vendor/llvm/llvm-trunk-r238337

369 files changed, 8422 insertions, 5652 deletions

diff --git a/test/CodeGen/Mips/2008-07-03-SRet.ll b/test/CodeGen/Mips/2008-07-03-SRet.ll
index afec7f65d607..6313ec4af356 100644
--- a/test/CodeGen/Mips/2008-07-03-SRet.ll
+++ b/test/CodeGen/Mips/2008-07-03-SRet.ll
@@ -7,11 +7,11 @@ entry:
 ; CHECK: sw ${{[0-9]+}}, {{[0-9]+}}($4)
 ; CHECK: sw ${{[0-9]+}}, {{[0-9]+}}($4)
 ; CHECK: sw ${{[0-9]+}}, {{[0-9]+}}($4)
-  getelementptr %struct.sret0* %agg.result, i32 0, i32 0    ; <i32*>:0 [#uses=1]
+  getelementptr %struct.sret0, %struct.sret0* %agg.result, i32 0, i32 0    ; <i32*>:0 [#uses=1]
   store i32 %dummy, i32* %0, align 4
-  getelementptr %struct.sret0* %agg.result, i32 0, i32 1    ; <i32*>:1 [#uses=1]
+  getelementptr %struct.sret0, %struct.sret0* %agg.result, i32 0, i32 1    ; <i32*>:1 [#uses=1]
   store i32 %dummy, i32* %1, align 4
-  getelementptr %struct.sret0* %agg.result, i32 0, i32 2    ; <i32*>:2 [#uses=1]
+  getelementptr %struct.sret0, %struct.sret0* %agg.result, i32 0, i32 2    ; <i32*>:2 [#uses=1]
   store i32 %dummy, i32* %2, align 4
   ret void
 }
diff --git a/test/CodeGen/Mips/2008-07-15-InternalConstant.ll b/test/CodeGen/Mips/2008-07-15-InternalConstant.ll
index 29a7b5c3761a..d7e8f5c2d03f 100644
--- a/test/CodeGen/Mips/2008-07-15-InternalConstant.ll
+++ b/test/CodeGen/Mips/2008-07-15-InternalConstant.ll
@@ -8,7 +8,7 @@ entry:
 ; CHECK: foo
 ; CHECK: %hi(.str)
 ; CHECK: %lo(.str)
-	ret i8* getelementptr ([10 x i8]* @.str, i32 0, i32 0)
+	ret i8* getelementptr ([10 x i8], [10 x i8]* @.str, i32 0, i32 0)
 }
 
 define i32* @bar() nounwind  {
@@ -16,7 +16,7 @@ entry:
 ; CHECK: bar
 ; CHECK: %hi(i0)
 ; CHECK: %lo(i0)
-  ret i32* getelementptr ([5 x i32]* @i0, i32 0, i32 0)
+  ret i32* getelementptr ([5 x i32], [5 x i32]* @i0, i32 0, i32 0)
 }
 
 ; CHECK: rodata.str1.4,"aMS",@progbits
diff --git a/test/CodeGen/Mips/2008-07-15-SmallSection.ll b/test/CodeGen/Mips/2008-07-15-SmallSection.ll
index cbc3ecf5edc8..08d99d899103 100644
--- a/test/CodeGen/Mips/2008-07-15-SmallSection.ll
+++ b/test/CodeGen/Mips/2008-07-15-SmallSection.ll
@@ -22,13 +22,13 @@ target triple = "mipsallegrexel-unknown-psp-elf"
 
 define i8* @A0() nounwind {
 entry:
-	ret i8* getelementptr ([8 x i8]* @s0, i32 0, i32 0)
+	ret i8* getelementptr ([8 x i8], [8 x i8]* @s0, i32 0, i32 0)
 }
 
 define i32 @A1() nounwind {
 entry:
-  load i32* getelementptr (%struct.anon* @foo, i32 0, i32 0), align 8 
-  load i32* getelementptr (%struct.anon* @foo, i32 0, i32 1), align 4 
+  load i32, i32* getelementptr (%struct.anon, %struct.anon* @foo, i32 0, i32 0), align 8 
+  load i32, i32* getelementptr (%struct.anon, %struct.anon* @foo, i32 0, i32 1), align 4 
   add i32 %1, %0
   ret i32 %2
 }
diff --git a/test/CodeGen/Mips/2008-08-01-AsmInline.ll b/test/CodeGen/Mips/2008-08-01-AsmInline.ll
index 3c1bb39b4340..5edba029502a 100644
--- a/test/CodeGen/Mips/2008-08-01-AsmInline.ll
+++ b/test/CodeGen/Mips/2008-08-01-AsmInline.ll
@@ -1,5 +1,5 @@
 ; RUN: llc -march=mips -mcpu=mips32 < %s | FileCheck %s
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=n64 < %s | FileCheck %s
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi=n64 < %s | FileCheck %s
 
 %struct.DWstruct = type { i32, i32 }
 
@@ -26,8 +26,8 @@ entry:
 define void @foo0() nounwind {
 entry:
 ; CHECK: addu
-  %0 = load i32* @gi1, align 4
-  %1 = load i32* @gi0, align 4
+  %0 = load i32, i32* @gi1, align 4
+  %1 = load i32, i32* @gi0, align 4
   %2 = tail call i32 asm "addu $0, $1, $2", "=r,r,r"(i32 %0, i32 %1) nounwind
   store i32 %2, i32* @gi2, align 4
   ret void
@@ -36,7 +36,7 @@ entry:
 define void @foo2() nounwind {
 entry:
 ; CHECK: neg.s
-  %0 = load float* @gf1, align 4
+  %0 = load float, float* @gf1, align 4
   %1 = tail call float asm "neg.s $0, $1", "=f,f"(float %0) nounwind
   store float %1, float* @gf0, align 4
   ret void
@@ -45,7 +45,7 @@ entry:
 define void @foo3() nounwind {
 entry:
 ; CHECK: neg.d
-  %0 = load double* @gd1, align 8
+  %0 = load double, double* @gd1, align 8
   %1 = tail call double asm "neg.d $0, $1", "=f,f"(double %0) nounwind
   store double %1, double* @gd0, align 8
   ret void
@@ -64,7 +64,7 @@ define void @foo4() {
 entry:
   %0 = tail call i32 asm sideeffect "ulh $0,16($$sp)\0A\09", "=r,~{$2}"()
   store i32 %0, i32* @gi2, align 4
-  %1 = load float* @gf0, align 4
+  %1 = load float, float* @gf0, align 4
   %2 = tail call double asm sideeffect "cvt.d.s $0, $1\0A\09", "=f,f,~{$f0}"(float %1)
   store double %2, double* @gd0, align 8
   ret void
diff --git a/test/CodeGen/Mips/2008-08-03-ReturnDouble.ll b/test/CodeGen/Mips/2008-08-03-ReturnDouble.ll
index c41d5213c178..592e574a3622 100644
--- a/test/CodeGen/Mips/2008-08-03-ReturnDouble.ll
+++ b/test/CodeGen/Mips/2008-08-03-ReturnDouble.ll
@@ -7,12 +7,12 @@ entry:
         %retval = alloca double         ; <double*> [#uses=3]
         store double 0.000000e+00, double* %retval
         %r = alloca double              ; <double*> [#uses=1]
-        load double* %r         ; <double>:0 [#uses=1]
+        load double, double* %r         ; <double>:0 [#uses=1]
         store double %0, double* %retval
         br label %return
 
 return:         ; preds = %entry
-        load double* %retval            ; <double>:1 [#uses=1]
+        load double, double* %retval            ; <double>:1 [#uses=1]
         ret double %1
 }
 
diff --git a/test/CodeGen/Mips/2008-10-13-LegalizerBug.ll b/test/CodeGen/Mips/2008-10-13-LegalizerBug.ll
index 18f5b3d7ff7b..eaf6ddc911e3 100644
--- a/test/CodeGen/Mips/2008-10-13-LegalizerBug.ll
+++ b/test/CodeGen/Mips/2008-10-13-LegalizerBug.ll
@@ -7,8 +7,8 @@ entry:
 
 continue.outer:         ; preds = %case4, %entry
         %p.0.ph.rec = phi i32 [ 0, %entry ], [ %indvar.next, %case4 ]          ; <i32> [#uses=2]
-        %p.0.ph = getelementptr i8* %0, i32 %p.0.ph.rec         ; <i8*> [#uses=1]
-        %1 = load i8* %p.0.ph           ; <i8> [#uses=1]
+        %p.0.ph = getelementptr i8, i8* %0, i32 %p.0.ph.rec         ; <i8*> [#uses=1]
+        %1 = load i8, i8* %p.0.ph           ; <i8> [#uses=1]
         switch i8 %1, label %infloop [
                 i8 0, label %return.split
                 i8 76, label %case4
diff --git a/test/CodeGen/Mips/2008-11-10-xint_to_fp.ll b/test/CodeGen/Mips/2008-11-10-xint_to_fp.ll
index 9c4838a87e51..789f7ee3d4b4 100644
--- a/test/CodeGen/Mips/2008-11-10-xint_to_fp.ll
+++ b/test/CodeGen/Mips/2008-11-10-xint_to_fp.ll
@@ -1,4 +1,4 @@
-; RUN: llc -march=mips -soft-float < %s
+; RUN: llc -march=mips -mattr=+soft-float < %s
 ; PR2667
 target datalayout = "e-p:32:32:32-i1:8:8-i8:8:32-i16:16:32-i32:32:32-i64:32:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64"
 	%struct._Bigint = type { %struct._Bigint*, i32, i32, i32, i32, [1 x i32] }
@@ -13,16 +13,16 @@ target datalayout = "e-p:32:32:32-i1:8:8-i8:8:32-i16:16:32-i32:32:32-i64:32:64-f
 define double @_erand48_r(%struct._reent* %r, i16* %xseed) nounwind {
 entry:
 	tail call void @__dorand48( %struct._reent* %r, i16* %xseed ) nounwind
-	load i16* %xseed, align 2		; <i16>:0 [#uses=1]
+	load i16, i16* %xseed, align 2		; <i16>:0 [#uses=1]
 	uitofp i16 %0 to double		; <double>:1 [#uses=1]
 	tail call double @ldexp( double %1, i32 -48 ) nounwind		; <double>:2 [#uses=1]
-	getelementptr i16* %xseed, i32 1		; <i16*>:3 [#uses=1]
-	load i16* %3, align 2		; <i16>:4 [#uses=1]
+	getelementptr i16, i16* %xseed, i32 1		; <i16*>:3 [#uses=1]
+	load i16, i16* %3, align 2		; <i16>:4 [#uses=1]
 	uitofp i16 %4 to double		; <double>:5 [#uses=1]
 	tail call double @ldexp( double %5, i32 -32 ) nounwind		; <double>:6 [#uses=1]
 	fadd double %2, %6		; <double>:7 [#uses=1]
-	getelementptr i16* %xseed, i32 2		; <i16*>:8 [#uses=1]
-	load i16* %8, align 2		; <i16>:9 [#uses=1]
+	getelementptr i16, i16* %xseed, i32 2		; <i16*>:8 [#uses=1]
+	load i16, i16* %8, align 2		; <i16>:9 [#uses=1]
 	uitofp i16 %9 to double		; <double>:10 [#uses=1]
 	tail call double @ldexp( double %10, i32 -16 ) nounwind		; <double>:11 [#uses=1]
 	fadd double %7, %11		; <double>:12 [#uses=1]
@@ -35,18 +35,18 @@ declare double @ldexp(double, i32)
 
 define double @erand48(i16* %xseed) nounwind {
 entry:
-	load %struct._reent** @_impure_ptr, align 4		; <%struct._reent*>:0 [#uses=1]
+	load %struct._reent*, %struct._reent** @_impure_ptr, align 4		; <%struct._reent*>:0 [#uses=1]
 	tail call void @__dorand48( %struct._reent* %0, i16* %xseed ) nounwind
-	load i16* %xseed, align 2		; <i16>:1 [#uses=1]
+	load i16, i16* %xseed, align 2		; <i16>:1 [#uses=1]
 	uitofp i16 %1 to double		; <double>:2 [#uses=1]
 	tail call double @ldexp( double %2, i32 -48 ) nounwind		; <double>:3 [#uses=1]
-	getelementptr i16* %xseed, i32 1		; <i16*>:4 [#uses=1]
-	load i16* %4, align 2		; <i16>:5 [#uses=1]
+	getelementptr i16, i16* %xseed, i32 1		; <i16*>:4 [#uses=1]
+	load i16, i16* %4, align 2		; <i16>:5 [#uses=1]
 	uitofp i16 %5 to double		; <double>:6 [#uses=1]
 	tail call double @ldexp( double %6, i32 -32 ) nounwind		; <double>:7 [#uses=1]
 	fadd double %3, %7		; <double>:8 [#uses=1]
-	getelementptr i16* %xseed, i32 2		; <i16*>:9 [#uses=1]
-	load i16* %9, align 2		; <i16>:10 [#uses=1]
+	getelementptr i16, i16* %xseed, i32 2		; <i16*>:9 [#uses=1]
+	load i16, i16* %9, align 2		; <i16>:10 [#uses=1]
 	uitofp i16 %10 to double		; <double>:11 [#uses=1]
 	tail call double @ldexp( double %11, i32 -16 ) nounwind		; <double>:12 [#uses=1]
 	fadd double %8, %12		; <double>:13 [#uses=1]
diff --git a/test/CodeGen/Mips/2009-11-16-CstPoolLoad.ll b/test/CodeGen/Mips/2009-11-16-CstPoolLoad.ll
index c3791dfc7ce6..f736ddd0def6 100644
--- a/test/CodeGen/Mips/2009-11-16-CstPoolLoad.ll
+++ b/test/CodeGen/Mips/2009-11-16-CstPoolLoad.ll
@@ -1,9 +1,9 @@
 ; RUN: llc -march=mipsel -relocation-model=pic < %s | FileCheck %s -check-prefix=PIC-O32
 ; RUN: llc -march=mipsel -relocation-model=static < %s | FileCheck %s -check-prefix=STATIC-O32
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=-n64,n32 -relocation-model=pic < %s | FileCheck %s -check-prefix=PIC-N32
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=-n64,n32 -relocation-model=static < %s | FileCheck %s -check-prefix=STATIC-N32
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=-n64,n64 -relocation-model=pic < %s | FileCheck %s -check-prefix=PIC-N64
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=-n64,n64 -relocation-model=static < %s | FileCheck %s -check-prefix=STATIC-N64
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi n32 -relocation-model=pic < %s | FileCheck %s -check-prefix=PIC-N32
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi n32 -relocation-model=static < %s | FileCheck %s -check-prefix=STATIC-N32
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi n64 -relocation-model=pic < %s | FileCheck %s -check-prefix=PIC-N64
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi n64 -relocation-model=static < %s | FileCheck %s -check-prefix=STATIC-N64
 
 define float @h() nounwind readnone {
 entry:
diff --git a/test/CodeGen/Mips/2010-07-20-Switch.ll b/test/CodeGen/Mips/2010-07-20-Switch.ll
index 5c840775cf9e..fd0254e9f5ec 100644
--- a/test/CodeGen/Mips/2010-07-20-Switch.ll
+++ b/test/CodeGen/Mips/2010-07-20-Switch.ll
@@ -15,7 +15,7 @@ define i32 @main() nounwind readnone {
 entry:
   %x = alloca i32, align 4                        ; <i32*> [#uses=2]
   store volatile i32 2, i32* %x, align 4
-  %0 = load volatile i32* %x, align 4             ; <i32> [#uses=1]
+  %0 = load volatile i32, i32* %x, align 4             ; <i32> [#uses=1]
 ; STATIC-O32: sll $[[R0:[0-9]+]], ${{[0-9]+}}, 2
 ; STATIC-O32: lui $[[R1:[0-9]+]], %hi($JTI0_0)
 ; STATIC-O32: addu $[[R2:[0-9]+]], $[[R0]], $[[R1]]
diff --git a/test/CodeGen/Mips/2012-12-12-ExpandMemcpy.ll b/test/CodeGen/Mips/2012-12-12-ExpandMemcpy.ll
index 9d4daee696db..24bcfaee8bad 100644
--- a/test/CodeGen/Mips/2012-12-12-ExpandMemcpy.ll
+++ b/test/CodeGen/Mips/2012-12-12-ExpandMemcpy.ll
@@ -4,7 +4,7 @@
 
 define void @t(i8* %ptr) {
 entry:
-  tail call void @llvm.memcpy.p0i8.p0i8.i64(i8* %ptr, i8* getelementptr inbounds ([7 x i8]* @.str, i64 0, i64 0), i64 7, i32 1, i1 false)
+  tail call void @llvm.memcpy.p0i8.p0i8.i64(i8* %ptr, i8* getelementptr inbounds ([7 x i8], [7 x i8]* @.str, i64 0, i64 0), i64 7, i32 1, i1 false)
   ret void
 }
 
diff --git a/test/CodeGen/Mips/Fast-ISel/br1.ll b/test/CodeGen/Mips/Fast-ISel/br1.ll
index 579a77f88fef..11842ddc4188 100644
--- a/test/CodeGen/Mips/Fast-ISel/br1.ll
+++ b/test/CodeGen/Mips/Fast-ISel/br1.ll
@@ -1,6 +1,6 @@
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
 ; RUN:     < %s | FileCheck %s
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32 \
 ; RUN:     < %s | FileCheck %s
 
 @b = global i32 1, align 4
@@ -10,7 +10,7 @@
 ; Function Attrs: nounwind
 define void @br() #0 {
 entry:
-  %0 = load i32* @b, align 4
+  %0 = load i32, i32* @b, align 4
   %tobool = icmp eq i32 %0, 0
   br i1 %tobool, label %if.end, label %if.then
 
diff --git a/test/CodeGen/Mips/Fast-ISel/callabi.ll b/test/CodeGen/Mips/Fast-ISel/callabi.ll
index e76d7a74bd0e..8f5d68b41f66 100644
--- a/test/CodeGen/Mips/Fast-ISel/callabi.ll
+++ b/test/CodeGen/Mips/Fast-ISel/callabi.ll
@@ -1,477 +1,538 @@
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
-; RUN:     < %s | FileCheck %s
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
-; RUN:     < %s | FileCheck %s
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
-; RUN:     < %s | FileCheck %s -check-prefix=mips32r2
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
-; RUN:     < %s | FileCheck %s -check-prefix=mips32
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
-; RUN:     < %s | FileCheck %s -check-prefix=CHECK2
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
-; RUN:     < %s | FileCheck %s -check-prefix=CHECK2
+; RUN: llc -march=mipsel -mcpu=mips32 -O0 \
+; RUN:     -mips-fast-isel -relocation-model=pic -fast-isel-abort=1 < %s | \
+; RUN:     FileCheck %s -check-prefix=ALL -check-prefix=32R1
+; RUN: llc -march=mipsel -mcpu=mips32r2 -O0 \
+; RUN:     -mips-fast-isel -relocation-model=pic -fast-isel-abort=1 < %s | \
+; RUN:     FileCheck %s -check-prefix=ALL -check-prefix=32R2
 
+declare void @xb(i8)
 
-@c1 = global i8 -45, align 1
-@uc1 = global i8 27, align 1
-@s1 = global i16 -1789, align 2
-@us1 = global i16 1256, align 2
+define void @cxb() {
+  ; ALL-LABEL:    cxb:
+
+  ; ALL:            addiu   $[[T0:[0-9]+]], $zero, 10
+
+  ; 32R1:           sll     $[[T1:[0-9]+]], $[[T0]], 24
+  ; 32R1:           sra     $4, $[[T1]], 24
+
+  ; 32R2:           seb     $4, $[[T0]]
+  call void @xb(i8 10)
+  ret void
+}
+
+declare void @xh(i16)
+
+define void @cxh() {
+  ; ALL-LABEL:    cxh:
+
+  ; ALL:            addiu   $[[T0:[0-9]+]], $zero, 10
 
-; Function Attrs: nounwind
-define void @cxi() #0 {
-entry:
-; CHECK-LABEL:  cxi
+  ; 32R1:           sll     $[[T1:[0-9]+]], $[[T0]], 16
+  ; 32R1:           sra     $4, $[[T1]], 16
+
+  ; 32R2:           seh     $4, $[[T0]]
+  call void @xh(i16 10)
+  ret void
+}
+
+declare void @xi(i32)
+
+define void @cxi() {
+  ; ALL-LABEL:    cxi:
+
+  ; ALL-DAG:        addiu   $4, $zero, 10
+  ; ALL-DAG:        lw      $25, %got(xi)(${{[0-9]+}})
+  ; ALL:            jalr    $25
   call void @xi(i32 10)
-; CHECK-DAG:    addiu   $4, $zero, 10
-; CHECK-DAG:    lw      $25, %got(xi)(${{[0-9]+}})
-; CHECK:        jalr    $25
+  ret void
+}
+
+declare void @xbb(i8, i8)
+
+define void @cxbb() {
+  ; ALL-LABEL:    cxbb:
+
+  ; ALL-DAG:        addiu   $[[T0:[0-9]+]], $zero, 76
+  ; ALL-DAG:        addiu   $[[T1:[0-9]+]], $zero, 101
+
+  ; 32R1-DAG:       sll     $[[T2:[0-9]+]], $[[T0]], 24
+  ; 32R1-DAG:       sra     $[[T3:[0-9]+]], $[[T2]], 24
+  ; 32R1-DAG:       sll     $[[T4:[0-9]+]], $[[T1]], 24
+  ; 32R1-DAG:       sra     $[[T5:[0-9]+]], $[[T4]], 24
 
+  ; 32R2-DAG:       seb     $4, $[[T0]]
+  ; 32R2-DAG:       seb     $5, $[[T1]]
+  call void @xbb(i8 76, i8 101)
   ret void
 }
 
-declare void @xi(i32) #1
+declare void @xhh(i16, i16)
 
-; Function Attrs: nounwind
-define void @cxii() #0 {
-entry:
-; CHECK-LABEL:  cxii
+define void @cxhh() {
+  ; ALL-LABEL:    cxhh:
+
+  ; ALL-DAG:        addiu   $[[T0:[0-9]+]], $zero, 76
+  ; ALL-DAG:        addiu   $[[T1:[0-9]+]], $zero, 101
+
+  ; 32R1-DAG:       sll     $[[T2:[0-9]+]], $[[T0]], 16
+  ; 32R1-DAG:       sra     $[[T3:[0-9]+]], $[[T2]], 16
+  ; 32R1-DAG:       sll     $[[T4:[0-9]+]], $[[T1]], 16
+  ; 32R1-DAG:       sra     $[[T5:[0-9]+]], $[[T4]], 16
+
+  ; 32R2-DAG:       seh     $4, $[[T0]]
+  ; 32R2-DAG:       seh     $5, $[[T1]]
+  call void @xhh(i16 76, i16 101)
+  ret void
+}
+
+declare void @xii(i32, i32)
+
+define void @cxii() {
+  ; ALL-LABEL:    cxii:
+
+  ; ALL-DAG:        addiu   $4, $zero, 746
+  ; ALL-DAG:        addiu   $5, $zero, 892
+  ; ALL-DAG:        lw      $25, %got(xii)(${{[0-9]+}})
+  ; ALL:            jalr    $25
   call void @xii(i32 746, i32 892)
-; CHECK-DAG:    addiu   $4, $zero, 746
-; CHECK-DAG:    addiu   $5, $zero, 892
-; CHECK-DAG:    lw      $25, %got(xii)(${{[0-9]+}})
-; CHECK:        jalr    $25
+  ret void
+}
+
+declare void @xccc(i8, i8, i8)
+
+define void @cxccc() {
+  ; ALL-LABEL:    cxccc:
 
+  ; ALL-DAG:        addiu   $[[T0:[0-9]+]], $zero, 88
+  ; ALL-DAG:        addiu   $[[T1:[0-9]+]], $zero, 44
+  ; ALL-DAG:        addiu   $[[T2:[0-9]+]], $zero, 11
+
+  ; 32R1-DAG:       sll     $[[T3:[0-9]+]], $[[T0]], 24
+  ; 32R1-DAG:       sra     $4, $[[T3]], 24
+  ; 32R1-DAG:       sll     $[[T4:[0-9]+]], $[[T1]], 24
+  ; 32R1-DAG:       sra     $5, $[[T4]], 24
+  ; 32R1-DAG:       sll     $[[T5:[0-9]+]], $[[T2]], 24
+  ; 32R1-DAG:       sra     $6, $[[T5]], 24
+
+  ; 32R2-DAG:       seb     $4, $[[T0]]
+  ; 32R2-DAG:       seb     $5, $[[T1]]
+  ; 32R2-DAG:       seb     $6, $[[T2]]
+  call void @xccc(i8 88, i8 44, i8 11)
   ret void
 }
 
-declare void @xii(i32, i32) #1
+declare void @xhhh(i16, i16, i16)
+
+define void @cxhhh() {
+  ; ALL-LABEL:    cxhhh:
 
-; Function Attrs: nounwind
-define void @cxiii() #0 {
-entry:
-; CHECK-LABEL:  cxiii
+  ; ALL-DAG:        addiu   $[[T0:[0-9]+]], $zero, 88
+  ; ALL-DAG:        addiu   $[[T1:[0-9]+]], $zero, 44
+  ; ALL-DAG:        addiu   $[[T2:[0-9]+]], $zero, 11
+
+  ; 32R1-DAG:       sll     $[[T3:[0-9]+]], $[[T0]], 16
+  ; 32R1-DAG:       sra     $4, $[[T3]], 16
+  ; 32R1-DAG:       sll     $[[T4:[0-9]+]], $[[T1]], 16
+  ; 32R1-DAG:       sra     $5, $[[T4]], 16
+  ; 32R1-DAG:       sll     $[[T5:[0-9]+]], $[[T2]], 16
+  ; 32R1-DAG:       sra     $6, $[[T5]], 16
+
+  ; 32R2-DAG:       seh     $4, $[[T0]]
+  ; 32R2-DAG:       seh     $5, $[[T1]]
+  ; 32R2-DAG:       seh     $6, $[[T2]]
+  call void @xhhh(i16 88, i16 44, i16 11)
+  ret void
+}
+
+declare void @xiii(i32, i32, i32)
+
+define void @cxiii() {
+  ; ALL-LABEL:    cxiii:
+
+  ; ALL-DAG:        addiu   $4, $zero, 88
+  ; ALL-DAG:        addiu   $5, $zero, 44
+  ; ALL-DAG:        addiu   $6, $zero, 11
+  ; ALL-DAG:        lw      $25, %got(xiii)(${{[0-9]+}})
+  ; ALL:            jalr    $25
   call void @xiii(i32 88, i32 44, i32 11)
-; CHECK-DAG:    addiu   $4, $zero, 88
-; CHECK-DAG:    addiu   $5, $zero, 44
-; CHECK-DAG:    addiu   $6, $zero, 11
-; CHECK-DAG:    lw      $25, %got(xiii)(${{[0-9]+}})
-; CHECK:        jalr    $25
   ret void
 }
 
-declare void @xiii(i32, i32, i32) #1
+declare void @xcccc(i8, i8, i8, i8)
 
-; Function Attrs: nounwind
-define void @cxiiii() #0 {
-entry:
-; CHECK-LABEL:  cxiiii
-  call void @xiiii(i32 167, i32 320, i32 97, i32 14)
-; CHECK-DAG:    addiu   $4, $zero, 167
-; CHECK-DAG:    addiu   $5, $zero, 320
-; CHECK-DAG:    addiu   $6, $zero, 97
-; CHECK-DAG:    addiu   $7, $zero, 14
-; CHECK-DAG:    lw      $25, %got(xiiii)(${{[0-9]+}})
-; CHECK:        jalr    $25
+define void @cxcccc() {
+  ; ALL-LABEL:    cxcccc:
 
