable as primary base
+    // Non-empty, nv-size not a multiple of 16.
+    void *cPointer1;
+    void *cPointer2;
+  };
+
+  // Proof of concept that the non-virtual components of B do not have
+  // to be 16-byte-aligned.
+  struct D : C, B {};
+
+  // For the following tests, we want to assign into a variable whose
+  // alignment is high enough that it will absolutely not be the
+  // constraint on the memcpy alignment.
+  typedef __attribute__((aligned(64))) Array AlignedArray;
+
+  // CHECK-LABEL: @_ZN5test11aERNS_1AE
+  void a(A &a) {
+    // CHECK: [[RESULT:%.*]] = alloca [[ARRAY:%.*]], align 64
+    // CHECK: [[A_P:%.*]] = load [[A:%.*]]*, [[A]]**
+    // CHECK: [[ARRAY_P:%.*]] = getelementptr inbounds [[A]], [[A]]* [[A_P]], i32 0, i32 0
+    // CHECK: [[T0:%.*]] = bitcast [[ARRAY]]* [[RESULT]] to i8*
+    // CHECK: [[T1:%.*]] = bitcast [[ARRAY]]* [[ARRAY_P]] to i8*
+    // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[T0]], i8* [[T1]], i64 16, i32 16, i1 false)
+    AlignedArray result = a.aArray;
+  }
+
+  // CHECK-LABEL: @_ZN5test11bERNS_1BE
+  void b(B &b) {
+    // CHECK: [[RESULT:%.*]] = alloca [[ARRAY]], align 64
+    // CHECK: [[B_P:%.*]] = load [[B:%.*]]*, [[B]]**
+    // CHECK: [[VPTR_P:%.*]] = bitcast [[B]]* [[B_P]] to i8**
+    // CHECK: [[VPTR:%.*]] = load i8*, i8** [[VPTR_P]], align 8
+    // CHECK: [[T0:%.*]] = getelementptr i8, i8* [[VPTR]], i64 -24
+    // CHECK: [[OFFSET_P:%.*]] = bitcast i8* [[T0]] to i64*
+    // CHECK: [[OFFSET:%.*]] = load i64, i64* [[OFFSET_P]], align 8
+    // CHECK: [[T0:%.*]] = bitcast [[B]]* [[B_P]] to i8*
+    // CHECK: [[T1:%.*]] = getelementptr inbounds i8, i8* [[T0]], i64 [[OFFSET]]
+    // CHECK: [[A_P:%.*]] = bitcast i8* [[T1]] to [[A]]*
+    // CHECK: [[ARRAY_P:%.*]] = getelementptr inbounds [[A]], [[A]]* [[A_P]], i32 0, i32 0
+    // CHECK: [[T0:%.*]] = bitcast [[ARRAY]]* [[RESULT]] to i8*
+    // CHECK: [[T1:%.*]] = bitcast [[ARRAY]]* [[ARRAY_P]] to i8*
+    // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[T0]], i8* [[T1]], i64 16, i32 16, i1 false)
+    AlignedArray result = b.aArray;
+  }
+
+  // CHECK-LABEL: @_ZN5test11cERNS_1BE
+  void c(B &b) {
+    // CHECK: [[RESULT:%.*]] = alloca [[ARRAY]], align 64
+    // CHECK: [[B_P:%.*]] = load [[B]]*, [[B]]**
+    // CHECK: [[ARRAY_P:%.*]] = getelementptr inbounds [[B]], [[B]]* [[B_P]], i32 0, i32 2
+    // CHECK: [[T0:%.*]] = bitcast [[ARRAY]]* [[RESULT]] to i8*
+    // CHECK: [[T1:%.*]] = bitcast [[ARRAY]]* [[ARRAY_P]] to i8*
+    // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[T0]], i8* [[T1]], i64 16, i32 8, i1 false)
+    AlignedArray result = b.bArray;
+  }
+
+  // CHECK-LABEL: @_ZN5test11dEPNS_1BE
+  void d(B *b) {
+    // CHECK: [[RESULT:%.*]] = alloca [[ARRAY]], align 64
+    // CHECK: [[B_P:%.*]] = load [[B]]*, [[B]]**
+    // CHECK: [[ARRAY_P:%.*]] = getelementptr inbounds [[B]], [[B]]* [[B_P]], i32 0, i32 2
+    // CHECK: [[T0:%.*]] = bitcast [[ARRAY]]* [[RESULT]] to i8*
+    // CHECK: [[T1:%.*]] = bitcast [[ARRAY]]* [[ARRAY_P]] to i8*
+    // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[T0]], i8* [[T1]], i64 16, i32 8, i1 false)
+    AlignedArray result = b->bArray;
+  }
+
+  // CHECK-LABEL: @_ZN5test11eEv
+  void e() {
+    // CHECK: [[B_P:%.*]] = alloca [[B]], align 16
+    // CHECK: [[RESULT:%.*]] = alloca [[ARRAY]], align 64
+    // CHECK: [[ARRAY_P:%.*]] = getelementptr inbounds [[B]], [[B]]* [[B_P]], i32 0, i32 2
+    // CHECK: [[T0:%.*]] = bitcast [[ARRAY]]* [[RESULT]] to i8*
+    // CHECK: [[T1:%.*]] = bitcast [[ARRAY]]* [[ARRAY_P]] to i8*
+    // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[T0]], i8* [[T1]], i64 16, i32 16, i1 false)
+    B b;
+    AlignedArray result = b.bArray;
+  }
+
+  // CHECK-LABEL: @_ZN5test11fEv
+  void f() {
+    // TODO: we should devirtualize this derived-to-base conversion.
+    // CHECK: [[D_P:%.*]] = alloca [[D:%.*]], align 16
+    // CHECK: [[RESULT:%.*]] = alloca [[ARRAY]], align 64
+    // CHECK: [[VPTR_P:%.*]] = bitcast [[D]]* [[D_P]] to i8**
+    // CHECK: [[VPTR:%.*]] = load i8*, i8** [[VPTR_P]], align 16
+    // CHECK: [[T0:%.*]] = getelementptr i8, i8* [[VPTR]], i64 -24
+    // CHECK: [[OFFSET_P:%.*]] = bitcast i8* [[T0]] to i64*
+    // CHECK: [[OFFSET:%.*]] = load i64, i64* [[OFFSET_P]], align 8
+    // CHECK: [[T0:%.*]] = bitcast [[D]]* [[D_P]] to i8*
+    // CHECK: [[T1:%.*]] = getelementptr inbounds i8, i8* [[T0]], i64 [[OFFSET]]
+    // CHECK: [[A_P:%.*]] = bitcast i8* [[T1]] to [[A]]*
+    // CHECK: [[ARRAY_P:%.*]] = getelementptr inbounds [[A]], [[A]]* [[A_P]], i32 0, i32 0
+    // CHECK: [[T0:%.*]] = bitcast [[ARRAY]]* [[RESULT]] to i8*
+    // CHECK: [[T1:%.*]] = bitcast [[ARRAY]]* [[ARRAY_P]] to i8*
+    // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[T0]], i8* [[T1]], i64 16, i32 16, i1 false)
+    D d;
+    AlignedArray result = d.aArray;
+  }
+
+  // CHECK-LABEL: @_ZN5test11gEv
+  void g() {
+    // CHECK: [[D_P:%.*]] = alloca [[D]], align 16
+    // CHECK: [[RESULT:%.*]] = alloca [[ARRAY]], align 64
+    // CHECK: [[T0:%.*]] = bitcast [[D]]* [[D_P]] to i8*
+    // CHECK: [[T1:%.*]] = getelementptr inbounds i8, i8* [[T0]], i64 24
+    // CHECK: [[B_P:%.*]] = bitcast i8* [[T1]] to [[B:%.*]]*
+    // CHECK: [[ARRAY_P:%.*]] = getelementptr inbounds [[B]], [[B]]* [[B_P]], i32 0, i32 2
+    // CHECK: [[T0:%.*]] = bitcast [[ARRAY]]* [[RESULT]] to i8*
+    // CHECK: [[T1:%.*]] = bitcast [[ARRAY]]* [[ARRAY_P]] to i8*
+    // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[T0]], i8* [[T1]], i64 16, i32 8, i1 false)
+    D d;
+    AlignedArray result = d.bArray;
+  }
+}
diff --git a/test/CodeGenCXX/arm.cpp b/test/CodeGenCXX/arm.cpp
index 0fadfe97b4..11ae6b24c2 100644
--- a/test/CodeGenCXX/arm.cpp
+++ b/test/CodeGenCXX/arm.cpp
@@ -152,8 +152,8 @@ namespace test3 {
   void e(A *x) {
     // CHECK-LABEL: define void @_ZN5test31eEPNS_1AE(
     // CHECK: icmp eq {{.*}}, null
-    // CHECK: getelementptr {{.*}}, i64 -8
-    // CHECK: getelementptr {{.*}}, i64 4
+    // CHECK: getelementptr {{.*}}, i32 -8
+    // CHECK: getelementptr {{.*}}, i32 4
     // CHECK: bitcast {{.*}} to i32*
     // CHECK: load
     // CHECK: invoke {{.*}} @_ZN5test31AD1Ev
@@ -164,8 +164,8 @@ namespace test3 {
   void f(A (*x)[20]) {
     // CHECK-LABEL: define void @_ZN5test31fEPA20_NS_1AE(
     // CHECK: icmp eq {{.*}}, null
-    // CHECK: getelementptr {{.*}}, i64 -8
-    // CHECK: getelementptr {{.*}}, i64 4
+    // CHECK: getelementptr {{.*}}, i32 -8
+    // CHECK: getelementptr {{.*}}, i32 4
     // CHECK: bitcast {{.*}} to i32*
     // CHECK: load
     // CHECK: invoke {{.*}} @_ZN5test31AD1Ev
@@ -223,8 +223,8 @@ namespace test4 {
 
   void e(A *x) {
     // CHECK-LABEL: define void @_ZN5test41eEPNS_1AE(
-    // CHECK: [[ALLOC:%.*]] = getelementptr inbounds {{.*}}, i64 -8
-    // CHECK: getelementptr inbounds {{.*}}, i64 4
+    // CHECK: [[ALLOC:%.*]] = getelementptr inbounds {{.*}}, i32 -8
+    // CHECK: getelementptr inbounds {{.*}}, i32 4
     // CHECK: bitcast
     // CHECK: [[T0:%.*]] = load i32, i32*
     // CHECK: [[T1:%.*]] = mul i32 4, [[T0]]
@@ -235,8 +235,8 @@ namespace test4 {
 
   void f(A (*x)[20]) {
     // CHECK-LABEL: define void @_ZN5test41fEPA20_NS_1AE(
-    // CHECK: [[ALLOC:%.*]] = getelementptr inbounds {{.*}}, i64 -8
-    // CHECK: getelementptr inbounds {{.*}}, i64 4
+    // CHECK: [[ALLOC:%.*]] = getelementptr inbounds {{.*}}, i32 -8
+    // CHECK: getelementptr inbounds {{.*}}, i32 4
     // CHECK: bitcast
     // CHECK: [[T0:%.*]] = load i32, i32*
     // CHECK: [[T1:%.*]] = mul i32 4, [[T0]]
@@ -293,7 +293,7 @@ namespace test7 {
 
   // CHECK-LABEL: define void @_ZN5test74testEv() {{.*}} personality i8* bitcast (i32 (...)* @__gxx_personality_v0 to i8*)
   void test() {
-    // CHECK:      [[T0:%.*]] = load atomic i8, i8* bitcast (i32* @_ZGVZN5test74testEvE1x to i8*) acquire, align 1
+    // CHECK:      [[T0:%.*]] = load atomic i8, i8* bitcast (i32* @_ZGVZN5test74testEvE1x to i8*) acquire, align 4
     // CHECK-NEXT: [[T1:%.*]] = and i8 [[T0]], 1
     // CHECK-NEXT: [[T2:%.*]] = icmp eq i8 [[T1]], 0
     // CHECK-NEXT: br i1 [[T2]]
@@ -328,7 +328,7 @@ namespace test8 {
 
   // CHECK-LABEL: define void @_ZN5test84testEv() {{.*}} personality i8* bitcast (i32 (...)* @__gxx_personality_v0 to i8*)
   void test() {
-    // CHECK:      [[T0:%.*]] = load atomic i8, i8* bitcast (i32* @_ZGVZN5test84testEvE1x to i8*) acquire, align 1
+    // CHECK:      [[T0:%.*]] = load atomic i8, i8* bitcast (i32* @_ZGVZN5test84testEvE1x to i8*) acquire, align 4
     // CHECK-NEXT: [[T1:%.*]] = and i8 [[T0]], 1
     // CHECK-NEXT: [[T2:%.*]] = icmp eq i8 [[T1]], 0
     // CHECK-NEXT: br i1 [[T2]]
@@ -388,7 +388,7 @@ namespace test9 {
 // CHECK-NEXT: store i32 16, i32* [[T0]]
 // CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds i32, i32* [[T0]], i32 1
 // CHECK-NEXT: store i32 [[N]], i32* [[T1]]
-// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds i8, i8* [[ALLOC]], i64 16
+// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds i8, i8* [[ALLOC]], i32 16
 // CHECK-NEXT: bitcast i8* [[T0]] to [[TEST9]]*
 //   Array allocation follows.
 
