diff --git a/builtins/target-avx-x2.ll b/builtins/target-avx-x2.ll
index 13c4335d..69b9d381 100644
--- a/builtins/target-avx-x2.ll
+++ b/builtins/target-avx-x2.ll
@@ -158,13 +158,13 @@ declare <8 x float> @llvm.x86.avx.max.ps.256(<8 x float>, <8 x float>) nounwind
 declare <8 x float> @llvm.x86.avx.min.ps.256(<8 x float>, <8 x float>) nounwind readnone
 
 define <16 x float> @__max_varying_float(<16 x float>,
-                                                  <16 x float>) nounwind readonly alwaysinline {
+                                         <16 x float>) nounwind readonly alwaysinline {
   binary8to16(call, float, @llvm.x86.avx.max.ps.256, %0, %1)
   ret <16 x float> %call
 }
 
 define <16 x float> @__min_varying_float(<16 x float>,
-                                                  <16 x float>) nounwind readonly alwaysinline {
+                                         <16 x float>) nounwind readonly alwaysinline {
   binary8to16(call, float, @llvm.x86.avx.min.ps.256, %0, %1)
   ret <16 x float> %call
 }
diff --git a/builtins/target-avx.ll b/builtins/target-avx.ll
index 608d2dcd..bc8c64a7 100644
--- a/builtins/target-avx.ll
+++ b/builtins/target-avx.ll
@@ -158,13 +158,13 @@ declare <8 x float> @llvm.x86.avx.max.ps.256(<8 x float>, <8 x float>) nounwind
 declare <8 x float> @llvm.x86.avx.min.ps.256(<8 x float>, <8 x float>) nounwind readnone
 
 define <8 x float> @__max_varying_float(<8 x float>,
-                                                 <8 x float>) nounwind readonly alwaysinline {
+                                        <8 x float>) nounwind readonly alwaysinline {
   %call = call <8 x float> @llvm.x86.avx.max.ps.256(<8 x float> %0, <8 x float> %1)
   ret <8 x float> %call
 }
 
 define <8 x float> @__min_varying_float(<8 x float>,
-                                                 <8 x float>) nounwind readonly alwaysinline {
+                                        <8 x float>) nounwind readonly alwaysinline {
   %call = call <8 x float> @llvm.x86.avx.min.ps.256(<8 x float> %0, <8 x float> %1)
   ret <8 x float> %call
 }
diff --git a/builtins/target-generic-1.ll b/builtins/target-generic-1.ll
index 5ced9da9..39870c75 100755
--- a/builtins/target-generic-1.ll
+++ b/builtins/target-generic-1.ll
@@ -48,7 +48,7 @@ gen_scatter(1, i64)
 
 
 define  <1 x i8> @__vselect_i8(<1 x i8>, <1 x i8> ,
-                                         <1 x i32> %mask) nounwind readnone alwaysinline {
+                               <1 x i32> %mask) nounwind readnone alwaysinline {
 ;  %mv = trunc <1 x i32> %mask to <1 x i8>
 ;  %notmask = xor <1 x i8> %mv, <i8 -1>
 ;  %cleared_old = and <1 x i8> %0, %notmask
@@ -69,7 +69,7 @@ define  <1 x i8> @__vselect_i8(<1 x i8>, <1 x i8> ,
 }
 
 define  <1 x i16> @__vselect_i16(<1 x i16>, <1 x i16> ,
-                                         <1 x i32> %mask) nounwind readnone alwaysinline {
+                                 <1 x i32> %mask) nounwind readnone alwaysinline {
 ;  %mv = trunc <1 x i32> %mask to <1 x i16>
 ;  %notmask = xor <1 x i16> %mv, <i16 -1>
 ;  %cleared_old = and <1 x i16> %0, %notmask
@@ -91,7 +91,7 @@ define  <1 x i16> @__vselect_i16(<1 x i16>, <1 x i16> ,
 
 
 define  <1 x i32> @__vselect_i32(<1 x i32>, <1 x i32> ,
-                                         <1 x i32> %mask) nounwind readnone alwaysinline {
+                                 <1 x i32> %mask) nounwind readnone alwaysinline {
 ;  %notmask = xor <1 x i32> %mask, <i32 -1>
 ;  %cleared_old = and <1 x i32> %0, %notmask
 ;  %masked_new = and <1 x i32> %1, %mask
@@ -109,8 +109,9 @@ define  <1 x i32> @__vselect_i32(<1 x i32>, <1 x i32> ,
    ret <1 x i32> %r
 