+  ; ALL-DAG:        addiu   $[[T0:[0-9]+]], $zero, 88
+  ; ALL-DAG:        addiu   $[[T1:[0-9]+]], $zero, 44
+  ; ALL-DAG:        addiu   $[[T2:[0-9]+]], $zero, 11
+  ; ALL-DAG:        addiu   $[[T3:[0-9]+]], $zero, 33
+
+  ; FIXME: We should avoid the unnecessary spill/reload here.
+
+  ; 32R1-DAG:       sll     $[[T4:[0-9]+]], $[[T0]], 24
+  ; 32R1-DAG:       sra     $[[T5:[0-9]+]], $[[T4]], 24
+  ; 32R1-DAG:       sw      $4, 16($sp)
+  ; 32R1-DAG:       move    $4, $[[T5]]
+  ; 32R1-DAG:       sll     $[[T6:[0-9]+]], $[[T1]], 24
+  ; 32R1-DAG:       sra     $5, $[[T6]], 24
+  ; 32R1-DAG:       sll     $[[T7:[0-9]+]], $[[T2]], 24
+  ; 32R1-DAG:       sra     $6, $[[T7]], 24
+  ; 32R1:           lw      $[[T8:[0-9]+]], 16($sp)
+  ; 32R1:           sll     $[[T9:[0-9]+]], $[[T8]], 24
+  ; 32R1:           sra     $7, $[[T9]], 24
+
+  ; 32R2-DAG:       seb     $[[T4:[0-9]+]], $[[T0]]
+  ; 32R2-DAG:       sw      $4, 16($sp)
+  ; 32R2-DAG:       move    $4, $[[T4]]
+  ; 32R2-DAG:       seb     $5, $[[T1]]
+  ; 32R2-DAG:       seb     $6, $[[T2]]
+  ; 32R2-DAG:       lw      $[[T5:[0-9]+]], 16($sp)
+  ; 32R2:           seb     $7, $[[T5]]
+  call void @xcccc(i8 88, i8 44, i8 11, i8 33)
+  ret void
+}
+
+declare void @xhhhh(i16, i16, i16, i16)
+
+define void @cxhhhh() {
+  ; ALL-LABEL:    cxhhhh:
+
+  ; ALL-DAG:        addiu   $[[T0:[0-9]+]], $zero, 88
+  ; ALL-DAG:        addiu   $[[T1:[0-9]+]], $zero, 44
+  ; ALL-DAG:        addiu   $[[T2:[0-9]+]], $zero, 11
+  ; ALL-DAG:        addiu   $[[T3:[0-9]+]], $zero, 33
+
+  ; FIXME: We should avoid the unnecessary spill/reload here.
+
+  ; 32R1-DAG:       sll     $[[T4:[0-9]+]], $[[T0]], 16
+  ; 32R1-DAG:       sra     $[[T5:[0-9]+]], $[[T4]], 16
+  ; 32R1-DAG:       sw      $4, 16($sp)
+  ; 32R1-DAG:       move    $4, $[[T5]]
+  ; 32R1-DAG:       sll     $[[T6:[0-9]+]], $[[T1]], 16
+  ; 32R1-DAG:       sra     $5, $[[T6]], 16
+  ; 32R1-DAG:       sll     $[[T7:[0-9]+]], $[[T2]], 16
+  ; 32R1-DAG:       sra     $6, $[[T7]], 16
+  ; 32R1:           lw      $[[T8:[0-9]+]], 16($sp)
+  ; 32R1:           sll     $[[T9:[0-9]+]], $[[T8]], 16
+  ; 32R1:           sra     $7, $[[T9]], 16
+
+  ; 32R2-DAG:       seh     $[[T4:[0-9]+]], $[[T0]]
+  ; 32R2-DAG:       sw      $4, 16($sp)
+  ; 32R2-DAG:       move    $4, $[[T4]]
+  ; 32R2-DAG:       seh     $5, $[[T1]]
+  ; 32R2-DAG:       seh     $6, $[[T2]]
+  ; 32R2-DAG:       lw      $[[T5:[0-9]+]], 16($sp)
+  ; 32R2:           seh     $7, $[[T5]]
+  call void @xhhhh(i16 88, i16 44, i16 11, i16 33)
+  ret void
+}
+
+declare void @xiiii(i32, i32, i32, i32)
+
+define void @cxiiii() {
+  ; ALL-LABEL:    cxiiii:
+
+  ; ALL-DAG:        addiu   $4, $zero, 167
+  ; ALL-DAG:        addiu   $5, $zero, 320
+  ; ALL-DAG:        addiu   $6, $zero, 97
+  ; ALL-DAG:        addiu   $7, $zero, 14
+  ; ALL-DAG:        lw      $25, %got(xiiii)(${{[0-9]+}})
+  ; ALL:            jalr    $25
+  call void @xiiii(i32 167, i32 320, i32 97, i32 14)
   ret void
 }
 
-declare void @xiiii(i32, i32, i32, i32) #1
+@c1 = global i8 -45, align 1
+@uc1 = global i8 27, align 1
+@s1 = global i16 -1789, align 2
+@us1 = global i16 1256, align 2
+
+define void @cxiiiiconv() {
+  ; ALL-LABEL:    cxiiiiconv:
 
-; Function Attrs: nounwind
-define void @cxiiiiconv() #0 {
-entry:
-; CHECK-LABEL: cxiiiiconv
-; mips32r2-LABEL:  cxiiiiconv
-; mips32-LABEL:  cxiiiiconv
-  %0 = load i8* @c1, align 1
-  %conv = sext i8 %0 to i32
-  %1 = load i8* @uc1, align 1
-  %conv1 = zext i8 %1 to i32
-  %2 = load i16* @s1, align 2
-  %conv2 = sext i16 %2 to i32
-  %3 = load i16* @us1, align 2
-  %conv3 = zext i16 %3 to i32
+  ; ALL:            addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
+  ; ALL-DAG:        lw      $[[REG_C1_ADDR:[0-9]+]], %got(c1)($[[REG_GP]])
+  ; ALL-DAG:        lbu     $[[REG_C1:[0-9]+]], 0($[[REG_C1_ADDR]])
+  ; 32R1-DAG:       sll     $[[REG_C1_1:[0-9]+]], $[[REG_C1]], 24
+  ; 32R1-DAG:       sra     $4, $[[REG_C1_1]], 24
+  ; 32R2-DAG:       seb     $4, $[[REG_C1]]
+  ; FIXME: andi is superfulous
+  ; ALL-DAG:        lw      $[[REG_UC1_ADDR:[0-9]+]], %got(uc1)($[[REG_GP]])
+  ; ALL-DAG:        lbu     $[[REG_UC1:[0-9]+]], 0($[[REG_UC1_ADDR]])
+  ; ALL-DAG:        andi    $5, $[[REG_UC1]], 255
+  ; ALL-DAG:        lw      $[[REG_S1_ADDR:[0-9]+]], %got(s1)($[[REG_GP]])
+  ; ALL-DAG:        lhu     $[[REG_S1:[0-9]+]], 0($[[REG_S1_ADDR]])
+  ; 32R1-DAG:       sll     $[[REG_S1_1:[0-9]+]], $[[REG_S1]], 16
+  ; 32R1-DAG:       sra     $6, $[[REG_S1_1]], 16
+  ; 32R2-DAG:       seh     $6, $[[REG_S1]]
+  ; FIXME andi is superfulous
+  ; ALL-DAG:        lw      $[[REG_US1_ADDR:[0-9]+]], %got(us1)($[[REG_GP]])
+  ; ALL-DAG:        lhu     $[[REG_US1:[0-9]+]], 0($[[REG_US1_ADDR]])
+  ; ALL-DAG:        andi    $7, $[[REG_US1]], 65535
+  ; ALL:            jalr    $25
+  %1 = load i8, i8* @c1, align 1
+  %conv = sext i8 %1 to i32
+  %2 = load i8, i8* @uc1, align 1
+  %conv1 = zext i8 %2 to i32
+  %3 = load i16, i16* @s1, align 2
+  %conv2 = sext i16 %3 to i32
+  %4 = load i16, i16* @us1, align 2
+  %conv3 = zext i16 %4 to i32
   call void @xiiii(i32 %conv, i32 %conv1, i32 %conv2, i32 %conv3)
-; CHECK:        addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; mips32r2:     addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; mips32:       addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; mips32r2-DAG:         lw      $[[REG_C1_ADDR:[0-9]+]], %got(c1)($[[REG_GP]])
-; mips32r2-DAG: lbu     $[[REG_C1:[0-9]+]], 0($[[REG_C1_ADDR]])
-; mips32r2-DAG  seb     $3, $[[REG_C1]]
-; mips32-DAG:   lw      $[[REG_C1_ADDR:[0-9]+]], %got(c1)($[[REG_GP]])
-; mips32-DAG:   lbu     $[[REG_C1:[0-9]+]], 0($[[REG_C1_ADDR]])
-; mips32-DAG:   sll     $[[REG_C1_1:[0-9]+]], $[[REG_C1]], 24
-; mips32-DAG:   sra     $4, $[[REG_C1_1]], 24
-; CHECK-DAG:    lw      $[[REG_UC1_ADDR:[0-9]+]], %got(uc1)($[[REG_GP]])
-; CHECK-DAG:    lbu     $[[REG_UC1:[0-9]+]], 0($[[REG_UC1_ADDR]])
-; FIXME andi is superfulous
-; CHECK-DAG:    andi    $5, $[[REG_UC1]], 255
-; mips32r2-DAG:         lw      $[[REG_S1_ADDR:[0-9]+]], %got(s1)($[[REG_GP]])
-; mips32r2-DAG: lhu     $[[REG_S1:[0-9]+]], 0($[[REG_S1_ADDR]])
-; mips32r2-DAG: seh     $6, $[[REG_S1]]
-; mips32-DAG:   lw      $[[REG_S1_ADDR:[0-9]+]], %got(s1)($[[REG_GP]])
-; mips32-DAG:   lhu     $[[REG_S1:[0-9]+]], 0($[[REG_S1_ADDR]])
-; mips32-DAG:   sll     $[[REG_S1_1:[0-9]+]], $[[REG_S1]], 16
-; mips32-DAG:   sra     $6, $[[REG_S1_1]], 16
-; CHECK-DAG:    lw      $[[REG_US1_ADDR:[0-9]+]], %got(us1)($[[REG_GP]])
-; CHECK-DAG:    lhu     $[[REG_US1:[0-9]+]], 0($[[REG_US1_ADDR]])
-; FIXME andi is superfulous
-; CHECK-DAG:    andi    $7, $[[REG_US1]], 65535
-; mips32r2:     jalr    $25
-; mips32r2:     jalr    $25
-; CHECK:        jalr    $25
   ret void
 }
 
-; Function Attrs: nounwind
-define void @cxf() #0 {
-entry:
-; CHECK-LABEL:  cxf
+declare void @xf(float)
+
+define void @cxf() {
+  ; ALL-LABEL:    cxf:
+
+  ; ALL:            addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
+  ; ALL:            lui     $[[REG_FPCONST_1:[0-9]+]], 17886
+  ; ALL:            ori     $[[REG_FPCONST:[0-9]+]], $[[REG_FPCONST_1]], 17067
+  ; ALL:            mtc1    $[[REG_FPCONST]], $f12
+  ; ALL:            lw      $25, %got(xf)($[[REG_GP]])
+  ; ALL:            jalr    $25
   call void @xf(float 0x40BBC85560000000)
-; CHECK:        addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; CHECK:        lui     $[[REG_FPCONST_1:[0-9]+]], 17886
-; CHECK:        ori     $[[REG_FPCONST:[0-9]+]], $[[REG_FPCONST_1]], 17067
-; CHECK: mtc1   $[[REG_FPCONST]], $f12
-; CHECK:        lw      $25, %got(xf)($[[REG_GP]])
-; CHECK:        jalr    $25
   ret void
 }
 
-declare void @xf(float) #1
+declare void @xff(float, float)
 
-; Function Attrs: nounwind
-define void @cxff() #0 {
-entry:
-; CHECK-LABEL:  cxff
+define void @cxff() {
+  ; ALL-LABEL:    cxff:
+
+  ; ALL:            addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
+  ; ALL-DAG:        lui     $[[REG_FPCONST_1:[0-9]+]], 16314
+  ; ALL-DAG:        ori     $[[REG_FPCONST:[0-9]+]], $[[REG_FPCONST_1]], 21349
+  ; ALL-DAG:        mtc1    $[[REG_FPCONST]], $f12
+  ; ALL-DAG:        lui     $[[REG_FPCONST_2:[0-9]+]], 16593
+  ; ALL-DAG:        ori     $[[REG_FPCONST_3:[0-9]+]], $[[REG_FPCONST_2]], 24642
+  ; ALL-DAG:        mtc1    $[[REG_FPCONST_3]], $f14
+  ; ALL-DAG:        lw      $25, %got(xff)($[[REG_GP]])
+  ; ALL:            jalr    $25
   call void @xff(float 0x3FF74A6CA0000000, float 0x401A2C0840000000)
-; CHECK:        addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; CHECK-DAG:    lui     $[[REG_FPCONST_1:[0-9]+]], 16314
-; CHECK-DAG:    ori     $[[REG_FPCONST:[0-9]+]], $[[REG_FPCONST_1]], 21349
-; CHECK-DAG: mtc1       $[[REG_FPCONST]], $f12
-; CHECK-DAG:    lui     $[[REG_FPCONST_2:[0-9]+]], 16593
-; CHECK-DAG:    ori     $[[REG_FPCONST_3:[0-9]+]], $[[REG_FPCONST_2]], 24642
-; CHECK-DAG: mtc1       $[[REG_FPCONST_3]], $f14
-; CHECK:        lw      $25, %got(xff)($[[REG_GP]])
-; CHECK:        jalr    $25
   ret void
 }
 
-declare void @xff(float, float) #1
+declare void @xfi(float, i32)
 
-; Function Attrs: nounwind
-define void @cxfi() #0 {
-entry:
-; CHECK-LABEL: cxfi
-  call void @xfi(float 0x4013906240000000, i32 102)
-; CHECK:        addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; CHECK-DAG:    lui     $[[REG_FPCONST_1:[0-9]+]], 16540
-; CHECK-DAG:    ori     $[[REG_FPCONST:[0-9]+]], $[[REG_FPCONST_1]], 33554
-; CHECK-DAG: mtc1       $[[REG_FPCONST]], $f12
-; CHECK-DAG:    addiu   $5, $zero, 102
-; CHECK:        lw      $25, %got(xfi)($[[REG_GP]])
-; CHECK:        jalr    $25
+define void @cxfi() {
+  ; ALL-LABEL:    cxfi:
 
+  ; ALL:            addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
+  ; ALL-DAG:        lui     $[[REG_FPCONST_1:[0-9]+]], 16540
+  ; ALL-DAG:        ori     $[[REG_FPCONST:[0-9]+]], $[[REG_FPCONST_1]], 33554
+  ; ALL-DAG:        mtc1    $[[REG_FPCONST]], $f12
+  ; ALL-DAG:        addiu   $5, $zero, 102
+  ; ALL-DAG:        lw      $25, %got(xfi)($[[REG_GP]])
+  ; ALL:            jalr    $25
+  call void @xfi(float 0x4013906240000000, i32 102)
   ret void
 }
 
-declare void @xfi(float, i32) #1
+declare void @xfii(float, i32, i32)
 
-; Function Attrs: nounwind
-define void @cxfii() #0 {
-entry:
-; CHECK-LABEL: cxfii
+define void @cxfii() {
+  ; ALL-LABEL:    cxfii:
+
+  ; ALL:            addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
+  ; ALL-DAG:        lui     $[[REG_FPCONST_1:[0-9]+]], 17142
+  ; ALL-DAG:        ori     $[[REG_FPCONST:[0-9]+]], $[[REG_FPCONST_1]], 16240
+  ; ALL-DAG:        mtc1    $[[REG_FPCONST]], $f12
+  ; ALL-DAG:        addiu   $5, $zero, 9993
+  ; ALL-DAG:        addiu   $6, $zero, 10922
+  ; ALL-DAG:        lw      $25, %got(xfii)($[[REG_GP]])
+  ; ALL:            jalr    $25
   call void @xfii(float 0x405EC7EE00000000, i32 9993, i32 10922)
-; CHECK:        addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; CHECK-DAG:    lui     $[[REG_FPCONST_1:[0-9]+]], 17142
-; CHECK-DAG:    ori     $[[REG_FPCONST:[0-9]+]], $[[REG_FPCONST_1]], 16240
-; CHECK-DAG: mtc1       $[[REG_FPCONST]], $f12
-; CHECK-DAG:    addiu   $5, $zero, 9993
-; CHECK-DAG:    addiu   $6, $zero, 10922
-; CHECK:        lw      $25, %got(xfii)($[[REG_GP]])
-; CHECK:        jalr    $25
   ret void
 }
 
-declare void @xfii(float, i32, i32) #1
+declare void @xfiii(float, i32, i32, i32)
+
+define void @cxfiii() {
+  ; ALL-LABEL:    cxfiii:
 
-; Function Attrs: nounwind
-define void @cxfiii() #0 {
-entry:
-; CHECK-LABEL: cxfiii
+  ; ALL:            addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
+  ; ALL-DAG:        lui     $[[REG_FPCONST_1:[0-9]+]], 17120
+  ; ALL-DAG:        ori     $[[REG_FPCONST:[0-9]+]], $[[REG_FPCONST_1]], 14681
+  ; ALL-DAG:        mtc1    $[[REG_FPCONST]], $f12
+  ; ALL-DAG:        addiu   $5, $zero, 3948
+  ; ALL-DAG:        lui     $[[REG_I_1:[0-9]+]], 1
+  ; ALL-DAG:        ori     $6, $[[REG_I_1]], 23475
+  ; ALL-DAG:        lui     $[[REG_I_2:[0-9]+]], 1
+  ; ALL-DAG:        ori     $7, $[[REG_I_2]], 45686
+  ; ALL-DAG:        lw      $25, %got(xfiii)($[[REG_GP]])
+  ; ALL:            jalr    $25
   call void @xfiii(float 0x405C072B20000000, i32 3948, i32 89011, i32 111222)
-; CHECK:        addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; CHECK-DAG:    lui     $[[REG_FPCONST_1:[0-9]+]], 17120
-; CHECK-DAG:    ori     $[[REG_FPCONST:[0-9]+]], $[[REG_FPCONST_1]], 14681
-; CHECK-DAG: mtc1       $[[REG_FPCONST]], $f12
-; CHECK-DAG:    addiu   $5, $zero, 3948
-; CHECK-DAG:    lui     $[[REG_I_1:[0-9]+]], 1
-; CHECK-DAG:    ori     $6, $[[REG_I_1]], 23475
-; CHECK-DAG:    lui     $[[REG_I_2:[0-9]+]], 1
-; CHECK-DAG:    ori     $7, $[[REG_I_2]], 45686
-; CHECK:        lw      $25, %got(xfiii)($[[REG_GP]])
-; CHECK:        jalr    $25
   ret void
 }
 
-declare void @xfiii(float, i32, i32, i32) #1
+declare void @xd(double)
+
+define void @cxd() {
+  ; ALL-LABEL:    cxd:
 
-; Function Attrs: nounwind
-define void @cxd() #0 {
-entry:
-; mips32r2-LABEL: cxd:
-; mips32-LABEL: cxd:
+  ; ALL:            addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
+  ; ALL-DAG:        lui     $[[REG_FPCONST_1:[0-9]+]], 16514
+  ; ALL-DAG:        ori     $[[REG_FPCONST_2:[0-9]+]], $[[REG_FPCONST_1]], 48037
+  ; ALL-DAG:        lui     $[[REG_FPCONST_3:[0-9]+]], 58195
+  ; ALL-DAG:        ori     $[[REG_FPCONST_4:[0-9]+]], $[[REG_FPCONST_3]], 63439
+  ; ALL-DAG:        mtc1    $[[REG_FPCONST_4]], $f12
+  ; 32R1-DAG:       mtc1    $[[REG_FPCONST_2]], $f13
+  ; 32R2-DAG:       mthc1   $[[REG_FPCONST_2]], $f12
+  ; ALL-DAG:        lw      $25, %got(xd)($[[REG_GP]])
+  ; ALL:            jalr    $25
   call void @xd(double 5.994560e+02)
-; mips32:       addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; mips32-DAG:   lui     $[[REG_FPCONST_1:[0-9]+]], 16514
-; mips32-DAG:   ori     $[[REG_FPCONST_2:[0-9]+]], $[[REG_FPCONST_1]], 48037
-; mips32-DAG:   lui     $[[REG_FPCONST_3:[0-9]+]], 58195
-; mips32-DAG:   ori     $[[REG_FPCONST_4:[0-9]+]], $[[REG_FPCONST_3]], 63439
-; mips32-DAG:    mtc1   $[[REG_FPCONST_4]], $f12
-; mips32-DAG:       mtc1        $[[REG_FPCONST_2]], $f13
-; mips32-DAG:   lw      $25, %got(xd)($[[REG_GP]])
-; mips32:       jalr    $25
-; mips32r2:     addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; mips32r2-DAG: lui     $[[REG_FPCONST_1:[0-9]+]], 16514
-; mips32r2-DAG: ori     $[[REG_FPCONST_2:[0-9]+]], $[[REG_FPCONST_1]], 48037
-; mips32r2-DAG: lui     $[[REG_FPCONST_3:[0-9]+]], 58195
-; mips32r2-DAG: ori     $[[REG_FPCONST_4:[0-9]+]], $[[REG_FPCONST_3]], 63439
-; mips32r2-DAG: mtc1    $[[REG_FPCONST_4]], $f12
-; mips32r2-DAG: mthc1   $[[REG_FPCONST_2]], $f12
-; mips32r2-DAG: lw      $25, %got(xd)($[[REG_GP]])
-; mips32r2 :    jalr    $25
   ret void
 }
 
-declare void @xd(double) #1
+declare void @xdd(double, double)
 
-; Function Attrs: nounwind
-define void @cxdd() #0 {
-; mips32r2-LABEL: cxdd:
-; mips32-LABEL: cxdd:
-entry:
+define void @cxdd() {
+  ; ALL-LABEL:    cxdd:
+
+  ; ALL:            addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
+  ; ALL-DAG:        lui     $[[REG_FPCONST_1:[0-9]+]], 16531
+  ; ALL-DAG:        ori     $[[REG_FPCONST_2:[0-9]+]], $[[REG_FPCONST_1]], 19435
+  ; ALL-DAG:        lui     $[[REG_FPCONST_3:[0-9]+]], 34078
+  ; ALL-DAG:        ori     $[[REG_FPCONST_4:[0-9]+]], $[[REG_FPCONST_3]], 47186
+  ; ALL-DAG:        mtc1    $[[REG_FPCONST_4]], $f12
+  ; 32R1-DAG:       mtc1    $[[REG_FPCONST_2]], $f13
+  ; 32R2-DAG:       mthc1   $[[REG_FPCONST_2]], $f12
+  ; ALL-DAG:        lui     $[[REG_FPCONST_1:[0-9]+]], 16629
+  ; ALL-DAG:        ori     $[[REG_FPCONST_2:[0-9]+]], $[[REG_FPCONST_1]], 45873
+  ; ALL-DAG:        lui     $[[REG_FPCONST_3:[0-9]+]], 63438
+  ; ALL-DAG:        ori     $[[REG_FPCONST_4:[0-9]+]], $[[REG_FPCONST_3]], 55575
+  ; ALL-DAG:        mtc1    $[[REG_FPCONST_4]], $f14
+  ; 32R1-DAG:       mtc1    $[[REG_FPCONST_2]], $f15
+  ; 32R2-DAG:       mthc1   $[[REG_FPCONST_2]], $f14
+  ; ALL-DAG:        lw      $25, %got(xdd)($[[REG_GP]])
+  ; ALL:            jalr    $25
   call void @xdd(double 1.234980e+03, double 0x40F5B331F7CED917)
-; mips32:       addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; mips32-DAG:   lui     $[[REG_FPCONST_1:[0-9]+]], 16531
-; mips32-DAG:   ori     $[[REG_FPCONST_2:[0-9]+]], $[[REG_FPCONST_1]], 19435
-; mips32-DAG:   lui     $[[REG_FPCONST_3:[0-9]+]], 34078
-; mips32-DAG:   ori     $[[REG_FPCONST_4:[0-9]+]], $[[REG_FPCONST_3]], 47186
-; mips32-DAG:   mtc1    $[[REG_FPCONST_4]], $f12
-; mips32-DAG:   mtc1    $[[REG_FPCONST_2]], $f13
-; mips32-DAG:   lui     $[[REG_FPCONST_1:[0-9]+]], 16629
-; mips32-DAG:   ori     $[[REG_FPCONST_2:[0-9]+]], $[[REG_FPCONST_1]], 45873
-; mips32-DAG:   lui     $[[REG_FPCONST_3:[0-9]+]], 63438
-; mips32-DAG:   ori     $[[REG_FPCONST_4:[0-9]+]], $[[REG_FPCONST_3]], 55575
-; mips32-DAG:   mtc1    $[[REG_FPCONST_4]], $f14
-; mips32-DAG:   mtc1    $[[REG_FPCONST_2]], $f15
-; mips32-DAG:   lw      $25, %got(xdd)($[[REG_GP]])
-; mips32:       jalr    $25
-; mips32r2:     addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; mips32r2-DAG: lui     $[[REG_FPCONST_1:[0-9]+]], 16531
-; mips32r2-DAG: ori     $[[REG_FPCONST_2:[0-9]+]], $[[REG_FPCONST_1]], 19435
-; mips32r2-DAG: lui     $[[REG_FPCONST_3:[0-9]+]], 34078
-; mips32r2-DAG: ori     $[[REG_FPCONST_4:[0-9]+]], $[[REG_FPCONST_3]], 47186
-; mips32r2-DAG: mtc1    $[[REG_FPCONST_4]], $f12
-; mips32r2-DAG: mthc1   $[[REG_FPCONST_2]], $f12
-; mips32r2-DAG: lui     $[[REG_FPCONST_1:[0-9]+]], 16629
-; mips32r2-DAG: ori     $[[REG_FPCONST_2:[0-9]+]], $[[REG_FPCONST_1]], 45873
-; mips32r2-DAG: lui     $[[REG_FPCONST_3:[0-9]+]], 63438
-; mips32r2-DAG: ori     $[[REG_FPCONST_4:[0-9]+]], $[[REG_FPCONST_3]], 55575
-; mips32r2-DAG: mtc1    $[[REG_FPCONST_4]], $f14
-; mips32r2-DAG: mthc1   $[[REG_FPCONST_2]], $f14
-; mips32r2-DAG: lw      $25, %got(xdd)($[[REG_GP]])
-; mips32r2 :    jalr    $25
   ret void
 }
 
-declare void @xdd(double, double) #1
+declare void @xif(i32, float)
 
-; Function Attrs: nounwind
-define void @cxif() #0 {
-entry:
-; CHECK-LABEL: cxif:
-  call void @xif(i32 345, float 0x407BCE5A20000000)
-; CHECK-DAG:    addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; CHECK-DAG:    addiu   $4, $zero, 345
-; CHECK-DAG:    lui     $[[REGF_1:[0-9]+]], 17374
-; CHECK-DAG:    ori     $[[REGF_2:[0-9]+]], $[[REGF_1]], 29393
-; CHECK-DAG:    mtc1    $[[REGF_2]], $f[[REGF_3:[0-9]+]]
-; CHECK-DAG:    mfc1    $5, $f[[REGF_3]]
-; CHECK-DAG:    lw      $25, %got(xif)($[[REG_GP]])
-; CHECK:        jalr    $25
+define void @cxif() {
+  ; ALL-LABEL:    cxif:
 
+  ; ALL:            addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
+  ; ALL-DAG:        addiu   $4, $zero, 345
+  ; ALL-DAG:        lui     $[[REGF_1:[0-9]+]], 17374
+  ; ALL-DAG:        ori     $[[REGF_2:[0-9]+]], $[[REGF_1]], 29393
+  ; ALL-DAG:        mtc1    $[[REGF_2]], $f[[REGF_3:[0-9]+]]
+  ; ALL-DAG:        mfc1    $5, $f[[REGF_3]]
+  ; ALL-DAG:        lw      $25, %got(xif)($[[REG_GP]])
+  ; ALL:            jalr    $25
+  call void @xif(i32 345, float 0x407BCE5A20000000)
   ret void
 }
 