@@ -400,8 +400,8 @@ namespace test9 {
 // CHECK-NEXT: [[T0:%.*]] = icmp eq [[TEST9]]* [[BEGIN]], null
 // CHECK-NEXT: br i1 [[T0]],
 // CHECK:      [[T0:%.*]] = bitcast [[TEST9]]* [[BEGIN]] to i8*
-// CHECK-NEXT: [[ALLOC:%.*]] = getelementptr inbounds i8, i8* [[T0]], i64 -16
-// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds i8, i8* [[ALLOC]], i64 4
+// CHECK-NEXT: [[ALLOC:%.*]] = getelementptr inbounds i8, i8* [[T0]], i32 -16
+// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds i8, i8* [[ALLOC]], i32 4
 // CHECK-NEXT: [[T1:%.*]] = bitcast i8* [[T0]] to i32*
 // CHECK-NEXT: [[N:%.*]] = load i32, i32* [[T1]]
 // CHECK-NEXT: [[END:%.*]] = getelementptr inbounds [[TEST9]], [[TEST9]]* [[BEGIN]], i32 [[N]]
diff --git a/test/CodeGenCXX/cxx0x-initializer-stdinitializerlist.cpp b/test/CodeGenCXX/cxx0x-initializer-stdinitializerlist.cpp
index 6d5d3971bd..311edaabb5 100644
--- a/test/CodeGenCXX/cxx0x-initializer-stdinitializerlist.cpp
+++ b/test/CodeGenCXX/cxx0x-initializer-stdinitializerlist.cpp
@@ -502,7 +502,7 @@ namespace B19773010 {
   }
   void f2() {
     // CHECK-LABEL: @_ZN9B197730102f2Ev
-    // CHECK: store %"struct.B19773010::pair"* getelementptr inbounds ([1 x %"struct.B19773010::pair"], [1 x %"struct.B19773010::pair"]* bitcast ([1 x { i8*, i32 }]* @_ZGRZN9B197730102f2EvE1p_ to [1 x %"struct.B19773010::pair"]*), i64 0, i64 0), %"struct.B19773010::pair"** getelementptr inbounds ([2 x %"class.std::initializer_list.10"], [2 x %"class.std::initializer_list.10"]* @_ZZN9B197730102f2EvE1p, i64 0, i64 1, i32 0), align 8
+    // CHECK: store %"struct.B19773010::pair"* getelementptr inbounds ([1 x %"struct.B19773010::pair"], [1 x %"struct.B19773010::pair"]* bitcast ([1 x { i8*, i32 }]* @_ZGRZN9B197730102f2EvE1p_ to [1 x %"struct.B19773010::pair"]*), i64 0, i64 0), %"struct.B19773010::pair"** getelementptr inbounds ([2 x %"class.std::initializer_list.10"], [2 x %"class.std::initializer_list.10"]* @_ZZN9B197730102f2EvE1p, i64 0, i64 1, i32 0), align 16
     static std::initializer_list<pair<const char *, E>> a, p[2] =
         {a, {{"", ENUM_CONSTANT}}};
   }
diff --git a/test/CodeGenCXX/cxx11-initializer-array-new.cpp b/test/CodeGenCXX/cxx11-initializer-array-new.cpp
index 2beb44ecf3..c662190ff3 100644
--- a/test/CodeGenCXX/cxx11-initializer-array-new.cpp
+++ b/test/CodeGenCXX/cxx11-initializer-array-new.cpp
@@ -28,7 +28,7 @@ void *p = new S[2][3]{ { 1, 2, 3 }, { 4, 5, 6 } };
 //
 // { 4, 5, 6 }
 //
-// CHECK: %[[S_1:.*]] = getelementptr inbounds [3 x %[[S]]], [3 x %[[S]]]* %[[S_0]], i32 1
+// CHECK: %[[S_1:.*]] = getelementptr inbounds [3 x %[[S]]], [3 x %[[S]]]* %[[S_0]], i64 1
 //
 // CHECK: %[[S_1_0:.*]] = getelementptr inbounds [3 x %[[S]]], [3 x %[[S]]]* %[[S_1]], i64 0, i64 0
 // CHECK: call void @_ZN1SC1Ei(%[[S]]* %[[S_1_0]], i32 4)
@@ -72,7 +72,7 @@ void *q = new S[n][3]{ { 1, 2, 3 }, { 4, 5, 6 } };
 //
 // { 4, 5, 6 }
 //
-// CHECK: %[[S_1:.*]] = getelementptr inbounds [3 x %[[S]]], [3 x %[[S]]]* %[[S_0]], i32 1
+// CHECK: %[[S_1:.*]] = getelementptr inbounds [3 x %[[S]]], [3 x %[[S]]]* %[[S_0]], i64 1
 //
 // CHECK: %[[S_1_0:.*]] = getelementptr inbounds [3 x %[[S]]], [3 x %[[S]]]* %[[S_1]], i64 0, i64 0
 // CHECK: call void @_ZN1SC1Ei(%[[S]]* %[[S_1_0]], i32 4)
@@ -83,7 +83,7 @@ void *q = new S[n][3]{ { 1, 2, 3 }, { 4, 5, 6 } };
 //
 // And the rest.
 //
-// CHECK: %[[S_2:.*]] = getelementptr inbounds [3 x %[[S]]], [3 x %[[S]]]* %[[S_1]], i32 1
+// CHECK: %[[S_2:.*]] = getelementptr inbounds [3 x %[[S]]], [3 x %[[S]]]* %[[S_1]], i64 1
 // CHECK: %[[S_2_AS_S:.*]] = bitcast [3 x %[[S]]]* %[[S_2]] to %[[S]]*
 //
 // CHECK: %[[REST:.*]] = sub i64 %[[ELTS]], 6
@@ -135,7 +135,7 @@ void *r = new T[n][3]{ { 1, 2, 3 }, { 4, 5, 6 } };
 //
 // { 4, 5, 6 }
 //
-// CHECK: %[[T_1:.*]] = getelementptr inbounds [3 x %[[T]]], [3 x %[[T]]]* %[[T_0]], i32 1
+// CHECK: %[[T_1:.*]] = getelementptr inbounds [3 x %[[T]]], [3 x %[[T]]]* %[[T_0]], i64 1
 //
 // CHECK: %[[T_1_0:.*]] = getelementptr inbounds [3 x %[[T]]], [3 x %[[T]]]* %[[T_1]], i64 0, i64 0
 // CHECK: %[[T_1_0_0:.*]] = getelementptr inbounds %[[T]], %[[T]]* %[[T_1_0]], i32 0, i32 0
@@ -149,7 +149,7 @@ void *r = new T[n][3]{ { 1, 2, 3 }, { 4, 5, 6 } };
 //
 // And the rest gets memset to 0.
 //
-// CHECK: %[[T_2:.*]] = getelementptr inbounds [3 x %[[T]]], [3 x %[[T]]]* %[[T_1]], i32 1
+// CHECK: %[[T_2:.*]] = getelementptr inbounds [3 x %[[T]]], [3 x %[[T]]]* %[[T_1]], i64 1
 // CHECK: %[[T_2_AS_T:.*]] = bitcast [3 x %[[T]]]* %[[T_2]] to %[[T]]*
 //
 // CHECK: %[[SIZE:.*]] = sub i64 %{{.*}}, 24
diff --git a/test/CodeGenCXX/delete-two-arg.cpp b/test/CodeGenCXX/delete-two-arg.cpp
index e5a4cfa3ee..85275b3eb1 100644
--- a/test/CodeGenCXX/delete-two-arg.cpp
+++ b/test/CodeGenCXX/delete-two-arg.cpp
@@ -30,7 +30,7 @@ namespace test2 {
     // CHECK:      [[NEW:%.*]] = call noalias i8* @_Znaj(i32 44)
     // CHECK-NEXT: [[T0:%.*]] = bitcast i8* [[NEW]] to i32*
     // CHECK-NEXT: store i32 10, i32* [[T0]]
-    // CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds i8, i8* [[NEW]], i64 4
+    // CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds i8, i8* [[NEW]], i32 4
     // CHECK-NEXT: [[T2:%.*]] = bitcast i8* [[T1]] to [[A]]*
     // CHECK-NEXT: ret [[A]]* [[T2]]
     return ::new A[10];
@@ -44,7 +44,7 @@ namespace test2 {
     // CHECK-NEXT: [[T1:%.*]] = icmp eq [[A]]* [[T0]], null
     // CHECK-NEXT: br i1 [[T1]],
     // CHECK:      [[T2:%.*]] = bitcast [[A]]* [[T0]] to i8*
-    // CHECK-NEXT: [[T3:%.*]] = getelementptr inbounds i8, i8* [[T2]], i64 -4
+    // CHECK-NEXT: [[T3:%.*]] = getelementptr inbounds i8, i8* [[T2]], i32 -4
     // CHECK-NEXT: [[T4:%.*]] = bitcast i8* [[T3]] to i32*
     // CHECK-NEXT: [[T5:%.*]] = load i32, i32* [[T4]]
     // CHECK-NEXT: call void @_ZdaPv(i8* [[T3]])
diff --git a/test/CodeGenCXX/lambda-expressions.cpp b/test/CodeGenCXX/lambda-expressions.cpp
index 28a8841b60..2ea0561f9e 100644
--- a/test/CodeGenCXX/lambda-expressions.cpp
+++ b/test/CodeGenCXX/lambda-expressions.cpp
@@ -81,7 +81,7 @@ int g() {
 };
 
 // PR14773
-// CHECK: [[ARRVAL:%[0-9a-zA-Z]*]] = load i32, i32* getelementptr inbounds ([0 x i32], [0 x i32]* @_ZZ14staticarrayrefvE5array, i32 0, i64 0), align 4
+// CHECK: [[ARRVAL:%[0-9a-zA-Z]*]] = load i32, i32* getelementptr inbounds ([0 x i32], [0 x i32]* @_ZZ14staticarrayrefvE5array, i64 0, i64 0), align 4
 // CHECK-NEXT: store i32 [[ARRVAL]]
 void staticarrayref(){
   static int array[] = {};
diff --git a/test/CodeGenCXX/microsoft-abi-array-cookies.cpp b/test/CodeGenCXX/microsoft-abi-array-cookies.cpp
index 62ead4fb69..75c0621347 100644
--- a/test/CodeGenCXX/microsoft-abi-array-cookies.cpp
+++ b/test/CodeGenCXX/microsoft-abi-array-cookies.cpp
@@ -28,12 +28,12 @@ void check_array_cookies_simple() {
 // 46 = 42 + size of cookie (4)
 // CHECK: [[COOKIE:%.*]] = bitcast i8* [[ALLOCATED]] to i32*
 // CHECK: store i32 42, i32* [[COOKIE]]
-// CHECK: [[ARRAY:%.*]] = getelementptr inbounds i8, i8* [[ALLOCATED]], i64 4
+// CHECK: [[ARRAY:%.*]] = getelementptr inbounds i8, i8* [[ALLOCATED]], i32 4
 // CHECK: bitcast i8* [[ARRAY]] to [[CLASS:%.*]]*
 
   delete [] array;
 // CHECK: [[ARRAY_AS_CHAR:%.*]] = bitcast [[CLASS]]* {{%.*}} to i8*
-// CHECK: getelementptr inbounds i8, i8* [[ARRAY_AS_CHAR]], i64 -4
+// CHECK: getelementptr inbounds i8, i8* [[ARRAY_AS_CHAR]], i32 -4
 }
 
 struct __attribute__((aligned(8))) ClassWithAlignment {
@@ -50,12 +50,12 @@ void check_array_cookies_aligned() {
 //   344 = 42*8 + size of cookie (8, due to alignment)
 // CHECK: [[COOKIE:%.*]] = bitcast i8* [[ALLOCATED]] to i32*
 // CHECK: store i32 42, i32* [[COOKIE]]
-// CHECK: [[ARRAY:%.*]] = getelementptr inbounds i8, i8* [[ALLOCATED]], i64 8
+// CHECK: [[ARRAY:%.*]] = getelementptr inbounds i8, i8* [[ALLOCATED]], i32 8
 // CHECK: bitcast i8* [[ARRAY]] to [[CLASS:%.*]]*
 
   delete [] array;
 // CHECK: [[ARRAY_AS_CHAR:%.*]] = bitcast [[CLASS]]*
-// CHECK: getelementptr inbounds i8, i8* [[ARRAY_AS_CHAR]], i64 -8
+// CHECK: getelementptr inbounds i8, i8* [[ARRAY_AS_CHAR]], i32 -8
 }
 
 namespace PR23990 {
diff --git a/test/CodeGenCXX/microsoft-abi-eh-cleanups.cpp b/test/CodeGenCXX/microsoft-abi-eh-cleanups.cpp
index 68f1430ace..686b2b0994 100644
--- a/test/CodeGenCXX/microsoft-abi-eh-cleanups.cpp
+++ b/test/CodeGenCXX/microsoft-abi-eh-cleanups.cpp
@@ -162,7 +162,7 @@ C::C() { foo(); }
 // WIN32-NOT:  load
 // WIN32:      bitcast %"struct.crash_on_partial_destroy::C"* %{{.*}} to i8*
 // WIN32-NOT:  load
-// WIN32:      getelementptr inbounds i8, i8* %{{.*}}, i64 4
+// WIN32:      getelementptr inbounds i8, i8* %{{.*}}, i32 4
 // WIN32-NOT:  load
 // WIN32:      bitcast i8* %{{.*}} to %"struct.crash_on_partial_destroy::A"*
 // WIN32:      call x86_thiscallcc void @"\01??1A@crash_on_partial_destroy@@UAE@XZ"
diff --git a/test/CodeGenCXX/microsoft-abi-structors.cpp b/test/CodeGenCXX/microsoft-abi-structors.cpp
index 88400e7c0b..0e44efa8d0 100644
--- a/test/CodeGenCXX/microsoft-abi-structors.cpp
+++ b/test/CodeGenCXX/microsoft-abi-structors.cpp
@@ -161,7 +161,7 @@ C::~C() {
 // CHECK:   load %"struct.dtor_in_second_nvbase::C"*, %"struct.dtor_in_second_nvbase::C"** %{{.*}}
 //      Now we this-adjust before calling ~B.
 // CHECK:   bitcast %"struct.dtor_in_second_nvbase::C"* %{{.*}} to i8*
-// CHECK:   getelementptr inbounds i8, i8* %{{.*}}, i64 4
+// CHECK:   getelementptr inbounds i8, i8* %{{.*}}, i32 4
 // CHECK:   bitcast i8* %{{.*}} to %"struct.dtor_in_second_nvbase::B"*
 // CHECK:   call x86_thiscallcc void @"\01??1B@dtor_in_second_nvbase@@UAE@XZ"
 // CHECK:       (%"struct.dtor_in_second_nvbase::B"* %{{.*}})
@@ -246,11 +246,11 @@ C::C() {
   //
   // CHECK: [[INIT_VBASES]]
   // CHECK-NEXT: %[[this_i8:.*]] = bitcast %"struct.constructors::C"* %{{.*}} to i8*
-  // CHECK-NEXT: %[[vbptr_off:.*]] = getelementptr inbounds i8, i8* %[[this_i8]], i64 0
+  // CHECK-NEXT: %[[vbptr_off:.*]] = getelementptr inbounds i8, i8* %[[this_i8]], i32 0
   // CHECK-NEXT: %[[vbptr:.*]] = bitcast i8* %[[vbptr_off]] to i32**
   // CHECK-NEXT: store i32* getelementptr inbounds ([2 x i32], [2 x i32]* @"\01??_8C@constructors@@7B@", i32 0, i32 0), i32** %[[vbptr]]
   // CHECK-NEXT: bitcast %"struct.constructors::C"* %{{.*}} to i8*
-  // CHECK-NEXT: getelementptr inbounds i8, i8* %{{.*}}, i64 4
+  // CHECK-NEXT: getelementptr inbounds i8, i8* %{{.*}}, i32 4
   // CHECK-NEXT: bitcast i8* %{{.*}} to %"struct.constructors::A"*
   // CHECK-NEXT: call x86_thiscallcc %"struct.constructors::A"* @"\01??0A@constructors@@QAE@XZ"(%"struct.constructors::A"* %{{.*}})
   // CHECK-NEXT: br label %[[SKIP_VBASES]]
@@ -281,11 +281,11 @@ D::D() {
   //
   // CHECK: [[INIT_VBASES]]
   // CHECK-NEXT: %[[this_i8:.*]] = bitcast %"struct.constructors::D"* %{{.*}} to i8*
-  // CHECK-NEXT: %[[vbptr_off:.*]] = getelementptr inbounds i8, i8* %[[this_i8]], i64 0
+  // CHECK-NEXT: %[[vbptr_off:.*]] = getelementptr inbounds i8, i8* %[[this_i8]], i32 0
   // CHECK-NEXT: %[[vbptr:.*]] = bitcast i8* %[[vbptr_off]] to i32**
   // CHECK-NEXT: store i32* getelementptr inbounds ([2 x i32], [2 x i32]* @"\01??_8D@constructors@@7B@", i32 0, i32 0), i32** %[[vbptr]]
   // CHECK-NEXT: bitcast %"struct.constructors::D"* %{{.*}} to i8*
-  // CHECK-NEXT: getelementptr inbounds i8, i8* %{{.*}}, i64 4
+  // CHECK-NEXT: getelementptr inbounds i8, i8* %{{.*}}, i32 4
   // CHECK-NEXT: bitcast i8* %{{.*}} to %"struct.constructors::A"*
   // CHECK-NEXT: call x86_thiscallcc %"struct.constructors::A"* @"\01??0A@constructors@@QAE@XZ"(%"struct.constructors::A"* %{{.*}})
   // CHECK-NEXT: br label %[[SKIP_VBASES]]
@@ -308,14 +308,14 @@ E::E() {
   //
   // CHECK: [[INIT_VBASES]]
   // CHECK-NEXT: %[[this_i8:.*]] = bitcast %"struct.constructors::E"* %{{.*}} to i8*
-  // CHECK-NEXT: %[[offs:.*]] = getelementptr inbounds i8, i8* %[[this_i8]], i64 0
+  // CHECK-NEXT: %[[offs:.*]] = getelementptr inbounds i8, i8* %[[this_i8]], i32 0
   // CHECK-NEXT: %[[vbptr_E:.*]] = bitcast i8* %[[offs]] to i32**
   // CHECK-NEXT: store i32* getelementptr inbounds ([3 x i32], [3 x i32]* @"\01??_8E@constructors@@7B01@@", i32 0, i32 0), i32** %[[vbptr_E]]
-  // CHECK-NEXT: %[[offs:.*]] = getelementptr inbounds i8, i8* %[[this_i8]], i64 4
+  // CHECK-NEXT: %[[offs:.*]] = getelementptr inbounds i8, i8* %[[this_i8]], i32 4
   // CHECK-NEXT: %[[vbptr_C:.*]] = bitcast i8* %[[offs]] to i32**
   // CHECK-NEXT: store i32* getelementptr inbounds ([2 x i32], [2 x i32]* @"\01??_8E@constructors@@7BC@1@@", i32 0, i32 0), i32** %[[vbptr_C]]
   // CHECK-NEXT: bitcast %"struct.constructors::E"* %{{.*}} to i8*
-  // CHECK-NEXT: getelementptr inbounds i8, i8* %{{.*}}, i64 4
+  // CHECK-NEXT: getelementptr inbounds i8, i8* %{{.*}}, i32 4
   // CHECK-NEXT: bitcast i8* %{{.*}} to %"struct.constructors::A"*
   // CHECK-NEXT: call x86_thiscallcc %"struct.constructors::A"* @"\01??0A@constructors@@QAE@XZ"(%"struct.constructors::A"* %{{.*}})
   // CHECK: call x86_thiscallcc %"struct.constructors::C"* @"\01??0C@constructors@@QAE@XZ"(%"struct.constructors::C"* %{{.*}}, i32 0)
diff --git a/test/CodeGenCXX/microsoft-abi-try-throw.cpp b/test/CodeGenCXX/microsoft-abi-try-throw.cpp
index fed3976171..93d6a689ce 100644
--- a/test/CodeGenCXX/microsoft-abi-try-throw.cpp
+++ b/test/CodeGenCXX/microsoft-abi-try-throw.cpp
@@ -26,7 +26,7 @@ int main() {
   }
 #endif
 #ifdef THROW
-  // THROW: store i32 42, i32* %[[mem_for_throw:.*]]
+  // THROW: store i32 42, i32* %[[mem_for_throw:.*]], align 4
   // THROW: %[[cast:.*]] = bitcast i32* %[[mem_for_throw]] to i8*
   // THROW: call void @_CxxThrowException(i8* %[[cast]], %eh.ThrowInfo* @_TI1H)
   throw int(42);
diff --git a/test/CodeGenCXX/microsoft-abi-virtual-inheritance-vtordisps.cpp b/test/CodeGenCXX/microsoft-abi-virtual-inheritance-vtordisps.cpp
index 204da8db15..bb73f8773c 100644
--- a/test/CodeGenCXX/microsoft-abi-virtual-inheritance-vtordisps.cpp
+++ b/test/CodeGenCXX/microsoft-abi-virtual-inheritance-vtordisps.cpp
@@ -26,7 +26,7 @@ D::D() {}  // Forces vftable emission.
 // CHECK-LABEL: define linkonce_odr x86_thiscallcc void @"\01?f@D@@$4PPPPPPPM@A@AEXXZ"
 // CHECK: %[[ECX:.*]] = load %struct.D*, %struct.D** %{{.*}}
 // CHECK: %[[ECX_i8:.*]] = bitcast %struct.D* %[[ECX]] to i8*
-// CHECK: %[[VTORDISP_PTR_i8:.*]] = getelementptr i8, i8* %[[ECX_i8]], i32 -4
+// CHECK: %[[VTORDISP_PTR_i8:.*]] = getelementptr inbounds i8, i8* %[[ECX_i8]], i32 -4
 // CHECK: %[[VTORDISP_PTR:.*]] = bitcast i8* %[[VTORDISP_PTR_i8]] to i32*
 // CHECK: %[[VTORDISP:.*]] = load i32, i32* %[[VTORDISP_PTR]]
 // CHECK: %[[VTORDISP_NEG:.*]] = sub i32 0, %[[VTORDISP]]
@@ -37,7 +37,7 @@ D::D() {}  // Forces vftable emission.
 // CHECK-LABEL: define linkonce_odr x86_thiscallcc void @"\01?f@D@@$4PPPPPPPI@3AEXXZ"
 // CHECK: %[[ECX:.*]] = load %struct.D*, %struct.D** %{{.*}}
 // CHECK: %[[ECX_i8:.*]] = bitcast %struct.D* %[[ECX]] to i8*
-// CHECK: %[[VTORDISP_PTR_i8:.*]] = getelementptr i8, i8* %[[ECX_i8]], i32 -8
+// CHECK: %[[VTORDISP_PTR_i8:.*]] = getelementptr inbounds i8, i8* %[[ECX_i8]], i32 -8
 // CHECK: %[[VTORDISP_PTR:.*]] = bitcast i8* %[[VTORDISP_PTR_i8]] to i32*
 // CHECK: %[[VTORDISP:.*]] = load i32, i32* %[[VTORDISP_PTR]]
 // CHECK: %[[VTORDISP_NEG:.*]] = sub i32 0, %[[VTORDISP]]
@@ -66,7 +66,7 @@ G::G() {}  // Forces vftable emission.
 // CHECK-LABEL: define linkonce_odr x86_thiscallcc void @"\01?f@E@@$R4BA@M@PPPPPPPM@7AEXXZ"(i8*)
 // CHECK: %[[ECX:.*]] = load %struct.E*, %struct.E** %{{.*}}
 // CHECK: %[[ECX_i8:.*]] = bitcast %struct.E* %[[ECX]] to i8*
-// CHECK: %[[VTORDISP_PTR_i8:.*]] = getelementptr i8, i8* %[[ECX_i8]], i32 -4
+// CHECK: %[[VTORDISP_PTR_i8:.*]] = getelementptr inbounds i8, i8* %[[ECX_i8]], i32 -4
 // CHECK: %[[VTORDISP_PTR:.*]] = bitcast i8* %[[VTORDISP_PTR_i8]] to i32*
 // CHECK: %[[VTORDISP:.*]] = load i32, i32* %[[VTORDISP_PTR]]
 // CHECK: %[[VTORDISP_NEG:.*]] = sub i32 0, %[[VTORDISP]]
diff --git a/test/CodeGenCXX/microsoft-abi-virtual-inheritance.cpp b/test/CodeGenCXX/microsoft-abi-virtual-inheritance.cpp
index b868d1f0b5..9a0b011783 100644
--- a/test/CodeGenCXX/microsoft-abi-virtual-inheritance.cpp
+++ b/test/CodeGenCXX/microsoft-abi-virtual-inheritance.cpp
@@ -91,7 +91,7 @@ B::~B() {
   // CHECK2: %[[B:.*]] = bitcast i8* %[[B_i8]] to %struct.B*
   // CHECK2: call x86_thiscallcc void @"\01??1B@@UAE@XZ"(%struct.B* %[[B]])
   // CHECK2: %[[THIS_i8:.*]] = bitcast %struct.B* %[[THIS]] to i8*
-  // CHECK2: %[[VBASE_i8:.*]] = getelementptr inbounds i8, i8* %[[THIS_i8]], i64 8
+  // CHECK2: %[[VBASE_i8:.*]] = getelementptr inbounds i8, i8* %[[THIS_i8]], i32 8
   // CHECK2: %[[VBASE:.*]] = bitcast i8* %[[VBASE_i8]] to %struct.VBase*
   // CHECK2: call x86_thiscallcc void @"\01??1VBase@@UAE@XZ"(%struct.VBase* %[[VBASE]])
   // CHECK2: ret
@@ -290,7 +290,7 @@ D::~D() {
   // CHECK: store %"struct.diamond::D"* %[[THIS]], %"struct.diamond::D"** %[[THIS_VAL:.*]], align 4
   // CHECK: %[[THIS:.*]] = load %"struct.diamond::D"*, %"struct.diamond::D"** %[[THIS_VAL]]
   // CHECK: %[[D_i8:.*]] = bitcast %"struct.diamond::D"* %[[THIS]] to i8*
-  // CHECK: %[[C_i8:.*]] = getelementptr inbounds i8, i8* %[[D_i8]], i64 4
+  // CHECK: %[[C_i8:.*]] = getelementptr inbounds i8, i8* %[[D_i8]], i32 4
   // CHECK: %[[C:.*]] = bitcast i8* %[[C_i8]] to %"struct.diamond::C"*
   // CHECK: %[[C_i8:.*]] = bitcast %"struct.diamond::C"* %[[C]] to i8*
   // CHECK: %[[ARG_i8:.*]] = getelementptr i8, i8* %{{.*}}, i32 16
diff --git a/test/CodeGenCXX/static-init-wasm.cpp b/test/CodeGenCXX/static-init-wasm.cpp
index e842509af4..2d187b5c05 100644
--- a/test/CodeGenCXX/static-init-wasm.cpp
+++ b/test/CodeGenCXX/static-init-wasm.cpp
@@ -17,7 +17,7 @@ void g() {
   static int a = f();
 }
 // WEBASSEMBLY32-LABEL: @_Z1gv()
-// WEBASSEMBLY32:       %[[R0:.+]] = load atomic i8, i8* bitcast (i32* @_ZGVZ1gvE1a to i8*) acquire, align 1
+// WEBASSEMBLY32:       %[[R0:.+]] = load atomic i8, i8* bitcast (i32* @_ZGVZ1gvE1a to i8*) acquire, align 4
 // WEBASSEMBLY32-NEXT:  %[[R1:.+]] = and i8 %[[R0]], 1
 // WEBASSEMBLY32-NEXT:  %[[R2:.+]] = icmp eq i8 %[[R1]], 0
 // WEBASSEMBLY32-NEXT:  br i1 %[[R2]], label %[[CHECK:.+]], label %[[END:.+]]
@@ -27,7 +27,7 @@ void g() {
 // WEBASSEMBLY32:       call void @__cxa_guard_release
 //
 // WEBASSEMBLY64-LABEL: @_Z1gv()
-// WEBASSEMBLY64:       %[[R0:.+]] = load atomic i8, i8* bitcast (i64* @_ZGVZ1gvE1a to i8*) acquire, align 1
+// WEBASSEMBLY64:       %[[R0:.+]] = load atomic i8, i8* bitcast (i64* @_ZGVZ1gvE1a to i8*) acquire, align 8
 // WEBASSEMBLY64-NEXT:  %[[R1:.+]] = and i8 %[[R0]], 1
 // WEBASSEMBLY64-NEXT:  %[[R2:.+]] = icmp eq i8 %[[R1]], 0
 // WEBASSEMBLY64-NEXT:  br i1 %[[R2]], label %[[CHECK:.+]], label %[[END:.+]]
diff --git a/test/CodeGenCXX/static-init.cpp b/test/CodeGenCXX/static-init.cpp
index 255251e707..541f6416ef 100644
--- a/test/CodeGenCXX/static-init.cpp
+++ b/test/CodeGenCXX/static-init.cpp
@@ -6,8 +6,8 @@
 