 }
+
 define  <1 x i64> @__vselect_i64(<1 x i64>, <1 x i64> ,
-                                         <1 x i32> %mask) nounwind readnone alwaysinline {
+                                 <1 x i32> %mask) nounwind readnone alwaysinline {
 ;  %newmask = zext <1 x i32> %mask to <1 x i64>
 ;  %notmask = xor <1 x i64> %newmask, <i64 -1>
 ;  %cleared_old = and <1 x i64> %0, %notmask
@@ -131,7 +132,7 @@ define  <1 x i64> @__vselect_i64(<1 x i64>, <1 x i64> ,
 }
 
 define  <1 x float> @__vselect_float(<1 x float>, <1 x float>,
-                                             <1 x i32> %mask) nounwind readnone alwaysinline {
+                                     <1 x i32> %mask) nounwind readnone alwaysinline {
 ;  %v0 = bitcast <1 x float> %0 to <1 x i32>
 ;  %v1 = bitcast <1 x float> %1 to <1 x i32>
 ;  %r = call <1 x i32> @__vselect_i32(<1 x i32> %v0, <1 x i32> %v1, <1 x i32> %mask)
diff --git a/builtins/target-sse2-x2.ll b/builtins/target-sse2-x2.ll
index 65d30939..56145a0f 100644
--- a/builtins/target-sse2-x2.ll
+++ b/builtins/target-sse2-x2.ll
@@ -398,7 +398,7 @@ define double @__reduce_max_double(<8 x double>) nounwind readnone {
 }
 
 define <4 x i64> @__add_varying_int64(<4 x i64>,
-                                               <4 x i64>) nounwind readnone alwaysinline {
+                                      <4 x i64>) nounwind readnone alwaysinline {
   %r = add <4 x i64> %0, %1
   ret <4 x i64> %r
 }
diff --git a/builtins/target-sse4-x2.ll b/builtins/target-sse4-x2.ll
index 1ac6b3e5..99e66e36 100644
--- a/builtins/target-sse4-x2.ll
+++ b/builtins/target-sse4-x2.ll
@@ -221,13 +221,13 @@ define <8 x i32> @__max_varying_int32(<8 x i32>, <8 x i32>) nounwind readonly al
 ; unsigned int min/max
 
 define <8 x i32> @__min_varying_uint32(<8 x i32>,
-                                                <8 x i32>) nounwind readonly alwaysinline {
+                                       <8 x i32>) nounwind readonly alwaysinline {
   binary4to8(call, i32, @llvm.x86.sse41.pminud, %0, %1)
   ret <8 x i32> %call
 }
 
 define <8 x i32> @__max_varying_uint32(<8 x i32>,
-                                                <8 x i32>) nounwind readonly alwaysinline {
+                                       <8 x i32>) nounwind readonly alwaysinline {
   binary4to8(call, i32, @llvm.x86.sse41.pmaxud, %0, %1)
   ret <8 x i32> %call
 }
@@ -325,7 +325,7 @@ define double @__reduce_max_double(<8 x double>) nounwind readnone {
 }
 
 define <4 x i64> @__add_varying_int64(<4 x i64>,
-                                               <4 x i64>) nounwind readnone alwaysinline {
+                                      <4 x i64>) nounwind readnone alwaysinline {
   %r = add <4 x i64> %0, %1
   ret <4 x i64> %r
 }
diff --git a/builtins/util.m4 b/builtins/util.m4
index 59185942..c53851e5 100644
--- a/builtins/util.m4
+++ b/builtins/util.m4
@@ -1934,7 +1934,7 @@ define i8 @__extract_int8(<WIDTH x i8>, i32) nounwind readnone alwaysinline {
 }
 
 define <WIDTH x i8> @__insert_int8(<WIDTH x i8>, i32, 
-                                           i8) nounwind readnone alwaysinline {
+                                   i8) nounwind readnone alwaysinline {
   %insert = insertelement <WIDTH x i8> %0, i8 %2, i32 %1
   ret <WIDTH x i8> %insert
 }
@@ -1945,7 +1945,7 @@ define i16 @__extract_int16(<WIDTH x i16>, i32) nounwind readnone alwaysinline {
 }
 
 define <WIDTH x i16> @__insert_int16(<WIDTH x i16>, i32, 
-                                           i16) nounwind readnone alwaysinline {
+                                     i16) nounwind readnone alwaysinline {
   %insert = insertelement <WIDTH x i16> %0, i16 %2, i32 %1
   ret <WIDTH x i16> %insert
 }
@@ -1956,7 +1956,7 @@ define i32 @__extract_int32(<WIDTH x i32>, i32) nounwind readnone alwaysinline {
 }
 
 define <WIDTH x i32> @__insert_int32(<WIDTH x i32>, i32, 
-                                           i32) nounwind readnone alwaysinline {
+                                     i32) nounwind readnone alwaysinline {
   %insert = insertelement <WIDTH x i32> %0, i32 %2, i32 %1
   ret <WIDTH x i32> %insert
 }
@@ -1967,7 +1967,7 @@ define i64 @__extract_int64(<WIDTH x i64>, i32) nounwind readnone alwaysinline {
 }
 
 define <WIDTH x i64> @__insert_int64(<WIDTH x i64>, i32, 
-                                           i64) nounwind readnone alwaysinline {
+                                     i64) nounwind readnone alwaysinline {
   %insert = insertelement <WIDTH x i64> %0, i64 %2, i32 %1
   ret <WIDTH x i64> %insert
 }