-declare void @xif(i32, float) #1
+declare void @xiff(i32, float, float)
 
-; Function Attrs: nounwind
-define void @cxiff() #0 {
-entry:
-; CHECK-LABEL: cxiff:
-; CHECK2-LABEL: cxiff:
+define void @cxiff() {
+  ; ALL-LABEL:    cxiff:
+
+  ; ALL:            addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
+  ; ALL-DAG:        addiu   $4, $zero, 12239
+  ; ALL-DAG:        lui     $[[REGF0_1:[0-9]+]], 17526
+  ; ALL-DAG:        ori     $[[REGF0_2:[0-9]+]], $[[REGF0_1]], 55706
+  ; ALL-DAG:        mtc1    $[[REGF0_2]], $f[[REGF0_3:[0-9]+]]
+  ; ALL-DAG:        lui     $[[REGF1_1:[0-9]+]], 16543
+  ; ALL-DAG:        ori     $[[REGF1_2:[0-9]+]], $[[REGF1_1]], 65326
+  ; ALL:            mtc1    $[[REGF1_2]], $f[[REGF1_3:[0-9]+]]
+  ; ALL-DAG:        mfc1    $5, $f[[REGF0_3]]
+  ; ALL-DAG:        mfc1    $6, $f[[REGF1_3]]
+  ; ALL-DAG:        lw      $25, %got(xiff)($[[REG_GP]])
+  ; ALL:            jalr    $25
   call void @xiff(i32 12239, float 0x408EDB3340000000, float 0x4013FFE5C0000000)
-; We need to do the two floating point parameters in a separate
-; check because we can't control the ordering of parts of the sequence
-;;
-; CHECK:        addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; CHECK:        addiu   $4, $zero, 12239
-; CHECK2:       addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; CHECK2:       addiu   $4, $zero, 12239
-; CHECK:        lui     $[[REGF_1:[0-9]+]], 17526
-; CHECK:        ori     $[[REGF_2:[0-9]+]], $[[REGF_1]], 55706
-; CHECK:        mtc1    $[[REGF_2]], $f[[REGF_3:[0-9]+]]
-; CHECK:        mfc1    $5, $f[[REGF_3]]
-; CHECK2:       lui     $[[REGF2_1:[0-9]+]], 16543
-; CHECK2:       ori     $[[REGF2_2:[0-9]+]], $[[REGF2_1]], 65326
-; CHECK2:       mtc1    $[[REGF2_2]], $f[[REGF2_3:[0-9]+]]
-; CHECK2:       mfc1    $6, $f[[REGF2_3]]
-; CHECK:        lw      $25, %got(xiff)($[[REG_GP]])
-; CHECK2:       lw      $25, %got(xiff)($[[REG_GP]])
-; CHECK:        jalr    $25
-; CHECK2:       jalr    $25
   ret void
 }
 
-declare void @xiff(i32, float, float) #1
+declare void @xifi(i32, float, i32)
 
-; Function Attrs: nounwind
-define void @cxifi() #0 {
-entry:
-; CHECK: cxifi:
-  call void @xifi(i32 887, float 0x402277CEE0000000, i32 888)
-; CHECK-DAG:    addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; CHECK-DAG:    addiu   $4, $zero, 887
-; CHECK-DAG:    lui     $[[REGF_1:[0-9]+]], 16659
-; CHECK-DAG:    ori     $[[REGF_2:[0-9]+]], $[[REGF_1]], 48759
-; CHECK-DAG:    mtc1    $[[REGF_2]], $f[[REGF_3:[0-9]+]]
-; CHECK-DAG:    mfc1    $5, $f[[REGF_3]]
-; CHECk-DAG:    addiu   $6, $zero, 888
-; CHECK-DAG:    lw      $25, %got(xifi)($[[REG_GP]])
-; CHECK:        jalr    $25
+define void @cxifi() {
+  ; ALL-LABEL:    cxifi:
 
+  ; ALL:            addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
+  ; ALL-DAG:        addiu   $4, $zero, 887
+  ; ALL-DAG:        lui     $[[REGF_1:[0-9]+]], 16659
+  ; ALL-DAG:        ori     $[[REGF_2:[0-9]+]], $[[REGF_1]], 48759
+  ; ALL-DAG:        mtc1    $[[REGF_2]], $f[[REGF_3:[0-9]+]]
+  ; ALL-DAG:        mfc1    $5, $f[[REGF_3]]
+  ; ALL-DAG:        addiu   $6, $zero, 888
+  ; ALL-DAG:        lw      $25, %got(xifi)($[[REG_GP]])
+  ; ALL:            jalr    $25
+  call void @xifi(i32 887, float 0x402277CEE0000000, i32 888)
   ret void
 }
 
-declare void @xifi(i32, float, i32) #1
+declare void @xifif(i32, float, i32, float)
 
-; Function Attrs: nounwind
-define void @cxifif() #0 {
-entry:
-; CHECK: cxifif:
-; CHECK2: cxifif:
-  call void @xifif(i32 67774, float 0x408EE0FBE0000000, i32 9991, float 0x40B15C8CC0000000)
-; CHECK-DAG:    addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; CHECK-DAG:    addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; CHECK-DAG:    lui     $[[REGI:[0-9]+]], 1
-; CHECK-DAG:    ori     $4, $[[REGI]], 2238
-; CHECK-DAG:    lui     $[[REGF_1:[0-9]+]], 17527
-; CHECK-DAG:    ori     $[[REGF_2:[0-9]+]], $[[REGF_1]], 2015
-; CHECK-DAG:    mtc1    $[[REGF_2]], $f[[REGF_3:[0-9]+]]
-; CHECK-DAG:    mfc1    $5, $f[[REGF_3]]
-; CHECk-DAG:    addiu   $6, $zero, 888
-; CHECK2:       lui     $[[REGF2_1:[0-9]+]], 17802
-; CHECK2:       ori     $[[REGF2_2:[0-9]+]], $[[REGF2_1]], 58470
-; CHECK2:       mtc1    $[[REGF2_2]], $f[[REGF2_3:[0-9]+]]
-; CHECK2:       mfc1    $7, $f[[REGF2_3]]
-; CHECK:        lw      $25, %got(xifif)($[[REG_GP]])
-; CHECK2:       lw      $25, %got(xifif)($[[REG_GP]])
-; CHECK2:       jalr    $25
-; CHECK:        jalr    $25
+define void @cxifif() {
+  ; ALL-LABEL:    cxifif:
 
+  ; ALL:            addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
+  ; ALL-DAG:        lui     $[[REGI:[0-9]+]], 1
+  ; ALL-DAG:        ori     $4, $[[REGI]], 2238
+  ; ALL-DAG:        lui     $[[REGF0_1:[0-9]+]], 17527
+  ; ALL-DAG:        ori     $[[REGF0_2:[0-9]+]], $[[REGF0_1]], 2015
+  ; ALL-DAG:        mtc1    $[[REGF0_2]], $f[[REGF0_3:[0-9]+]]
+  ; ALL-DAG:        addiu   $6, $zero, 9991
+  ; ALL-DAG:        lui     $[[REGF1_1:[0-9]+]], 17802
+  ; ALL-DAG:        ori     $[[REGF1_2:[0-9]+]], $[[REGF1_1]], 58470
+  ; ALL:            mtc1    $[[REGF1_2]], $f[[REGF1_3:[0-9]+]]
+  ; ALL-DAG:        mfc1    $5, $f[[REGF0_3]]
+  ; ALL-DAG:        mfc1    $7, $f[[REGF1_3]]
+  ; ALL-DAG:        lw      $25, %got(xifif)($[[REG_GP]])
+  ; ALL:            jalr    $25
+  call void @xifif(i32 67774, float 0x408EE0FBE0000000,
+                   i32 9991, float 0x40B15C8CC0000000)
   ret void
 }
 
-declare void @xifif(i32, float, i32, float) #1
+declare void @xiffi(i32, float, float, i32)
 
-; Function Attrs: nounwind
-define void @cxiffi() #0 {
-entry:
-; CHECK-label: cxiffi:
-; CHECK2-label: cxiffi:
-  call void @xiffi(i32 45, float 0x3FF6666660000000, float 0x408F333340000000, i32 234)
-; CHECK-DAG:    addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; CHECK-DAG:    addiu   $4, $zero, 45
-; CHECK2-DAG:   addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; CHECK2-DAG:   addiu   $4, $zero, 45
-; CHECK-DAG:    lui     $[[REGF_1:[0-9]+]], 16307
-; CHECK-DAG:    ori     $[[REGF_2:[0-9]+]], $[[REGF_1]], 13107
-; CHECK-DAG:    mtc1    $[[REGF_2]], $f[[REGF_3:[0-9]+]]
-; CHECK-DAG:    mfc1    $5, $f[[REGF_3]]
-; CHECK2:       lui     $[[REGF2_1:[0-9]+]], 17529
-; CHECK2:       ori     $[[REGF2_2:[0-9]+]], $[[REGF2_1]], 39322
-; CHECK2:       mtc1    $[[REGF2_2]], $f[[REGF2_3:[0-9]+]]
-; CHECK2:       mfc1    $6, $f[[REGF2_3]]
-; CHECK-DAG:    lw      $25, %got(xiffi)($[[REG_GP]])
-; CHECK-DAG:    addiu   $7, $zero, 234
-; CHECK2-DAG:   lw      $25, %got(xiffi)($[[REG_GP]])
-; CHECK:        jalr    $25
-; CHECK2:       jalr    $25
+define void @cxiffi() {
+  ; ALL-LABEL:    cxiffi:
 
+  ; ALL:            addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
+  ; ALL-DAG:        addiu   $4, $zero, 45
+  ; ALL-DAG:        lui     $[[REGF0_1:[0-9]+]], 16307
+  ; ALL-DAG:        ori     $[[REGF0_2:[0-9]+]], $[[REGF0_1]], 13107
+  ; ALL-DAG:        mtc1    $[[REGF0_2]], $f[[REGF0_3:[0-9]+]]
+  ; ALL-DAG:        lui     $[[REGF1_1:[0-9]+]], 17529
+  ; ALL-DAG:        ori     $[[REGF1_2:[0-9]+]], $[[REGF1_1]], 39322
+  ; ALL:            mtc1    $[[REGF1_2]], $f[[REGF1_3:[0-9]+]]
+  ; ALL-DAG:        addiu   $7, $zero, 234
+  ; ALL-DAG:        mfc1    $5, $f[[REGF0_3]]
+  ; ALL-DAG:        mfc1    $6, $f[[REGF1_3]]
+  ; ALL-DAG:        lw      $25, %got(xiffi)($[[REG_GP]])
+  ; ALL:            jalr    $25
+  call void @xiffi(i32 45, float 0x3FF6666660000000,
+                   float 0x408F333340000000, i32 234)
   ret void
 }
 
-declare void @xiffi(i32, float, float, i32) #1
+declare void @xifii(i32, float, i32, i32)
 
-; Function Attrs: nounwind
-define void @cxifii() #0 {
-entry:
-; CHECK-DAG:    cxifii:
-  call void @xifii(i32 12239, float 0x408EDB3340000000, i32 998877, i32 1234)
-; CHECK-DAG:    addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
-; CHECK-DAG:    addiu   $4, $zero, 12239
-; CHECK-DAG:    lui     $[[REGF_1:[0-9]+]], 17526
-; CHECK-DAG:    ori     $[[REGF_2:[0-9]+]], $[[REGF_1]], 55706
-; CHECK-DAG:    mtc1    $[[REGF_2]], $f[[REGF_3:[0-9]+]]
-; CHECK-DAG:    mfc1    $5, $f[[REGF_3]]
-; CHECK-DAG:    lui     $[[REGI2:[0-9]+]], 15
-; CHECK-DAG:    ori     $6, $[[REGI2]], 15837
-; CHECk-DAG:    addiu   $7, $zero, 1234
-; CHECK-DAG:    lw      $25, %got(xifii)($[[REG_GP]])
-; CHECK:        jalr    $25
-  ret void
-}
-
-declare void @xifii(i32, float, i32, i32) #1
-
-; FIXME: this function will not pass yet. 
-; Function Attrs: nounwind
-; define void @cxfid() #0 {
-;entry:
-;  call void @xfid(float 0x4013B851E0000000, i32 811123, double 0x40934BFF487FCB92)
-;  ret void
-;}
-
-declare void @xfid(float, i32, double) #1
+define void @cxifii() {
+  ; ALL-LABEL:    cxifii:
 
-; Function Attrs: nounwind
-define void @g() #0 {
-entry:
-  call void @cxi()
-  call void @cxii()
-  call void @cxiii()
-  call void @cxiiii()
-  call void @cxiiiiconv()
-  call void @cxf()
-  call void @cxff()
-  call void @cxd()
-  call void @cxfi()
-  call void @cxfii()
-  call void @cxfiii()
-  call void @cxdd()
-  call void @cxif()
-  call void @cxiff()
-  call void @cxifi()
-  call void @cxifii()
-  call void @cxifif()
-  call void @cxiffi()
+  ; ALL-DAG:    addu    $[[REG_GP:[0-9]+]], ${{[0-9]+}}, ${{[0-9+]}}
+  ; ALL-DAG:    addiu   $4, $zero, 12239
+  ; ALL-DAG:    lui     $[[REGF_1:[0-9]+]], 17526
+  ; ALL-DAG:    ori     $[[REGF_2:[0-9]+]], $[[REGF_1]], 55706
+  ; ALL-DAG:    mtc1    $[[REGF_2]], $f[[REGF_3:[0-9]+]]
+  ; ALL-DAG:    mfc1    $5, $f[[REGF_3]]
+  ; ALL-DAG:    lui     $[[REGI2:[0-9]+]], 15
+  ; ALL-DAG:    ori     $6, $[[REGI2]], 15837
+  ; ALL-DAG:    addiu   $7, $zero, 1234
+  ; ALL-DAG:    lw      $25, %got(xifii)($[[REG_GP]])
+  ; ALL:        jalr    $25
+  call void @xifii(i32 12239, float 0x408EDB3340000000, i32 998877, i32 1234)
   ret void
 }
-
-
-attributes #0 = { nounwind "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
-attributes #1 = { "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
-
-!llvm.ident = !{!0}
-
-!0 = !{!"clang version 3.6.0 (gitosis@dmz-portal.mips.com:clang 43992fe7b17de5553ac06d323cb80cc6723a9ae3) (gitosis@dmz-portal.mips.com:llvm.git 0834e6839eb170197c81bb02e916258d1527e312)"}
diff --git a/test/CodeGen/Mips/Fast-ISel/constexpr-address.ll b/test/CodeGen/Mips/Fast-ISel/constexpr-address.ll
new file mode 100644
index 000000000000..df60d8071836
--- /dev/null
+++ b/test/CodeGen/Mips/Fast-ISel/constexpr-address.ll
@@ -0,0 +1,18 @@
+; RUN: llc -march=mipsel -mcpu=mips32 -relocation-model=pic \
+; RUN:     -fast-isel=true -mips-fast-isel -fast-isel-abort=1 < %s | FileCheck %s
+; RUN: llc -march=mipsel -mcpu=mips32r2 -relocation-model=pic \
+; RUN:     -fast-isel=true -mips-fast-isel -fast-isel-abort=1 < %s | FileCheck %s
+
+@ARR = external global [10 x i32], align 4
+
+define void @foo() {
+; CHECK-LABEL: foo
+
+; CHECK-DAG:    lw      $[[ARR:[0-9]+]], %got(ARR)({{.*}})
+; CHECK-DAG:    addiu   $[[T0:[0-9]+]], $zero, 12345
+; CHECK:        sw      $[[T0]], 8($[[ARR]])
+
+entry:
+  store i32 12345, i32* getelementptr inbounds ([10 x i32], [10 x i32]* @ARR, i32 0, i32 2), align 4
+  ret void
+}
diff --git a/test/CodeGen/Mips/Fast-ISel/fastalloca.ll b/test/CodeGen/Mips/Fast-ISel/fastalloca.ll
new file mode 100644
index 000000000000..b4a9f1ce7ab0
--- /dev/null
+++ b/test/CodeGen/Mips/Fast-ISel/fastalloca.ll
@@ -0,0 +1,32 @@
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
+; RUN:     < %s | FileCheck %s
+
+%struct.x = type { i32 }
+
+@i = common global i32 0, align 4
+
+define i32 @foobar(i32 signext %x) {
+entry:
+; CHECK-LABEL: foobar:
+  %retval = alloca i32, align 4
+  %x.addr = alloca i32, align 4
+  %a = alloca %struct.x, align 4
+  %c = alloca %struct.x*, align 4
+  store i32 %x, i32* %x.addr, align 4
+  %x1 = getelementptr inbounds %struct.x, %struct.x* %a, i32 0, i32 0
+  %0 = load i32, i32* %x.addr, align 4
+  store i32 %0, i32* %x1, align 4
+  store %struct.x* %a, %struct.x** %c, align 4
+  %1 = load %struct.x*, %struct.x** %c, align 4
+  %x2 = getelementptr inbounds %struct.x, %struct.x* %1, i32 0, i32 0
+  %2 = load i32, i32* %x2, align 4
+  store i32 %2, i32* @i, align 4
+  %3 = load i32, i32* %retval
+; CHECK-DAG:    lw      $[[I_ADDR:[0-9]+]], %got(i)($[[REG_GP:[0-9]+]])
+; CHECK-DAG:    addiu   $[[A_ADDR:[0-9]+]], $sp, 8
+; CHECK-DAG:    sw      $[[A_ADDR]], [[A_ADDR_FI:[0-9]+]]($sp)
+; CHECK-DAG:    lw      $[[A_ADDR2:[0-9]+]], [[A_ADDR_FI]]($sp)
+; CHECK-DAG:    lw      $[[A_X:[0-9]+]], 0($[[A_ADDR2]])
+; CHECK-DAG:    sw      $[[A_X]], 0($[[I_ADDR]])
+  ret i32 %3
+}
diff --git a/test/CodeGen/Mips/Fast-ISel/fpcmpa.ll b/test/CodeGen/Mips/Fast-ISel/fpcmpa.ll
index c72b1e70c718..72de888b26e0 100644
--- a/test/CodeGen/Mips/Fast-ISel/fpcmpa.ll
+++ b/test/CodeGen/Mips/Fast-ISel/fpcmpa.ll
@@ -1,6 +1,6 @@
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
 ; RUN:     < %s | FileCheck %s
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32 \
 ; RUN:     < %s | FileCheck %s
 
 @f1 = common global float 0.000000e+00, align 4
@@ -12,8 +12,8 @@
 ; Function Attrs: nounwind
 define void @feq1()  {
 entry:
-  %0 = load float* @f1, align 4
-  %1 = load float* @f2, align 4
+  %0 = load float, float* @f1, align 4
+  %1 = load float, float* @f2, align 4
   %cmp = fcmp oeq float %0, %1
 ; CHECK-LABEL:  feq1:
 ; CHECK-DAG:    lw      $[[REG_F2_GOT:[0-9]+]], %got(f2)(${{[0-9]+}})
@@ -33,8 +33,8 @@ entry:
 ; Function Attrs: nounwind
 define void @fne1()  {
 entry:
-  %0 = load float* @f1, align 4
-  %1 = load float* @f2, align 4
+  %0 = load float, float* @f1, align 4
+  %1 = load float, float* @f2, align 4
   %cmp = fcmp une float %0, %1
 ; CHECK-LABEL:  fne1:
 ; CHECK-DAG:    lw      $[[REG_F2_GOT:[0-9]+]], %got(f2)(${{[0-9]+}})
@@ -53,8 +53,8 @@ entry:
 ; Function Attrs: nounwind
 define void @flt1()  {
 entry:
-  %0 = load float* @f1, align 4
-  %1 = load float* @f2, align 4
+  %0 = load float, float* @f1, align 4
+  %1 = load float, float* @f2, align 4
   %cmp = fcmp olt float %0, %1
 ; CHECK-LABEL:  flt1:
 ; CHECK-DAG:    lw      $[[REG_F2_GOT:[0-9]+]], %got(f2)(${{[0-9]+}})
@@ -74,8 +74,8 @@ entry:
 ; Function Attrs: nounwind
 define void @fgt1()  {
 entry:
-  %0 = load float* @f1, align 4
-  %1 = load float* @f2, align 4
+  %0 = load float, float* @f1, align 4
+  %1 = load float, float* @f2, align 4
   %cmp = fcmp ogt float %0, %1
 ; CHECK-LABEL: fgt1:
 ; CHECK-DAG:    lw      $[[REG_F2_GOT:[0-9]+]], %got(f2)(${{[0-9]+}})
@@ -94,8 +94,8 @@ entry:
 ; Function Attrs: nounwind
 define void @fle1()  {
 entry:
-  %0 = load float* @f1, align 4
-  %1 = load float* @f2, align 4
+  %0 = load float, float* @f1, align 4
+  %1 = load float, float* @f2, align 4
   %cmp = fcmp ole float %0, %1
 ; CHECK-LABEL:  fle1:
 ; CHECK-DAG:    lw      $[[REG_F2_GOT:[0-9]+]], %got(f2)(${{[0-9]+}})
@@ -114,8 +114,8 @@ entry:
 ; Function Attrs: nounwind
 define void @fge1()  {
 entry:
-  %0 = load float* @f1, align 4
-  %1 = load float* @f2, align 4
+  %0 = load float, float* @f1, align 4
+  %1 = load float, float* @f2, align 4
   %cmp = fcmp oge float %0, %1
 ; CHECK-LABEL:  fge1:
 ; CHECK-DAG:    lw      $[[REG_F2_GOT:[0-9]+]], %got(f2)(${{[0-9]+}})
@@ -134,8 +134,8 @@ entry:
 ; Function Attrs: nounwind
 define void @deq1()  {
 entry:
-  %0 = load double* @d1, align 8
-  %1 = load double* @d2, align 8
+  %0 = load double, double* @d1, align 8
+  %1 = load double, double* @d2, align 8
   %cmp = fcmp oeq double %0, %1
 ; CHECK-LABEL:  deq1:
 ; CHECK-DAG:    lw      $[[REG_D2_GOT:[0-9]+]], %got(d2)(${{[0-9]+}})
@@ -154,8 +154,8 @@ entry:
 ; Function Attrs: nounwind
 define void @dne1()  {
 entry:
-  %0 = load double* @d1, align 8
-  %1 = load double* @d2, align 8
+  %0 = load double, double* @d1, align 8
+  %1 = load double, double* @d2, align 8
   %cmp = fcmp une double %0, %1
 ; CHECK-LABEL:  dne1:
 ; CHECK-DAG:    lw      $[[REG_D2_GOT:[0-9]+]], %got(d2)(${{[0-9]+}})
@@ -174,8 +174,8 @@ entry:
 ; Function Attrs: nounwind
 define void @dlt1()  {
 entry:
-  %0 = load double* @d1, align 8
-  %1 = load double* @d2, align 8
+  %0 = load double, double* @d1, align 8
+  %1 = load double, double* @d2, align 8
   %cmp = fcmp olt double %0, %1
 ; CHECK-LABEL:  dlt1:
 ; CHECK-DAG:    lw      $[[REG_D2_GOT:[0-9]+]], %got(d2)(${{[0-9]+}})
@@ -194,8 +194,8 @@ entry:
 ; Function Attrs: nounwind
 define void @dgt1()  {
 entry:
-  %0 = load double* @d1, align 8
-  %1 = load double* @d2, align 8
+  %0 = load double, double* @d1, align 8
+  %1 = load double, double* @d2, align 8
   %cmp = fcmp ogt double %0, %1
 ; CHECK-LABEL:  dgt1:
 ; CHECK-DAG:    lw      $[[REG_D2_GOT:[0-9]+]], %got(d2)(${{[0-9]+}})
@@ -214,8 +214,8 @@ entry:
 ; Function Attrs: nounwind
 define void @dle1()  {
 entry:
-  %0 = load double* @d1, align 8
-  %1 = load double* @d2, align 8
+  %0 = load double, double* @d1, align 8
+  %1 = load double, double* @d2, align 8
   %cmp = fcmp ole double %0, %1
 ; CHECK-LABEL:  dle1:
 ; CHECK-DAG:    lw      $[[REG_D2_GOT:[0-9]+]], %got(d2)(${{[0-9]+}})
@@ -234,8 +234,8 @@ entry:
 ; Function Attrs: nounwind
 define void @dge1()  {
 entry:
-  %0 = load double* @d1, align 8
-  %1 = load double* @d2, align 8
+  %0 = load double, double* @d1, align 8
+  %1 = load double, double* @d2, align 8
   %cmp = fcmp oge double %0, %1
 ; CHECK-LABEL:  dge1:
 ; CHECK-DAG:    lw      $[[REG_D2_GOT:[0-9]+]], %got(d2)(${{[0-9]+}})
diff --git a/test/CodeGen/Mips/Fast-ISel/fpext.ll b/test/CodeGen/Mips/Fast-ISel/fpext.ll
index 98aca756c58f..5ac22490ff02 100644
--- a/test/CodeGen/Mips/Fast-ISel/fpext.ll
+++ b/test/CodeGen/Mips/Fast-ISel/fpext.ll
@@ -1,6 +1,6 @@
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
 ; RUN:     < %s | FileCheck %s
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32 \
 ; RUN:     < %s | FileCheck %s
 
 @f = global float 0x40147E6B80000000, align 4
@@ -10,7 +10,7 @@
 ; Function Attrs: nounwind
 define void @dv() #0 {
 entry:
-  %0 = load float* @f, align 4
+  %0 = load float, float* @f, align 4
   %conv = fpext float %0 to double
 ; CHECK: cvt.d.s  $f{{[0-9]+}}, $f{{[0-9]+}}
   store double %conv, double* @d_f, align 8
diff --git a/test/CodeGen/Mips/Fast-ISel/fpintconv.ll b/test/CodeGen/Mips/Fast-ISel/fpintconv.ll
index 846726a868b3..a94ef5081539 100644
--- a/test/CodeGen/Mips/Fast-ISel/fpintconv.ll
+++ b/test/CodeGen/Mips/Fast-ISel/fpintconv.ll
@@ -1,6 +1,6 @@
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
 ; RUN:     < %s | FileCheck %s
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32 \
 ; RUN:     < %s | FileCheck %s
 
 
@@ -14,7 +14,7 @@
 define void @ifv() {
 entry:
 ; CHECK-LABEL:   .ent  ifv
-  %0 = load float* @f, align 4
+  %0 = load float, float* @f, align 4
   %conv = fptosi float %0 to i32
 ; CHECK:   trunc.w.s  $f[[REG:[0-9]+]], $f{{[0-9]+}}
 ; CHECK:   mfc1	${{[0-9]+}}, $f[[REG]]
@@ -26,7 +26,7 @@ entry:
 define void @idv() {
 entry:
 ; CHECK-LABEL:   .ent  idv
-  %0 = load double* @d, align 8
+  %0 = load double, double* @d, align 8
   %conv = fptosi double %0 to i32
 ; CHECK:   trunc.w.d  $f[[REG:[0-9]+]], $f{{[0-9]+}}
 ; CHECK:   mfc1	${{[0-9]+}}, $f[[REG]]
diff --git a/test/CodeGen/Mips/Fast-ISel/fptrunc.ll b/test/CodeGen/Mips/Fast-ISel/fptrunc.ll
index d843dee5a8c9..2eec4c3ef547 100644
--- a/test/CodeGen/Mips/Fast-ISel/fptrunc.ll
+++ b/test/CodeGen/Mips/Fast-ISel/fptrunc.ll
@@ -1,6 +1,6 @@
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
 ; RUN:     < %s | FileCheck %s
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32 \
 ; RUN:     < %s | FileCheck %s
 