 // CHECK: @_ZZN5test31BC1EvE1u = internal global { i8, [3 x i8] } { i8 97, [3 x i8] undef }, align 4
 
-// CHECK: @_ZZ2h2vE1i = linkonce_odr global i32 0, comdat, align
-// CHECK: @_ZGVZ2h2vE1i = linkonce_odr global i64 0, comdat{{$}}
+// CHECK: @_ZZ2h2vE1i = linkonce_odr global i32 0, comdat, align 4
+// CHECK: @_ZGVZ2h2vE1i = linkonce_odr global i64 0, comdat, align 8{{$}}
 // CHECK: @_ZZN5test1L6getvarEiE3var = internal constant [4 x i32] [i32 1, i32 0, i32 2, i32 4], align 16
 // CHECK: @_ZZN5test414useStaticLocalEvE3obj = linkonce_odr global %"struct.test4::HasVTable" zeroinitializer, comdat, align 8
 
@@ -17,7 +17,7 @@ struct A {
 };
 
 void f() {
-  // CHECK: load atomic i8, i8* bitcast (i64* @_ZGVZ1fvE1a to i8*) acquire, align 1
+  // CHECK: load atomic i8, i8* bitcast (i64* @_ZGVZ1fvE1a to i8*) acquire, align 8
   // CHECK: call i32 @__cxa_guard_acquire
   // CHECK: call void @_ZN1AC1Ev
   // CHECK: call i32 @__cxa_atexit(void (i8*)* bitcast (void (%struct.A*)* @_ZN1AD1Ev to void (i8*)*), i8* getelementptr inbounds (%struct.A, %struct.A* @_ZZ1fvE1a, i32 0, i32 0), i8* @__dso_handle)
diff --git a/test/CodeGenCXX/vararg-non-pod-ms-compat.cpp b/test/CodeGenCXX/vararg-non-pod-ms-compat.cpp
index 530a4284b4..64b1c525a6 100644
--- a/test/CodeGenCXX/vararg-non-pod-ms-compat.cpp
+++ b/test/CodeGenCXX/vararg-non-pod-ms-compat.cpp
@@ -18,7 +18,7 @@ void test(X x) {
   // X64: alloca %struct.X
 
   // X64: %[[agg:[^ ]*]] = alloca %struct.X
-  // X64: %[[valptr:[^ ]*]] = getelementptr %struct.X, %struct.X* %[[agg]], i32 0, i32 0
+  // X64: %[[valptr:[^ ]*]] = getelementptr inbounds %struct.X, %struct.X* %[[agg]], i32 0, i32 0
   // X64: %[[val:[^ ]*]] = load i32, i32* %[[valptr]]
   // X64: call void (...) @"\01?vararg@@YAXZZ"(i32 %[[val]])
 
diff --git a/test/CodeGenCXX/wasm-args-returns.cpp b/test/CodeGenCXX/wasm-args-returns.cpp
index 4ef085f6ba..2b80430014 100644
--- a/test/CodeGenCXX/wasm-args-returns.cpp
+++ b/test/CodeGenCXX/wasm-args-returns.cpp
@@ -42,17 +42,17 @@ struct copy_ctor {
     copy_ctor(copy_ctor const&);
 };
 test(copy_ctor);
-// CHECK: define void @_Z7forward9copy_ctor(%struct.copy_ctor* noalias sret %{{.*}}, %struct.copy_ctor* align 8 %{{.*}})
+// CHECK: define void @_Z7forward9copy_ctor(%struct.copy_ctor* noalias sret %{{.*}}, %struct.copy_ctor* %{{.*}})
 //
 // CHECK: declare %struct.copy_ctor* @_ZN9copy_ctorC1ERKS_(%struct.copy_ctor* returned, %struct.copy_ctor* dereferenceable(8))
 //
 // CHECK: define void @_Z14test_copy_ctorv()
 // CHECK: %[[tmp:.*]] = alloca %struct.copy_ctor, align 8
 // CHECK: call void @_Z13def_copy_ctorv(%struct.copy_ctor* nonnull sret %[[tmp]])
-// CHECK: call void @_Z3use9copy_ctor(%struct.copy_ctor* nonnull align 8 %[[tmp]])
+// CHECK: call void @_Z3use9copy_ctor(%struct.copy_ctor* nonnull %[[tmp]])
 // CHECK: ret void
 //
-// CHECK: declare void @_Z3use9copy_ctor(%struct.copy_ctor* align 8)
+// CHECK: declare void @_Z3use9copy_ctor(%struct.copy_ctor*)
 // CHECK: declare void @_Z13def_copy_ctorv(%struct.copy_ctor* sret)
 
 struct __attribute__((aligned(16))) aligned_copy_ctor {
@@ -60,17 +60,17 @@ struct __attribute__((aligned(16))) aligned_copy_ctor {
     aligned_copy_ctor(aligned_copy_ctor const&);
 };
 test(aligned_copy_ctor);
-// CHECK: define void @_Z7forward17aligned_copy_ctor(%struct.aligned_copy_ctor* noalias sret %{{.*}}, %struct.aligned_copy_ctor* align 16 %{{.*}})
+// CHECK: define void @_Z7forward17aligned_copy_ctor(%struct.aligned_copy_ctor* noalias sret %{{.*}}, %struct.aligned_copy_ctor* %{{.*}})
 //
 // CHECK: declare %struct.aligned_copy_ctor* @_ZN17aligned_copy_ctorC1ERKS_(%struct.aligned_copy_ctor* returned, %struct.aligned_copy_ctor* dereferenceable(16))
 //
 // CHECK: define void @_Z22test_aligned_copy_ctorv()
 // CHECK: %[[tmp:.*]] = alloca %struct.aligned_copy_ctor, align 16
 // CHECK: call void @_Z21def_aligned_copy_ctorv(%struct.aligned_copy_ctor* nonnull sret %[[tmp]])
-// CHECK: call void @_Z3use17aligned_copy_ctor(%struct.aligned_copy_ctor* nonnull align 16 %[[tmp]])
+// CHECK: call void @_Z3use17aligned_copy_ctor(%struct.aligned_copy_ctor* nonnull %[[tmp]])
 // CHECK: ret void
 //
-// CHECK: declare void @_Z3use17aligned_copy_ctor(%struct.aligned_copy_ctor* align 16)
+// CHECK: declare void @_Z3use17aligned_copy_ctor(%struct.aligned_copy_ctor*)
 // CHECK: declare void @_Z21def_aligned_copy_ctorv(%struct.aligned_copy_ctor* sret)
 
 struct empty {};
diff --git a/test/CodeGenObjC/arc-captured-32bit-block-var-layout-2.m b/test/CodeGenObjC/arc-captured-32bit-block-var-layout-2.m
index 6ab02a916e..10feda938d 100644
--- a/test/CodeGenObjC/arc-captured-32bit-block-var-layout-2.m
+++ b/test/CodeGenObjC/arc-captured-32bit-block-var-layout-2.m
@@ -11,31 +11,30 @@ int main() {
   NSString *strong;
   unsigned long long eightByte = 0x8001800181818181ull;
   // Test1
-// CHECK: block variable layout: BL_NON_OBJECT_WORD:3, BL_STRONG:1, BL_OPERATOR:0
+  // CHECK: Inline block variable layout: 0x0100, BL_STRONG:1, BL_OPERATOR:0
   void (^block1)() = ^{ printf("%#llx", eightByte); NSLog(@"%@", strong); };
 
   // Test2
   int i = 1;
-// CHECK:  block variable layout: BL_NON_OBJECT_WORD:3, BL_STRONG:1, BL_OPERATOR:0
+  // CHECK: Inline block variable layout: 0x0100, BL_STRONG:1, BL_OPERATOR:0
   void (^block2)() = ^{ printf("%#llx, %d", eightByte, i); NSLog(@"%@", strong); };
 
   //  Test3
   char ch = 'a';
-// CHECK: block variable layout: BL_NON_OBJECT_WORD:3, BL_STRONG:1, BL_OPERATOR:0
+  // CHECK: Inline block variable layout: 0x0100, BL_STRONG:1, BL_OPERATOR:0
   void (^block3)() = ^{ printf("%c %#llx", ch, eightByte); NSLog(@"%@", strong); };
 
   // Test4
   unsigned long fourByte = 0x8001ul;
-// block variable layout: BL_NON_OBJECT_WORD:1, BL_STRONG:1, BL_OPERATOR:0
-// CHECK: Inline instruction for block variable layout: 0x0100
+  // CHECK: Inline block variable layout: 0x0100, BL_STRONG:1, BL_OPERATOR:0
   void (^block4)() = ^{ printf("%c %#lx", ch, fourByte); NSLog(@"%@", strong); };
 
   // Test5
-// CHECK: block variable layout: BL_NON_OBJECT_WORD:3, BL_STRONG:1, BL_OPERATOR:0
+  // CHECK: Inline block variable layout: 0x0100, BL_STRONG:1, BL_OPERATOR:0
   void (^block5)() = ^{ NSLog(@"%@", strong); printf("%c %#llx", ch, eightByte); };
 
   // Test6
-// CHECK: block variable layout: BL_OPERATOR:0
+  // CHECK: Block variable layout: BL_OPERATOR:0
   void (^block6)() = ^{ printf("%#llx", eightByte); };
 }
 
diff --git a/test/CodeGenObjC/arc-captured-32bit-block-var-layout.m b/test/CodeGenObjC/arc-captured-32bit-block-var-layout.m
index 45a894c381..d1b5783427 100644
--- a/test/CodeGenObjC/arc-captured-32bit-block-var-layout.m
+++ b/test/CodeGenObjC/arc-captured-32bit-block-var-layout.m
@@ -33,7 +33,7 @@ void f() {
 // and a descriptor pointer).
 