 @d = global double 0x40147E6B74DF0446, align 8
@@ -10,7 +10,7 @@
 ; Function Attrs: nounwind
 define void @fv() #0 {
 entry:
-  %0 = load double* @d, align 8
+  %0 = load double, double* @d, align 8
   %conv = fptrunc double %0 to float
 ; CHECK: cvt.s.d  $f{{[0-9]+}}, $f{{[0-9]+}}
   store float %conv, float* @f, align 4
diff --git a/test/CodeGen/Mips/Fast-ISel/icmpa.ll b/test/CodeGen/Mips/Fast-ISel/icmpa.ll
index bd41a2911dc4..670a8d5cfb4e 100644
--- a/test/CodeGen/Mips/Fast-ISel/icmpa.ll
+++ b/test/CodeGen/Mips/Fast-ISel/icmpa.ll
@@ -1,6 +1,6 @@
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
 ; RUN:     < %s | FileCheck %s
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32 \
 ; RUN:     < %s | FileCheck %s
 
 @c = global i32 4, align 4
@@ -14,8 +14,8 @@ define void @eq()  {
 entry:
 ; CHECK-LABEL:  .ent  eq
 
-  %0 = load i32* @c, align 4
-  %1 = load i32* @d, align 4
+  %0 = load i32, i32* @c, align 4
+  %1 = load i32, i32* @d, align 4
   %cmp = icmp eq i32 %0, %1
   %conv = zext i1 %cmp to i32
 ; CHECK-DAG:  lw	$[[REG_D_GOT:[0-9+]]], %got(d)(${{[0-9]+}})
@@ -35,8 +35,8 @@ entry:
 define void @ne()  {
 entry:
 ; CHECK-LABEL:  .ent  ne
-  %0 = load i32* @c, align 4
-  %1 = load i32* @d, align 4
+  %0 = load i32, i32* @c, align 4
+  %1 = load i32, i32* @d, align 4
   %cmp = icmp ne i32 %0, %1
   %conv = zext i1 %cmp to i32
 ; CHECK-DAG:  lw	$[[REG_D_GOT:[0-9+]]], %got(d)(${{[0-9]+}})
@@ -56,8 +56,8 @@ entry:
 define void @ugt()  {
 entry:
 ; CHECK-LABEL:  .ent  ugt
-  %0 = load i32* @uc, align 4
-  %1 = load i32* @ud, align 4
+  %0 = load i32, i32* @uc, align 4
+  %1 = load i32, i32* @ud, align 4
   %cmp = icmp ugt i32 %0, %1
   %conv = zext i1 %cmp to i32
 ; CHECK-DAG:  lw	$[[REG_UD_GOT:[0-9+]]], %got(ud)(${{[0-9]+}})
@@ -76,8 +76,8 @@ entry:
 define void @ult()  {
 entry:
 ; CHECK-LABEL:  .ent  ult
-  %0 = load i32* @uc, align 4
-  %1 = load i32* @ud, align 4
+  %0 = load i32, i32* @uc, align 4
+  %1 = load i32, i32* @ud, align 4
   %cmp = icmp ult i32 %0, %1
   %conv = zext i1 %cmp to i32
 ; CHECK-DAG:  lw	$[[REG_UD_GOT:[0-9+]]], %got(ud)(${{[0-9]+}})
@@ -95,8 +95,8 @@ entry:
 define void @uge()  {
 entry:
 ; CHECK-LABEL:  .ent  uge
-  %0 = load i32* @uc, align 4
-  %1 = load i32* @ud, align 4
+  %0 = load i32, i32* @uc, align 4
+  %1 = load i32, i32* @ud, align 4
   %cmp = icmp uge i32 %0, %1
   %conv = zext i1 %cmp to i32
 ; CHECK-DAG:  lw	$[[REG_UD_GOT:[0-9+]]], %got(ud)(${{[0-9]+}})
@@ -115,8 +115,8 @@ entry:
 define void @ule()  {
 entry:
 ; CHECK-LABEL:  .ent  ule
-  %0 = load i32* @uc, align 4
-  %1 = load i32* @ud, align 4
+  %0 = load i32, i32* @uc, align 4
+  %1 = load i32, i32* @ud, align 4
   %cmp = icmp ule i32 %0, %1
   %conv = zext i1 %cmp to i32
 ; CHECK-DAG:  lw	$[[REG_UD_GOT:[0-9+]]], %got(ud)(${{[0-9]+}})
@@ -135,8 +135,8 @@ entry:
 define void @sgt()  {
 entry:
 ; CHECK-LABEL:  .ent sgt
-  %0 = load i32* @c, align 4
-  %1 = load i32* @d, align 4
+  %0 = load i32, i32* @c, align 4
+  %1 = load i32, i32* @d, align 4
   %cmp = icmp sgt i32 %0, %1
   %conv = zext i1 %cmp to i32
 ; CHECK-DAG:  lw	$[[REG_D_GOT:[0-9+]]], %got(d)(${{[0-9]+}})
@@ -154,8 +154,8 @@ entry:
 define void @slt()  {
 entry:
 ; CHECK-LABEL:  .ent slt
-  %0 = load i32* @c, align 4
-  %1 = load i32* @d, align 4
+  %0 = load i32, i32* @c, align 4
+  %1 = load i32, i32* @d, align 4
   %cmp = icmp slt i32 %0, %1
   %conv = zext i1 %cmp to i32
 ; CHECK-DAG:  lw	$[[REG_D_GOT:[0-9+]]], %got(d)(${{[0-9]+}})
@@ -173,8 +173,8 @@ entry:
 define void @sge()  {
 entry:
 ; CHECK-LABEL:  .ent sge
-  %0 = load i32* @c, align 4
-  %1 = load i32* @d, align 4
+  %0 = load i32, i32* @c, align 4
+  %1 = load i32, i32* @d, align 4
   %cmp = icmp sge i32 %0, %1
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @b1, align 4
@@ -193,8 +193,8 @@ entry:
 define void @sle()  {
 entry:
 ; CHECK-LABEL:  .ent sle
-  %0 = load i32* @c, align 4
-  %1 = load i32* @d, align 4
+  %0 = load i32, i32* @c, align 4
+  %1 = load i32, i32* @d, align 4
   %cmp = icmp sle i32 %0, %1
   %conv = zext i1 %cmp to i32
 ; CHECK-DAG:  lw	$[[REG_D_GOT:[0-9+]]], %got(d)(${{[0-9]+}})
diff --git a/test/CodeGen/Mips/Fast-ISel/loadstore2.ll b/test/CodeGen/Mips/Fast-ISel/loadstore2.ll
index d84478b9c5a9..3daf03d681cb 100644
--- a/test/CodeGen/Mips/Fast-ISel/loadstore2.ll
+++ b/test/CodeGen/Mips/Fast-ISel/loadstore2.ll
@@ -4,9 +4,9 @@ target triple = "mips--linux-gnu"
 
 @c2 = common global i8 0, align 1
 @c1 = common global i8 0, align 1
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
 ; RUN:     < %s | FileCheck %s
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32 \
 ; RUN:     < %s | FileCheck %s
 
 @s2 = common global i16 0, align 2
@@ -21,7 +21,7 @@ target triple = "mips--linux-gnu"
 ; Function Attrs: nounwind
 define void @cfoo() #0 {
 entry:
-  %0 = load i8* @c2, align 1
+  %0 = load i8, i8* @c2, align 1
   store i8 %0, i8* @c1, align 1
 ; CHECK-LABEL:	cfoo:
 ; CHECK:	lbu	$[[REGc:[0-9]+]], 0(${{[0-9]+}})
@@ -34,7 +34,7 @@ entry:
 ; Function Attrs: nounwind
 define void @sfoo() #0 {
 entry:
-  %0 = load i16* @s2, align 2
+  %0 = load i16, i16* @s2, align 2
   store i16 %0, i16* @s1, align 2
 ; CHECK-LABEL:	sfoo:
 ; CHECK:	lhu	$[[REGs:[0-9]+]], 0(${{[0-9]+}})
@@ -46,7 +46,7 @@ entry:
 ; Function Attrs: nounwind
 define void @ifoo() #0 {
 entry:
-  %0 = load i32* @i2, align 4
+  %0 = load i32, i32* @i2, align 4
   store i32 %0, i32* @i1, align 4
 ; CHECK-LABEL:	ifoo:
 ; CHECK:	lw	$[[REGi:[0-9]+]], 0(${{[0-9]+}})
@@ -58,7 +58,7 @@ entry:
 ; Function Attrs: nounwind
 define void @ffoo() #0 {
 entry:
-  %0 = load float* @f2, align 4
+  %0 = load float, float* @f2, align 4
   store float %0, float* @f1, align 4
 ; CHECK-LABEL:	ffoo:
 ; CHECK:	lwc1	$f[[REGf:[0-9]+]], 0(${{[0-9]+}})
@@ -71,7 +71,7 @@ entry:
 ; Function Attrs: nounwind
 define void @dfoo() #0 {
 entry:
-  %0 = load double* @d2, align 8
+  %0 = load double, double* @d2, align 8
   store double %0, double* @d1, align 8
 ; CHECK-LABEL:        dfoo:
 ; CHECK:        ldc1    $f[[REGd:[0-9]+]], 0(${{[0-9]+}})
diff --git a/test/CodeGen/Mips/Fast-ISel/loadstoreconv.ll b/test/CodeGen/Mips/Fast-ISel/loadstoreconv.ll
index f7f2c6481b3c..acba132b28e1 100644
--- a/test/CodeGen/Mips/Fast-ISel/loadstoreconv.ll
+++ b/test/CodeGen/Mips/Fast-ISel/loadstoreconv.ll
@@ -1,10 +1,10 @@
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
 ; RUN:     < %s | FileCheck %s
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32 \
 ; RUN:     < %s | FileCheck %s
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
 ; RUN:     < %s | FileCheck %s -check-prefix=mips32r2
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32 \
 ; RUN:     < %s | FileCheck %s -check-prefix=mips32
 
 @b2 = global i8 0, align 1
@@ -28,11 +28,11 @@
 define void @_Z3b_iv()  {
 entry:
 ; CHECK-LABEL:   .ent  _Z3b_iv
-  %0 = load i8* @b1, align 1
+  %0 = load i8, i8* @b1, align 1
   %tobool = trunc i8 %0 to i1
   %frombool = zext i1 %tobool to i8
   store i8 %frombool, i8* @b2, align 1
-  %1 = load i8* @b2, align 1
+  %1 = load i8, i8* @b2, align 1
   %tobool1 = trunc i8 %1 to i1
   %conv = zext i1 %tobool1 to i32
   store i32 %conv, i32* @i, align 4
@@ -51,10 +51,10 @@ define void @_Z4uc_iv()  {
 entry:
 ; CHECK-LABEL:  .ent  _Z4uc_iv
 
-  %0 = load i8* @uc1, align 1
+  %0 = load i8, i8* @uc1, align 1
   %conv = zext i8 %0 to i32
   store i32 %conv, i32* @i, align 4
-  %1 = load i8* @uc2, align 1
+  %1 = load i8, i8* @uc2, align 1
   %conv1 = zext i8 %1 to i32
 ; CHECK:   lbu  $[[REG1:[0-9]+]], 0(${{[0-9]+}})
 ; CHECK:  andi  ${{[0-9]+}}, $[[REG1]], 255
@@ -71,10 +71,10 @@ entry:
 ; mips32r2-LABEL:  .ent  _Z4sc_iv
 ; mips32-LABEL:  .ent  _Z4sc_iv
 
-  %0 = load i8* @sc1, align 1
+  %0 = load i8, i8* @sc1, align 1
   %conv = sext i8 %0 to i32
   store i32 %conv, i32* @i, align 4
-  %1 = load i8* @sc2, align 1
+  %1 = load i8, i8* @sc2, align 1
   %conv1 = sext i8 %1 to i32
   store i32 %conv1, i32* @j, align 4
 ; mips32r2:  lbu  $[[REG1:[0-9]+]], 0(${{[0-9]+}})
@@ -91,10 +91,10 @@ entry:
 define void @_Z4us_iv()  {
 entry:
 ; CHECK-LABEL:  .ent  _Z4us_iv
-  %0 = load i16* @us1, align 2
+  %0 = load i16, i16* @us1, align 2
   %conv = zext i16 %0 to i32
   store i32 %conv, i32* @i, align 4
-  %1 = load i16* @us2, align 2
+  %1 = load i16, i16* @us2, align 2
   %conv1 = zext i16 %1 to i32
   store i32 %conv1, i32* @j, align 4
   ret void
@@ -109,10 +109,10 @@ entry:
 ; mips32r2-LABEL:  .ent  _Z4ss_iv
 ; mips32=LABEL:  .ent  _Z4ss_iv
 
-  %0 = load i16* @ss1, align 2
+  %0 = load i16, i16* @ss1, align 2
   %conv = sext i16 %0 to i32
   store i32 %conv, i32* @i, align 4
-  %1 = load i16* @ss2, align 2
+  %1 = load i16, i16* @ss2, align 2
   %conv1 = sext i16 %1 to i32
   store i32 %conv1, i32* @j, align 4
 ; mips32r2:  lhu  $[[REG1:[0-9]+]], 0(${{[0-9]+}})
@@ -129,7 +129,7 @@ entry:
 define void @_Z4b_ssv()  {
 entry:
 ; CHECK-LABEL:  .ent  _Z4b_ssv
-  %0 = load i8* @b2, align 1
+  %0 = load i8, i8* @b2, align 1
   %tobool = trunc i8 %0 to i1
   %conv = zext i1 %tobool to i16
   store i16 %conv, i16* @ssi, align 2
@@ -143,10 +143,10 @@ entry:
 define void @_Z5uc_ssv()  {
 entry:
 ; CHECK-LABEL:  .ent  _Z5uc_ssv
-  %0 = load i8* @uc1, align 1
+  %0 = load i8, i8* @uc1, align 1
   %conv = zext i8 %0 to i16
   store i16 %conv, i16* @ssi, align 2
-  %1 = load i8* @uc2, align 1
+  %1 = load i8, i8* @uc2, align 1
   %conv1 = zext i8 %1 to i16
 ; CHECK:   lbu  $[[REG1:[0-9]+]], 0(${{[0-9]+}})
 ; CHECK:  andi  ${{[0-9]+}}, $[[REG1]], 255
@@ -161,10 +161,10 @@ define void @_Z5sc_ssv()  {
 entry:
 ; mips32r2-LABEL:  .ent  _Z5sc_ssv
 ; mips32-LABEL:  .ent  _Z5sc_ssv
-  %0 = load i8* @sc1, align 1
+  %0 = load i8, i8* @sc1, align 1
   %conv = sext i8 %0 to i16
   store i16 %conv, i16* @ssi, align 2
-  %1 = load i8* @sc2, align 1
+  %1 = load i8, i8* @sc2, align 1
   %conv1 = sext i8 %1 to i16
   store i16 %conv1, i16* @ssj, align 2
 ; mips32r2:  lbu  $[[REG1:[0-9]+]], 0(${{[0-9]+}})
diff --git a/test/CodeGen/Mips/Fast-ISel/loadstrconst.ll b/test/CodeGen/Mips/Fast-ISel/loadstrconst.ll
index 93cf4c15a2f5..9f644ecd1875 100644
--- a/test/CodeGen/Mips/Fast-ISel/loadstrconst.ll
+++ b/test/CodeGen/Mips/Fast-ISel/loadstrconst.ll
@@ -1,6 +1,6 @@
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
 ; RUN:     < %s | FileCheck %s
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32 \
 ; RUN:     < %s | FileCheck %s
 
 @.str = private unnamed_addr constant [6 x i8] c"hello\00", align 1
@@ -9,7 +9,7 @@
 ; Function Attrs: nounwind
 define void @foo() #0 {
 entry:
-  store i8* getelementptr inbounds ([6 x i8]* @.str, i32 0, i32 0), i8** @s, align 4
+  store i8* getelementptr inbounds ([6 x i8], [6 x i8]* @.str, i32 0, i32 0), i8** @s, align 4
   ret void
 ; CHECK:        .ent    foo
 ; CHECK:        lw      $[[REG1:[0-9]+]], %got($.str)(${{[0-9]+}})
diff --git a/test/CodeGen/Mips/Fast-ISel/logopm.ll b/test/CodeGen/Mips/Fast-ISel/logopm.ll
new file mode 100644
index 000000000000..0f0c3bf9e1dc
--- /dev/null
+++ b/test/CodeGen/Mips/Fast-ISel/logopm.ll
@@ -0,0 +1,606 @@
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -fast-isel -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 < %s | FileCheck %s
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -fast-isel -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32 < %s | FileCheck %s
+
+@ub1 = common global i8 0, align 1
+@ub2 = common global i8 0, align 1
+@ub3 = common global i8 0, align 1
+@uc1 = common global i8 0, align 1
+@uc2 = common global i8 0, align 1
+@uc3 = common global i8 0, align 1
+@us1 = common global i16 0, align 2
+@us2 = common global i16 0, align 2
+@us3 = common global i16 0, align 2
+@ub = common global i8 0, align 1
+@uc = common global i8 0, align 1
+@us = common global i16 0, align 2
+@.str = private unnamed_addr constant [4 x i8] c"%i\0A\00", align 1
+@ui = common global i32 0, align 4
+@ui1 = common global i32 0, align 4
+@ui2 = common global i32 0, align 4
+@ui3 = common global i32 0, align 4
+
+; Function Attrs: noinline nounwind
+define void @andUb() #0 {
+entry:
+  %0 = load i8, i8* @ub1, align 1
+  %1 = load i8, i8* @ub2, align 1
+  %conv0 = trunc i8 %0 to i1
+  %conv1 = trunc i8 %1 to i1
+  %and0 = and i1 %conv1, %conv0
+  %conv3 = zext i1 %and0 to i8
+  store i8 %conv3, i8* @ub, align 1, !tbaa !2
+; CHECK-LABEL:  .ent    andUb
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UB_ADDR:[0-9]+]], %got(ub)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UB2_ADDR:[0-9]+]], %got(ub2)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UB1_ADDR:[0-9]+]], %got(ub1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UB1:[0-9]+]], 0($[[UB1_ADDR]])
+; CHECK-DAG:    lbu     $[[UB2:[0-9]+]], 0($[[UB2_ADDR]])
+; CHECK-DAG:    and     $[[RES1:[0-9]+]], $[[UB2]], $[[UB1]]
+; CHECK:        andi    $[[RES:[0-9]+]], $[[RES1]], 1
+; CHECK:        sb      $[[RES]], 0($[[UB_ADDR]])
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @andUb0() #0 {
+entry:
+  %0 = load i8, i8* @ub1, align 1, !tbaa !2
+  %conv = trunc i8 %0 to i1
+  %and = and i1 %conv, 0
+  %conv1 = zext i1 %and to i8
+  store i8 %conv1, i8* @ub, align 1, !tbaa !2
+; CHECK-LABEL:  .ent    andUb0
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UB_ADDR:[0-9]+]], %got(ub)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UB1_ADDR:[0-9]+]], %got(ub1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UB1:[0-9]+]], 0($[[UB1_ADDR]])
+; CHECK-DAG:    and     $[[RES1:[0-9]+]], $[[UB1]], $zero
+; CHECK:        andi    $[[RES:[0-9]+]], $[[RES1]], 1
+; CHECK:        sb      $[[RES]], 0($[[UB_ADDR]])
+; CHECK:        .end    andUb0
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @andUb1() #0 {
+; clang uses i8 constants for booleans, so we test with an i8 1.
+entry:
+  %x = load i8, i8* @ub1, align 1, !tbaa !2
+  %and = and i8 %x, 1
+  %conv = trunc i8 %and to i1
+  %conv1 = zext i1 %conv to i8
+  store i8 %conv1, i8* @ub, align 1, !tbaa !2
+; CHECK-LABEL:  .ent    andUb1
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UB_ADDR:[0-9]+]], %got(ub)($[[REG_GP]])
+; CHECK-DAG:    addiu   $[[CONST:[0-9]+]], $zero, 1
+; CHECK-DAG:    lw      $[[UB1_ADDR:[0-9]+]], %got(ub1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UB1:[0-9]+]], 0($[[UB1_ADDR]])
+; CHECK-DAG:    and     $[[RES1:[0-9]+]], $[[UB1]], $[[CONST]]
+; CHECK:        andi    $[[RES:[0-9]+]], $[[RES1]], 1
+; CHECK:        sb      $[[RES]], 0($[[UB_ADDR]])
+; CHECK:        .end    andUb1
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @orUb() #0 {
+entry:
+  %0 = load i8, i8* @ub1, align 1
+  %1 = load i8, i8* @ub2, align 1
+  %conv0 = trunc i8 %0 to i1
+  %conv1 = trunc i8 %1 to i1
+  %or0 = or i1 %conv1, %conv0
+  %conv3 = zext i1 %or0 to i8
+  store i8 %conv3, i8* @ub, align 1, !tbaa !2
+; CHECK-LABEL:  .ent    orUb
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UB_ADDR:[0-9]+]], %got(ub)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UB2_ADDR:[0-9]+]], %got(ub2)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UB1_ADDR:[0-9]+]], %got(ub1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UB1:[0-9]+]], 0($[[UB1_ADDR]])
+; CHECK-DAG:    lbu     $[[UB2:[0-9]+]], 0($[[UB2_ADDR]])
+; CHECK-DAG:    or      $[[RES1:[0-9]+]], $[[UB2]], $[[UB1]]
+; CHECK:        andi    $[[RES:[0-9]+]], $[[RES1]], 1
+; CHECK:        sb      $[[RES]], 0($[[UB_ADDR]])
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @orUb0() #0 {
+entry:
+  %0 = load i8, i8* @ub1, align 1, !tbaa !2
+  %conv = trunc i8 %0 to i1
+  %or = or i1 %conv, 0
+  %conv1 = zext i1 %or to i8
+  store i8 %conv1, i8* @ub, align 1, !tbaa !2
+; CHECK-LABEL:  .ent    orUb0
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UB_ADDR:[0-9]+]], %got(ub)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UB1_ADDR:[0-9]+]], %got(ub1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UB1:[0-9]+]], 0($[[UB1_ADDR]])
+; CHECK:        andi    $[[RES:[0-9]+]], $[[UB1]], 1
+; CHECK:        sb      $[[RES]], 0($[[UB_ADDR]])
+; CHECK:        .end    orUb0
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @orUb1() #0 {
+entry:
+  %x = load i8, i8* @ub1, align 1, !tbaa !2
+  %or = or i8 %x, 1
+  %conv = trunc i8 %or to i1
+  %conv1 = zext i1 %conv to i8
+  store i8 %conv1, i8* @ub, align 1, !tbaa !2
+; CHECK-LABEL:  .ent    orUb1
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UB_ADDR:[0-9]+]], %got(ub)($[[REG_GP]])
+; CHECK-DAG:    addiu   $[[CONST:[0-9]+]], $zero, 1
+; CHECK-DAG:    lw      $[[UB1_ADDR:[0-9]+]], %got(ub1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UB1:[0-9]+]], 0($[[UB1_ADDR]])
+; CHECK-DAG:    or      $[[RES1:[0-9]+]], $[[UB1]], $[[CONST]]
+; CHECK:        andi    $[[RES:[0-9]+]], $[[RES1]], 1
+; CHECK:        sb      $[[RES]], 0($[[UB_ADDR]])
+; CHECK:        .end    orUb1
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @xorUb() #0 {
+entry:
+  %0 = load i8, i8* @ub1, align 1
+  %1 = load i8, i8* @ub2, align 1
+  %conv0 = trunc i8 %0 to i1
+  %conv1 = trunc i8 %1 to i1
+  %xor0 = xor i1 %conv1, %conv0
+  %conv3 = zext i1 %xor0 to i8
+  store i8 %conv3, i8* @ub, align 1, !tbaa !2
+; CHECK-LABEL: .ent    xorUb
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UB_ADDR:[0-9]+]], %got(ub)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UB2_ADDR:[0-9]+]], %got(ub2)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UB1_ADDR:[0-9]+]], %got(ub1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UB1:[0-9]+]], 0($[[UB1_ADDR]])
+; CHECK-DAG:    lbu     $[[UB2:[0-9]+]], 0($[[UB2_ADDR]])
+; CHECK-DAG:    xor     $[[RES1:[0-9]+]], $[[UB2]], $[[UB1]]
+; CHECK:        andi    $[[RES:[0-9]+]], $[[RES1]], 1
+; CHECK:        sb      $[[RES]], 0($[[UB_ADDR]])
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @xorUb0() #0 {
+entry:
+  %0 = load i8, i8* @ub1, align 1, !tbaa !2
+  %conv = trunc i8 %0 to i1
+  %xor = xor i1 %conv, 0
+  %conv1 = zext i1 %xor to i8
+  store i8 %conv1, i8* @ub, align 1, !tbaa !2
+; CHECK-LABEL:  .ent    xorUb0
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UB_ADDR:[0-9]+]], %got(ub)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UB1_ADDR:[0-9]+]], %got(ub1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UB1:[0-9]+]], 0($[[UB1_ADDR]])
+; CHECK-DAG:    xor     $[[RES1:[0-9]+]], $[[UB1]], $zero
+; CHECK:        andi    $[[RES:[0-9]+]], $[[RES1]], 1
+; CHECK:        sb      $[[RES]], 0($[[UB_ADDR]])
+; CHECK:        .end    xorUb0
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @xorUb1() #0 {
+entry:
+  %x = load i8, i8* @ub1, align 1, !tbaa !2
+  %xor = xor i8 1, %x
+  %conv = trunc i8 %xor to i1
+  %conv1 = zext i1 %conv to i8
+  store i8 %conv1, i8* @ub, align 1, !tbaa !2
+; CHECK-LABEL:  .ent    xorUb1
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UB_ADDR:[0-9]+]], %got(ub)($[[REG_GP]])
+; CHECK-DAG:    addiu   $[[CONST:[0-9]+]], $zero, 1
+; CHECK-DAG:    lw      $[[UB1_ADDR:[0-9]+]], %got(ub1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UB1:[0-9]+]], 0($[[UB1_ADDR]])
+; CHECK-DAG:    xor     $[[RES1:[0-9]+]], $[[UB1]], $[[CONST]]
+; CHECK:        andi    $[[RES:[0-9]+]], $[[RES1]], 1
+; CHECK:        sb      $[[RES]], 0($[[UB_ADDR]])
+; CHECK:        .end    xorUb1
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @andUc() #0 {
+entry:
+  %0 = load i8, i8* @uc1, align 1, !tbaa !2
+  %1 = load i8, i8* @uc2, align 1, !tbaa !2
+  %and3 = and i8 %1, %0
+  store i8 %and3, i8* @uc, align 1, !tbaa !2
+; CHECK-LABEL:  .ent    andUc
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UC_ADDR:[0-9]+]], %got(uc)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UC2_ADDR:[0-9]+]], %got(uc2)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UC1_ADDR:[0-9]+]], %got(uc1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UC1:[0-9]+]], 0($[[UC1_ADDR]])
+; CHECK-DAG:    lbu     $[[UC2:[0-9]+]], 0($[[UC2_ADDR]])
+; CHECK-DAG:    and     $[[RES:[0-9]+]], $[[UC2]], $[[UB1]]
+; CHECK:        sb      $[[RES]], 0($[[UC_ADDR]])
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @andUc0() #0 {
+entry:
+  %0 = load i8, i8* @uc1, align 1, !tbaa !2
+  %and = and i8 %0, 67
+  store i8 %and, i8* @uc, align 1, !tbaa !2
+; CHECK-LABEL:  .ent    andUc0
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UC_ADDR:[0-9]+]], %got(uc)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UC1_ADDR:[0-9]+]], %got(uc1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UC1:[0-9]+]], 0($[[UC1_ADDR]])
+; CHECK-DAG:    addiu   $[[CONST_67:[0-9]+]], $zero, 67
+; CHECK-DAG:    and     $[[RES:[0-9]+]], $[[UC1]], $[[CONST_67]]
+; CHECK:        sb      $[[RES]], 0($[[UC_ADDR]])
+; CHECK:        .end    andUc0
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @andUc1() #0 {
+entry:
+  %0 = load i8, i8* @uc1, align 1, !tbaa !2
+  %and = and i8 %0, 167
+  store i8 %and, i8* @uc, align 1, !tbaa !2
+; CHECK-LABEL:  .ent    andUc1
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UC_ADDR:[0-9]+]], %got(uc)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UC1_ADDR:[0-9]+]], %got(uc1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UC1:[0-9]+]], 0($[[UC1_ADDR]])
+; CHECK-DAG:    addiu   $[[CONST_Neg89:[0-9]+]], $zero, -89
+; CHECK-DAG:    and     $[[RES:[0-9]+]], $[[UC1]], $[[CONST_Neg89]]
+; CHECK:        sb      $[[RES]], 0($[[UC_ADDR]])
+; CHECK:        .end    andUc1
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @orUc() #0 {
+entry:
+  %0 = load i8, i8* @uc1, align 1, !tbaa !2
+  %1 = load i8, i8* @uc2, align 1, !tbaa !2
+  %or3 = or i8 %1, %0
+  store i8 %or3, i8* @uc, align 1, !tbaa !2
+; CHECK-LABEL:  .ent    orUc
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UC_ADDR:[0-9]+]], %got(uc)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UC2_ADDR:[0-9]+]], %got(uc2)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UC1_ADDR:[0-9]+]], %got(uc1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UC1:[0-9]+]], 0($[[UC1_ADDR]])
+; CHECK-DAG:    lbu     $[[UC2:[0-9]+]], 0($[[UC2_ADDR]])
+; CHECK-DAG:    or      $[[RES:[0-9]+]], $[[UC2]], $[[UC1]]
+; CHECK:        sb      $[[RES]], 0($[[UC_ADDR]])
+; CHECK:        .end    orUc
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @orUc0() #0 {
+entry:
+  %0 = load i8, i8* @uc1, align 1, !tbaa !2
+   %or = or i8 %0, 69
+  store i8 %or, i8* @uc, align 1, !tbaa !2
+; CHECK-LABEL:  .ent    orUc0
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UC_ADDR:[0-9]+]], %got(uc)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UC1_ADDR:[0-9]+]], %got(uc1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UC1:[0-9]+]], 0($[[UC1_ADDR]])
+; CHECK-DAG:    addiu   $[[CONST_69:[0-9]+]], $zero, 69
+; CHECK-DAG:    or      $[[RES:[0-9]+]], $[[UC1]], $[[CONST_69]]
+; CHECK:        sb      $[[RES]], 0($[[UC_ADDR]])
+; CHECK:        .end    orUc0
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @orUc1() #0 {
+entry:
+  %0 = load i8, i8* @uc1, align 1, !tbaa !2
+  %or = or i8 %0, 238
+  store i8 %or, i8* @uc, align 1, !tbaa !2
+; CHECK-LABEL:  .ent    orUc1
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UC_ADDR:[0-9]+]], %got(uc)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UC1_ADDR:[0-9]+]], %got(uc1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UC1:[0-9]+]], 0($[[UC1_ADDR]])
+; CHECK-DAG:    addiu   $[[CONST_neg18:[0-9]+]], $zero, -18
+; CHECK-DAG:    or      $[[RES:[0-9]+]], $[[UC1]], $[[CONST_neg18]]
+; CHECK:        sb      $[[RES]], 0($[[UC_ADDR]])
+; CHECK:        .end    orUc1
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @xorUc() #0 {
+entry:
+  %0 = load i8, i8* @uc1, align 1, !tbaa !2
+  %1 = load i8, i8* @uc2, align 1, !tbaa !2
+  %xor3 = xor i8 %1, %0
+  store i8 %xor3, i8* @uc, align 1, !tbaa !2
+; CHECK-LABEL: .ent    xorUc
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UC_ADDR:[0-9]+]], %got(uc)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UC2_ADDR:[0-9]+]], %got(uc2)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UC1_ADDR:[0-9]+]], %got(uc1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UC1:[0-9]+]], 0($[[UC1_ADDR]])
+; CHECK-DAG:    lbu     $[[UC2:[0-9]+]], 0($[[UC2_ADDR]])
+; CHECK-DAG:    xor     $[[RES:[0-9]+]], $[[UC2]], $[[UC1]]
+; CHECK:        sb      $[[RES]], 0($[[UC_ADDR]])
+; CHECK:        .end    xorUc
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @xorUc0() #0 {
+entry:
+  %0 = load i8, i8* @uc1, align 1, !tbaa !2
+  %xor = xor i8 %0, 23
+  store i8 %xor, i8* @uc, align 1, !tbaa !2
+; CHECK-LABEL:  .ent    xorUc0
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UC_ADDR:[0-9]+]], %got(uc)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UC1_ADDR:[0-9]+]], %got(uc1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UC1:[0-9]+]], 0($[[UC1_ADDR]])
+; CHECK-DAG:    addiu   $[[CONST_23:[0-9]+]], $zero, 23
+; CHECK-DAG:    xor     $[[RES:[0-9]+]], $[[UC1]], $[[CONST_23]]
+; CHECK:        sb      $[[RES]], 0($[[UC_ADDR]])
+; CHECK:        .end    xorUc0
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @xorUc1() #0 {
+entry:
+  %0 = load i8, i8* @uc1, align 1, !tbaa !2
+  %xor = xor i8 %0, 120
+  store i8 %xor, i8* @uc, align 1, !tbaa !2
+; CHECK-LABEL:  .ent    xorUc1
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[UC_ADDR:[0-9]+]], %got(uc)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[UC1_ADDR:[0-9]+]], %got(uc1)($[[REG_GP]])
+; CHECK-DAG:    lbu     $[[UC1:[0-9]+]], 0($[[UC1_ADDR]])
+; CHECK-DAG:    addiu   $[[CONST_120:[0-9]+]], $zero, 120
+; CHECK-DAG:    xor     $[[RES:[0-9]+]], $[[UC1]], $[[CONST_120]]
+; CHECK:        sb      $[[RES]], 0($[[UC_ADDR]])
+; CHECK:        .end    xorUc1
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @andUs() #0 {
+entry:
+  %0 = load i16, i16* @us1, align 2, !tbaa !5
+  %1 = load i16, i16* @us2, align 2, !tbaa !5
+  %and3 = and i16 %1, %0
+  store i16 %and3, i16* @us, align 2, !tbaa !5
+; CHECK-LABEL:  .ent    andUs
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[US_ADDR:[0-9]+]], %got(us)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[US2_ADDR:[0-9]+]], %got(us2)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[US1_ADDR:[0-9]+]], %got(us1)($[[REG_GP]])
+; CHECK-DAG:    lhu     $[[US1:[0-9]+]], 0($[[US1_ADDR]])
+; CHECK-DAG:    lhu     $[[US2:[0-9]+]], 0($[[US2_ADDR]])
+; CHECK-DAG:    and     $[[RES:[0-9]+]], $[[US2]], $[[UB1]]
+; CHECK:        sh      $[[RES]], 0($[[US_ADDR]])
+; CHECK:        .end andUs
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @andUs0() #0 {
+entry:
+  %0 = load i16, i16* @us1, align 2, !tbaa !5
+  %and = and i16 %0, 4660
+  store i16 %and, i16* @us, align 2, !tbaa !5
+; CHECK-LABEL: .ent    andUs0
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[US_ADDR:[0-9]+]], %got(us)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[US1_ADDR:[0-9]+]], %got(us1)($[[REG_GP]])
+; CHECK-DAG:    lhu     $[[US1:[0-9]+]], 0($[[US1_ADDR]])
+; CHECK-DAG:    addiu   $[[CONST_4660:[0-9]+]], $zero, 4660
+; CHECK-DAG:    and     $[[RES:[0-9]+]], $[[US1]], $[[CONST_4660]]
+; CHECK:        sh      $[[RES]], 0($[[US_ADDR]])
+; CHECK:        .end    andUs0
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @andUs1() #0 {
+entry:
+  %0 = load i16, i16* @us1, align 2, !tbaa !5
+  %and = and i16 %0, 61351
+  store i16 %and, i16* @us, align 2, !tbaa !5
+; CHECK-LABEL:  .ent    andUs1
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[US_ADDR:[0-9]+]], %got(us)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[US1_ADDR:[0-9]+]], %got(us1)($[[REG_GP]])
+; CHECK-DAG:    lhu     $[[US1:[0-9]+]], 0($[[US1_ADDR]])
+; CHECK-DAG:    addiu   $[[CONST_Neg4185:[0-9]+]], $zero, -4185
+; CHECK-DAG:    and     $[[RES:[0-9]+]], $[[US1]], $[[CONST_Neg4185]]
+; CHECK:        sh      $[[RES]], 0($[[US_ADDR]])
+; CHECK:        .end    andUs1
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @orUs() #0 {
+entry:
+  %0 = load i16, i16* @us1, align 2, !tbaa !5
+  %1 = load i16, i16* @us2, align 2, !tbaa !5
+  %or3 = or i16 %1, %0
+  store i16 %or3, i16* @us, align 2, !tbaa !5
+; CHECK-LABEL:  .ent    orUs
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[US_ADDR:[0-9]+]], %got(us)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[US2_ADDR:[0-9]+]], %got(us2)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[US1_ADDR:[0-9]+]], %got(us1)($[[REG_GP]])
+; CHECK-DAG:    lhu     $[[US1:[0-9]+]], 0($[[US1_ADDR]])
+; CHECK-DAG:    lhu     $[[US2:[0-9]+]], 0($[[US2_ADDR]])
+; CHECK-DAG:    or      $[[RES:[0-9]+]], $[[US2]], $[[US1]]
+; CHECK:        sh      $[[RES]], 0($[[US_ADDR]])
+; CHECK:        .end    orUs
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @orUs0() #0 {
+entry:
+  %0 = load i16, i16* @us1, align 2, !tbaa !5
+  %or = or i16 %0, 17666
+  store i16 %or, i16* @us, align 2, !tbaa !5
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @orUs1() #0 {
+entry:
+  %0 = load i16, i16* @us1, align 2, !tbaa !5
+  %or = or i16 %0, 60945
+  store i16 %or, i16* @us, align 2, !tbaa !5
+; CHECK-LABEL:  .ent    orUs1
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[US_ADDR:[0-9]+]], %got(us)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[US1_ADDR:[0-9]+]], %got(us1)($[[REG_GP]])
+; CHECK-DAG:    lhu     $[[US1:[0-9]+]], 0($[[US1_ADDR]])
+; CHECK-DAG:    addiu   $[[CONST_neg4591:[0-9]+]], $zero, -4591
+; CHECK-DAG:    or      $[[RES:[0-9]+]], $[[US1]], $[[CONST_neg4591]]
+; CHECK:        sh      $[[RES]], 0($[[US_ADDR]])
+; CHECK:        .end    orUs1
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @xorUs() #0 {
+entry:
+  %0 = load i16, i16* @us1, align 2, !tbaa !5
+  %1 = load i16, i16* @us2, align 2, !tbaa !5
+  %xor3 = xor i16 %1, %0
+  store i16 %xor3, i16* @us, align 2, !tbaa !5
+; CHECK-LABEL:  .ent    xorUs
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[US_ADDR:[0-9]+]], %got(us)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[US2_ADDR:[0-9]+]], %got(us2)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[US1_ADDR:[0-9]+]], %got(us1)($[[REG_GP]])
+; CHECK-DAG:    lhu     $[[US1:[0-9]+]], 0($[[US1_ADDR]])
+; CHECK-DAG:    lhu     $[[US2:[0-9]+]], 0($[[US2_ADDR]])
+; CHECK-DAG:    xor     $[[RES:[0-9]+]], $[[US2]], $[[US1]]
+; CHECK:        sh      $[[RES]], 0($[[US_ADDR]])
+; CHECK:        .end    xorUs
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @xorUs0() #0 {
+entry:
+  %0 = load i16, i16* @us1, align 2, !tbaa !5
+  %xor = xor i16 %0, 6062
+  store i16 %xor, i16* @us, align 2, !tbaa !5
+; CHECK-LABEL:  .ent    xorUs0
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[US_ADDR:[0-9]+]], %got(us)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[US1_ADDR:[0-9]+]], %got(us1)($[[REG_GP]])
+; CHECK-DAG:    lhu     $[[US1:[0-9]+]], 0($[[US1_ADDR]])
+; CHECK-DAG:    addiu   $[[CONST_6062:[0-9]+]], $zero, 6062
+; CHECK-DAG:    xor     $[[RES:[0-9]+]], $[[US1]], $[[CONST_6062]]
+; CHECK:        sh      $[[RES]], 0($[[US_ADDR]])
+; CHECK:        .end    xorUs0
+
+  ret void
+}
+
+; Function Attrs: noinline nounwind
+define void @xorUs1() #0 {
+entry:
+  %0 = load i16, i16* @us1, align 2, !tbaa !5
+  %xor = xor i16 %0, 60024
+  store i16 %xor, i16* @us, align 2, !tbaa !5
+; CHECK-LABEL:  .ent    xorUs1
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[US_ADDR:[0-9]+]], %got(us)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[US1_ADDR:[0-9]+]], %got(us1)($[[REG_GP]])
+; CHECK-DAG:    lhu     $[[US1:[0-9]+]], 0($[[US1_ADDR]])
+; CHECK-DAG:    addiu   $[[CONST_Neg5512:[0-9]+]], $zero, -5512
+; CHECK-DAG:    xor     $[[RES:[0-9]+]], $[[US1]], $[[CONST_Neg5512]]
+; CHECK:        sh      $[[RES]], 0($[[US_ADDR]])
+; CHECK:        .end    xorUs1
+  ret void
+}
+
+attributes #0 = { noinline nounwind "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
+attributes #1 = { nounwind "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
+attributes #2 = { nounwind }
+
+!llvm.module.flags = !{!0}
+!llvm.ident = !{!1}
+
+!0 = !{i32 1, !"PIC Level", i32 2}
+!1 = !{!"clang version 3.7.0 (trunk)"}
+!2 = !{!3, !3, i64 0}
+!3 = !{!"omnipotent char", !4, i64 0}
+!4 = !{!"Simple C/C++ TBAA"}
+!5 = !{!6, !6, i64 0}
+!6 = !{!"short", !3, i64 0}
diff --git a/test/CodeGen/Mips/Fast-ISel/nullvoid.ll b/test/CodeGen/Mips/Fast-ISel/nullvoid.ll
index c847561d0278..5fa3f13ace4c 100644
--- a/test/CodeGen/Mips/Fast-ISel/nullvoid.ll
+++ b/test/CodeGen/Mips/Fast-ISel/nullvoid.ll
@@ -1,6 +1,6 @@
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
 ; RUN:     < %s | FileCheck %s
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32 \
 ; RUN:     < %s | FileCheck %s
 
 ; Function Attrs: nounwind
diff --git a/test/CodeGen/Mips/Fast-ISel/overflt.ll b/test/CodeGen/Mips/Fast-ISel/overflt.ll
new file mode 100644
index 000000000000..57f991e23d95
--- /dev/null
+++ b/test/CodeGen/Mips/Fast-ISel/overflt.ll
@@ -0,0 +1,64 @@
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
+; RUN:     < %s | FileCheck %s
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32 \
+; RUN:     < %s | FileCheck %s
+
+@x = common global [128000 x float] zeroinitializer, align 4
+@y = global float* getelementptr inbounds ([128000 x float], [128000 x float]* @x, i32 0, i32 0), align 4
+@result = common global float 0.000000e+00, align 4
+@.str = private unnamed_addr constant [5 x i8] c"%f \0A\00", align 1
+
+; Function Attrs: nounwind
+define void @foo() {
+entry:
+; CHECK-LABEL:   .ent  foo
+  %0 = load float*, float** @y, align 4
+  %arrayidx = getelementptr inbounds float, float* %0, i32 64000
+  store float 5.500000e+00, float* %arrayidx, align 4
+; CHECK:        lui     $[[REG_FPCONST_INT:[0-9]+]], 16560
+; CHECK:        mtc1    $[[REG_FPCONST_INT]], $f[[REG_FPCONST:[0-9]+]]
+; CHECK:        lw      $[[REG_Y_GOT:[0-9]+]], %got(y)(${{[0-9]+}})
+; CHECK:        lw      $[[REG_Y:[0-9]+]], 0($[[REG_Y_GOT]])
+; CHECK:        lui     $[[REG_IDX_UPPER:[0-9]+]], 3
+; CHECK:        ori     $[[REG_IDX:[0-9]+]], $[[REG_IDX_UPPER]], 59392
+; CHECK:        addu    $[[REG_Y_IDX:[0-9]+]], $[[REG_IDX]], $[[REG_Y]]
+; CHECK:        swc1    $f[[REG_FPCONST]], 0($[[REG_Y_IDX]])
+  ret void
+; CHECK-LABEL:   .end  foo
+}
+
+; Function Attrs: nounwind
+define void @goo() {
+entry:
+; CHECK-LABEL:   .ent  goo
+  %0 = load float*, float** @y, align 4
+  %arrayidx = getelementptr inbounds float, float* %0, i32 64000
+  %1 = load float, float* %arrayidx, align 4
+  store float %1, float* @result, align 4
+; CHECK-DAG:    lw      $[[REG_RESULT:[0-9]+]], %got(result)(${{[0-9]+}})
+; CHECK-DAG:    lw      $[[REG_Y_GOT:[0-9]+]], %got(y)(${{[0-9]+}})
+; CHECK-DAG:    lw      $[[REG_Y:[0-9]+]], 0($[[REG_Y_GOT]])
+; CHECK-DAG:    lui     $[[REG_IDX_UPPER:[0-9]+]], 3
+; CHECK-DAG:    ori     $[[REG_IDX:[0-9]+]], $[[REG_IDX_UPPER]], 59392
+; CHECK-DAG:    addu    $[[REG_Y_IDX:[0-9]+]], $[[REG_IDX]], $[[REG_Y]]
+; CHECK-DAG:    lwc1    $f[[Y_IDX:[0-9]+]], 0($[[REG_Y_IDX]])
+; CHECK-DAG:    swc1    $f[[Y_IDX]], 0($[[REG_RESULT]])
+; CHECK-LABEL:   .end  goo
+  ret void
+}
+
+; 
+; Original C code for test.
+;
+;float x[128000];
+;float *y = x;
+;float result;
+
+
+;void foo() {
+;  y[64000] = 5.5;
+;}
+
+;void goo() {
+;  result = y[64000];
+;}
diff --git a/test/CodeGen/Mips/Fast-ISel/retabi.ll b/test/CodeGen/Mips/Fast-ISel/retabi.ll
new file mode 100644
index 000000000000..03119b827eb6
--- /dev/null
+++ b/test/CodeGen/Mips/Fast-ISel/retabi.ll
@@ -0,0 +1,108 @@
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
+; RUN:     < %s | FileCheck %s
+
+@i = global i32 75, align 4
+@s = global i16 -345, align 2
+@c = global i8 118, align 1
+@f = global float 0x40BE623360000000, align 4
+@d = global double 1.298330e+03, align 8
+
+; Function Attrs: nounwind
+define i32 @reti() {
+entry:
+; CHECK-LABEL: reti:
+  %0 = load i32, i32* @i, align 4
+  ret i32 %0
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK:        lw      $[[REG_I_ADDR:[0-9]+]], %got(i)($[[REG_GP]])
+; CHECK:        lw      $2, 0($[[REG_I_ADDR]])
+; CHECK:        jr      $ra
+}
+
+; Function Attrs: nounwind
+define i16 @retus() {
+entry:
+; CHECK-LABEL: retus:
+  %0 = load i16, i16* @s, align 2
+  ret i16 %0
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK:        lw      $[[REG_S_ADDR:[0-9]+]], %got(s)($[[REG_GP]])
+; CHECK:        lhu     $2, 0($[[REG_S_ADDR]])
+; CHECK:        jr      $ra
+}
+
+; Function Attrs: nounwind
+define signext i16 @rets() {
+entry:
+; CHECK-LABEL: rets:
+  %0 = load i16, i16* @s, align 2
+  ret i16 %0
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK:        lw      $[[REG_S_ADDR:[0-9]+]], %got(s)($[[REG_GP]])
+; CHECK:        lhu     $[[REG_S:[0-9]+]], 0($[[REG_S_ADDR]])
+; CHECK:        seh     $2, $[[REG_S]]
+; CHECK:        jr      $ra
+}
+
+; Function Attrs: nounwind
+define i8 @retuc() {
+entry:
+; CHECK-LABEL: retuc:
+  %0 = load i8, i8* @c, align 1
+  ret i8 %0
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK:        lw      $[[REG_C_ADDR:[0-9]+]], %got(c)($[[REG_GP]])
+; CHECK:        lbu     $2, 0($[[REG_C_ADDR]])
+; CHECK:        jr      $ra
+}
+
+; Function Attrs: nounwind
+define signext i8 @retc() {
+entry:
+; CHECK-LABEL: retc:
+  %0 = load i8, i8* @c, align 1
+  ret i8 %0
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK:        lw      $[[REG_C_ADDR:[0-9]+]], %got(c)($[[REG_GP]])
+; CHECK:        lbu     $[[REG_C:[0-9]+]], 0($[[REG_C_ADDR]])
+; CHECK:        seb     $2, $[[REG_C]]
+; CHECK:        jr      $ra
+}
+
+; Function Attrs: nounwind
+define float @retf() {
+entry:
+; CHECK-LABEL: retf:
+  %0 = load float, float* @f, align 4
+  ret float %0
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK:        lw      $[[REG_F_ADDR:[0-9]+]], %got(f)($[[REG_GP]])
+; CHECK:        lwc1    $f0, 0($[[REG_F_ADDR]])
+; CHECK:        jr      $ra
+}
+
+; Function Attrs: nounwind
+define double @retd() {
+entry:
+; CHECK-LABEL: retd:
+  %0 = load double, double* @d, align 8
+  ret double %0
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK:        addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK:        addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK:        lw      $[[REG_D_ADDR:[0-9]+]], %got(d)($[[REG_GP]])
+; CHECK:        ldc1    $f0, 0($[[REG_D_ADDR]])
+; CHECK:        jr      $ra
+}
diff --git a/test/CodeGen/Mips/Fast-ISel/shftopm.ll b/test/CodeGen/Mips/Fast-ISel/shftopm.ll
new file mode 100644
index 000000000000..90ddd190be13
--- /dev/null
+++ b/test/CodeGen/Mips/Fast-ISel/shftopm.ll
@@ -0,0 +1,122 @@
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel \
+; RUN:     -fast-isel-abort=1 -mcpu=mips32r2  < %s | FileCheck %s
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel \
+; RUN:     -fast-isel-abort=1 -mcpu=mips32 < %s | FileCheck %s
+
+@s1 = global i16 -89, align 2
+@s2 = global i16 4, align 2
+@us1 = global i16 -503, align 2
+@us2 = global i16 5, align 2
+@s3 = common global i16 0, align 2
+@us3 = common global i16 0, align 2
+
+define void @sll() {
+entry:
+  %0 = load i16, i16* @s1, align 2
+  %1 = load i16, i16* @s2, align 2
+  %shl = shl i16 %0, %1
+  store i16 %shl, i16* @s3, align 2
+; CHECK-LABEL:  sll:
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK-DAG:    addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK-DAG:    addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[S3_ADDR:[0-9]+]], %got(s3)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[S2_ADDR:[0-9]+]], %got(s2)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[S1_ADDR:[0-9]+]], %got(s1)($[[REG_GP]])
+; CHECK-DAG:    lhu     $[[S1:[0-9]+]], 0($[[S1_ADDR]])
+; CHECK-DAG:    lhu     $[[S2:[0-9]+]], 0($[[S2_ADDR]])
+; CHECK:        sllv    $[[RES:[0-9]+]], $[[S1]], $[[S2]]
+; CHECK:        sh      $[[RES]], 0($[[S3_ADDR]])
+  ret void
+}
+
+define void @slli() {
+entry:
+  %0 = load i16, i16* @s1, align 2
+  %shl = shl i16 %0, 5
+  store i16 %shl, i16* @s3, align 2
+; CHECK-LABEL:  slli:
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK-DAG:    addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK-DAG:    addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[S3_ADDR:[0-9]+]], %got(s3)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[S1_ADDR:[0-9]+]], %got(s1)($[[REG_GP]])
+; CHECK-DAG:    lhu     $[[S1:[0-9]+]], 0($[[S1_ADDR]])
+; CHECK:        sll     $[[RES:[0-9]+]], $[[S1]],  5
+; CHECK:        sh      $[[RES]], 0($[[S3_ADDR]])
+  ret void
+}
+
+define void @srl() {
+entry:
+  %0 = load i16, i16* @us1, align 2
+  %1 = load i16, i16* @us2, align 2
+  %shr = lshr i16 %0, %1
+  store i16 %shr, i16* @us3, align 2
+  ret void
+; CHECK-LABEL:  srl:
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK-DAG:    addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK-DAG:    addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[US3_ADDR:[0-9]+]], %got(us3)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[US2_ADDR:[0-9]+]], %got(us2)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[US1_ADDR:[0-9]+]], %got(us1)($[[REG_GP]])
+; CHECK-DAG:    lhu     $[[US1:[0-9]+]], 0($[[US1_ADDR]])
+; CHECK-DAG:    lhu     $[[US2:[0-9]+]], 0($[[US2_ADDR]])
+; CHECK:        srlv    $[[RES:[0-9]+]], $[[US1]], $[[US2]]
+; CHECK:        sh      $[[RES]], 0($[[S3_ADDR]])
+}
+
+define void @srli() {
+entry:
+  %0 = load i16, i16* @us1, align 2
+  %shr = lshr i16 %0, 4
+  store i16 %shr, i16* @us3, align 2
+; CHECK-LABEL:  srli:
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK-DAG:    addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK-DAG:    addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[US3_ADDR:[0-9]+]], %got(us3)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[US1_ADDR:[0-9]+]], %got(us1)($[[REG_GP]])
+; CHECK-DAG:    lhu     $[[US1:[0-9]+]], 0($[[US1_ADDR]])
+; CHECK:        srl     $[[RES:[0-9]+]], $[[US1]], 4
+; CHECK:        sh      $[[RES]], 0($[[S3_ADDR]])
+  ret void
+}
+
+define void @sra() {
+entry:
+  %0 = load i16, i16* @s1, align 2
+  %1 = load i16, i16* @s2, align 2
+  %shr = ashr i16 %0, %1
+  store i16 %shr, i16* @s3, align 2
+; CHECK-LABEL:  sra:
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK-DAG:    addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK-DAG:    addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[S3_ADDR:[0-9]+]], %got(s3)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[S2_ADDR:[0-9]+]], %got(s2)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[S1_ADDR:[0-9]+]], %got(s1)($[[REG_GP]])
+; CHECK-DAG:    lhu     $[[S1:[0-9]+]], 0($[[S1_ADDR]])
+; CHECK-DAG:    lhu     $[[S2:[0-9]+]], 0($[[S2_ADDR]])
+; CHECK:        srav    $[[RES:[0-9]+]], $[[S1]], $[[S2]]
+; CHECK:        sh      $[[RES]], 0($[[S3_ADDR]])
+  ret void
+}
+
+define void @srai() {
+entry:
+  %0 = load i16, i16* @s1, align 2
+  %shr = ashr i16 %0, 2
+  store i16 %shr, i16* @s3, align 2
+; CHECK-LABEL:  srai:
+; CHECK:        lui     $[[REG_GPa:[0-9]+]], %hi(_gp_disp)
+; CHECK-DAG:    addiu   $[[REG_GPb:[0-9]+]], $[[REG_GPa]], %lo(_gp_disp)
+; CHECK-DAG:    addu    $[[REG_GP:[0-9]+]], $[[REG_GPb]], $25
+; CHECK-DAG:    lw      $[[S3_ADDR:[0-9]+]], %got(s3)($[[REG_GP]])
+; CHECK-DAG:    lw      $[[S1_ADDR:[0-9]+]], %got(s1)($[[REG_GP]])
+; CHECK-DAG:    lhu     $[[S1:[0-9]+]], 0($[[S1_ADDR]])
+; CHECK:        sra     $[[RES:[0-9]+]], $[[S1]], 2
+; CHECK:        sh      $[[RES]], 0($[[S3_ADDR]])
+  ret void
+}
diff --git a/test/CodeGen/Mips/Fast-ISel/shift.ll b/test/CodeGen/Mips/Fast-ISel/shift.ll
index 18fd5ac32d22..df1c82700d59 100644
--- a/test/CodeGen/Mips/Fast-ISel/shift.ll
+++ b/test/CodeGen/Mips/Fast-ISel/shift.ll
@@ -9,7 +9,7 @@ define i32 @main() nounwind uwtable {
 entry:
   %foo = alloca %struct.s, align 4
   %0 = bitcast %struct.s* %foo to i32*
-  %bf.load = load i32* %0, align 4
+  %bf.load = load i32, i32* %0, align 4
   %bf.lshr = lshr i32 %bf.load, 2
   %cmp = icmp ne i32 %bf.lshr, 2
   br i1 %cmp, label %if.then, label %if.end
diff --git a/test/CodeGen/Mips/Fast-ISel/simplestore.ll b/test/CodeGen/Mips/Fast-ISel/simplestore.ll
index 83e3f3f24274..bcb198b1a823 100644
--- a/test/CodeGen/Mips/Fast-ISel/simplestore.ll
+++ b/test/CodeGen/Mips/Fast-ISel/simplestore.ll
@@ -1,6 +1,6 @@
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
 ; RUN:     < %s | FileCheck %s
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32 \
 ; RUN:     < %s | FileCheck %s
 