 // Test 1
-// CHECK: Inline instruction for block variable layout: 0x0320
+// CHECK: Inline block variable layout: 0x0320, BL_STRONG:3, BL_BYREF:2, BL_OPERATOR:0
     void (^b)() = ^{
         byref_int = sh + ch+ch1+ch2 ;
         x(bar);
@@ -44,7 +44,7 @@ void f() {
     b();
 
 // Test 2
-// CHECK: Inline instruction for block variable layout: 0x0331
+// CHECK: Inline block variable layout: 0x0331, BL_STRONG:3, BL_BYREF:3, BL_WEAK:1, BL_OPERATOR:0
     void (^c)() = ^{
         byref_int = sh + ch+ch1+ch2 ;
         x(bar);
@@ -65,7 +65,7 @@ void g() {
   unsigned int i;
   NSString *y;
   NSString *z;
-// CHECK: Inline instruction for block variable layout: 0x0401
+// CHECK: Inline block variable layout: 0x0401, BL_STRONG:4, BL_WEAK:1, BL_OPERATOR:0
   void (^c)() = ^{
    int j = i + bletch;
    x(foo);
@@ -110,7 +110,7 @@ void h() {
 block variable layout: BL_NON_OBJECT_WORD:1, BL_UNRETAINE:1, BL_NON_OBJECT_WORD:1, 
                        BL_UNRETAINE:1, BL_NON_OBJECT_WORD:3, BL_BYREF:1, BL_OPERATOR:0
 */
-// CHECK: block variable layout: BL_BYREF:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_OPERATOR:0
+// CHECK: Block variable layout: BL_BYREF:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_OPERATOR:0
   void (^c)() = ^{
     x(s2.ui.o1);
     x(u2.o1);
@@ -125,7 +125,7 @@ void arr1() {
     __unsafe_unretained id unsafe_unretained_var[4];
  } imported_s;
 
-// CHECK: block variable layout: BL_UNRETAINED:4, BL_OPERATOR:0
+// CHECK: Block variable layout: BL_UNRETAINED:4, BL_OPERATOR:0
     void (^c)() = ^{
         x(imported_s.unsafe_unretained_var[2]);
     };    
@@ -140,7 +140,7 @@ void arr2() {
     __unsafe_unretained id unsafe_unretained_var[4];
  } imported_s;
 
-// CHECK: block variable layout: BL_NON_OBJECT_WORD:1, BL_UNRETAINED:4, BL_OPERATOR:0
+// CHECK: Block variable layout: BL_NON_OBJECT_WORD:1, BL_UNRETAINED:4, BL_OPERATOR:0
     void (^c)() = ^{
         x(imported_s.unsafe_unretained_var[2]);
     };    
@@ -155,7 +155,7 @@ void arr3() {
     __unsafe_unretained id unsafe_unretained_var[0];
  } imported_s;
 
-// CHECK: block variable layout: BL_OPERATOR:0
+// CHECK: Block variable layout: BL_OPERATOR:0
     void (^c)() = ^{
       int i = imported_s.a;
     };    
@@ -181,7 +181,7 @@ void arr4() {
     } f4[2][2];
   } captured_s;
 
-// CHECK: block variable layout: BL_UNRETAINED:3, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_OPERATOR:0
+// CHECK: Block variable layout: BL_UNRETAINED:3, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_OPERATOR:0
   void (^c)() = ^{
       id i = captured_s.f0.s_f1;
   };
@@ -199,7 +199,7 @@ void bf1() {
     int flag4: 24;
   } s;
 
-// CHECK:  block variable layout: BL_OPERATOR:0
+// CHECK:  Block variable layout: BL_OPERATOR:0
   int (^c)() = ^{
       return s.flag;
   };
@@ -212,7 +212,7 @@ void bf2() {
     int flag : 1;
   } s;
 
-// CHECK: block variable layout: BL_OPERATOR:0
+// CHECK: Block variable layout: BL_OPERATOR:0
   int (^c)() = ^{
       return s.flag;
   };
@@ -243,7 +243,7 @@ void bf3() {
         unsigned int _filler : 32;
     } _flags;
 
-// CHECK: block variable layout: BL_OPERATOR:0
+// CHECK: Block variable layout: BL_OPERATOR:0
   unsigned char (^c)() = ^{
       return _flags._draggedNodesAreDeletable;
   };
@@ -278,7 +278,7 @@ void bf4() {
         unsigned int _filler : 32;
     } _flags;
 
-// CHECK:  block variable layout: BL_OPERATOR:0
+// CHECK:  Block variable layout: BL_OPERATOR:0
   unsigned char (^c)() = ^{
       return _flags._draggedNodesAreDeletable;
   };
@@ -296,7 +296,7 @@ void bf5() {
         unsigned char flag1 : 1;
     } _flags;
 
-// CHECK:  block variable layout: BL_OPERATOR:0
+// CHECK:  Block variable layout: BL_OPERATOR:0
   unsigned char (^c)() = ^{
       return _flags.flag;
   };
@@ -313,7 +313,7 @@ void bf6() {
         unsigned char flag1 : 1;
     } _flags;
 
-// CHECK: block variable layout: BL_OPERATOR:0
+// CHECK: Block variable layout: BL_OPERATOR:0
   unsigned char (^c)() = ^{
       return _flags.flag;
   };
@@ -329,7 +329,7 @@ void Test7() {
     __weak id wid9, wid10, wid11, wid12;
     __weak id wid13, wid14, wid15, wid16;
     const id bar = (id) opaque_id();
-// CHECK: block variable layout: BL_STRONG:1, BL_WEAK:16, BL_OPERATOR:0
+// CHECK: Block variable layout: BL_STRONG:1, BL_WEAK:16, BL_OPERATOR:0
     void (^b)() = ^{
       x(bar);
       x(wid1);
@@ -364,7 +364,7 @@ __weak id wid;
     __weak id w9, w10, w11, w12;
     __weak id w13, w14, w15, w16;
     const id bar = (id) opaque_id();
-// CHECK: block variable layout: BL_STRONG:1, BL_WEAK:16, BL_WEAK:16, BL_WEAK:1, BL_OPERATOR:0
+// CHECK: Block variable layout: BL_STRONG:1, BL_WEAK:16, BL_WEAK:16, BL_WEAK:1, BL_OPERATOR:0
     void (^b)() = ^{
       x(bar);
       x(wid1);
diff --git a/test/CodeGenObjC/arc-captured-block-var-inlined-layout.m b/test/CodeGenObjC/arc-captured-block-var-inlined-layout.m
index 89e2b570b0..07b194da94 100644
--- a/test/CodeGenObjC/arc-captured-block-var-inlined-layout.m
+++ b/test/CodeGenObjC/arc-captured-block-var-inlined-layout.m
@@ -1,7 +1,7 @@
 // RUN: %clang_cc1 -fblocks -fobjc-arc -fobjc-runtime-has-weak -triple x86_64-apple-darwin -print-ivar-layout -emit-llvm -o /dev/null %s > %t-64.layout
-// RUN: FileCheck --input-file=%t-64.layout %s
+// RUN: FileCheck -check-prefix=CHECK -check-prefix=CHECK-64 --input-file=%t-64.layout %s
 // RUN: %clang_cc1 -fblocks -fobjc-arc -fobjc-runtime-has-weak -triple i386-apple-darwin -print-ivar-layout -emit-llvm -o /dev/null  %s > %t-32.layout
-// RUN: FileCheck -check-prefix=CHECK-i386 --input-file=%t-32.layout %s
+// RUN: FileCheck -check-prefix=CHECK -check-prefix=CHECK-32 --input-file=%t-32.layout %s
 // rdar://12184410
 
 void x(id y) {}
@@ -17,22 +17,19 @@ void f() {
     __block id byref_bab = (id)0;
     __block id bl_var1;
 
-// CHECK: Inline instruction for block variable layout: 0x0100
-// CHECK-i386: Inline instruction for block variable layout: 0x0100
+// CHECK: Inline block variable layout: 0x0100, BL_STRONG:1, BL_OPERATOR:0
     void (^b)() = ^{
         x(bar);
     };    
 
-// CHECK: Inline instruction for block variable layout: 0x0210
-// CHECK-i386: Inline instruction for block variable layout: 0x0210
+// CHECK: Inline block variable layout: 0x0210, BL_STRONG:2, BL_BYREF:1, BL_OPERATOR:0
     void (^c)() = ^{
         x(bar);
         x(baz);
         byref_int = 1;
     };    
 
-// CHECK: Inline instruction for block variable layout: 0x0230
-// CHECK-i386: Inline instruction for block variable layout: 0x0230
+// CHECK: Inline block variable layout: 0x0230, BL_STRONG:2, BL_BYREF:3, BL_OPERATOR:0
     void (^d)() = ^{
         x(bar);
         x(baz);
@@ -41,8 +38,7 @@ void f() {
         byref_bab = 0;
     };
 
-// CHECK: Inline instruction for block variable layout: 0x0231
-// CHECK-i386: Inline instruction for block variable layout: 0x0231
+// CHECK: Inline block variable layout: 0x0231, BL_STRONG:2, BL_BYREF:3, BL_WEAK:1, BL_OPERATOR:0
     __weak id wid;
     id (^e)() = ^{
         x(bar);
@@ -53,8 +49,7 @@ void f() {
         return wid;
     };
 
-// CHECK: Inline instruction for block variable layout: 0x0235
-// CHECK-i386: Inline instruction for block variable layout: 0x0235
+// CHECK: Inline block variable layout: 0x0235, BL_STRONG:2, BL_BYREF:3, BL_WEAK:5, BL_OPERATOR:0
     __weak id wid1, wid2, wid3, wid4;
     id (^f)() = ^{
         x(bar);
@@ -69,8 +64,7 @@ void f() {
         return wid;
     };
 
-// CHECK: Inline instruction for block variable layout: 0x035
-// CHECK-i386: Inline instruction for block variable layout: 0x035
+// CHECK: Inline block variable layout: 0x035, BL_BYREF:3, BL_WEAK:5, BL_OPERATOR:0
     id (^g)() = ^{
         byref_int = 1;
         bl_var1 = 0;
@@ -82,21 +76,18 @@ void f() {
         return wid;
     };
 
-// CHECK: Inline instruction for block variable layout: 0x01
-// CHECK-i386: Inline instruction for block variable layout: 0x01
+// CHECK: Inline block variable layout: 0x01, BL_WEAK:1, BL_OPERATOR:0
     id (^h)() = ^{
         return wid;
     };
 
-// CHECK: Inline instruction for block variable layout: 0x020
-// CHECK-i386: Inline instruction for block variable layout: 0x020
+// CHECK: Inline block variable layout: 0x020, BL_BYREF:2, BL_OPERATOR:0
     void (^ii)() = ^{
        byref_int = 1;
        byref_bab = 0;
     };
 
-// CHECK: Inline instruction for block variable layout: 0x0102
-// CHECK-i386: Inline instruction for block variable layout: 0x0102
+// CHECK: Inline block variable layout: 0x0102, BL_STRONG:1, BL_WEAK:2, BL_OPERATOR:0
     void (^jj)() = ^{
       x(bar);
       x(wid1);
@@ -114,8 +105,7 @@ int main() {
         __weak NSString *w1 = 0;
 
 
-// CHECK: Inline instruction for block variable layout: 0x0201
-// CHECK-i386: Inline instruction for block variable layout: 0x0201
+// CHECK: Inline block variable layout: 0x0201, BL_STRONG:2, BL_WEAK:1, BL_OPERATOR:0
         dispatch_block_t block2 = ^{
                 NSLog(@"%@, %@, %@", s1, w1, s2);
         };
diff --git a/test/CodeGenObjC/arc-captured-block-var-layout.m b/test/CodeGenObjC/arc-captured-block-var-layout.m
index 42e6060dcc..f8518d1c44 100644
--- a/test/CodeGenObjC/arc-captured-block-var-layout.m
+++ b/test/CodeGenObjC/arc-captured-block-var-layout.m
@@ -34,7 +34,7 @@ void f() {
 
 // Test 1
 // Inline instruction for block variable layout: 0x0320 (3 strong 2 byref)
-// CHECK-LP64: Inline instruction for block variable layout: 0x0320
+// CHECK-LP64: Inline block variable layout: 0x0320, BL_STRONG:3, BL_BYREF:2, BL_OPERATOR:0
     void (^b)() = ^{
         byref_int = sh + ch+ch1+ch2 ;
         x(bar);
@@ -46,7 +46,7 @@ void f() {
 
 // Test 2
 // Inline instruction for block variable layout: 0x0331 (3 strong 3 byref 1 weak)
-// CHECK-LP64: Inline instruction for block variable layout: 0x0331
+// CHECK-LP64: Inline block variable layout: 0x0331, BL_STRONG:3, BL_BYREF:3, BL_WEAK:1, BL_OPERATOR:0
     void (^c)() = ^{
         byref_int = sh + ch+ch1+ch2 ;
         x(bar);
@@ -68,7 +68,7 @@ void g() {
   NSString *y;
   NSString *z;
 // Inline instruction for block variable layout: 0x0401 (4 strong 0 byref 1 weak)
-// CHECK-LP64: Inline instruction for block variable layout: 0x0401
+// CHECK-LP64: Inline block variable layout: 0x0401, BL_STRONG:4, BL_WEAK:1, BL_OPERATOR:0
   void (^c)() = ^{
    int j = i + bletch;
    x(foo);
@@ -109,7 +109,7 @@ void h() {
   union U u2;
   __block id block_id;
 
-// CHECK-LP64: block variable layout: BL_BYREF:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_OPERATOR:0
+// CHECK-LP64: Block variable layout: BL_BYREF:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_OPERATOR:0
   void (^c)() = ^{
     x(s2.ui.o1);
     x(u2.o1);
@@ -124,7 +124,7 @@ void arr1() {
     __unsafe_unretained id unsafe_unretained_var[4];
  } imported_s;
 
-// CHECK-LP64: block variable layout: BL_UNRETAINED:4, BL_OPERATOR:0
+// CHECK-LP64: Block variable layout: BL_UNRETAINED:4, BL_OPERATOR:0
     void (^c)() = ^{
         x(imported_s.unsafe_unretained_var[2]);
     };    
@@ -139,7 +139,7 @@ void arr2() {
     __unsafe_unretained id unsafe_unretained_var[4];
  } imported_s;
 
-// CHECK-LP64: block variable layout: BL_NON_OBJECT_WORD:1, BL_UNRETAINED:4, BL_OPERATOR:0
+// CHECK-LP64: Block variable layout: BL_NON_OBJECT_WORD:1, BL_UNRETAINED:4, BL_OPERATOR:0
     void (^c)() = ^{
         x(imported_s.unsafe_unretained_var[2]);
     };    
@@ -154,7 +154,7 @@ void arr3() {
     __unsafe_unretained id unsafe_unretained_var[0];
  } imported_s;
 
-// CHECK-LP64: block variable layout: BL_OPERATOR:0
+// CHECK-LP64: Block variable layout: BL_OPERATOR:0
     void (^c)() = ^{
       int i = imported_s.a;
     };    
@@ -180,7 +180,7 @@ void arr4() {
     } f4[2][2];
   } captured_s;
 
-// CHECK-LP64: block variable layout: BL_UNRETAINED:3, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_OPERATOR:0
+// CHECK-LP64: Block variable layout: BL_UNRETAINED:3, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_NON_OBJECT_WORD:1, BL_UNRETAINED:1, BL_OPERATOR:0
   void (^c)() = ^{
       id i = captured_s.f0.s_f1;
   };
@@ -198,7 +198,7 @@ void bf1() {
     int flag4: 24;
   } s;
 
-// CHECK-LP64: block variable layout: BL_OPERATOR:0
+// CHECK-LP64: Block variable layout: BL_OPERATOR:0
   int (^c)() = ^{
       return s.flag;
   };
@@ -211,7 +211,7 @@ void bf2() {
     int flag : 1;
   } s;
 
-// CHECK-LP64: block variable layout: BL_OPERATOR:0
+// CHECK-LP64: Block variable layout: BL_OPERATOR:0
   int (^c)() = ^{
       return s.flag;
   };
@@ -242,7 +242,7 @@ void bf3() {
         unsigned int _filler : 32;
     } _flags;
 
-// CHECK-LP64: block variable layout: BL_OPERATOR:0
+// CHECK-LP64: Block variable layout: BL_OPERATOR:0
   unsigned char (^c)() = ^{
       return _flags._draggedNodesAreDeletable;
   };
@@ -277,7 +277,7 @@ void bf4() {
         unsigned int _filler : 32;
     } _flags;
 
-// CHECK-LP64: block variable layout: BL_OPERATOR:0
+// CHECK-LP64: Block variable layout: BL_OPERATOR:0
   unsigned char (^c)() = ^{
       return _flags._draggedNodesAreDeletable;
   };
@@ -295,7 +295,7 @@ void bf5() {
         unsigned char flag1 : 1;
     } _flags;
 
-// CHECK-LP64: block variable layout: BL_OPERATOR:0
+// CHECK-LP64: Block variable layout: BL_OPERATOR:0
   unsigned char (^c)() = ^{
       return _flags.flag;
   };
@@ -312,7 +312,7 @@ void bf6() {
         unsigned char flag1 : 1;
     } _flags;
 
-// CHECK-LP64: block variable layout: BL_OPERATOR:0
+// CHECK-LP64: Block variable layout: BL_OPERATOR:0
   unsigned char (^c)() = ^{
       return _flags.flag;
   };
@@ -328,7 +328,7 @@ void Test7() {
     __weak id wid9, wid10, wid11, wid12;
     __weak id wid13, wid14, wid15, wid16;
     const id bar = (id) opaque_id();
-// CHECK-LP64: block variable layout: BL_STRONG:1, BL_WEAK:16, BL_OPERATOR:0
+// CHECK-LP64: Block variable layout: BL_STRONG:1, BL_WEAK:16, BL_OPERATOR:0
     void (^b)() = ^{
       x(bar);
       x(wid1);
@@ -363,7 +363,7 @@ __weak id wid;
     __weak id w9, w10, w11, w12;
     __weak id w13, w14, w15, w16;
     const id bar = (id) opaque_id();
-// CHECK-LP64: block variable layout: BL_STRONG:1, BL_WEAK:16, BL_WEAK:16, BL_WEAK:1, BL_OPERATOR:0
+// CHECK-LP64: Block variable layout: BL_STRONG:1, BL_WEAK:16, BL_WEAK:16, BL_WEAK:1, BL_OPERATOR:0
     void (^b)() = ^{
       x(bar);
       x(wid1);
diff --git a/test/CodeGenObjC/arc-literals.m b/test/CodeGenObjC/arc-literals.m
index d107a28506..ab6c82b743 100644
--- a/test/CodeGenObjC/arc-literals.m
+++ b/test/CodeGenObjC/arc-literals.m
@@ -47,10 +47,10 @@ void test_array(id a, id b) {
   // CHECK: call i8* @objc_retain(i8*
 
   // Constructing the array
-  // CHECK:      [[T0:%.*]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[OBJECTS:%[A-Za-z0-9]+]], i32 0, i32 0
+  // CHECK:      [[T0:%.*]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[OBJECTS:%[A-Za-z0-9]+]], i64 0, i64 0
   // CHECK-NEXT: [[V0:%.*]] = load i8*, i8** [[A]],
   // CHECK-NEXT: store i8* [[V0]], i8** [[T0]]
-  // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[OBJECTS]], i32 0, i32 1
+  // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[OBJECTS]], i64 0, i64 1
   // CHECK-NEXT: [[V1:%.*]] = load i8*, i8** [[B]],
   // CHECK-NEXT: store i8* [[V1]], i8** [[T0]]
 