 @abcd = external global i32
diff --git a/test/CodeGen/Mips/Fast-ISel/simplestorefp1.ll b/test/CodeGen/Mips/Fast-ISel/simplestorefp1.ll
index 74723ae1beeb..f4b91d850255 100644
--- a/test/CodeGen/Mips/Fast-ISel/simplestorefp1.ll
+++ b/test/CodeGen/Mips/Fast-ISel/simplestorefp1.ll
@@ -1,10 +1,10 @@
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
 ; RUN:     < %s | FileCheck %s 
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32 \
 ; RUN:     < %s | FileCheck %s
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
 ; RUN:     < %s | FileCheck %s -check-prefix=mips32r2 
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32 \
 ; RUN:     < %s | FileCheck %s -check-prefix=mips32
 
 @f = common global float 0.000000e+00, align 4
diff --git a/test/CodeGen/Mips/Fast-ISel/simplestorei.ll b/test/CodeGen/Mips/Fast-ISel/simplestorei.ll
index 128e1de9cad0..83ccae0b1de5 100644
--- a/test/CodeGen/Mips/Fast-ISel/simplestorei.ll
+++ b/test/CodeGen/Mips/Fast-ISel/simplestorei.ll
@@ -1,6 +1,6 @@
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32r2 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32r2 \
 ; RUN:     < %s | FileCheck %s
-; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort -mcpu=mips32 \
+; RUN: llc -march=mipsel -relocation-model=pic -O0 -mips-fast-isel -fast-isel-abort=1 -mcpu=mips32 \
 ; RUN:     < %s | FileCheck %s
 
 @ijk = external global i32
diff --git a/test/CodeGen/Mips/abiflags32.ll b/test/CodeGen/Mips/abiflags32.ll
index e32d4a586ee3..39e2a90151e3 100644
--- a/test/CodeGen/Mips/abiflags32.ll
+++ b/test/CodeGen/Mips/abiflags32.ll
@@ -1,6 +1,6 @@
 ; RUN: llc -filetype=asm -mtriple mipsel-unknown-linux -mcpu=mips32 %s -o - | FileCheck %s
 ; RUN: llc -filetype=asm -mtriple mipsel-unknown-linux -mcpu=mips32 -mattr=fp64 %s -o - | FileCheck  -check-prefix=CHECK-64 %s
-; RUN: llc -filetype=asm -mtriple mipsel-unknown-linux -mcpu=mips64 -mattr=-n64,n32 %s -o - | FileCheck  -check-prefix=CHECK-64n %s
+; RUN: llc -filetype=asm -mtriple mipsel-unknown-linux -mcpu=mips64 -target-abi n32 %s -o - | FileCheck  -check-prefix=CHECK-64n %s
 
 ; CHECK: .nan    legacy
 ; We don't emit '.module fp=32' for compatibility with binutils 2.24 which
diff --git a/test/CodeGen/Mips/addi.ll b/test/CodeGen/Mips/addi.ll
index 01d409e521d7..b6af2ee45687 100644
--- a/test/CodeGen/Mips/addi.ll
+++ b/test/CodeGen/Mips/addi.ll
@@ -8,16 +8,16 @@
 
 define void @foo() nounwind {
 entry:
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %add = add nsw i32 %0, 5
   store i32 %add, i32* @i, align 4
-  %1 = load i32* @j, align 4
+  %1 = load i32, i32* @j, align 4
   %sub = sub nsw i32 %1, 5
   store i32 %sub, i32* @j, align 4
-  %2 = load i32* @k, align 4
+  %2 = load i32, i32* @k, align 4
   %add1 = add nsw i32 %2, 10000
   store i32 %add1, i32* @k, align 4
-  %3 = load i32* @l, align 4
+  %3 = load i32, i32* @l, align 4
   %sub2 = sub nsw i32 %3, 10000
   store i32 %sub2, i32* @l, align 4
 ; 16: 	addiu	${{[0-9]+}}, 5	# 16 bit inst
diff --git a/test/CodeGen/Mips/addressing-mode.ll b/test/CodeGen/Mips/addressing-mode.ll
index ea76dde82dc3..81e062062ecf 100644
--- a/test/CodeGen/Mips/addressing-mode.ll
+++ b/test/CodeGen/Mips/addressing-mode.ll
@@ -20,10 +20,10 @@ for.cond1.preheader:
 for.body3:
   %s.120 = phi i32 [ %s.022, %for.cond1.preheader ], [ %add7, %for.body3 ]
   %j.019 = phi i32 [ 0, %for.cond1.preheader ], [ %add8, %for.body3 ]
-  %arrayidx4 = getelementptr inbounds [256 x i32]* %a, i32 %i.021, i32 %j.019
-  %0 = load i32* %arrayidx4, align 4
-  %arrayidx6 = getelementptr inbounds [256 x i32]* %b, i32 %i.021, i32 %j.019
-  %1 = load i32* %arrayidx6, align 4
+  %arrayidx4 = getelementptr inbounds [256 x i32], [256 x i32]* %a, i32 %i.021, i32 %j.019
+  %0 = load i32, i32* %arrayidx4, align 4
+  %arrayidx6 = getelementptr inbounds [256 x i32], [256 x i32]* %b, i32 %i.021, i32 %j.019
+  %1 = load i32, i32* %arrayidx6, align 4
   %add = add i32 %0, %s.120
   %add7 = add i32 %add, %1
   %add8 = add nsw i32 %j.019, %m
diff --git a/test/CodeGen/Mips/align16.ll b/test/CodeGen/Mips/align16.ll
index 689ae8307f57..f385adfaa04c 100644
--- a/test/CodeGen/Mips/align16.ll
+++ b/test/CodeGen/Mips/align16.ll
@@ -15,13 +15,13 @@ entry:
   %x = alloca i32, align 8
   %zz = alloca i32, align 4
   %z = alloca i32, align 4
-  %0 = load i32* @i, align 4
-  %arrayidx = getelementptr inbounds [512 x i32]* %y, i32 0, i32 10
+  %0 = load i32, i32* @i, align 4
+  %arrayidx = getelementptr inbounds [512 x i32], [512 x i32]* %y, i32 0, i32 10
   store i32 %0, i32* %arrayidx, align 4
-  %1 = load i32* @i, align 4
+  %1 = load i32, i32* @i, align 4
   store i32 %1, i32* %x, align 8
   call void @p(i32* %x)
-  %arrayidx1 = getelementptr inbounds [512 x i32]* %y, i32 0, i32 10
+  %arrayidx1 = getelementptr inbounds [512 x i32], [512 x i32]* %y, i32 0, i32 10
   call void @p(i32* %arrayidx1)
   ret void
 }
diff --git a/test/CodeGen/Mips/alloca.ll b/test/CodeGen/Mips/alloca.ll
index fc7ef862a328..747a1362161d 100644
--- a/test/CodeGen/Mips/alloca.ll
+++ b/test/CodeGen/Mips/alloca.ll
@@ -9,7 +9,7 @@ entry:
 ; CHECK: move  $4, $[[T0]]
 ; CHECK: move  $4, $[[T2]]
   %tmp1 = alloca i8, i32 %size, align 4
-  %add.ptr = getelementptr inbounds i8* %tmp1, i32 5
+  %add.ptr = getelementptr inbounds i8, i8* %tmp1, i32 5
   store i8 97, i8* %add.ptr, align 1
   %tmp4 = alloca i8, i32 %size, align 4
   call void @foo2(double 1.000000e+00, double 2.000000e+00, i32 3) nounwind
@@ -39,17 +39,17 @@ entry:
 if.then:                                          ; preds = %entry
 ; CHECK: addiu $4, $[[T0]], 40
 
-  %add.ptr = getelementptr inbounds i8* %tmp1, i32 40
+  %add.ptr = getelementptr inbounds i8, i8* %tmp1, i32 40
   %1 = bitcast i8* %add.ptr to i32*
   call void @foo3(i32* %1) nounwind
-  %arrayidx15.pre = getelementptr inbounds i8* %tmp1, i32 12
+  %arrayidx15.pre = getelementptr inbounds i8, i8* %tmp1, i32 12
   %.pre = bitcast i8* %arrayidx15.pre to i32*
   br label %if.end
 
 if.else:                                          ; preds = %entry
 ; CHECK: addiu $4, $[[T0]], 12
 
-  %add.ptr5 = getelementptr inbounds i8* %tmp1, i32 12
+  %add.ptr5 = getelementptr inbounds i8, i8* %tmp1, i32 12
   %2 = bitcast i8* %add.ptr5 to i32*
   call void @foo3(i32* %2) nounwind
   br label %if.end
@@ -59,24 +59,24 @@ if.end:                                           ; preds = %if.else, %if.then
 ; CHECK: lw  $25, %call16(printf)
 
   %.pre-phi = phi i32* [ %2, %if.else ], [ %.pre, %if.then ]
-  %tmp7 = load i32* %0, align 4
-  %arrayidx9 = getelementptr inbounds i8* %tmp1, i32 4
+  %tmp7 = load i32, i32* %0, align 4
+  %arrayidx9 = getelementptr inbounds i8, i8* %tmp1, i32 4
   %3 = bitcast i8* %arrayidx9 to i32*
-  %tmp10 = load i32* %3, align 4
-  %arrayidx12 = getelementptr inbounds i8* %tmp1, i32 8
+  %tmp10 = load i32, i32* %3, align 4
+  %arrayidx12 = getelementptr inbounds i8, i8* %tmp1, i32 8
   %4 = bitcast i8* %arrayidx12 to i32*
-  %tmp13 = load i32* %4, align 4
-  %tmp16 = load i32* %.pre-phi, align 4
-  %arrayidx18 = getelementptr inbounds i8* %tmp1, i32 16
+  %tmp13 = load i32, i32* %4, align 4
+  %tmp16 = load i32, i32* %.pre-phi, align 4
+  %arrayidx18 = getelementptr inbounds i8, i8* %tmp1, i32 16
   %5 = bitcast i8* %arrayidx18 to i32*
-  %tmp19 = load i32* %5, align 4
-  %arrayidx21 = getelementptr inbounds i8* %tmp1, i32 20
+  %tmp19 = load i32, i32* %5, align 4
+  %arrayidx21 = getelementptr inbounds i8, i8* %tmp1, i32 20
   %6 = bitcast i8* %arrayidx21 to i32*
-  %tmp22 = load i32* %6, align 4
-  %arrayidx24 = getelementptr inbounds i8* %tmp1, i32 24
+  %tmp22 = load i32, i32* %6, align 4
+  %arrayidx24 = getelementptr inbounds i8, i8* %tmp1, i32 24
   %7 = bitcast i8* %arrayidx24 to i32*
-  %tmp25 = load i32* %7, align 4
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([22 x i8]* @.str, i32 0, i32 0), i32 %tmp7, i32 %tmp10, i32 %tmp13, i32 %tmp16, i32 %tmp19, i32 %tmp22, i32 %tmp25) nounwind
+  %tmp25 = load i32, i32* %7, align 4
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([22 x i8], [22 x i8]* @.str, i32 0, i32 0), i32 %tmp7, i32 %tmp10, i32 %tmp13, i32 %tmp16, i32 %tmp19, i32 %tmp22, i32 %tmp25) nounwind
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/alloca16.ll b/test/CodeGen/Mips/alloca16.ll
index 4f6059878c3b..be8cc740310b 100644
--- a/test/CodeGen/Mips/alloca16.ll
+++ b/test/CodeGen/Mips/alloca16.ll
@@ -12,7 +12,7 @@ define void @temp(i32 %foo) nounwind {
 entry:
   %foo.addr = alloca i32, align 4
   store i32 %foo, i32* %foo.addr, align 4
-  %0 = load i32* %foo.addr, align 4
+  %0 = load i32, i32* %foo.addr, align 4
   store i32 %0, i32* @t, align 4
   ret void
 }
@@ -28,46 +28,46 @@ entry:
   %sssi = alloca i32, align 4
   %ip = alloca i32*, align 4
   %sssj = alloca i32, align 4
-  %0 = load i32* @iiii, align 4
+  %0 = load i32, i32* @iiii, align 4
   store i32 %0, i32* %sssi, align 4
-  %1 = load i32* @kkkk, align 4
+  %1 = load i32, i32* @kkkk, align 4
   %mul = mul nsw i32 %1, 100
   %2 = alloca i8, i32 %mul
   %3 = bitcast i8* %2 to i32*
   store i32* %3, i32** %ip, align 4
-  %4 = load i32* @jjjj, align 4
+  %4 = load i32, i32* @jjjj, align 4
   store i32 %4, i32* %sssj, align 4
-  %5 = load i32* @jjjj, align 4
-  %6 = load i32* @iiii, align 4
-  %7 = load i32** %ip, align 4
-  %arrayidx = getelementptr inbounds i32* %7, i32 %6
+  %5 = load i32, i32* @jjjj, align 4
+  %6 = load i32, i32* @iiii, align 4
+  %7 = load i32*, i32** %ip, align 4
+  %arrayidx = getelementptr inbounds i32, i32* %7, i32 %6
   store i32 %5, i32* %arrayidx, align 4
-  %8 = load i32* @kkkk, align 4
-  %9 = load i32* @jjjj, align 4
-  %10 = load i32** %ip, align 4
-  %arrayidx1 = getelementptr inbounds i32* %10, i32 %9
+  %8 = load i32, i32* @kkkk, align 4
+  %9 = load i32, i32* @jjjj, align 4
+  %10 = load i32*, i32** %ip, align 4
+  %arrayidx1 = getelementptr inbounds i32, i32* %10, i32 %9
   store i32 %8, i32* %arrayidx1, align 4
-  %11 = load i32* @iiii, align 4
-  %12 = load i32* @kkkk, align 4
-  %13 = load i32** %ip, align 4
-  %arrayidx2 = getelementptr inbounds i32* %13, i32 %12
+  %11 = load i32, i32* @iiii, align 4
+  %12 = load i32, i32* @kkkk, align 4
+  %13 = load i32*, i32** %ip, align 4
+  %arrayidx2 = getelementptr inbounds i32, i32* %13, i32 %12
   store i32 %11, i32* %arrayidx2, align 4
-  %14 = load i32** %ip, align 4
-  %arrayidx3 = getelementptr inbounds i32* %14, i32 25
-  %15 = load i32* %arrayidx3, align 4
+  %14 = load i32*, i32** %ip, align 4
+  %arrayidx3 = getelementptr inbounds i32, i32* %14, i32 25
+  %15 = load i32, i32* %arrayidx3, align 4
   store i32 %15, i32* @riii, align 4
-  %16 = load i32** %ip, align 4
-  %arrayidx4 = getelementptr inbounds i32* %16, i32 35
-  %17 = load i32* %arrayidx4, align 4
+  %16 = load i32*, i32** %ip, align 4
+  %arrayidx4 = getelementptr inbounds i32, i32* %16, i32 35
+  %17 = load i32, i32* %arrayidx4, align 4
   store i32 %17, i32* @rjjj, align 4
-  %18 = load i32** %ip, align 4
-  %arrayidx5 = getelementptr inbounds i32* %18, i32 100
-  %19 = load i32* %arrayidx5, align 4
+  %18 = load i32*, i32** %ip, align 4
+  %arrayidx5 = getelementptr inbounds i32, i32* %18, i32 100
+  %19 = load i32, i32* %arrayidx5, align 4
   store i32 %19, i32* @rkkk, align 4
-  %20 = load i32* @t, align 4
-  %21 = load i32** %ip, align 4
-  %arrayidx6 = getelementptr inbounds i32* %21, i32 %20
-  %22 = load i32* %arrayidx6, align 4
+  %20 = load i32, i32* @t, align 4
+  %21 = load i32*, i32** %ip, align 4
+  %arrayidx6 = getelementptr inbounds i32, i32* %21, i32 %20
+  %22 = load i32, i32* %arrayidx6, align 4
 ; 16: 	addiu $sp, -16
   call void @temp(i32 %22)
 ; 16: 	addiu $sp, 16
diff --git a/test/CodeGen/Mips/analyzebranch.ll b/test/CodeGen/Mips/analyzebranch.ll
index 4b5d09778d79..d5ecaaeddc33 100644
--- a/test/CodeGen/Mips/analyzebranch.ll
+++ b/test/CodeGen/Mips/analyzebranch.ll
@@ -60,7 +60,7 @@ if.then:                                          ; preds = %entry
   unreachable
 
 if.end:                                           ; preds = %entry
-  tail call void (...)* @f2() nounwind
+  tail call void (...) @f2() nounwind
   ret void
 }
 
diff --git a/test/CodeGen/Mips/and1.ll b/test/CodeGen/Mips/and1.ll
index 4ff1204fe7ae..57076a4d4fcf 100644
--- a/test/CodeGen/Mips/and1.ll
+++ b/test/CodeGen/Mips/and1.ll
@@ -6,11 +6,11 @@
 
 define i32 @main() nounwind {
 entry:
-  %0 = load i32* @x, align 4
-  %1 = load i32* @y, align 4
+  %0 = load i32, i32* @x, align 4
+  %1 = load i32, i32* @y, align 4
   %and = and i32 %0, %1
 ; 16:	and	${{[0-9]+}}, ${{[0-9]+}}
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([7 x i8]* @.str, i32 0, i32 0), i32 %and)
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([7 x i8], [7 x i8]* @.str, i32 0, i32 0), i32 %and)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/atomic.ll b/test/CodeGen/Mips/atomic.ll
index ccfeb00967e3..031cce0b6074 100644
--- a/test/CodeGen/Mips/atomic.ll
+++ b/test/CodeGen/Mips/atomic.ll
@@ -54,7 +54,7 @@ define i32 @AtomicSwap32(i32 signext %newval) nounwind {
 entry:
   %newval.addr = alloca i32, align 4
   store i32 %newval, i32* %newval.addr, align 4
-  %tmp = load i32* %newval.addr, align 4
+  %tmp = load i32, i32* %newval.addr, align 4
   %0 = atomicrmw xchg i32* @x, i32 %tmp monotonic
   ret i32 %0
 
@@ -74,7 +74,7 @@ define i32 @AtomicCmpSwap32(i32 signext %oldval, i32 signext %newval) nounwind {
 entry:
   %newval.addr = alloca i32, align 4
   store i32 %newval, i32* %newval.addr, align 4
-  %tmp = load i32* %newval.addr, align 4
+  %tmp = load i32, i32* %newval.addr, align 4
   %0 = cmpxchg i32* @x, i32 %oldval, i32 %tmp monotonic monotonic
   %1 = extractvalue { i32, i1 } %0, 0
   ret i32 %1
@@ -429,7 +429,7 @@ entry:
 ; FIXME: At the moment, we don't seem to do addr+offset for any atomic load/store.
 define i32 @AtomicLoadAdd32_OffGt9Bit(i32 signext %incr) nounwind {
 entry:
-  %0 = atomicrmw add i32* getelementptr(i32* @x, i32 256), i32 %incr monotonic
+  %0 = atomicrmw add i32* getelementptr(i32, i32* @x, i32 256), i32 %incr monotonic
   ret i32 %0
 
 ; ALL-LABEL: AtomicLoadAdd32_OffGt9Bit:
diff --git a/test/CodeGen/Mips/atomicops.ll b/test/CodeGen/Mips/atomicops.ll
index c26415233d0b..0ff9f5c22a84 100644
--- a/test/CodeGen/Mips/atomicops.ll
+++ b/test/CodeGen/Mips/atomicops.ll
@@ -18,15 +18,15 @@ entry:
   store volatile i32 0, i32* %x, align 4
   %0 = atomicrmw add i32* %x, i32 1 seq_cst
   %add.i = add nsw i32 %0, 2
-  %1 = load volatile i32* %x, align 4
-  %call1 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([8 x i8]* @.str, i32 0, i32 0), i32 %add.i, i32 %1) nounwind
+  %1 = load volatile i32, i32* %x, align 4
+  %call1 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([8 x i8], [8 x i8]* @.str, i32 0, i32 0), i32 %add.i, i32 %1) nounwind
   %pair = cmpxchg i32* %x, i32 1, i32 2 seq_cst seq_cst
   %2 = extractvalue { i32, i1 } %pair, 0
-  %3 = load volatile i32* %x, align 4
-  %call2 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([8 x i8]* @.str, i32 0, i32 0), i32 %2, i32 %3) nounwind
+  %3 = load volatile i32, i32* %x, align 4
+  %call2 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([8 x i8], [8 x i8]* @.str, i32 0, i32 0), i32 %2, i32 %3) nounwind
   %4 = atomicrmw xchg i32* %x, i32 1 seq_cst
-  %5 = load volatile i32* %x, align 4
-  %call3 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([8 x i8]* @.str, i32 0, i32 0), i32 %4, i32 %5) nounwind
+  %5 = load volatile i32, i32* %x, align 4
+  %call3 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([8 x i8], [8 x i8]* @.str, i32 0, i32 0), i32 %4, i32 %5) nounwind
 ; 16-LABEL: main:
 ; 16:	lw	${{[0-9]+}}, %call16(__sync_synchronize)(${{[0-9]+}})
 ; 16: 	lw	${{[0-9]+}}, %call16(__sync_fetch_and_add_4)(${{[0-9]+}})
diff --git a/test/CodeGen/Mips/beqzc.ll b/test/CodeGen/Mips/beqzc.ll
index 4a294c2d817e..37bece884212 100644
--- a/test/CodeGen/Mips/beqzc.ll
+++ b/test/CodeGen/Mips/beqzc.ll
@@ -1,4 +1,4 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=cond-b-short
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=cond-b-short
 
 @i = global i32 0, align 4
 @j = common global i32 0, align 4
@@ -6,7 +6,7 @@
 ; Function Attrs: nounwind optsize
 define i32 @main() #0 {
 entry:
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %cmp = icmp eq i32 %0, 0
   %. = select i1 %cmp, i32 10, i32 55
   store i32 %., i32* @j, align 4
diff --git a/test/CodeGen/Mips/beqzc1.ll b/test/CodeGen/Mips/beqzc1.ll
index 8f929a8e3541..1f5575f099fa 100644
--- a/test/CodeGen/Mips/beqzc1.ll
+++ b/test/CodeGen/Mips/beqzc1.ll
@@ -1,4 +1,4 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=cond-b-short
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=cond-b-short
 
 @i = global i32 0, align 4
 @j = common global i32 0, align 4
@@ -6,7 +6,7 @@
 ; Function Attrs: nounwind optsize
 define i32 @main() #0 {
 entry:
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %cmp = icmp eq i32 %0, 0
   br i1 %cmp, label %if.then, label %if.end
 
diff --git a/test/CodeGen/Mips/biggot.ll b/test/CodeGen/Mips/biggot.ll
index da287eea6fd1..b56ce6ba87b1 100644
--- a/test/CodeGen/Mips/biggot.ll
+++ b/test/CodeGen/Mips/biggot.ll
@@ -20,7 +20,7 @@ entry:
 ; N64: daddu  $[[R3:[0-9]+]], $[[R2]], ${{[a-z0-9]+}}
 ; N64: ld  ${{[0-9]+}}, %call_lo(foo0)($[[R3]])
 
-  %0 = load i32* @v0, align 4
+  %0 = load i32, i32* @v0, align 4
   tail call void @foo0(i32 %0) nounwind
   ret void
 }
diff --git a/test/CodeGen/Mips/blockaddr.ll b/test/CodeGen/Mips/blockaddr.ll
index d6dc7e7b24b0..f74363702af5 100644
--- a/test/CodeGen/Mips/blockaddr.ll
+++ b/test/CodeGen/Mips/blockaddr.ll
@@ -1,9 +1,9 @@
 ; RUN: llc -march=mipsel -relocation-model=pic < %s | FileCheck %s -check-prefix=PIC-O32
 ; RUN: llc -march=mipsel -relocation-model=static < %s | FileCheck %s -check-prefix=STATIC-O32
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=-n64,n32 -relocation-model=pic < %s | FileCheck %s -check-prefix=PIC-N32
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=-n64,n32 -relocation-model=static < %s | FileCheck %s -check-prefix=STATIC-N32
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=-n64,n64 -relocation-model=pic < %s | FileCheck %s -check-prefix=PIC-N64
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=-n64,n64 -relocation-model=static < %s | FileCheck %s -check-prefix=STATIC-N64
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi n32 -relocation-model=pic < %s | FileCheck %s -check-prefix=PIC-N32
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi n32 -relocation-model=static < %s | FileCheck %s -check-prefix=STATIC-N32
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi n64 -relocation-model=pic < %s | FileCheck %s -check-prefix=PIC-N64
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi n64 -relocation-model=static < %s | FileCheck %s -check-prefix=STATIC-N64
 ; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips32 -mattr=+mips16 -relocation-model=static < %s | FileCheck %s -check-prefix=STATIC-MIPS16-1
 ; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips32 -mattr=+mips16 -relocation-model=static < %s | FileCheck %s -check-prefix=STATIC-MIPS16-2
 
diff --git a/test/CodeGen/Mips/brconeq.ll b/test/CodeGen/Mips/brconeq.ll
index 613391557efd..f555528bbb64 100644
--- a/test/CodeGen/Mips/brconeq.ll
+++ b/test/CodeGen/Mips/brconeq.ll
@@ -6,8 +6,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @i, align 4
-  %1 = load i32* @j, align 4
+  %0 = load i32, i32* @i, align 4
+  %1 = load i32, i32* @j, align 4
   %cmp = icmp eq i32 %0, %1
 ; 16:	cmp	${{[0-9]+}}, ${{[0-9]+}}
 ; 16:	bteqz	$[[LABEL:[0-9A-Ba-b_]+]]
diff --git a/test/CodeGen/Mips/brconeqk.ll b/test/CodeGen/Mips/brconeqk.ll
index 2c0e72dabd29..59edae82e5ad 100644
--- a/test/CodeGen/Mips/brconeqk.ll
+++ b/test/CodeGen/Mips/brconeqk.ll
@@ -5,7 +5,7 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %cmp = icmp eq i32 %0, 10
   br i1 %cmp, label %if.end, label %if.then
 ; 16:	cmpi	${{[0-9]+}}, {{[0-9]+}}
diff --git a/test/CodeGen/Mips/brconeqz.ll b/test/CodeGen/Mips/brconeqz.ll
index 5586e7b976da..22c566407528 100644
--- a/test/CodeGen/Mips/brconeqz.ll
+++ b/test/CodeGen/Mips/brconeqz.ll
@@ -5,7 +5,7 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %cmp = icmp eq i32 %0, 0
   br i1 %cmp, label %if.end, label %if.then
 ; 16:	beqz	${{[0-9]+}}, $[[LABEL:[0-9A-Ba-b_]+]]
diff --git a/test/CodeGen/Mips/brconge.ll b/test/CodeGen/Mips/brconge.ll
index 02f0a633b313..46d19847d9bc 100644
--- a/test/CodeGen/Mips/brconge.ll
+++ b/test/CodeGen/Mips/brconge.ll
@@ -8,8 +8,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @i, align 4
-  %1 = load i32* @j, align 4
+  %0 = load i32, i32* @i, align 4
+  %1 = load i32, i32* @j, align 4
   %cmp = icmp slt i32 %0, %1
   br i1 %cmp, label %if.then, label %if.end
 