@@ -83,16 +83,16 @@ void test_dictionary(id k1, id o1, id k2, id o2) {
   // CHECK: call i8* @objc_retain(i8*
 
   // Constructing the arrays
-  // CHECK:      [[T0:%.*]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[KEYS:%[A-Za-z0-9]+]], i32 0, i32 0
+  // CHECK:      [[T0:%.*]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[KEYS:%[A-Za-z0-9]+]], i64 0, i64 0
   // CHECK-NEXT: [[V0:%.*]] = load i8*, i8** [[K1]],
   // CHECK-NEXT: store i8* [[V0]], i8** [[T0]]
-  // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[OBJECTS:%[A-Za-z0-9]+]], i32 0, i32 0
+  // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[OBJECTS:%[A-Za-z0-9]+]], i64 0, i64 0
   // CHECK-NEXT: [[V1:%.*]] = load i8*, i8** [[O1]],
   // CHECK-NEXT: store i8* [[V1]], i8** [[T0]]
-  // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[KEYS]], i32 0, i32 1
+  // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[KEYS]], i64 0, i64 1
   // CHECK-NEXT: [[V2:%.*]] = load i8*, i8** [[K2]],
   // CHECK-NEXT: store i8* [[V2]], i8** [[T0]]
-  // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[OBJECTS]], i32 0, i32 1
+  // CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[OBJECTS]], i64 0, i64 1
   // CHECK-NEXT: [[V3:%.*]] = load i8*, i8** [[O2]],
   // CHECK-NEXT: store i8* [[V3]], i8** [[T0]]
 
@@ -128,7 +128,7 @@ void test_property(B *b) {
   // Retain parameter
   // CHECK: call i8* @objc_retain
 
-  // CHECK:      [[T0:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[OBJECTS:%.*]], i32 0, i32 0
+  // CHECK:      [[T0:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[OBJECTS:%.*]], i64 0, i64 0
 
   // Invoke 'prop'
   // CHECK:      [[SEL:%.*]] = load i8*, i8** @OBJC_SELECTOR_REFERENCES
diff --git a/test/CodeGenObjC/arc.m b/test/CodeGenObjC/arc.m
index 3aafefd43c..f756df1460 100644
--- a/test/CodeGenObjC/arc.m
+++ b/test/CodeGenObjC/arc.m
@@ -515,7 +515,7 @@ void test19() {
 
   // CHECK-NEXT: [[CALL:%.*]] = call i8* @test19_helper()
   // CHECK-NEXT: [[T1:%.*]] = call i8* @objc_retainAutoreleasedReturnValue(i8* [[CALL]]) [[NUW]]
-  // CHECK-NEXT: [[SLOT:%.*]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[X]], i32 0, i64 2
+  // CHECK-NEXT: [[SLOT:%.*]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[X]], i64 0, i64 2
   // CHECK-NEXT: [[T0:%.*]] = load i8*, i8** [[SLOT]]
   // CHECK-NEXT: store i8* [[T1]], i8** [[SLOT]]
   // CHECK-NEXT: call void @objc_release(i8* [[T0]]) [[NUW]]
@@ -556,7 +556,7 @@ void test20(unsigned n) {
   // Zero-initialize.
   // CHECK-NEXT: [[T0:%.*]] = bitcast i8** [[VLA]] to i8*
   // CHECK-NEXT: [[T1:%.*]] = mul nuw i64 [[DIM]], 8
-  // CHECK-NEXT: call void @llvm.memset.p0i8.i64(i8* [[T0]], i8 0, i64 [[T1]], i32 8, i1 false)
+  // CHECK-NEXT: call void @llvm.memset.p0i8.i64(i8* [[T0]], i8 0, i64 [[T1]], i32 16, i1 false)
 
   // Destroy.
   // CHECK-NEXT: [[END:%.*]] = getelementptr inbounds i8*, i8** [[VLA]], i64 [[DIM]]
@@ -599,7 +599,7 @@ void test21(unsigned n) {
   // CHECK-NEXT: [[T0:%.*]] = bitcast [3 x i8*]* [[VLA]] to i8*
   // CHECK-NEXT: [[T1:%.*]] = mul nuw i64 2, [[DIM]]
   // CHECK-NEXT: [[T2:%.*]] = mul nuw i64 [[T1]], 24
-  // CHECK-NEXT: call void @llvm.memset.p0i8.i64(i8* [[T0]], i8 0, i64 [[T2]], i32 8, i1 false)
+  // CHECK-NEXT: call void @llvm.memset.p0i8.i64(i8* [[T0]], i8 0, i64 [[T2]], i32 16, i1 false)
 
   // Destroy.
   // CHECK-NEXT: [[T0:%.*]] = mul nuw i64 2, [[DIM]]
diff --git a/test/CodeGenObjC/debug-info-block-captured-self.m b/test/CodeGenObjC/debug-info-block-captured-self.m
index b2ad97b65f..06dd9ea60e 100644
--- a/test/CodeGenObjC/debug-info-block-captured-self.m
+++ b/test/CodeGenObjC/debug-info-block-captured-self.m
@@ -53,13 +53,13 @@ typedef enum {
 // CHECK: define internal void {{.*}}_block_invoke{{.*}}
 // CHECK:        %[[MEM1:.*]] = alloca i8*, align 8
 // CHECK-NEXT:   %[[MEM2:.*]] = alloca i8*, align 8
+// CHECK-NEXT:   [[DBGADDR:%.*]] = alloca [[BLOCK_T:<{.*}>]]*, align 8
 // CHECK:        store i8* [[BLOCK_DESC:%.*]], i8** %[[MEM1]], align 8
 // CHECK:        %[[TMP0:.*]] = load i8*, i8** %[[MEM1]]
 // CHECK:        call void @llvm.dbg.value(metadata i8* %[[TMP0]], i64 0, metadata ![[BDMD:[0-9]+]], metadata !{{.*}})
 // CHECK:        call void @llvm.dbg.declare(metadata i8* [[BLOCK_DESC]], metadata ![[BDMD:[0-9]+]], metadata !{{.*}})
-// CHECK:        %[[TMP1:.*]] = bitcast
-// CHECK-NEXT:   store
-// CHECK:        call void @llvm.dbg.declare(metadata <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, %0* }>** {{[^,]*}}, metadata ![[SELF:.*]], metadata !{{.*}})
+// CHECK:        store [[BLOCK_T]]* {{%.*}}, [[BLOCK_T]]** [[DBGADDR]], align 8
+// CHECK:        call void @llvm.dbg.declare(metadata [[BLOCK_T]]** [[DBGADDR]], metadata ![[SELF:.*]], metadata !{{.*}})
 // make sure we are still in the same function
 // CHECK: define {{.*}}__copy_helper_block_
 // Metadata
diff --git a/test/CodeGenObjC/ivar-base-as-invariant-load.m b/test/CodeGenObjC/ivar-base-as-invariant-load.m
index 4a17eb16f6..a3201e0401 100644
--- a/test/CodeGenObjC/ivar-base-as-invariant-load.m
+++ b/test/CodeGenObjC/ivar-base-as-invariant-load.m
@@ -23,7 +23,7 @@
 
 @end
 
-// CHECK: [[T1:%.*]] = load i64, i64* @"OBJC_IVAR_$_A._flags", !invariant.load ![[MD_NUM:[0-9]+]]
-// CHECK: [[T2:%.*]] = load i64, i64* @"OBJC_IVAR_$_A._flags", !invariant.load ![[MD_NUM]]
-// CHECK: [[T3:%.*]] = load i64, i64* @"OBJC_IVAR_$_A._flags", !invariant.load ![[MD_NUM]]
+// CHECK: [[T1:%.*]] = load i64, i64* @"OBJC_IVAR_$_A._flags", align 8, !invariant.load ![[MD_NUM:[0-9]+]]
+// CHECK: [[T2:%.*]] = load i64, i64* @"OBJC_IVAR_$_A._flags", align 8, !invariant.load ![[MD_NUM]]
+// CHECK: [[T3:%.*]] = load i64, i64* @"OBJC_IVAR_$_A._flags", align 8, !invariant.load ![[MD_NUM]]
 //
diff --git a/test/CodeGenObjC/ivar-invariant.m b/test/CodeGenObjC/ivar-invariant.m
index b9c5bec930..eb1ba9a0d7 100644
--- a/test/CodeGenObjC/ivar-invariant.m
+++ b/test/CodeGenObjC/ivar-invariant.m
@@ -29,7 +29,7 @@
 @end
 
 // CHECK: define internal i8* @"\01-[Derived init]"
-// CHECK: [[IVAR:%.*]] = load i64, i64* @"OBJC_IVAR_$_Derived.member", !invariant.load
+// CHECK: [[IVAR:%.*]] = load i64, i64* @"OBJC_IVAR_$_Derived.member", align 8, !invariant.load
 
 void * variant_load_1(int i) {
     void *ptr;
@@ -41,7 +41,7 @@ void * variant_load_1(int i) {
 }
 
 // CHECK-LABEL: define i8* @variant_load_1(i32 %i)
-// CHECK: [[IVAR:%.*]] = load i64, i64* @"OBJC_IVAR_$_Derived.member"{{$}}
+// CHECK: [[IVAR:%.*]] = load i64, i64* @"OBJC_IVAR_$_Derived.member", align 8{{$}}
 
 @interface Container : Derived @end
 @implementation Container
@@ -51,8 +51,8 @@ void * variant_load_1(int i) {
 }
 @end
 
-// CHECK: define internal i8* @"\01-[Container invariant_load_1]"
-// CHECK: [[IVAR:%.*]] = load i64, i64* @"OBJC_IVAR_$_Derived.member", !invariant.load
+// CHECK-LABEL: define internal i8* @"\01-[Container invariant_load_1]"
+// CHECK: [[IVAR:%.*]] = load i64, i64* @"OBJC_IVAR_$_Derived.member", align 8, !invariant.load
 
 @interface ForBlock
 { 
diff --git a/test/CodeGenObjC/mrr-captured-block-var-inlined-layout.m b/test/CodeGenObjC/mrr-captured-block-var-inlined-layout.m
index 6ea656443a..c87140a87d 100644
--- a/test/CodeGenObjC/mrr-captured-block-var-inlined-layout.m
+++ b/test/CodeGenObjC/mrr-captured-block-var-inlined-layout.m
@@ -1,7 +1,7 @@
 // RUN: %clang_cc1 -fblocks -fobjc-runtime-has-weak -triple x86_64-apple-darwin -print-ivar-layout -emit-llvm -o /dev/null %s > %t-64.layout
-// RUN: FileCheck --input-file=%t-64.layout %s
+// RUN: FileCheck -check-prefix=CHECK -check-prefix=CHECK-64 --input-file=%t-64.layout %s
 // RUN: %clang_cc1 -fblocks -fobjc-runtime-has-weak -triple i386-apple-darwin -print-ivar-layout -emit-llvm -o /dev/null %s > %t-32.layout
-// RUN: FileCheck -check-prefix=CHECK-i386 --input-file=%t-32.layout %s
+// RUN: FileCheck -check-prefix=CHECK -check-prefix=CHECK-32 --input-file=%t-32.layout %s
 // rdar://12184410
 // rdar://12184410
 
@@ -20,15 +20,13 @@ void f() {
     __block id bl_var1;
 
 // block variable layout: BL_STRONG:1, BL_OPERATOR:0
-// CHECK: Inline instruction for block variable layout: 0x0100
-// CHECK-i386: Inline instruction for block variable layout: 0x0100
+// CHECK: Inline block variable layout: 0x0100, BL_STRONG:1, BL_OPERATOR:0
     void (^b)() = ^{
         x(bar);
     };    
 
 // block variable layout: BL_STRONG:2, BL_BYREF:1, BL_OPERATOR:0
-// CHECK: Inline instruction for block variable layout: 0x0210
-// CHECK-i386: Inline instruction for block variable layout: 0x0210
+// CHECK: Inline block variable layout: 0x0210, BL_STRONG:2, BL_BYREF:1, BL_OPERATOR:0
     void (^c)() = ^{
         x(bar);
         x(baz);
@@ -36,8 +34,7 @@ void f() {
     };    
 
 // block variable layout: BL_STRONG:2, BL_BYREF:3, BL_OPERATOR:0
-// CHECK: Inline instruction for block variable layout: 0x0230
-// CHECK-i386: Inline instruction for block variable layout: 0x0230
+// CHECK: Inline block variable layout: 0x0230, BL_STRONG:2, BL_BYREF:3, BL_OPERATOR:0
     void (^d)() = ^{
         x(bar);
         x(baz);
@@ -47,8 +44,7 @@ void f() {
     };
 
 // block variable layout: BL_STRONG:2, BL_BYREF:3, BL_OPERATOR:0
-// CHECK: Inline instruction for block variable layout: 0x0230
-// CHECK-i386: Inline instruction for block variable layout: 0x0230
+// CHECK: Inline block variable layout: 0x0230, BL_STRONG:2, BL_BYREF:3, BL_OPERATOR:0
     id (^e)() = ^{
         x(bar);
         x(baz);
@@ -58,8 +54,7 @@ void f() {
         return wid;
     };
 
-// CHECK: Inline instruction for block variable layout: 0x020
-// CHECK-i386: Inline instruction for block variable layout: 0x020
+// CHECK: Inline block variable layout: 0x020, BL_BYREF:2, BL_OPERATOR:0
     void (^ii)() = ^{
        byref_int = 1;
        byref_bab = 0;
diff --git a/test/CodeGenObjC/selector-ref-invariance.m b/test/CodeGenObjC/selector-ref-invariance.m
index 5758a1cd7f..18fb828d29 100644
--- a/test/CodeGenObjC/selector-ref-invariance.m
+++ b/test/CodeGenObjC/selector-ref-invariance.m
@@ -3,7 +3,7 @@
 // rdar://6027699
 
 void test(id x) {
-// CHECK: load i8*, i8** @OBJC_SELECTOR_REFERENCES_, !invariant.load
+// CHECK: load i8*, i8** @OBJC_SELECTOR_REFERENCES_, align 8, !invariant.load
 // CHECK: @objc_msgSend
   [x foo];
 }
diff --git a/test/CodeGenObjCXX/arc-new-delete.mm b/test/CodeGenObjCXX/arc-new-delete.mm
index 9a61f183c6..f853ea4366 100644
--- a/test/CodeGenObjCXX/arc-new-delete.mm
+++ b/test/CodeGenObjCXX/arc-new-delete.mm
@@ -1,4 +1,5 @@
-// RUN: %clang_cc1 -fobjc-arc -fobjc-runtime-has-weak -fblocks -triple x86_64-apple-darwin10.0.0 -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -fobjc-arc -fobjc-runtime-has-weak -fblocks -triple x86_64-apple-darwin10.0.0 -emit-llvm -o - %s | FileCheck %s -check-prefix=CHECK -check-prefix=UNOPT
+// RUN: %clang_cc1 -fobjc-arc -fobjc-runtime-has-weak -fblocks -triple x86_64-apple-darwin10.0.0 -emit-llvm -o - %s -O -disable-llvm-optzns | FileCheck %s -check-prefix=CHECK -check-prefix=OPT
 
 typedef __strong id strong_id;
 typedef __weak id weak_id;
@@ -6,8 +7,10 @@ typedef __weak id weak_id;
 // CHECK-LABEL: define void @_Z8test_newP11objc_object
 void test_new(id invalue) {
   // CHECK: [[INVALUEADDR:%.*]] = alloca i8*
-  // CHECK-NEXT: store i8* null, i8** [[INVALUEADDR]]
-  // CHECK-NEXT: call void @objc_storeStrong(i8** [[INVALUEADDR]], i8* [[INVALUE:%.*]])
+  // UNOPT-NEXT: store i8* null, i8** [[INVALUEADDR]]
+  // UNOPT-NEXT: call void @objc_storeStrong(i8** [[INVALUEADDR]], i8* [[INVALUE:%.*]])
+  // OPT-NEXT: [[T0:%.*]] = call i8* @objc_retain(i8* [[INVALUE:%.*]])
+  // OPT-NEXT: store i8* [[T0]], i8** [[INVALUEADDR]]
 
   // CHECK: call noalias i8* @_Znwm
   // CHECK-NEXT: {{bitcast i8\*.*to i8\*\*}}
@@ -15,7 +18,8 @@ void test_new(id invalue) {
   new strong_id;
   // CHECK: call noalias i8* @_Znwm
   // CHECK-NEXT: {{bitcast i8\*.*to i8\*\*}}
-  // CHECK-NEXT: store i8* null, i8**
+  // UNOPT-NEXT: store i8* null, i8**
+  // OPT-NEXT: call i8* @objc_initWeak(i8** {{.*}}, i8* null)
   new weak_id;
 
   // CHECK: call noalias i8* @_Znwm
@@ -24,7 +28,8 @@ void test_new(id invalue) {
   new __strong id;
   // CHECK: call noalias i8* @_Znwm
   // CHECK-NEXT: {{bitcast i8\*.*to i8\*\*}}
-  // CHECK-NEXT: store i8* null, i8**
+  // UNOPT-NEXT: store i8* null, i8**
+  // OPT-NEXT: call i8* @objc_initWeak(i8** {{.*}}, i8* null)
   new __weak id;
 
   // CHECK: call noalias i8* @_Znwm
@@ -36,7 +41,8 @@ void test_new(id invalue) {
   // CHECK: call i8* @objc_initWeak
   new __weak id(invalue);
 
-  // CHECK: call void @objc_storeStrong
+  // UNOPT: call void @objc_storeStrong
+  // OPT: call void @objc_release
   // CHECK: ret void
 }
 