@@ -22,7 +22,7 @@ if.then:                                          ; preds = %entry
   br label %if.end
 
 if.end:                                           ; preds = %if.then, %entry
-  %2 = load i32* @k, align 4
+  %2 = load i32, i32* @k, align 4
   %cmp1 = icmp slt i32 %0, %2
   br i1 %cmp1, label %if.then2, label %if.end3
 
diff --git a/test/CodeGen/Mips/brcongt.ll b/test/CodeGen/Mips/brcongt.ll
index 767b51b21b91..cefacb8318b0 100644
--- a/test/CodeGen/Mips/brcongt.ll
+++ b/test/CodeGen/Mips/brcongt.ll
@@ -7,8 +7,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @i, align 4
-  %1 = load i32* @j, align 4
+  %0 = load i32, i32* @i, align 4
+  %1 = load i32, i32* @j, align 4
   %cmp = icmp sgt i32 %0, %1
   br i1 %cmp, label %if.end, label %if.then
 ; 16:	slt	${{[0-9]+}}, ${{[0-9]+}}
diff --git a/test/CodeGen/Mips/brconle.ll b/test/CodeGen/Mips/brconle.ll
index 854b2481c6e6..e1f15ecb6b92 100644
--- a/test/CodeGen/Mips/brconle.ll
+++ b/test/CodeGen/Mips/brconle.ll
@@ -8,8 +8,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @j, align 4
-  %1 = load i32* @i, align 4
+  %0 = load i32, i32* @j, align 4
+  %1 = load i32, i32* @i, align 4
   %cmp = icmp sgt i32 %0, %1
   br i1 %cmp, label %if.then, label %if.end
 
@@ -22,7 +22,7 @@ if.then:                                          ; preds = %entry
   br label %if.end
 
 if.end:                                           ; preds = %if.then, %entry
-  %2 = load i32* @k, align 4
+  %2 = load i32, i32* @k, align 4
   %cmp1 = icmp sgt i32 %1, %2
   br i1 %cmp1, label %if.then2, label %if.end3
 
diff --git a/test/CodeGen/Mips/brconlt.ll b/test/CodeGen/Mips/brconlt.ll
index 931a3e8c7ba4..049f35c393fe 100644
--- a/test/CodeGen/Mips/brconlt.ll
+++ b/test/CodeGen/Mips/brconlt.ll
@@ -7,8 +7,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @j, align 4
-  %1 = load i32* @i, align 4
+  %0 = load i32, i32* @j, align 4
+  %1 = load i32, i32* @i, align 4
   %cmp = icmp slt i32 %0, %1
   br i1 %cmp, label %if.end, label %if.then
 
diff --git a/test/CodeGen/Mips/brconne.ll b/test/CodeGen/Mips/brconne.ll
index 5d5bde3fcf91..b260320b94e1 100644
--- a/test/CodeGen/Mips/brconne.ll
+++ b/test/CodeGen/Mips/brconne.ll
@@ -6,8 +6,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @j, align 4
-  %1 = load i32* @i, align 4
+  %0 = load i32, i32* @j, align 4
+  %1 = load i32, i32* @i, align 4
   %cmp = icmp eq i32 %0, %1
   br i1 %cmp, label %if.then, label %if.end
 ; 16:	cmp	${{[0-9]+}}, ${{[0-9]+}}
diff --git a/test/CodeGen/Mips/brconnek.ll b/test/CodeGen/Mips/brconnek.ll
index 6208d7c5a04b..778a5cce72b3 100644
--- a/test/CodeGen/Mips/brconnek.ll
+++ b/test/CodeGen/Mips/brconnek.ll
@@ -5,7 +5,7 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @j, align 4
+  %0 = load i32, i32* @j, align 4
   %cmp = icmp eq i32 %0, 5
   br i1 %cmp, label %if.then, label %if.end
 
diff --git a/test/CodeGen/Mips/brconnez.ll b/test/CodeGen/Mips/brconnez.ll
index 47db7901b517..754714b21daf 100644
--- a/test/CodeGen/Mips/brconnez.ll
+++ b/test/CodeGen/Mips/brconnez.ll
@@ -5,7 +5,7 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @j, align 4
+  %0 = load i32, i32* @j, align 4
   %cmp = icmp eq i32 %0, 0
   br i1 %cmp, label %if.then, label %if.end
 
diff --git a/test/CodeGen/Mips/brdelayslot.ll b/test/CodeGen/Mips/brdelayslot.ll
index 68341c1ba25b..0f46619b8272 100644
--- a/test/CodeGen/Mips/brdelayslot.ll
+++ b/test/CodeGen/Mips/brdelayslot.ll
@@ -54,18 +54,18 @@ declare void @foo4(double)
 
 define void @foo5(i32 %a) nounwind {
 entry:
-  %0 = load i32* @g2, align 4
+  %0 = load i32, i32* @g2, align 4
   %tobool = icmp eq i32 %a, 0
   br i1 %tobool, label %if.else, label %if.then
 
 if.then:
-  %1 = load i32* @g1, align 4
+  %1 = load i32, i32* @g1, align 4
   %add = add nsw i32 %1, %0
   store i32 %add, i32* @g1, align 4
   br label %if.end
 
 if.else:
-  %2 = load i32* @g3, align 4
+  %2 = load i32, i32* @g3, align 4
   %sub = sub nsw i32 %2, %0
   store i32 %sub, i32* @g3, align 4
   br label %if.end
@@ -99,9 +99,9 @@ declare void @foo7(double, float)
 define i32 @foo8(i32 %a) nounwind {
 entry:
   store i32 %a, i32* @g1, align 4
-  %0 = load void ()** @foo9, align 4
+  %0 = load void ()*, void ()** @foo9, align 4
   tail call void %0() nounwind
-  %1 = load i32* @g1, align 4
+  %1 = load i32, i32* @g1, align 4
   %add = add nsw i32 %1, %a
   ret i32 %add
 }
@@ -144,8 +144,8 @@ entry:
 for.body:                                         ; preds = %entry, %for.body
   %s.06 = phi i32 [ %add, %for.body ], [ 0, %entry ]
   %i.05 = phi i32 [ %inc, %for.body ], [ 0, %entry ]
-  %arrayidx = getelementptr inbounds i32* %a, i32 %i.05
-  %0 = load i32* %arrayidx, align 4
+  %arrayidx = getelementptr inbounds i32, i32* %a, i32 %i.05
+  %0 = load i32, i32* %arrayidx, align 4
   %add = add nsw i32 %0, %s.06
   %inc = add nsw i32 %i.05, 1
   %exitcond = icmp eq i32 %inc, %n
diff --git a/test/CodeGen/Mips/brind.ll b/test/CodeGen/Mips/brind.ll
index 4c591fa1bba1..a3e9b8011a2b 100644
--- a/test/CodeGen/Mips/brind.ll
+++ b/test/CodeGen/Mips/brind.ll
@@ -9,29 +9,29 @@
 
 define i32 @main() nounwind {
 entry:
-  %puts = tail call i32 @puts(i8* getelementptr inbounds ([2 x i8]* @str, i32 0, i32 0))
+  %puts = tail call i32 @puts(i8* getelementptr inbounds ([2 x i8], [2 x i8]* @str, i32 0, i32 0))
   br label %L1
 
 L1:                                               ; preds = %entry, %L3
   %i.0 = phi i32 [ 0, %entry ], [ %inc, %L3 ]
-  %puts5 = tail call i32 @puts(i8* getelementptr inbounds ([2 x i8]* @str5, i32 0, i32 0))
+  %puts5 = tail call i32 @puts(i8* getelementptr inbounds ([2 x i8], [2 x i8]* @str5, i32 0, i32 0))
   br label %L2
 
 L2:                                               ; preds = %L1, %L3
   %i.1 = phi i32 [ %i.0, %L1 ], [ %inc, %L3 ]
-  %puts6 = tail call i32 @puts(i8* getelementptr inbounds ([2 x i8]* @str6, i32 0, i32 0))
+  %puts6 = tail call i32 @puts(i8* getelementptr inbounds ([2 x i8], [2 x i8]* @str6, i32 0, i32 0))
   br label %L3
 
 L3:                                               ; preds = %L2, %L3
   %i.2 = phi i32 [ %i.1, %L2 ], [ %inc, %L3 ]
-  %puts7 = tail call i32 @puts(i8* getelementptr inbounds ([2 x i8]* @str7, i32 0, i32 0))
+  %puts7 = tail call i32 @puts(i8* getelementptr inbounds ([2 x i8], [2 x i8]* @str7, i32 0, i32 0))
   %inc = add i32 %i.2, 1
-  %arrayidx = getelementptr inbounds [5 x i8*]* @main.L, i32 0, i32 %i.2
-  %0 = load i8** %arrayidx, align 4
+  %arrayidx = getelementptr inbounds [5 x i8*], [5 x i8*]* @main.L, i32 0, i32 %i.2
+  %0 = load i8*, i8** %arrayidx, align 4
   indirectbr i8* %0, [label %L1, label %L2, label %L3, label %L4]
 ; 16: 	jrc	 ${{[0-9]+}}
 L4:                                               ; preds = %L3
-  %puts8 = tail call i32 @puts(i8* getelementptr inbounds ([2 x i8]* @str8, i32 0, i32 0))
+  %puts8 = tail call i32 @puts(i8* getelementptr inbounds ([2 x i8], [2 x i8]* @str8, i32 0, i32 0))
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/brsize3.ll b/test/CodeGen/Mips/brsize3.ll
index 3620868bb2f6..dad0d841d4c6 100644
--- a/test/CodeGen/Mips/brsize3.ll
+++ b/test/CodeGen/Mips/brsize3.ll
@@ -1,6 +1,6 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=b-no-short
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=b-no-short
 
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=b-long
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=b-long
 
 ; ModuleID = 'brsize3.c'
 target datalayout = "E-p:32:32:32-i1:8:8-i8:8:32-i16:16:32-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-n32-S64"
diff --git a/test/CodeGen/Mips/brsize3a.ll b/test/CodeGen/Mips/brsize3a.ll
index f05e21191925..e1cd5893ceda 100644
--- a/test/CodeGen/Mips/brsize3a.ll
+++ b/test/CodeGen/Mips/brsize3a.ll
@@ -1,4 +1,4 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=b-short
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=b-short
 
 ; ModuleID = 'brsize3.c'
 target datalayout = "E-p:32:32:32-i1:8:8-i8:8:32-i16:16:32-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-n32-S64"
diff --git a/test/CodeGen/Mips/cache-intrinsic.ll b/test/CodeGen/Mips/cache-intrinsic.ll
index 2fa411589596..987032eaeb89 100644
--- a/test/CodeGen/Mips/cache-intrinsic.ll
+++ b/test/CodeGen/Mips/cache-intrinsic.ll
@@ -10,10 +10,10 @@ define i32 @main() {
 entry:
   %retval = alloca i32, align 4
   store i32 0, i32* %retval
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([4 x i8]* @.str, i32 0, i32 0), i8* getelementptr inbounds ([32 x i8]* @buffer, i32 0, i32 0))
-  %call1 = call i8* @strcpy(i8* getelementptr inbounds ([32 x i8]* @buffer, i32 0, i32 0), i8* getelementptr inbounds ([25 x i8]* @.str1, i32 0, i32 0)) #3
-  call void @llvm.clear_cache(i8* getelementptr inbounds ([32 x i8]* @buffer, i32 0, i32 0), i8* getelementptr inbounds (i8* getelementptr inbounds ([32 x i8]* @buffer, i32 0, i32 0), i32 32)) #3
-  %call2 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([4 x i8]* @.str, i32 0, i32 0), i8* getelementptr inbounds ([32 x i8]* @buffer, i32 0, i32 0))
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([4 x i8], [4 x i8]* @.str, i32 0, i32 0), i8* getelementptr inbounds ([32 x i8], [32 x i8]* @buffer, i32 0, i32 0))
+  %call1 = call i8* @strcpy(i8* getelementptr inbounds ([32 x i8], [32 x i8]* @buffer, i32 0, i32 0), i8* getelementptr inbounds ([25 x i8], [25 x i8]* @.str1, i32 0, i32 0)) #3
+  call void @llvm.clear_cache(i8* getelementptr inbounds ([32 x i8], [32 x i8]* @buffer, i32 0, i32 0), i8* getelementptr inbounds (i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @buffer, i32 0, i32 0), i32 32)) #3
+  %call2 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([4 x i8], [4 x i8]* @.str, i32 0, i32 0), i8* getelementptr inbounds ([32 x i8], [32 x i8]* @buffer, i32 0, i32 0))
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/cconv/arguments-float.ll b/test/CodeGen/Mips/cconv/arguments-float.ll
index 14a3baa7f539..c81c7215e164 100644
--- a/test/CodeGen/Mips/cconv/arguments-float.ll
+++ b/test/CodeGen/Mips/cconv/arguments-float.ll
@@ -1,14 +1,14 @@
-; RUN: llc -march=mips -relocation-model=static -soft-float < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 --check-prefix=O32BE %s
-; RUN: llc -march=mipsel -relocation-model=static -soft-float < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 --check-prefix=O32LE %s
+; RUN: llc -march=mips -relocation-model=static -mattr=+soft-float < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 --check-prefix=O32BE %s
+; RUN: llc -march=mipsel -relocation-model=static -mattr=+soft-float < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 --check-prefix=O32LE %s
 
-; RUN-TODO: llc -march=mips64 -relocation-model=static -soft-float -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
-; RUN-TODO: llc -march=mips64el -relocation-model=static -soft-float -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64 -relocation-model=static -mattr=+soft-float -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64el -relocation-model=static -mattr=+soft-float -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
 
-; RUN: llc -march=mips64 -relocation-model=static -soft-float -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=NEW %s
-; RUN: llc -march=mips64el -relocation-model=static -soft-float -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=NEW %s
+; RUN: llc -march=mips64 -relocation-model=static -mattr=+soft-float -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=NEW %s
+; RUN: llc -march=mips64el -relocation-model=static -mattr=+soft-float -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=NEW %s
 
-; RUN: llc -march=mips64 -relocation-model=static -soft-float -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=NEW %s
-; RUN: llc -march=mips64el -relocation-model=static -soft-float -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=NEW %s
+; RUN: llc -march=mips64 -relocation-model=static -mattr=+soft-float -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=NEW %s
+; RUN: llc -march=mips64el -relocation-model=static -mattr=+soft-float -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=NEW %s
 
 ; Test the floating point arguments for all ABI's and byte orders as specified
 ; by section 5 of MD00305 (MIPS ABIs Described).
@@ -24,23 +24,23 @@
 define void @double_args(double %a, double %b, double %c, double %d, double %e,
                          double %f, double %g, double %h, double %i) nounwind {
 entry:
-        %0 = getelementptr [11 x double]* @doubles, i32 0, i32 1
+        %0 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 1
         store volatile double %a, double* %0
-        %1 = getelementptr [11 x double]* @doubles, i32 0, i32 2
+        %1 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 2
         store volatile double %b, double* %1
-        %2 = getelementptr [11 x double]* @doubles, i32 0, i32 3
+        %2 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 3
         store volatile double %c, double* %2
-        %3 = getelementptr [11 x double]* @doubles, i32 0, i32 4
+        %3 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 4
         store volatile double %d, double* %3
-        %4 = getelementptr [11 x double]* @doubles, i32 0, i32 5
+        %4 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 5
         store volatile double %e, double* %4
-        %5 = getelementptr [11 x double]* @doubles, i32 0, i32 6
+        %5 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 6
         store volatile double %f, double* %5
-        %6 = getelementptr [11 x double]* @doubles, i32 0, i32 7
+        %6 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 7
         store volatile double %g, double* %6
-        %7 = getelementptr [11 x double]* @doubles, i32 0, i32 8
+        %7 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 8
         store volatile double %h, double* %7
-        %8 = getelementptr [11 x double]* @doubles, i32 0, i32 9
+        %8 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 9
         store volatile double %i, double* %8
         ret void
 }
@@ -105,25 +105,25 @@ define void @float_args(float %a, float %b, float %c, float %d, float %e,
                         float %f, float %g, float %h, float %i, float %j)
                        nounwind {
 entry:
-        %0 = getelementptr [11 x float]* @floats, i32 0, i32 1
+        %0 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 1
         store volatile float %a, float* %0
-        %1 = getelementptr [11 x float]* @floats, i32 0, i32 2
+        %1 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 2
         store volatile float %b, float* %1
-        %2 = getelementptr [11 x float]* @floats, i32 0, i32 3
+        %2 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 3
         store volatile float %c, float* %2
-        %3 = getelementptr [11 x float]* @floats, i32 0, i32 4
+        %3 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 4
         store volatile float %d, float* %3
-        %4 = getelementptr [11 x float]* @floats, i32 0, i32 5
+        %4 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 5
         store volatile float %e, float* %4
-        %5 = getelementptr [11 x float]* @floats, i32 0, i32 6
+        %5 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 6
         store volatile float %f, float* %5
-        %6 = getelementptr [11 x float]* @floats, i32 0, i32 7
+        %6 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 7
         store volatile float %g, float* %6
-        %7 = getelementptr [11 x float]* @floats, i32 0, i32 8
+        %7 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 8
         store volatile float %h, float* %7
-        %8 = getelementptr [11 x float]* @floats, i32 0, i32 9
+        %8 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 9
         store volatile float %i, float* %8
-        %9 = getelementptr [11 x float]* @floats, i32 0, i32 10
+        %9 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 10
         store volatile float %j, float* %9
         ret void
 }
@@ -170,9 +170,9 @@ entry:
 
 define void @double_arg2(i8 %a, double %b) nounwind {
 entry:
-        %0 = getelementptr [11 x i8]* @bytes, i32 0, i32 1
+        %0 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 1
         store volatile i8 %a, i8* %0
-        %1 = getelementptr [11 x double]* @doubles, i32 0, i32 1
+        %1 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 1
         store volatile double %b, double* %1
         ret void
 }
@@ -197,9 +197,9 @@ entry:
 
 define void @float_arg2(i8 signext %a, float %b) nounwind {
 entry:
-        %0 = getelementptr [11 x i8]* @bytes, i32 0, i32 1
+        %0 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 1
         store volatile i8 %a, i8* %0
-        %1 = getelementptr [11 x float]* @floats, i32 0, i32 1
+        %1 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 1
         store volatile float %b, float* %1
         ret void
 }
diff --git a/test/CodeGen/Mips/cconv/arguments-fp128.ll b/test/CodeGen/Mips/cconv/arguments-fp128.ll
index c8cd8fd11e50..6c62609396c5 100644
--- a/test/CodeGen/Mips/cconv/arguments-fp128.ll
+++ b/test/CodeGen/Mips/cconv/arguments-fp128.ll
@@ -1,8 +1,8 @@
-; RUN: llc -march=mips64 -relocation-model=static -soft-float -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 %s
-; RUN: llc -march=mips64el -relocation-model=static -soft-float -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 %s
+; RUN: llc -march=mips64 -relocation-model=static -mattr=+soft-float -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 %s
+; RUN: llc -march=mips64el -relocation-model=static -mattr=+soft-float -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 %s
 
-; RUN: llc -march=mips64 -relocation-model=static -soft-float -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 %s
-; RUN: llc -march=mips64el -relocation-model=static -soft-float -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 %s
+; RUN: llc -march=mips64 -relocation-model=static -mattr=+soft-float -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 %s
+; RUN: llc -march=mips64el -relocation-model=static -mattr=+soft-float -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 %s
 
 ; Test the fp128 arguments for all ABI's and byte orders as specified
 ; by section 2 of the MIPSpro N32 Handbook.
@@ -13,15 +13,15 @@
 
 define void @ldouble_args(fp128 %a, fp128 %b, fp128 %c, fp128 %d, fp128 %e) nounwind {
 entry:
-        %0 = getelementptr [11 x fp128]* @ldoubles, i32 0, i32 1
+        %0 = getelementptr [11 x fp128], [11 x fp128]* @ldoubles, i32 0, i32 1
         store volatile fp128 %a, fp128* %0
-        %1 = getelementptr [11 x fp128]* @ldoubles, i32 0, i32 2
+        %1 = getelementptr [11 x fp128], [11 x fp128]* @ldoubles, i32 0, i32 2
         store volatile fp128 %b, fp128* %1
-        %2 = getelementptr [11 x fp128]* @ldoubles, i32 0, i32 3
+        %2 = getelementptr [11 x fp128], [11 x fp128]* @ldoubles, i32 0, i32 3
         store volatile fp128 %c, fp128* %2
-        %3 = getelementptr [11 x fp128]* @ldoubles, i32 0, i32 4
+        %3 = getelementptr [11 x fp128], [11 x fp128]* @ldoubles, i32 0, i32 4
         store volatile fp128 %d, fp128* %3
-        %4 = getelementptr [11 x fp128]* @ldoubles, i32 0, i32 5
+        %4 = getelementptr [11 x fp128], [11 x fp128]* @ldoubles, i32 0, i32 5
         store volatile fp128 %e, fp128* %4
         ret void
 }
diff --git a/test/CodeGen/Mips/cconv/arguments-hard-float-varargs.ll b/test/CodeGen/Mips/cconv/arguments-hard-float-varargs.ll
index 70ccf14c5450..9f1fe91ec172 100644
--- a/test/CodeGen/Mips/cconv/arguments-hard-float-varargs.ll
+++ b/test/CodeGen/Mips/cconv/arguments-hard-float-varargs.ll
@@ -1,14 +1,14 @@
 ; RUN: llc -march=mips -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 --check-prefix=O32BE %s
 ; RUN: llc -march=mipsel -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 --check-prefix=O32LE %s
 
-; RUN-TODO: llc -march=mips64 -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
-; RUN-TODO: llc -march=mips64el -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64 -relocation-model=static -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64el -relocation-model=static -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
 
-; RUN: llc -march=mips64 -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=N32 --check-prefix=NEW --check-prefix=NEWBE %s
-; RUN: llc -march=mips64el -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=N32 --check-prefix=NEW --check-prefix=NEWLE %s
+; RUN: llc -march=mips64 -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=N32 --check-prefix=NEW --check-prefix=NEWBE %s
+; RUN: llc -march=mips64el -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=N32 --check-prefix=NEW --check-prefix=NEWLE %s
 
-; RUN: llc -march=mips64 -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=N64 --check-prefix=NEW --check-prefix=NEWBE %s
-; RUN: llc -march=mips64el -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=N64 --check-prefix=NEW --check-prefix=NEWLE %s
+; RUN: llc -march=mips64 -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=N64 --check-prefix=NEW --check-prefix=NEWBE %s
+; RUN: llc -march=mips64el -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=N64 --check-prefix=NEW --check-prefix=NEWLE %s
 
 ; Test the effect of varargs on floating point types in the non-variable part
 ; of the argument list as specified by section 2 of the MIPSpro N32 Handbook.
@@ -25,14 +25,14 @@
 define void @double_args(double %a, ...)
                          nounwind {
 entry:
-        %0 = getelementptr [11 x double]* @doubles, i32 0, i32 1
+        %0 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 1
         store volatile double %a, double* %0
 
         %ap = alloca i8*
         %ap2 = bitcast i8** %ap to i8*
         call void @llvm.va_start(i8* %ap2)
         %b = va_arg i8** %ap, double
-        %1 = getelementptr [11 x double]* @doubles, i32 0, i32 2
+        %1 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 2
         store volatile double %b, double* %1
         call void @llvm.va_end(i8* %ap2)
         ret void
@@ -90,14 +90,14 @@ entry:
 
 define void @float_args(float %a, ...) nounwind {
 entry:
-        %0 = getelementptr [11 x float]* @floats, i32 0, i32 1
+        %0 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 1
         store volatile float %a, float* %0
 
         %ap = alloca i8*
         %ap2 = bitcast i8** %ap to i8*
         call void @llvm.va_start(i8* %ap2)
         %b = va_arg i8** %ap, float
-        %1 = getelementptr [11 x float]* @floats, i32 0, i32 2
+        %1 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 2
         store volatile float %b, float* %1
         call void @llvm.va_end(i8* %ap2)
         ret void
diff --git a/test/CodeGen/Mips/cconv/arguments-hard-float.ll b/test/CodeGen/Mips/cconv/arguments-hard-float.ll
index 9837f7ee5586..24148ed176db 100644
--- a/test/CodeGen/Mips/cconv/arguments-hard-float.ll
+++ b/test/CodeGen/Mips/cconv/arguments-hard-float.ll
@@ -1,14 +1,14 @@
 ; RUN: llc -march=mips -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 --check-prefix=O32BE %s
 ; RUN: llc -march=mipsel -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 --check-prefix=O32LE %s
 
-; RUN-TODO: llc -march=mips64 -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
-; RUN-TODO: llc -march=mips64el -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64 -relocation-model=static -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64el -relocation-model=static -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
 
-; RUN: llc -march=mips64 -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=NEW %s
-; RUN: llc -march=mips64el -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=NEW %s
+; RUN: llc -march=mips64 -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=NEW %s
+; RUN: llc -march=mips64el -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=NEW %s
 
-; RUN: llc -march=mips64 -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=NEW %s
-; RUN: llc -march=mips64el -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=NEW %s
+; RUN: llc -march=mips64 -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=NEW %s
+; RUN: llc -march=mips64el -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=NEW %s
 