@@ -57,8 +63,9 @@ void test_array_new() {
 // CHECK-LABEL: define void @_Z11test_deletePU8__strongP11objc_objectPU6__weakS0_
 void test_delete(__strong id *sptr, __weak id *wptr) {
   // CHECK: br i1
-  // CHECK: load i8*, i8**
-  // CHECK-NEXT: call void @objc_release
+  // UNOPT: call void @objc_storeStrong(i8** {{.*}}, i8* null)
+  // OPT: load i8*, i8**
+  // OPT-NEXT: call void @objc_release
   // CHECK: call void @_ZdlPv
   delete sptr;
 
@@ -77,7 +84,9 @@ void test_array_delete(__strong id *sptr, __weak id *wptr) {
   // CHECK-NEXT: icmp eq i8** [[BEGIN]], [[END]]
   // CHECK: [[PAST:%.*]] = phi i8** [ [[END]], {{%.*}} ], [ [[CUR:%.*]],
   // CHECK-NEXT: [[CUR]] = getelementptr inbounds i8*, i8** [[PAST]], i64 -1
-  // CHECK-NEXT: call void @objc_storeStrong(i8** [[CUR]], i8* null)
+  // UNOPT-NEXT: call void @objc_storeStrong(i8** [[CUR]], i8* null)
+  // OPT-NEXT: [[T0:%.*]] = load i8*, i8** [[CUR]]
+  // OPT-NEXT: objc_release(i8* [[T0]])
   // CHECK-NEXT: icmp eq i8** [[CUR]], [[BEGIN]]
   // CHECK: call void @_ZdaPv
   delete [] sptr;
diff --git a/test/CodeGenObjCXX/literals.mm b/test/CodeGenObjCXX/literals.mm
index 7089de23b9..4d1b501941 100644
--- a/test/CodeGenObjCXX/literals.mm
+++ b/test/CodeGenObjCXX/literals.mm
@@ -22,14 +22,14 @@ void test_array() {
   // Initializing first element
   // CHECK: [[PTR1:%.*]] = bitcast i8** [[ARR]] to i8*
   // CHECK-NEXT: call void @llvm.lifetime.start(i64 8, i8* [[PTR1]])
-  // CHECK: [[ELEMENT0:%[a-zA-Z0-9.]+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[OBJECTS]], i32 0, i32 0
+  // CHECK: [[ELEMENT0:%[a-zA-Z0-9.]+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[OBJECTS]], i64 0, i64 0
   // CHECK-NEXT: call void @_ZN1XC1Ev
   // CHECK-NEXT: [[OBJECT0:%[a-zA-Z0-9.]+]] = invoke i8* @_ZNK1XcvP11objc_objectEv
   // CHECK: [[RET0:%[a-zA-Z0-9.]+]] = call i8* @objc_retainAutoreleasedReturnValue(i8* [[OBJECT0]])
   // CHECK: store i8* [[RET0]], i8** [[ELEMENT0]]
   
   // Initializing the second element
-  // CHECK: [[ELEMENT1:%[a-zA-Z0-9.]+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[OBJECTS]], i32 0, i32 1
+  // CHECK: [[ELEMENT1:%[a-zA-Z0-9.]+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[OBJECTS]], i64 0, i64 1
   // CHECK-NEXT: invoke void @_ZN1YC1Ev
   // CHECK: [[OBJECT1:%[a-zA-Z0-9.]+]] = invoke i8* @_ZNK1YcvP11objc_objectEv
   // CHECK: [[RET1:%[a-zA-Z0-9.]+]] = call i8* @objc_retainAutoreleasedReturnValue(i8* [[OBJECT1]])
@@ -74,14 +74,14 @@ void test_array_instantiation() {
   // Initializing first element
   // CHECK:      [[PTR1:%.*]] = bitcast i8** [[ARR]] to i8*
   // CHECK-NEXT: call void @llvm.lifetime.start(i64 8, i8* [[PTR1]])
-  // CHECK: [[ELEMENT0:%[a-zA-Z0-9.]+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[OBJECTS]], i32 0, i32 0
+  // CHECK: [[ELEMENT0:%[a-zA-Z0-9.]+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[OBJECTS]], i64 0, i64 0
   // CHECK-NEXT: call void @_ZN1XC1Ev
   // CHECK-NEXT: [[OBJECT0:%[a-zA-Z0-9.]+]] = invoke i8* @_ZNK1XcvP11objc_objectEv
   // CHECK: [[RET0:%[a-zA-Z0-9.]+]] = call i8* @objc_retainAutoreleasedReturnValue(i8* [[OBJECT0]])
   // CHECK: store i8* [[RET0]], i8** [[ELEMENT0]]
   
   // Initializing the second element
-  // CHECK: [[ELEMENT1:%[a-zA-Z0-9.]+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[OBJECTS]], i32 0, i32 1
+  // CHECK: [[ELEMENT1:%[a-zA-Z0-9.]+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[OBJECTS]], i64 0, i64 1
   // CHECK-NEXT: invoke void @_ZN1YC1Ev
   // CHECK: [[OBJECT1:%[a-zA-Z0-9.]+]] = invoke i8* @_ZNK1YcvP11objc_objectEv
   // CHECK: [[RET1:%[a-zA-Z0-9.]+]] = call i8* @objc_retainAutoreleasedReturnValue(i8* [[OBJECT1]])
diff --git a/test/CodeGenObjCXX/property-lvalue-capture.mm b/test/CodeGenObjCXX/property-lvalue-capture.mm
index b800c39fb3..c5a753c8cf 100644
--- a/test/CodeGenObjCXX/property-lvalue-capture.mm
+++ b/test/CodeGenObjCXX/property-lvalue-capture.mm
@@ -24,10 +24,10 @@ typedef Quad2<double> Quad2d;
 }
 @end
 
-// CHECK:   [[TWO:%.*]] = load i8*, i8** @OBJC_SELECTOR_REFERENCES_, !invariant.load ![[MD_NUM:[0-9]+]]
+// CHECK:   [[TWO:%.*]] = load i8*, i8** @OBJC_SELECTOR_REFERENCES_, align 8, !invariant.load ![[MD_NUM:[0-9]+]]
 // CHECK:   [[THREE:%.*]] = bitcast [[ONET:%.*]]* [[ONE:%.*]] to i8*
 // CHECK:   [[CALL:%.*]] = call nonnull %struct.Quad2* bitcast (i8* (i8*, i8*, ...)* @objc_msgSend to %struct.Quad2* (i8*, i8*)*)(i8* [[THREE]], i8* [[TWO]])
-// CHECK:   [[FOUR:%.*]] = load i8*, i8** @OBJC_SELECTOR_REFERENCES_.2, !invariant.load ![[MD_NUM]]
+// CHECK:   [[FOUR:%.*]] = load i8*, i8** @OBJC_SELECTOR_REFERENCES_.2, align 8, !invariant.load ![[MD_NUM]]
 // CHECK:   [[FIVE:%.*]] = bitcast [[ONET]]* [[ZERO:%.*]] to i8*
 // CHECK:   call void bitcast (i8* (i8*, i8*, ...)* @objc_msgSend to void (i8*, i8*, %struct.Quad2*)*)(i8* [[FIVE]], i8* [[FOUR]], %struct.Quad2* nonnull [[CALL]])
 
@@ -47,7 +47,7 @@ void test(C *c, const A &a) {
 }
 
 // CHECK:   [[ONE1:%.*]] = load %struct.A*, %struct.A** [[AADDR:%.*]], align 8
-// CHECK:   [[TWO1:%.*]] = load i8*, i8** @OBJC_SELECTOR_REFERENCES_.5, !invariant.load ![[MD_NUM]]
+// CHECK:   [[TWO1:%.*]] = load i8*, i8** @OBJC_SELECTOR_REFERENCES_.5, align 8, !invariant.load ![[MD_NUM]]
 // CHECK:   [[THREE1:%.*]] = bitcast [[TWOT:%.*]]* [[ZERO1:%.*]] to i8*
 // CHECK:   call void bitcast (i8* (i8*, i8*, ...)* @objc_msgSend to void (i8*, i8*, %struct.A*)*)(i8* [[THREE1]], i8* [[TWO1]], %struct.A* dereferenceable({{[0-9]+}}) [[ONE1]])
 // CHECK:   store %struct.A* [[ONE1]], %struct.A** [[RESULT:%.*]], align 8
diff --git a/test/CodeGenObjCXX/property-object-conditional-exp.mm b/test/CodeGenObjCXX/property-object-conditional-exp.mm
index e3fc2d7096..cdee635a6f 100644
--- a/test/CodeGenObjCXX/property-object-conditional-exp.mm
+++ b/test/CodeGenObjCXX/property-object-conditional-exp.mm
@@ -24,10 +24,10 @@ extern "C" bool CGRectIsEmpty(CGRect);
 
 // CHECK: [[SRC:%.*]] = call { i8*, i32 } bitcast (i8* (i8*, i8*, ...)* @objc_msgSend
 // CHECK-NEXT: bitcast
-// CHECK-NEXT:getelementptr { i8*, i32 }, { i8*, i32 }* [[SRC:%.*]]
+// CHECK-NEXT:getelementptr inbounds { i8*, i32 }, { i8*, i32 }* [[SRC:%.*]]
 // CHECK-NEXT:extractvalue
 // CHECK-NEXT:store
-// CHECK-NEXT:getelementptr { i8*, i32 }, { i8*, i32 }* [[SRC:%.*]]
+// CHECK-NEXT:getelementptr inbounds { i8*, i32 }, { i8*, i32 }* [[SRC:%.*]]
 // CHECK-NEXT:extractvalue
 // CHECK-NEXT:store
   dataRect = CGRectIsEmpty(virtualBounds) ? self.bounds : virtualBounds;
diff --git a/test/OpenMP/for_reduction_codegen.cpp b/test/OpenMP/for_reduction_codegen.cpp
index 16a2213655..3d326bda0a 100644
--- a/test/OpenMP/for_reduction_codegen.cpp
+++ b/test/OpenMP/for_reduction_codegen.cpp
@@ -81,7 +81,7 @@ int main() {
     // LAMBDA: call void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
     // LAMBDA: call void @__kmpc_for_static_fini(
 
-    // LAMBDA: [[G_PRIV_REF:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[RED_LIST]], i32 0, i32 0
+    // LAMBDA: [[G_PRIV_REF:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[RED_LIST]], i64 0, i64 0
     // LAMBDA: [[BITCAST:%.+]] = bitcast double* [[G_PRIVATE_ADDR]] to i8*
     // LAMBDA: store i8* [[BITCAST]], i8** [[G_PRIV_REF]],
     // LAMBDA: call i32 @__kmpc_reduce(
@@ -143,7 +143,7 @@ int main() {
     // BLOCKS: call void {{%.+}}(i8
     // BLOCKS: call void @__kmpc_for_static_fini(
 
-    // BLOCKS: [[G_PRIV_REF:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[RED_LIST]], i32 0, i32 0
+    // BLOCKS: [[G_PRIV_REF:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[RED_LIST]], i64 0, i64 0
     // BLOCKS: [[BITCAST:%.+]] = bitcast double* [[G_PRIVATE_ADDR]] to i8*
     // BLOCKS: store i8* [[BITCAST]], i8** [[G_PRIV_REF]],
     // BLOCKS: call i32 @__kmpc_reduce(
@@ -243,16 +243,16 @@ int main() {
 
 // void *RedList[<n>] = {<ReductionVars>[0], ..., <ReductionVars>[<n>-1]};
 
-// CHECK: [[T_VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 0
+// CHECK: [[T_VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 0
 // CHECK: [[BITCAST:%.+]] = bitcast float* [[T_VAR_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[T_VAR_PRIV_REF]],
-// CHECK: [[VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 1
+// CHECK: [[VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 1
 // CHECK: [[BITCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[VAR_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[VAR_PRIV_REF]],
-// CHECK: [[VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 2
+// CHECK: [[VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 2
 // CHECK: [[BITCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[VAR1_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[VAR1_PRIV_REF]],
-// CHECK: [[T_VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 3
+// CHECK: [[T_VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 3
 // CHECK: [[BITCAST:%.+]] = bitcast float* [[T_VAR1_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[T_VAR1_PRIV_REF]],
 
@@ -392,38 +392,38 @@ int main() {
 // }
 // CHECK: define internal void [[REDUCTION_FUNC]](i8*, i8*)
 // t_var_lhs = (float*)lhs[0];
-// CHECK: [[T_VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS:%.+]], i32 0, i32 0
+// CHECK: [[T_VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS:%.+]], i64 0, i64 0
 // CHECK: [[T_VAR_RHS_VOID:%.+]] = load i8*, i8** [[T_VAR_RHS_REF]],
 // CHECK: [[T_VAR_RHS:%.+]] = bitcast i8* [[T_VAR_RHS_VOID]] to float*
 // t_var_rhs = (float*)rhs[0];
-// CHECK: [[T_VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS:%.+]], i32 0, i32 0
+// CHECK: [[T_VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS:%.+]], i64 0, i64 0
 // CHECK: [[T_VAR_LHS_VOID:%.+]] = load i8*, i8** [[T_VAR_LHS_REF]],
 // CHECK: [[T_VAR_LHS:%.+]] = bitcast i8* [[T_VAR_LHS_VOID]] to float*
 
 // var_lhs = (S<float>*)lhs[1];
-// CHECK: [[VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 1
+// CHECK: [[VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 1
 // CHECK: [[VAR_RHS_VOID:%.+]] = load i8*, i8** [[VAR_RHS_REF]],
 // CHECK: [[VAR_RHS:%.+]] = bitcast i8* [[VAR_RHS_VOID]] to [[S_FLOAT_TY]]*
 // var_rhs = (S<float>*)rhs[1];
-// CHECK: [[VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 1
+// CHECK: [[VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 1
 // CHECK: [[VAR_LHS_VOID:%.+]] = load i8*, i8** [[VAR_LHS_REF]],
 // CHECK: [[VAR_LHS:%.+]] = bitcast i8* [[VAR_LHS_VOID]] to [[S_FLOAT_TY]]*
 
 // var1_lhs = (S<float>*)lhs[2];
-// CHECK: [[VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 2
+// CHECK: [[VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 2
 // CHECK: [[VAR1_RHS_VOID:%.+]] = load i8*, i8** [[VAR1_RHS_REF]],
 // CHECK: [[VAR1_RHS:%.+]] = bitcast i8* [[VAR1_RHS_VOID]] to [[S_FLOAT_TY]]*
 // var1_rhs = (S<float>*)rhs[2];
-// CHECK: [[VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 2
+// CHECK: [[VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 2
 // CHECK: [[VAR1_LHS_VOID:%.+]] = load i8*, i8** [[VAR1_LHS_REF]],
 // CHECK: [[VAR1_LHS:%.+]] = bitcast i8* [[VAR1_LHS_VOID]] to [[S_FLOAT_TY]]*
 
 // t_var1_lhs = (float*)lhs[3];
-// CHECK: [[T_VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 3
+// CHECK: [[T_VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 3
 // CHECK: [[T_VAR1_RHS_VOID:%.+]] = load i8*, i8** [[T_VAR1_RHS_REF]],
 // CHECK: [[T_VAR1_RHS:%.+]] = bitcast i8* [[T_VAR1_RHS_VOID]] to float*
 // t_var1_rhs = (float*)rhs[3];
-// CHECK: [[T_VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 3
+// CHECK: [[T_VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 3
 // CHECK: [[T_VAR1_LHS_VOID:%.+]] = load i8*, i8** [[T_VAR1_LHS_REF]],
 // CHECK: [[T_VAR1_LHS:%.+]] = bitcast i8* [[T_VAR1_LHS_VOID]] to float*
 
@@ -516,16 +516,16 @@ int main() {
 
 // void *RedList[<n>] = {<ReductionVars>[0], ..., <ReductionVars>[<n>-1]};
 
-// CHECK: [[T_VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 0
+// CHECK: [[T_VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 0
 // CHECK: [[BITCAST:%.+]] = bitcast i{{[0-9]+}}* [[T_VAR_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[T_VAR_PRIV_REF]],
-// CHECK: [[VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 1
+// CHECK: [[VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 1
 // CHECK: [[BITCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[VAR_PRIV_REF]],
-// CHECK: [[VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 2
+// CHECK: [[VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 2
 // CHECK: [[BITCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR1_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[VAR1_PRIV_REF]],
-// CHECK: [[T_VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 3
+// CHECK: [[T_VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 3
 // CHECK: [[BITCAST:%.+]] = bitcast i{{[0-9]+}}* [[T_VAR1_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[T_VAR1_PRIV_REF]],
 
@@ -634,38 +634,38 @@ int main() {
 // }
 // CHECK: define internal void [[REDUCTION_FUNC]](i8*, i8*)
 // t_var_lhs = (i{{[0-9]+}}*)lhs[0];
-// CHECK: [[T_VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS:%.+]], i32 0, i32 0
+// CHECK: [[T_VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS:%.+]], i64 0, i64 0
 // CHECK: [[T_VAR_RHS_VOID:%.+]] = load i8*, i8** [[T_VAR_RHS_REF]],
 // CHECK: [[T_VAR_RHS:%.+]] = bitcast i8* [[T_VAR_RHS_VOID]] to i{{[0-9]+}}*
 // t_var_rhs = (i{{[0-9]+}}*)rhs[0];
-// CHECK: [[T_VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS:%.+]], i32 0, i32 0
+// CHECK: [[T_VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS:%.+]], i64 0, i64 0
 // CHECK: [[T_VAR_LHS_VOID:%.+]] = load i8*, i8** [[T_VAR_LHS_REF]],
 // CHECK: [[T_VAR_LHS:%.+]] = bitcast i8* [[T_VAR_LHS_VOID]] to i{{[0-9]+}}*
 
 // var_lhs = (S<i{{[0-9]+}}>*)lhs[1];
-// CHECK: [[VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 1
+// CHECK: [[VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 1
 // CHECK: [[VAR_RHS_VOID:%.+]] = load i8*, i8** [[VAR_RHS_REF]],
 // CHECK: [[VAR_RHS:%.+]] = bitcast i8* [[VAR_RHS_VOID]] to [[S_INT_TY]]*
 // var_rhs = (S<i{{[0-9]+}}>*)rhs[1];
-// CHECK: [[VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 1
+// CHECK: [[VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 1
 // CHECK: [[VAR_LHS_VOID:%.+]] = load i8*, i8** [[VAR_LHS_REF]],
 // CHECK: [[VAR_LHS:%.+]] = bitcast i8* [[VAR_LHS_VOID]] to [[S_INT_TY]]*
 
 // var1_lhs = (S<i{{[0-9]+}}>*)lhs[2];
-// CHECK: [[VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 2
+// CHECK: [[VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 2
 // CHECK: [[VAR1_RHS_VOID:%.+]] = load i8*, i8** [[VAR1_RHS_REF]],
 // CHECK: [[VAR1_RHS:%.+]] = bitcast i8* [[VAR1_RHS_VOID]] to [[S_INT_TY]]*
 // var1_rhs = (S<i{{[0-9]+}}>*)rhs[2];
-// CHECK: [[VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 2
+// CHECK: [[VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 2
 // CHECK: [[VAR1_LHS_VOID:%.+]] = load i8*, i8** [[VAR1_LHS_REF]],
 // CHECK: [[VAR1_LHS:%.+]] = bitcast i8* [[VAR1_LHS_VOID]] to [[S_INT_TY]]*
 
 // t_var1_lhs = (i{{[0-9]+}}*)lhs[3];
-// CHECK: [[T_VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 3
+// CHECK: [[T_VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 3
 // CHECK: [[T_VAR1_RHS_VOID:%.+]] = load i8*, i8** [[T_VAR1_RHS_REF]],
 // CHECK: [[T_VAR1_RHS:%.+]] = bitcast i8* [[T_VAR1_RHS_VOID]] to i{{[0-9]+}}*
 // t_var1_rhs = (i{{[0-9]+}}*)rhs[3];
-// CHECK: [[T_VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 3
+// CHECK: [[T_VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 3
 // CHECK: [[T_VAR1_LHS_VOID:%.+]] = load i8*, i8** [[T_VAR1_LHS_REF]],
 // CHECK: [[T_VAR1_LHS:%.+]] = bitcast i8* [[T_VAR1_LHS_VOID]] to i{{[0-9]+}}*
 
diff --git a/test/OpenMP/parallel_codegen.cpp b/test/OpenMP/parallel_codegen.cpp
index 6486e44a9d..37d1679357 100644
--- a/test/OpenMP/parallel_codegen.cpp
+++ b/test/OpenMP/parallel_codegen.cpp
@@ -49,7 +49,7 @@ int main (int argc, char **argv) {
 // CHECK-DEBUG-DAG:   [[LOC_2_ADDR:%.+]] = alloca %ident_t
 // CHECK-DEBUG:       [[KMPC_LOC_VOIDPTR:%.+]] = bitcast %ident_t* [[LOC_2_ADDR]] to i8*
 // CHECK-DEBUG-NEXT:  [[KMPC_DEFAULT_LOC_VOIDPTR:%.+]] = bitcast %ident_t* [[DEF_LOC_2]] to i8*
-// CHECK-DEBUG-NEXT:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[KMPC_LOC_VOIDPTR]], i8* [[KMPC_DEFAULT_LOC_VOIDPTR]], i64 ptrtoint (%ident_t* getelementptr (%ident_t, %ident_t* null, i32 1) to i64), i32 8, i1 false)
+// CHECK-DEBUG-NEXT:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[KMPC_LOC_VOIDPTR]], i8* [[KMPC_DEFAULT_LOC_VOIDPTR]], i64 24, i32 8, i1 false)
 // CHECK-DEBUG:       [[ARGC_REF:%.+]] = getelementptr inbounds %struct.anon, %struct.anon* [[AGG_CAPTURED]], i32 0, i32 0
 // CHECK-DEBUG-NEXT:  store i32* {{%[a-z0-9.]+}}, i32** [[ARGC_REF]]
 // CHECK-DEBUG-NEXT:  [[KMPC_LOC_PSOURCE_REF:%.+]] = getelementptr inbounds %ident_t, %ident_t* [[LOC_2_ADDR]], i32 0, i32 4
@@ -108,7 +108,7 @@ int main (int argc, char **argv) {
 // CHECK-DEBUG-DAG:   [[LOC_2_ADDR:%.+]] = alloca %ident_t
 // CHECK-DEBUG:       [[KMPC_LOC_VOIDPTR:%.+]] = bitcast %ident_t* [[LOC_2_ADDR]] to i8*
 // CHECK-DEBUG-NEXT:  [[KMPC_DEFAULT_LOC_VOIDPTR:%.+]] = bitcast %ident_t* [[DEF_LOC_2]] to i8*
-// CHECK-DEBUG-NEXT:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[KMPC_LOC_VOIDPTR]], i8* [[KMPC_DEFAULT_LOC_VOIDPTR]], i64 ptrtoint (%ident_t* getelementptr (%ident_t, %ident_t* null, i32 1) to i64), i32 8, i1 false)
+// CHECK-DEBUG-NEXT:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[KMPC_LOC_VOIDPTR]], i8* [[KMPC_DEFAULT_LOC_VOIDPTR]], i64 24, i32 8, i1 false)
 // CHECK-DEBUG:       [[ARGC_REF:%.+]] = getelementptr inbounds %struct.anon.0, %struct.anon.0* [[AGG_CAPTURED]], i32 0, i32 0
 // CHECK-DEBUG-NEXT:  store i8*** {{%[a-z0-9.]+}}, i8**** [[ARGC_REF]]
 // CHECK-DEBUG-NEXT:  [[KMPC_LOC_PSOURCE_REF:%.+]] = getelementptr inbounds %ident_t, %ident_t* [[LOC_2_ADDR]], i32 0, i32 4
diff --git a/test/OpenMP/parallel_reduction_codegen.cpp b/test/OpenMP/parallel_reduction_codegen.cpp
index 04d19ebea8..d04676010c 100644
--- a/test/OpenMP/parallel_reduction_codegen.cpp
+++ b/test/OpenMP/parallel_reduction_codegen.cpp
@@ -74,7 +74,7 @@ int main() {
     // LAMBDA: store i{{[0-9]+}}* [[G_PRIVATE_ADDR]], i{{[0-9]+}}** [[G_PRIVATE_ADDR_REF]]
     // LAMBDA: call void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
 
-    // LAMBDA: [[G_PRIV_REF:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[RED_LIST]], i32 0, i32 0
+    // LAMBDA: [[G_PRIV_REF:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[RED_LIST]], i64 0, i64 0
     // LAMBDA: [[BITCAST:%.+]] = bitcast i32* [[G_PRIVATE_ADDR]] to i8*
     // LAMBDA: store i8* [[BITCAST]], i8** [[G_PRIV_REF]],
     // LAMBDA: call i32 @__kmpc_reduce_nowait(
@@ -136,7 +136,7 @@ int main() {
     // BLOCKS-NOT: [[G]]{{[[^:word:]]}}
     // BLOCKS: call void {{%.+}}(i8
 
-    // BLOCKS: [[G_PRIV_REF:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[RED_LIST]], i32 0, i32 0
+    // BLOCKS: [[G_PRIV_REF:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[RED_LIST]], i64 0, i64 0
     // BLOCKS: [[BITCAST:%.+]] = bitcast i32* [[G_PRIVATE_ADDR]] to i8*
     // BLOCKS: store i8* [[BITCAST]], i8** [[G_PRIV_REF]],
     // BLOCKS: call i32 @__kmpc_reduce_nowait(
@@ -226,16 +226,16 @@ int main() {
 
 // void *RedList[<n>] = {<ReductionVars>[0], ..., <ReductionVars>[<n>-1]};
 
-// CHECK: [[T_VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 0
+// CHECK: [[T_VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 0
 // CHECK: [[BITCAST:%.+]] = bitcast float* [[T_VAR_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[T_VAR_PRIV_REF]],
-// CHECK: [[VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 1
+// CHECK: [[VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 1
 // CHECK: [[BITCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[VAR_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[VAR_PRIV_REF]],
-// CHECK: [[VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 2
+// CHECK: [[VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 2
 // CHECK: [[BITCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[VAR1_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[VAR1_PRIV_REF]],
-// CHECK: [[T_VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 3
+// CHECK: [[T_VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 3
 // CHECK: [[BITCAST:%.+]] = bitcast float* [[T_VAR1_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[T_VAR1_PRIV_REF]],
 
@@ -373,38 +373,38 @@ int main() {
 // }
 // CHECK: define internal void [[REDUCTION_FUNC]](i8*, i8*)
 // t_var_lhs = (float*)lhs[0];
-// CHECK: [[T_VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS:%.+]], i32 0, i32 0
+// CHECK: [[T_VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS:%.+]], i64 0, i64 0
 // CHECK: [[T_VAR_RHS_VOID:%.+]] = load i8*, i8** [[T_VAR_RHS_REF]],
 // CHECK: [[T_VAR_RHS:%.+]] = bitcast i8* [[T_VAR_RHS_VOID]] to float*
 // t_var_rhs = (float*)rhs[0];
-// CHECK: [[T_VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS:%.+]], i32 0, i32 0
+// CHECK: [[T_VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS:%.+]], i64 0, i64 0
 // CHECK: [[T_VAR_LHS_VOID:%.+]] = load i8*, i8** [[T_VAR_LHS_REF]],
 // CHECK: [[T_VAR_LHS:%.+]] = bitcast i8* [[T_VAR_LHS_VOID]] to float*
 
 // var_lhs = (S<float>*)lhs[1];
-// CHECK: [[VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 1
+// CHECK: [[VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 1
 // CHECK: [[VAR_RHS_VOID:%.+]] = load i8*, i8** [[VAR_RHS_REF]],
 // CHECK: [[VAR_RHS:%.+]] = bitcast i8* [[VAR_RHS_VOID]] to [[S_FLOAT_TY]]*
 // var_rhs = (S<float>*)rhs[1];
-// CHECK: [[VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 1
+// CHECK: [[VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 1
 // CHECK: [[VAR_LHS_VOID:%.+]] = load i8*, i8** [[VAR_LHS_REF]],
 // CHECK: [[VAR_LHS:%.+]] = bitcast i8* [[VAR_LHS_VOID]] to [[S_FLOAT_TY]]*
 
 // var1_lhs = (S<float>*)lhs[2];
-// CHECK: [[VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 2
+// CHECK: [[VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 2
 // CHECK: [[VAR1_RHS_VOID:%.+]] = load i8*, i8** [[VAR1_RHS_REF]],
 // CHECK: [[VAR1_RHS:%.+]] = bitcast i8* [[VAR1_RHS_VOID]] to [[S_FLOAT_TY]]*
 // var1_rhs = (S<float>*)rhs[2];
-// CHECK: [[VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 2
+// CHECK: [[VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 2
 // CHECK: [[VAR1_LHS_VOID:%.+]] = load i8*, i8** [[VAR1_LHS_REF]],
 // CHECK: [[VAR1_LHS:%.+]] = bitcast i8* [[VAR1_LHS_VOID]] to [[S_FLOAT_TY]]*
 
 // t_var1_lhs = (float*)lhs[3];
-// CHECK: [[T_VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 3
+// CHECK: [[T_VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 3
 // CHECK: [[T_VAR1_RHS_VOID:%.+]] = load i8*, i8** [[T_VAR1_RHS_REF]],
 // CHECK: [[T_VAR1_RHS:%.+]] = bitcast i8* [[T_VAR1_RHS_VOID]] to float*
 // t_var1_rhs = (float*)rhs[3];
-// CHECK: [[T_VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 3
+// CHECK: [[T_VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 3
 // CHECK: [[T_VAR1_LHS_VOID:%.+]] = load i8*, i8** [[T_VAR1_LHS_REF]],
 // CHECK: [[T_VAR1_LHS:%.+]] = bitcast i8* [[T_VAR1_LHS_VOID]] to float*
 
@@ -488,16 +488,16 @@ int main() {
 
 // void *RedList[<n>] = {<ReductionVars>[0], ..., <ReductionVars>[<n>-1]};
 
-// CHECK: [[T_VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 0
+// CHECK: [[T_VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 0
 // CHECK: [[BITCAST:%.+]] = bitcast i{{[0-9]+}}* [[T_VAR_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[T_VAR_PRIV_REF]],
-// CHECK: [[VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 1
+// CHECK: [[VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 1
 // CHECK: [[BITCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[VAR_PRIV_REF]],
-// CHECK: [[VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 2
+// CHECK: [[VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 2
 // CHECK: [[BITCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR1_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[VAR1_PRIV_REF]],
-// CHECK: [[T_VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 3
+// CHECK: [[T_VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 3
 // CHECK: [[BITCAST:%.+]] = bitcast i{{[0-9]+}}* [[T_VAR1_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[T_VAR1_PRIV_REF]],
 
@@ -608,38 +608,38 @@ int main() {
 // }
 // CHECK: define internal void [[REDUCTION_FUNC]](i8*, i8*)
 // t_var_lhs = (i{{[0-9]+}}*)lhs[0];
-// CHECK: [[T_VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS:%.+]], i32 0, i32 0
+// CHECK: [[T_VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS:%.+]], i64 0, i64 0
 // CHECK: [[T_VAR_RHS_VOID:%.+]] = load i8*, i8** [[T_VAR_RHS_REF]],
 // CHECK: [[T_VAR_RHS:%.+]] = bitcast i8* [[T_VAR_RHS_VOID]] to i{{[0-9]+}}*
 // t_var_rhs = (i{{[0-9]+}}*)rhs[0];
-// CHECK: [[T_VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS:%.+]], i32 0, i32 0
+// CHECK: [[T_VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS:%.+]], i64 0, i64 0
 // CHECK: [[T_VAR_LHS_VOID:%.+]] = load i8*, i8** [[T_VAR_LHS_REF]],
 // CHECK: [[T_VAR_LHS:%.+]] = bitcast i8* [[T_VAR_LHS_VOID]] to i{{[0-9]+}}*
 
 // var_lhs = (S<i{{[0-9]+}}>*)lhs[1];
-// CHECK: [[VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 1
+// CHECK: [[VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 1
 // CHECK: [[VAR_RHS_VOID:%.+]] = load i8*, i8** [[VAR_RHS_REF]],
 // CHECK: [[VAR_RHS:%.+]] = bitcast i8* [[VAR_RHS_VOID]] to [[S_INT_TY]]*
 // var_rhs = (S<i{{[0-9]+}}>*)rhs[1];
-// CHECK: [[VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 1
+// CHECK: [[VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 1
 // CHECK: [[VAR_LHS_VOID:%.+]] = load i8*, i8** [[VAR_LHS_REF]],
 // CHECK: [[VAR_LHS:%.+]] = bitcast i8* [[VAR_LHS_VOID]] to [[S_INT_TY]]*
 
 // var1_lhs = (S<i{{[0-9]+}}>*)lhs[2];
-// CHECK: [[VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 2
+// CHECK: [[VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 2
 // CHECK: [[VAR1_RHS_VOID:%.+]] = load i8*, i8** [[VAR1_RHS_REF]],
 // CHECK: [[VAR1_RHS:%.+]] = bitcast i8* [[VAR1_RHS_VOID]] to [[S_INT_TY]]*
 // var1_rhs = (S<i{{[0-9]+}}>*)rhs[2];
-// CHECK: [[VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 2
+// CHECK: [[VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 2
 // CHECK: [[VAR1_LHS_VOID:%.+]] = load i8*, i8** [[VAR1_LHS_REF]],
 // CHECK: [[VAR1_LHS:%.+]] = bitcast i8* [[VAR1_LHS_VOID]] to [[S_INT_TY]]*
 
 // t_var1_lhs = (i{{[0-9]+}}*)lhs[3];
-// CHECK: [[T_VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 3
+// CHECK: [[T_VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 3
 // CHECK: [[T_VAR1_RHS_VOID:%.+]] = load i8*, i8** [[T_VAR1_RHS_REF]],
 // CHECK: [[T_VAR1_RHS:%.+]] = bitcast i8* [[T_VAR1_RHS_VOID]] to i{{[0-9]+}}*
 // t_var1_rhs = (i{{[0-9]+}}*)rhs[3];
-// CHECK: [[T_VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 3
+// CHECK: [[T_VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 3
 // CHECK: [[T_VAR1_LHS_VOID:%.+]] = load i8*, i8** [[T_VAR1_LHS_REF]],
 // CHECK: [[T_VAR1_LHS:%.+]] = bitcast i8* [[T_VAR1_LHS_VOID]] to i{{[0-9]+}}*
 
diff --git a/test/OpenMP/sections_reduction_codegen.cpp b/test/OpenMP/sections_reduction_codegen.cpp
index 4d404dbcc6..3c2525790b 100644
--- a/test/OpenMP/sections_reduction_codegen.cpp
+++ b/test/OpenMP/sections_reduction_codegen.cpp
@@ -74,7 +74,7 @@ int main() {
     // LAMBDA: call void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
     // LAMBDA: call void @__kmpc_for_static_fini(
 
-    // LAMBDA: [[G_PRIV_REF:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[RED_LIST]], i32 0, i32 0
+    // LAMBDA: [[G_PRIV_REF:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[RED_LIST]], i64 0, i64 0
     // LAMBDA: [[BITCAST:%.+]] = bitcast double* [[G_PRIVATE_ADDR]] to i8*
     // LAMBDA: store i8* [[BITCAST]], i8** [[G_PRIV_REF]],
     // LAMBDA: call i32 @__kmpc_reduce(
@@ -135,7 +135,7 @@ int main() {
     // BLOCKS: call void {{%.+}}(i8
     // BLOCKS: call void @__kmpc_for_static_fini(
 
-    // BLOCKS: [[G_PRIV_REF:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[RED_LIST]], i32 0, i32 0
+    // BLOCKS: [[G_PRIV_REF:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[RED_LIST]], i64 0, i64 0
     // BLOCKS: [[BITCAST:%.+]] = bitcast double* [[G_PRIVATE_ADDR]] to i8*
     // BLOCKS: store i8* [[BITCAST]], i8** [[G_PRIV_REF]],
     // BLOCKS: call i32 @__kmpc_reduce(
@@ -277,16 +277,16 @@ int main() {
 
 // void *RedList[<n>] = {<ReductionVars>[0], ..., <ReductionVars>[<n>-1]};
 
-// CHECK: [[T_VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 0
+// CHECK: [[T_VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 0
 // CHECK: [[BITCAST:%.+]] = bitcast i{{[0-9]+}}* [[T_VAR_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[T_VAR_PRIV_REF]],
-// CHECK: [[VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 1
+// CHECK: [[VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 1
 // CHECK: [[BITCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[VAR_PRIV_REF]],
-// CHECK: [[VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 2
+// CHECK: [[VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 2
 // CHECK: [[BITCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR1_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[VAR1_PRIV_REF]],
-// CHECK: [[T_VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 3
+// CHECK: [[T_VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 3
 // CHECK: [[BITCAST:%.+]] = bitcast i{{[0-9]+}}* [[T_VAR1_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[T_VAR1_PRIV_REF]],
 
@@ -395,38 +395,38 @@ int main() {
 // }
 // CHECK: define internal void [[REDUCTION_FUNC]](i8*, i8*)
 // t_var_lhs = (i{{[0-9]+}}*)lhs[0];
-// CHECK: [[T_VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS:%.+]], i32 0, i32 0
+// CHECK: [[T_VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS:%.+]], i64 0, i64 0
 // CHECK: [[T_VAR_RHS_VOID:%.+]] = load i8*, i8** [[T_VAR_RHS_REF]],
 // CHECK: [[T_VAR_RHS:%.+]] = bitcast i8* [[T_VAR_RHS_VOID]] to i{{[0-9]+}}*
 // t_var_rhs = (i{{[0-9]+}}*)rhs[0];
-// CHECK: [[T_VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS:%.+]], i32 0, i32 0
+// CHECK: [[T_VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS:%.+]], i64 0, i64 0
 // CHECK: [[T_VAR_LHS_VOID:%.+]] = load i8*, i8** [[T_VAR_LHS_REF]],
 // CHECK: [[T_VAR_LHS:%.+]] = bitcast i8* [[T_VAR_LHS_VOID]] to i{{[0-9]+}}*
 
 // var_lhs = (S<i{{[0-9]+}}>*)lhs[1];
-// CHECK: [[VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 1
+// CHECK: [[VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 1
 // CHECK: [[VAR_RHS_VOID:%.+]] = load i8*, i8** [[VAR_RHS_REF]],
 // CHECK: [[VAR_RHS:%.+]] = bitcast i8* [[VAR_RHS_VOID]] to [[S_INT_TY]]*
 // var_rhs = (S<i{{[0-9]+}}>*)rhs[1];
-// CHECK: [[VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 1
+// CHECK: [[VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 1
 // CHECK: [[VAR_LHS_VOID:%.+]] = load i8*, i8** [[VAR_LHS_REF]],
 // CHECK: [[VAR_LHS:%.+]] = bitcast i8* [[VAR_LHS_VOID]] to [[S_INT_TY]]*
 
 // var1_lhs = (S<i{{[0-9]+}}>*)lhs[2];
-// CHECK: [[VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 2
+// CHECK: [[VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 2
 // CHECK: [[VAR1_RHS_VOID:%.+]] = load i8*, i8** [[VAR1_RHS_REF]],
 // CHECK: [[VAR1_RHS:%.+]] = bitcast i8* [[VAR1_RHS_VOID]] to [[S_INT_TY]]*
 // var1_rhs = (S<i{{[0-9]+}}>*)rhs[2];
-// CHECK: [[VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 2
+// CHECK: [[VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 2
 // CHECK: [[VAR1_LHS_VOID:%.+]] = load i8*, i8** [[VAR1_LHS_REF]],
 // CHECK: [[VAR1_LHS:%.+]] = bitcast i8* [[VAR1_LHS_VOID]] to [[S_INT_TY]]*
 
 // t_var1_lhs = (i{{[0-9]+}}*)lhs[3];
-// CHECK: [[T_VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 3
+// CHECK: [[T_VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 3
 // CHECK: [[T_VAR1_RHS_VOID:%.+]] = load i8*, i8** [[T_VAR1_RHS_REF]],
 // CHECK: [[T_VAR1_RHS:%.+]] = bitcast i8* [[T_VAR1_RHS_VOID]] to i{{[0-9]+}}*
 // t_var1_rhs = (i{{[0-9]+}}*)rhs[3];
-// CHECK: [[T_VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 3
+// CHECK: [[T_VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 3
 // CHECK: [[T_VAR1_LHS_VOID:%.+]] = load i8*, i8** [[T_VAR1_LHS_REF]],
 // CHECK: [[T_VAR1_LHS:%.+]] = bitcast i8* [[T_VAR1_LHS_VOID]] to i{{[0-9]+}}*
 
diff --git a/test/OpenMP/task_codegen.cpp b/test/OpenMP/task_codegen.cpp
index 16a97d2ed7..433578d681 100644
--- a/test/OpenMP/task_codegen.cpp
+++ b/test/OpenMP/task_codegen.cpp
@@ -53,29 +53,29 @@ int main() {
 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[SHAREDS_REF]], i8* [[BITCAST]], i64 8, i32 8, i1 false)
 // CHECK: [[DESTRUCTORS_REF_PTR:%.+]] = getelementptr inbounds [[KMP_TASK_T]], [[KMP_TASK_T]]* [[TASK_PTR]], i32 0, i32 3
 // CHECK: store i32 (i32, i8*)* null, i32 (i32, i8*)** [[DESTRUCTORS_REF_PTR]]
-// CHECK: getelementptr inbounds [4 x [[KMP_DEPEND_INFO]]], [4 x [[KMP_DEPEND_INFO]]]* %{{[^,]+}}, i32 0, i32 0
-// CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 0
-// CHECK: store i64 ptrtoint (i32* @{{.+}} to i64), i64*
-// CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 1
-// CHECK: store i64 4, i64*
-// CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 2
-// CHECK: store i8 1, i8*
-// CHECK: getelementptr inbounds [4 x [[KMP_DEPEND_INFO]]], [4 x [[KMP_DEPEND_INFO]]]* %{{[^,]+}}, i32 0, i32 1
-// CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 0
+// CHECK: [[DEP:%.*]] = getelementptr inbounds [4 x [[KMP_DEPEND_INFO]]], [4 x [[KMP_DEPEND_INFO]]]* [[DEPENDENCIES:%.*]], i64 0, i64 0
+// CHECK: [[T0:%.*]] = getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* [[DEP]], i32 0, i32 0
+// CHECK: store i64 ptrtoint (i32* @{{.+}} to i64), i64* [[T0]]
+// CHECK: [[T0:%.*]] = getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* [[DEP]], i32 0, i32 1
+// CHECK: store i64 4, i64* [[T0]]
+// CHECK: [[T0:%.*]] = getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* [[DEP]], i32 0, i32 2
+// CHECK: store i8 1, i8* [[T0]]
+// CHECK: [[DEP:%.*]] = getelementptr inbounds [4 x [[KMP_DEPEND_INFO]]], [4 x [[KMP_DEPEND_INFO]]]* [[DEPENDENCIES]], i64 0, i64 1
+// CHECK: [[T0:%.*]] = getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* [[DEP]], i32 0, i32 0
 // CHECK: ptrtoint i8* [[B]] to i64
-// CHECK: store i64 %{{[^,]+}}, i64*
-// CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 1
-// CHECK: store i64 1, i64*
-// CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 2
-// CHECK: store i8 1, i8*
-// CHECK: getelementptr inbounds [4 x [[KMP_DEPEND_INFO]]], [4 x [[KMP_DEPEND_INFO]]]* %{{[^,]+}}, i32 0, i32 2
-// CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 0
+// CHECK: store i64 %{{[^,]+}}, i64* [[T0]]
+// CHECK: [[T0:%.*]] = getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* [[DEP]], i32 0, i32 1
+// CHECK: store i64 1, i64* [[T0]]
+// CHECK: [[T0:%.*]] = getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* [[DEP]], i32 0, i32 2
+// CHECK: store i8 1, i8* [[T0]]
+// CHECK: [[DEP:%.*]] = getelementptr inbounds [4 x [[KMP_DEPEND_INFO]]], [4 x [[KMP_DEPEND_INFO]]]* [[DEPENDENCIES]], i64 0, i64 2
+// CHECK: [[T0:%.*]] = getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* [[DEP]], i32 0, i32 0
 // CHECK: ptrtoint [2 x [[STRUCT_S]]]* [[S]] to i64
-// CHECK: store i64 %{{[^,]+}}, i64*
-// CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 1
-// CHECK: store i64 8, i64*
-// CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 2
-// CHECK: store i8 1, i8*
+// CHECK: store i64 %{{[^,]+}}, i64* [[T0]]
+// CHECK: [[T0:%.*]] = getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* [[DEP]], i32 0, i32 1
+// CHECK: store i64 8, i64* [[T0]]
+// CHECK: [[T0:%.*]] = getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* [[DEP]], i32 0, i32 2
+// CHECK: store i8 1, i8* [[T0]]
 // CHECK: [[IDX1:%.+]] = mul nsw i64 0, [[A_VAL:%.+]]
 // CHECK: [[START:%.+]] = getelementptr inbounds i32, i32* %{{.+}}, i64 [[IDX1]]
 // CHECK: [[IDX1:%.+]] = mul nsw i64 9, [[A_VAL]]
@@ -84,16 +84,16 @@ int main() {
 // CHECK: [[START_INT:%.+]] = ptrtoint i32* [[START]] to i64
 // CHECK: [[END_INT:%.+]] = ptrtoint i32* [[END1]] to i64
 // CHECK: [[SIZEOF:%.+]] = sub nuw i64 [[END_INT]], [[START_INT]]
-// CHECK: getelementptr inbounds [4 x [[KMP_DEPEND_INFO]]], [4 x [[KMP_DEPEND_INFO]]]* %{{[^,]+}}, i32 0, i32 3
-// CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 0
-// CHECK: ptrtoint i32* [[START]] to i64
-// CHECK: store i64 %{{[^,]+}}, i64*
-// CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 1
-// CHECK: store i64 [[SIZEOF]], i64*
-// CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 2
-// CHECK: store i8 1, i8*
-// CHECK: getelementptr inbounds [4 x [[KMP_DEPEND_INFO]]], [4 x [[KMP_DEPEND_INFO]]]* %{{[^,]+}}, i32 0, i32 0
-// CHECK: bitcast [[KMP_DEPEND_INFO]]* %{{.+}} to i8*
+// CHECK: [[DEP:%.*]] = getelementptr inbounds [4 x [[KMP_DEPEND_INFO]]], [4 x [[KMP_DEPEND_INFO]]]* %{{[^,]+}}, i64 0, i64 3
+// CHECK: [[T0:%.*]] = getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* [[DEP]], i32 0, i32 0
+// CHECK: [[T1:%.*]] = ptrtoint i32* [[START]] to i64
+// CHECK: store i64 [[T1]], i64* [[T0]]
+// CHECK: [[T0:%.*]] = getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 1
+// CHECK: store i64 [[SIZEOF]], i64* [[T0]]
+// CHECK: [[T0:%.*]] = getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 2
+// CHECK: store i8 1, i8* [[T0]]
+// CHECK: [[DEPS:%.*]] = getelementptr inbounds [4 x [[KMP_DEPEND_INFO]]], [4 x [[KMP_DEPEND_INFO]]]* [[DEPENDENCIES]], i32 0, i32 0
+// CHECK: bitcast [[KMP_DEPEND_INFO]]* [[DEPS]] to i8*
 // CHECK: call i32 @__kmpc_omp_task_with_deps([[IDENT_T]]* @{{.+}}, i32 [[GTID]], i8* [[ORIG_TASK_PTR]], i32 4, i8* %{{[^,]+}}, i32 0, i8* null)
 #pragma omp task shared(a, s) depend(in : a, b, s, arr[:])
   {
@@ -111,8 +111,8 @@ int main() {
 // CHECK: [[ORIG_TASK_PTR:%.+]] = call i8* @__kmpc_omp_task_alloc([[IDENT_T]]* @{{.+}}, i32 [[GTID]], i32 0, i64 32, i64 1,
 // CHECK: [[DESTRUCTORS_REF_PTR:%.+]] = getelementptr inbounds [[KMP_TASK_T]]{{.*}}* {{%.+}}, i32 0, i32 3
 // CHECK: store i32 (i32, i8*)* null, i32 (i32, i8*)** [[DESTRUCTORS_REF_PTR]]
-// CHECK: getelementptr inbounds [2 x [[STRUCT_S]]], [2 x [[STRUCT_S]]]* [[S]], i32 0, i64 0
-// CHECK: getelementptr inbounds [2 x [[KMP_DEPEND_INFO]]], [2 x [[KMP_DEPEND_INFO]]]* %{{[^,]+}}, i32 0, i32 0
+// CHECK: getelementptr inbounds [2 x [[STRUCT_S]]], [2 x [[STRUCT_S]]]* [[S]], i64 0, i64 0
+// CHECK: getelementptr inbounds [2 x [[KMP_DEPEND_INFO]]], [2 x [[KMP_DEPEND_INFO]]]* %{{[^,]+}}, i64 0, i64 0
 // CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 0
 // CHECK: ptrtoint [[STRUCT_S]]* %{{.+}} to i64
 // CHECK: store i64 %{{[^,]+}}, i64*
@@ -134,7 +134,7 @@ int main() {
 // CHECK: [[START_INT:%.+]] = ptrtoint i32* [[START1]] to i64
 // CHECK: [[END_INT:%.+]] = ptrtoint i32* [[END2]] to i64
 // CHECK: [[SIZEOF:%.+]] = sub nuw i64 [[END_INT]], [[START_INT]]
-// CHECK: getelementptr inbounds [2 x [[KMP_DEPEND_INFO]]], [2 x [[KMP_DEPEND_INFO]]]* %{{[^,]+}}, i32 0, i32 1
+// CHECK: getelementptr inbounds [2 x [[KMP_DEPEND_INFO]]], [2 x [[KMP_DEPEND_INFO]]]* %{{[^,]+}}, i64 0, i64 1
 // CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 0
 // CHECK: ptrtoint i32* [[START1]] to i64
 // CHECK: store i64 %{{[^,]+}}, i64*
@@ -152,15 +152,15 @@ int main() {
 // CHECK: [[ORIG_TASK_PTR:%.+]] = call i8* @__kmpc_omp_task_alloc([[IDENT_T]]* @{{.+}}, i32 [[GTID]], i32 3, i64 32, i64 1,
 // CHECK: [[DESTRUCTORS_REF_PTR:%.+]] = getelementptr inbounds [[KMP_TASK_T]]{{.*}}* {{%.+}}, i32 0, i32 3
 // CHECK: store i32 (i32, i8*)* null, i32 (i32, i8*)** [[DESTRUCTORS_REF_PTR]]
-// CHECK: getelementptr inbounds [3 x [[KMP_DEPEND_INFO]]], [3 x [[KMP_DEPEND_INFO]]]* %{{[^,]+}}, i32 0, i32 0
+// CHECK: getelementptr inbounds [3 x [[KMP_DEPEND_INFO]]], [3 x [[KMP_DEPEND_INFO]]]* %{{[^,]+}}, i64 0, i64 0
 // CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 0
 // CHECK: store i64 ptrtoint (i32* @{{.+}} to i64), i64*
 // CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 1
 // CHECK: store i64 4, i64*
 // CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 2
 // CHECK: store i8 3, i8*
-// CHECK: getelementptr inbounds [2 x [[STRUCT_S]]], [2 x [[STRUCT_S]]]* [[S]], i32 0, i64 1
-// CHECK: getelementptr inbounds [3 x [[KMP_DEPEND_INFO]]], [3 x [[KMP_DEPEND_INFO]]]* %{{[^,]+}}, i32 0, i32 1
+// CHECK: getelementptr inbounds [2 x [[STRUCT_S]]], [2 x [[STRUCT_S]]]* [[S]], i64 0, i64 1
+// CHECK: getelementptr inbounds [3 x [[KMP_DEPEND_INFO]]], [3 x [[KMP_DEPEND_INFO]]]* %{{[^,]+}}, i64 0, i64 1
 // CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 0
 // CHECK: ptrtoint [[STRUCT_S]]* %{{.+}} to i64
 // CHECK: store i64 %{{[^,]+}}, i64*
@@ -184,7 +184,7 @@ int main() {
 // CHECK: [[START_INT:%.+]] = ptrtoint i32* [[START1]] to i64
 // CHECK: [[END_INT:%.+]] = ptrtoint i32* [[END2]] to i64
 // CHECK: [[SIZEOF:%.+]] = sub nuw i64 [[END_INT]], [[START_INT]]
-// CHECK: getelementptr inbounds [3 x [[KMP_DEPEND_INFO]]], [3 x [[KMP_DEPEND_INFO]]]* %{{[^,]+}}, i32 0, i32 2
+// CHECK: getelementptr inbounds [3 x [[KMP_DEPEND_INFO]]], [3 x [[KMP_DEPEND_INFO]]]* %{{[^,]+}}, i64 0, i64 2
 // CHECK: getelementptr inbounds [[KMP_DEPEND_INFO]], [[KMP_DEPEND_INFO]]* %{{[^,]+}}, i32 0, i32 0
 // CHECK: ptrtoint i32* [[START1]] to i64
 // CHECK: store i64 %{{[^,]+}}, i64*