 ; Test the floating point arguments for all ABI's and byte orders as specified
 ; by section 5 of MD00305 (MIPS ABIs Described).
@@ -24,23 +24,23 @@
 define void @double_args(double %a, double %b, double %c, double %d, double %e,
                          double %f, double %g, double %h, double %i) nounwind {
 entry:
-        %0 = getelementptr [11 x double]* @doubles, i32 0, i32 1
+        %0 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 1
         store volatile double %a, double* %0
-        %1 = getelementptr [11 x double]* @doubles, i32 0, i32 2
+        %1 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 2
         store volatile double %b, double* %1
-        %2 = getelementptr [11 x double]* @doubles, i32 0, i32 3
+        %2 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 3
         store volatile double %c, double* %2
-        %3 = getelementptr [11 x double]* @doubles, i32 0, i32 4
+        %3 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 4
         store volatile double %d, double* %3
-        %4 = getelementptr [11 x double]* @doubles, i32 0, i32 5
+        %4 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 5
         store volatile double %e, double* %4
-        %5 = getelementptr [11 x double]* @doubles, i32 0, i32 6
+        %5 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 6
         store volatile double %f, double* %5
-        %6 = getelementptr [11 x double]* @doubles, i32 0, i32 7
+        %6 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 7
         store volatile double %g, double* %6
-        %7 = getelementptr [11 x double]* @doubles, i32 0, i32 8
+        %7 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 8
         store volatile double %h, double* %7
-        %8 = getelementptr [11 x double]* @doubles, i32 0, i32 9
+        %8 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 9
         store volatile double %i, double* %8
         ret void
 }
@@ -87,23 +87,23 @@ entry:
 define void @float_args(float %a, float %b, float %c, float %d, float %e,
                         float %f, float %g, float %h, float %i) nounwind {
 entry:
-        %0 = getelementptr [11 x float]* @floats, i32 0, i32 1
+        %0 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 1
         store volatile float %a, float* %0
-        %1 = getelementptr [11 x float]* @floats, i32 0, i32 2
+        %1 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 2
         store volatile float %b, float* %1
-        %2 = getelementptr [11 x float]* @floats, i32 0, i32 3
+        %2 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 3
         store volatile float %c, float* %2
-        %3 = getelementptr [11 x float]* @floats, i32 0, i32 4
+        %3 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 4
         store volatile float %d, float* %3
-        %4 = getelementptr [11 x float]* @floats, i32 0, i32 5
+        %4 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 5
         store volatile float %e, float* %4
-        %5 = getelementptr [11 x float]* @floats, i32 0, i32 6
+        %5 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 6
         store volatile float %f, float* %5
-        %6 = getelementptr [11 x float]* @floats, i32 0, i32 7
+        %6 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 7
         store volatile float %g, float* %6
-        %7 = getelementptr [11 x float]* @floats, i32 0, i32 8
+        %7 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 8
         store volatile float %h, float* %7
-        %8 = getelementptr [11 x float]* @floats, i32 0, i32 9
+        %8 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 9
         store volatile float %i, float* %8
         ret void
 }
@@ -153,9 +153,9 @@ entry:
 
 define void @double_arg2(i8 %a, double %b) nounwind {
 entry:
-        %0 = getelementptr [11 x i8]* @bytes, i32 0, i32 1
+        %0 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 1
         store volatile i8 %a, i8* %0
-        %1 = getelementptr [11 x double]* @doubles, i32 0, i32 1
+        %1 = getelementptr [11 x double], [11 x double]* @doubles, i32 0, i32 1
         store volatile double %b, double* %1
         ret void
 }
@@ -184,9 +184,9 @@ entry:
 
 define void @float_arg2(i8 %a, float %b) nounwind {
 entry:
-        %0 = getelementptr [11 x i8]* @bytes, i32 0, i32 1
+        %0 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 1
         store volatile i8 %a, i8* %0
-        %1 = getelementptr [11 x float]* @floats, i32 0, i32 1
+        %1 = getelementptr [11 x float], [11 x float]* @floats, i32 0, i32 1
         store volatile float %b, float* %1
         ret void
 }
diff --git a/test/CodeGen/Mips/cconv/arguments-hard-fp128.ll b/test/CodeGen/Mips/cconv/arguments-hard-fp128.ll
index 5e3f403495f5..26eb569f865d 100644
--- a/test/CodeGen/Mips/cconv/arguments-hard-fp128.ll
+++ b/test/CodeGen/Mips/cconv/arguments-hard-fp128.ll
@@ -1,8 +1,8 @@
-; RUN: llc -march=mips64 -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 %s
-; RUN: llc -march=mips64el -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 %s
+; RUN: llc -march=mips64 -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 %s
+; RUN: llc -march=mips64el -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 %s
 
-; RUN: llc -march=mips64 -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 %s
-; RUN: llc -march=mips64el -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 %s
+; RUN: llc -march=mips64 -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 %s
+; RUN: llc -march=mips64el -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 %s
 
 ; Test the fp128 arguments for all ABI's and byte orders as specified
 ; by section 2 of the MIPSpro N32 Handbook.
@@ -13,15 +13,15 @@
 
 define void @ldouble_args(fp128 %a, fp128 %b, fp128 %c, fp128 %d, fp128 %e) nounwind {
 entry:
-        %0 = getelementptr [11 x fp128]* @ldoubles, i32 0, i32 1
+        %0 = getelementptr [11 x fp128], [11 x fp128]* @ldoubles, i32 0, i32 1
         store volatile fp128 %a, fp128* %0
-        %1 = getelementptr [11 x fp128]* @ldoubles, i32 0, i32 2
+        %1 = getelementptr [11 x fp128], [11 x fp128]* @ldoubles, i32 0, i32 2
         store volatile fp128 %b, fp128* %1
-        %2 = getelementptr [11 x fp128]* @ldoubles, i32 0, i32 3
+        %2 = getelementptr [11 x fp128], [11 x fp128]* @ldoubles, i32 0, i32 3
         store volatile fp128 %c, fp128* %2
-        %3 = getelementptr [11 x fp128]* @ldoubles, i32 0, i32 4
+        %3 = getelementptr [11 x fp128], [11 x fp128]* @ldoubles, i32 0, i32 4
         store volatile fp128 %d, fp128* %3
-        %4 = getelementptr [11 x fp128]* @ldoubles, i32 0, i32 5
+        %4 = getelementptr [11 x fp128], [11 x fp128]* @ldoubles, i32 0, i32 5
         store volatile fp128 %e, fp128* %4
         ret void
 }
diff --git a/test/CodeGen/Mips/cconv/arguments-small-structures-bigger-than-32bits.ll b/test/CodeGen/Mips/cconv/arguments-small-structures-bigger-than-32bits.ll
index d17290e552e0..087a0515f379 100644
--- a/test/CodeGen/Mips/cconv/arguments-small-structures-bigger-than-32bits.ll
+++ b/test/CodeGen/Mips/cconv/arguments-small-structures-bigger-than-32bits.ll
@@ -52,8 +52,8 @@ entry:
   %0 = bitcast { i48 }* %s1_1.coerce to i8*
   %1 = bitcast %struct.S1* %s1_1 to i8*
   call void @llvm.memcpy.p0i8.p0i8.i64(i8* %0, i8* %1, i64 6, i32 0, i1 false)
-  %2 = getelementptr { i48 }* %s1_1.coerce, i32 0, i32 0
-  %3 = load i48* %2, align 1
+  %2 = getelementptr { i48 }, { i48 }* %s1_1.coerce, i32 0, i32 0
+  %3 = load i48, i48* %2, align 1
   call void @fS1(i48 inreg %3)
   ret void
  ; ALL-LABEL: f1:
@@ -69,8 +69,8 @@ entry:
   %0 = bitcast { i40 }* %s2_1.coerce to i8*
   %1 = bitcast %struct.S2* %s2_1 to i8*
   call void @llvm.memcpy.p0i8.p0i8.i64(i8* %0, i8* %1, i64 5, i32 0, i1 false)
-  %2 = getelementptr { i40 }* %s2_1.coerce, i32 0, i32 0
-  %3 = load i40* %2, align 1
+  %2 = getelementptr { i40 }, { i40 }* %s2_1.coerce, i32 0, i32 0
+  %3 = load i40, i40* %2, align 1
   call void @fS2(i40 inreg %3)
   ret void
  ; ALL-LABEL: f2:
diff --git a/test/CodeGen/Mips/cconv/arguments-struct.ll b/test/CodeGen/Mips/cconv/arguments-struct.ll
index c1bc84ee7a04..ee6bfaeb9537 100644
--- a/test/CodeGen/Mips/cconv/arguments-struct.ll
+++ b/test/CodeGen/Mips/cconv/arguments-struct.ll
@@ -1,14 +1,14 @@
 ; RUN: llc -mtriple=mips-unknown-linux-gnu -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32-BE %s
 ; RUN: llc -mtriple=mipsel-unknown-linux-gnu -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32-LE %s
 
-; RUN-TODO: llc -mtriple=mips64-unknown-linux-gnu -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32-BE %s
-; RUN-TODO: llc -mtriple=mips64el-unknown-linux-gnu -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32-LE %s
+; RUN-TODO: llc -mtriple=mips64-unknown-linux-gnu -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32-BE %s
+; RUN-TODO: llc -mtriple=mips64el-unknown-linux-gnu -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32-LE %s
 
-; RUN: llc -mtriple=mips64-unknown-linux-gnu -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=NEW-BE %s
-; RUN: llc -mtriple=mips64el-unknown-linux-gnu -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=NEW-LE %s
+; RUN: llc -mtriple=mips64-unknown-linux-gnu -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=NEW-BE %s
+; RUN: llc -mtriple=mips64el-unknown-linux-gnu -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=NEW-LE %s
 
-; RUN: llc -mtriple=mips64-unknown-linux-gnu -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=NEW-BE %s
-; RUN: llc -mtriple=mips64el-unknown-linux-gnu -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=NEW-LE %s
+; RUN: llc -mtriple=mips64-unknown-linux-gnu -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=NEW-BE %s
+; RUN: llc -mtriple=mips64el-unknown-linux-gnu -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=NEW-LE %s
 
 ; Test small structures for all ABI's and byte orders.
 ;
@@ -19,7 +19,7 @@
 
 define void @s_i8(i8 inreg %a) nounwind {
 entry:
-	store i8 %a, i8* getelementptr inbounds ([2 x i8]* @bytes, i32 0, i32 1)
+	store i8 %a, i8* getelementptr inbounds ([2 x i8], [2 x i8]* @bytes, i32 0, i32 1)
         ret void
 }
 
diff --git a/test/CodeGen/Mips/cconv/arguments-varargs-small-structs-byte.ll b/test/CodeGen/Mips/cconv/arguments-varargs-small-structs-byte.ll
index 458b124c9927..ba3aeb598f50 100644
--- a/test/CodeGen/Mips/cconv/arguments-varargs-small-structs-byte.ll
+++ b/test/CodeGen/Mips/cconv/arguments-varargs-small-structs-byte.ll
@@ -140,11 +140,11 @@ define void @smallStruct_1b(%struct.SmallStruct_1b* %ss) #0 {
 entry:
   %ss.addr = alloca %struct.SmallStruct_1b*, align 8
   store %struct.SmallStruct_1b* %ss, %struct.SmallStruct_1b** %ss.addr, align 8
-  %0 = load %struct.SmallStruct_1b** %ss.addr, align 8
+  %0 = load %struct.SmallStruct_1b*, %struct.SmallStruct_1b** %ss.addr, align 8
   %1 = bitcast %struct.SmallStruct_1b* %0 to { i8 }*
-  %2 = getelementptr { i8 }* %1, i32 0, i32 0
-  %3 = load i8* %2, align 1
-  call void (i8*, ...)* @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8]* @.str, i32 0, i32 0), i8 inreg %3)
+  %2 = getelementptr { i8 }, { i8 }* %1, i32 0, i32 0
+  %3 = load i8, i8* %2, align 1
+  call void (i8*, ...) @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @.str, i32 0, i32 0), i8 inreg %3)
   ret void
  ; CHECK-LABEL: smallStruct_1b: 
  ; CHECK: dsll $[[R1:[0-9]+]], $[[R2:[0-9]+]], 56
@@ -154,11 +154,11 @@ define void @smallStruct_2b(%struct.SmallStruct_2b* %ss) #0 {
 entry:
   %ss.addr = alloca %struct.SmallStruct_2b*, align 8
   store %struct.SmallStruct_2b* %ss, %struct.SmallStruct_2b** %ss.addr, align 8
-  %0 = load %struct.SmallStruct_2b** %ss.addr, align 8
+  %0 = load %struct.SmallStruct_2b*, %struct.SmallStruct_2b** %ss.addr, align 8
   %1 = bitcast %struct.SmallStruct_2b* %0 to { i16 }*
-  %2 = getelementptr { i16 }* %1, i32 0, i32 0
-  %3 = load i16* %2, align 1
-  call void (i8*, ...)* @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8]* @.str, i32 0, i32 0), i16 inreg %3)
+  %2 = getelementptr { i16 }, { i16 }* %1, i32 0, i32 0
+  %3 = load i16, i16* %2, align 1
+  call void (i8*, ...) @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @.str, i32 0, i32 0), i16 inreg %3)
   ret void
  ; CHECK-LABEL: smallStruct_2b:
  ; CHECK: dsll $[[R1:[0-9]+]], $[[R2:[0-9]+]], 48
@@ -169,13 +169,13 @@ entry:
   %ss.addr = alloca %struct.SmallStruct_3b*, align 8
   %.coerce = alloca { i24 }
   store %struct.SmallStruct_3b* %ss, %struct.SmallStruct_3b** %ss.addr, align 8
-  %0 = load %struct.SmallStruct_3b** %ss.addr, align 8
+  %0 = load %struct.SmallStruct_3b*, %struct.SmallStruct_3b** %ss.addr, align 8
   %1 = bitcast { i24 }* %.coerce to i8*
   %2 = bitcast %struct.SmallStruct_3b* %0 to i8*
   call void @llvm.memcpy.p0i8.p0i8.i64(i8* %1, i8* %2, i64 3, i32 0, i1 false)
-  %3 = getelementptr { i24 }* %.coerce, i32 0, i32 0
-  %4 = load i24* %3, align 1
-  call void (i8*, ...)* @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8]* @.str, i32 0, i32 0), i24 inreg %4)
+  %3 = getelementptr { i24 }, { i24 }* %.coerce, i32 0, i32 0
+  %4 = load i24, i24* %3, align 1
+  call void (i8*, ...) @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @.str, i32 0, i32 0), i24 inreg %4)
   ret void
  ; CHECK-LABEL: smallStruct_3b:
  ; CHECK: dsll $[[R1:[0-9]+]], $[[R2:[0-9]+]], 40
@@ -187,11 +187,11 @@ define void @smallStruct_4b(%struct.SmallStruct_4b* %ss) #0 {
 entry:
   %ss.addr = alloca %struct.SmallStruct_4b*, align 8
   store %struct.SmallStruct_4b* %ss, %struct.SmallStruct_4b** %ss.addr, align 8
-  %0 = load %struct.SmallStruct_4b** %ss.addr, align 8
+  %0 = load %struct.SmallStruct_4b*, %struct.SmallStruct_4b** %ss.addr, align 8
   %1 = bitcast %struct.SmallStruct_4b* %0 to { i32 }*
-  %2 = getelementptr { i32 }* %1, i32 0, i32 0
-  %3 = load i32* %2, align 1
-  call void (i8*, ...)* @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8]* @.str, i32 0, i32 0), i32 inreg %3)
+  %2 = getelementptr { i32 }, { i32 }* %1, i32 0, i32 0
+  %3 = load i32, i32* %2, align 1
+  call void (i8*, ...) @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @.str, i32 0, i32 0), i32 inreg %3)
   ret void
  ; CHECK-LABEL: smallStruct_4b:
  ; CHECK: dsll $[[R1:[0-9]+]], $[[R2:[0-9]+]], 32
@@ -202,13 +202,13 @@ entry:
   %ss.addr = alloca %struct.SmallStruct_5b*, align 8
   %.coerce = alloca { i40 }
   store %struct.SmallStruct_5b* %ss, %struct.SmallStruct_5b** %ss.addr, align 8
-  %0 = load %struct.SmallStruct_5b** %ss.addr, align 8
+  %0 = load %struct.SmallStruct_5b*, %struct.SmallStruct_5b** %ss.addr, align 8
   %1 = bitcast { i40 }* %.coerce to i8*
   %2 = bitcast %struct.SmallStruct_5b* %0 to i8*
   call void @llvm.memcpy.p0i8.p0i8.i64(i8* %1, i8* %2, i64 5, i32 0, i1 false)
-  %3 = getelementptr { i40 }* %.coerce, i32 0, i32 0
-  %4 = load i40* %3, align 1
-  call void (i8*, ...)* @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8]* @.str, i32 0, i32 0), i40 inreg %4)
+  %3 = getelementptr { i40 }, { i40 }* %.coerce, i32 0, i32 0
+  %4 = load i40, i40* %3, align 1
+  call void (i8*, ...) @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @.str, i32 0, i32 0), i40 inreg %4)
   ret void
  ; CHECK-LABEL: smallStruct_5b:
  ; CHECK: dsll $[[R1:[0-9]+]], $[[R2:[0-9]+]], 24
@@ -219,13 +219,13 @@ entry:
   %ss.addr = alloca %struct.SmallStruct_6b*, align 8
   %.coerce = alloca { i48 }
   store %struct.SmallStruct_6b* %ss, %struct.SmallStruct_6b** %ss.addr, align 8
-  %0 = load %struct.SmallStruct_6b** %ss.addr, align 8
+  %0 = load %struct.SmallStruct_6b*, %struct.SmallStruct_6b** %ss.addr, align 8
   %1 = bitcast { i48 }* %.coerce to i8*
   %2 = bitcast %struct.SmallStruct_6b* %0 to i8*
   call void @llvm.memcpy.p0i8.p0i8.i64(i8* %1, i8* %2, i64 6, i32 0, i1 false)
-  %3 = getelementptr { i48 }* %.coerce, i32 0, i32 0
-  %4 = load i48* %3, align 1
-  call void (i8*, ...)* @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8]* @.str, i32 0, i32 0), i48 inreg %4)
+  %3 = getelementptr { i48 }, { i48 }* %.coerce, i32 0, i32 0
+  %4 = load i48, i48* %3, align 1
+  call void (i8*, ...) @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @.str, i32 0, i32 0), i48 inreg %4)
   ret void
  ; CHECK-LABEL: smallStruct_6b:
  ; CHECK: dsll $[[R1:[0-9]+]], $[[R2:[0-9]+]], 16
@@ -236,13 +236,13 @@ entry:
   %ss.addr = alloca %struct.SmallStruct_7b*, align 8
   %.coerce = alloca { i56 }
   store %struct.SmallStruct_7b* %ss, %struct.SmallStruct_7b** %ss.addr, align 8
-  %0 = load %struct.SmallStruct_7b** %ss.addr, align 8
+  %0 = load %struct.SmallStruct_7b*, %struct.SmallStruct_7b** %ss.addr, align 8
   %1 = bitcast { i56 }* %.coerce to i8*
   %2 = bitcast %struct.SmallStruct_7b* %0 to i8*
   call void @llvm.memcpy.p0i8.p0i8.i64(i8* %1, i8* %2, i64 7, i32 0, i1 false)
-  %3 = getelementptr { i56 }* %.coerce, i32 0, i32 0
-  %4 = load i56* %3, align 1
-  call void (i8*, ...)* @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8]* @.str, i32 0, i32 0), i56 inreg %4)
+  %3 = getelementptr { i56 }, { i56 }* %.coerce, i32 0, i32 0
+  %4 = load i56, i56* %3, align 1
+  call void (i8*, ...) @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @.str, i32 0, i32 0), i56 inreg %4)
   ret void
  ; CHECK-LABEL: smallStruct_7b:
  ; CHECK: dsll $[[R1:[0-9]+]], $[[R2:[0-9]+]], 8
@@ -252,11 +252,11 @@ define void @smallStruct_8b(%struct.SmallStruct_8b* %ss) #0 {
 entry:
   %ss.addr = alloca %struct.SmallStruct_8b*, align 8
   store %struct.SmallStruct_8b* %ss, %struct.SmallStruct_8b** %ss.addr, align 8
-  %0 = load %struct.SmallStruct_8b** %ss.addr, align 8
+  %0 = load %struct.SmallStruct_8b*, %struct.SmallStruct_8b** %ss.addr, align 8
   %1 = bitcast %struct.SmallStruct_8b* %0 to { i64 }*
-  %2 = getelementptr { i64 }* %1, i32 0, i32 0
-  %3 = load i64* %2, align 1
-  call void (i8*, ...)* @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8]* @.str, i32 0, i32 0), i64 inreg %3)
+  %2 = getelementptr { i64 }, { i64 }* %1, i32 0, i32 0
+  %3 = load i64, i64* %2, align 1
+  call void (i8*, ...) @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @.str, i32 0, i32 0), i64 inreg %3)
   ret void
  ; CHECK-LABEL: smallStruct_8b:
  ; CHECK-NOT: dsll
@@ -267,15 +267,15 @@ entry:
   %ss.addr = alloca %struct.SmallStruct_9b*, align 8
   %.coerce = alloca { i64, i8 }
   store %struct.SmallStruct_9b* %ss, %struct.SmallStruct_9b** %ss.addr, align 8
-  %0 = load %struct.SmallStruct_9b** %ss.addr, align 8
+  %0 = load %struct.SmallStruct_9b*, %struct.SmallStruct_9b** %ss.addr, align 8
   %1 = bitcast { i64, i8 }* %.coerce to i8*
   %2 = bitcast %struct.SmallStruct_9b* %0 to i8*
   call void @llvm.memcpy.p0i8.p0i8.i64(i8* %1, i8* %2, i64 9, i32 0, i1 false)
-  %3 = getelementptr { i64, i8 }* %.coerce, i32 0, i32 0
-  %4 = load i64* %3, align 1
-  %5 = getelementptr { i64, i8 }* %.coerce, i32 0, i32 1
-  %6 = load i8* %5, align 1
-  call void (i8*, ...)* @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8]* @.str, i32 0, i32 0), i64 inreg %4, i8 inreg %6)
+  %3 = getelementptr { i64, i8 }, { i64, i8 }* %.coerce, i32 0, i32 0
+  %4 = load i64, i64* %3, align 1
+  %5 = getelementptr { i64, i8 }, { i64, i8 }* %.coerce, i32 0, i32 1
+  %6 = load i8, i8* %5, align 1
+  call void (i8*, ...) @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @.str, i32 0, i32 0), i64 inreg %4, i8 inreg %6)
   ret void
  ; CHECK-LABEL: smallStruct_9b:
  ; CHECK: dsll $[[R1:[0-9]+]], $[[R2:[0-9]+]], 56
diff --git a/test/CodeGen/Mips/cconv/arguments-varargs-small-structs-combinations.ll b/test/CodeGen/Mips/cconv/arguments-varargs-small-structs-combinations.ll
index 899a3e8ff0a1..74d3d859ed75 100644
--- a/test/CodeGen/Mips/cconv/arguments-varargs-small-structs-combinations.ll
+++ b/test/CodeGen/Mips/cconv/arguments-varargs-small-structs-combinations.ll
@@ -74,11 +74,11 @@ define void @smallStruct_1b1s(%struct.SmallStruct_1b1s* %ss) #0 {
 entry:
   %ss.addr = alloca %struct.SmallStruct_1b1s*, align 8
   store %struct.SmallStruct_1b1s* %ss, %struct.SmallStruct_1b1s** %ss.addr, align 8
-  %0 = load %struct.SmallStruct_1b1s** %ss.addr, align 8
+  %0 = load %struct.SmallStruct_1b1s*, %struct.SmallStruct_1b1s** %ss.addr, align 8
   %1 = bitcast %struct.SmallStruct_1b1s* %0 to { i32 }*
-  %2 = getelementptr { i32 }* %1, i32 0, i32 0
-  %3 = load i32* %2, align 1
-  call void (i8*, ...)* @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8]* @.str, i32 0, i32 0), i32 inreg %3)
+  %2 = getelementptr { i32 }, { i32 }* %1, i32 0, i32 0
+  %3 = load i32, i32* %2, align 1
+  call void (i8*, ...) @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @.str, i32 0, i32 0), i32 inreg %3)
   ret void
  ; CHECK-LABEL: smallStruct_1b1s:
  ; CHECK: dsll $[[R1:[0-9]+]], $[[R2:[0-9]+]], 32
@@ -88,11 +88,11 @@ define void @smallStruct_1b1i(%struct.SmallStruct_1b1i* %ss) #0 {
 entry:
   %ss.addr = alloca %struct.SmallStruct_1b1i*, align 8
   store %struct.SmallStruct_1b1i* %ss, %struct.SmallStruct_1b1i** %ss.addr, align 8
-  %0 = load %struct.SmallStruct_1b1i** %ss.addr, align 8
+  %0 = load %struct.SmallStruct_1b1i*, %struct.SmallStruct_1b1i** %ss.addr, align 8
   %1 = bitcast %struct.SmallStruct_1b1i* %0 to { i64 }*
-  %2 = getelementptr { i64 }* %1, i32 0, i32 0
-  %3 = load i64* %2, align 1
-  call void (i8*, ...)* @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8]* @.str, i32 0, i32 0), i64 inreg %3)
+  %2 = getelementptr { i64 }, { i64 }* %1, i32 0, i32 0
+  %3 = load i64, i64* %2, align 1
+  call void (i8*, ...) @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @.str, i32 0, i32 0), i64 inreg %3)
   ret void
  ; CHECK-LABEL: smallStruct_1b1i:
  ; CHECK-NOT: dsll
@@ -103,13 +103,13 @@ entry:
   %ss.addr = alloca %struct.SmallStruct_1b1s1b*, align 8
   %.coerce = alloca { i48 }
   store %struct.SmallStruct_1b1s1b* %ss, %struct.SmallStruct_1b1s1b** %ss.addr, align 8
-  %0 = load %struct.SmallStruct_1b1s1b** %ss.addr, align 8
+  %0 = load %struct.SmallStruct_1b1s1b*, %struct.SmallStruct_1b1s1b** %ss.addr, align 8
   %1 = bitcast { i48 }* %.coerce to i8*
   %2 = bitcast %struct.SmallStruct_1b1s1b* %0 to i8*
   call void @llvm.memcpy.p0i8.p0i8.i64(i8* %1, i8* %2, i64 6, i32 0, i1 false)
-  %3 = getelementptr { i48 }* %.coerce, i32 0, i32 0
-  %4 = load i48* %3, align 1
-  call void (i8*, ...)* @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8]* @.str, i32 0, i32 0), i48 inreg %4)
+  %3 = getelementptr { i48 }, { i48 }* %.coerce, i32 0, i32 0
+  %4 = load i48, i48* %3, align 1
+  call void (i8*, ...) @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @.str, i32 0, i32 0), i48 inreg %4)
   ret void
  ; CHECK-LABEL: smallStruct_1b1s1b:
  ; CHECK: dsll $[[R1:[0-9]+]], $[[R2:[0-9]+]], 16
@@ -121,11 +121,11 @@ define void @smallStruct_1s1i(%struct.SmallStruct_1s1i* %ss) #0 {
 entry:
   %ss.addr = alloca %struct.SmallStruct_1s1i*, align 8
   store %struct.SmallStruct_1s1i* %ss, %struct.SmallStruct_1s1i** %ss.addr, align 8
-  %0 = load %struct.SmallStruct_1s1i** %ss.addr, align 8
+  %0 = load %struct.SmallStruct_1s1i*, %struct.SmallStruct_1s1i** %ss.addr, align 8
   %1 = bitcast %struct.SmallStruct_1s1i* %0 to { i64 }*
-  %2 = getelementptr { i64 }* %1, i32 0, i32 0
-  %3 = load i64* %2, align 1
-  call void (i8*, ...)* @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8]* @.str, i32 0, i32 0), i64 inreg %3)
+  %2 = getelementptr { i64 }, { i64 }* %1, i32 0, i32 0
+  %3 = load i64, i64* %2, align 1
+  call void (i8*, ...) @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @.str, i32 0, i32 0), i64 inreg %3)
   ret void
  ; CHECK-LABEL: smallStruct_1s1i:
  ; CHECK-NOT: dsll
@@ -136,13 +136,13 @@ entry:
   %ss.addr = alloca %struct.SmallStruct_3b1s*, align 8
   %.coerce = alloca { i48 }
   store %struct.SmallStruct_3b1s* %ss, %struct.SmallStruct_3b1s** %ss.addr, align 8
-  %0 = load %struct.SmallStruct_3b1s** %ss.addr, align 8
+  %0 = load %struct.SmallStruct_3b1s*, %struct.SmallStruct_3b1s** %ss.addr, align 8
   %1 = bitcast { i48 }* %.coerce to i8*
   %2 = bitcast %struct.SmallStruct_3b1s* %0 to i8*
   call void @llvm.memcpy.p0i8.p0i8.i64(i8* %1, i8* %2, i64 6, i32 0, i1 false)
-  %3 = getelementptr { i48 }* %.coerce, i32 0, i32 0
-  %4 = load i48* %3, align 1
-  call void (i8*, ...)* @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8]* @.str, i32 0, i32 0), i48 inreg %4)
+  %3 = getelementptr { i48 }, { i48 }* %.coerce, i32 0, i32 0
+  %4 = load i48, i48* %3, align 1
+  call void (i8*, ...) @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @.str, i32 0, i32 0), i48 inreg %4)
   ret void
  ; CHECK-LABEL: smallStruct_3b1s:
  ; CHECK: dsll $[[R1:[0-9]+]], $[[R2:[0-9]+]], 16
diff --git a/test/CodeGen/Mips/cconv/arguments-varargs-small-structs-multiple-args.ll b/test/CodeGen/Mips/cconv/arguments-varargs-small-structs-multiple-args.ll
index 1f7362523346..a4ac5e7bd8a6 100644
--- a/test/CodeGen/Mips/cconv/arguments-varargs-small-structs-multiple-args.ll
+++ b/test/CodeGen/Mips/cconv/arguments-varargs-small-structs-multiple-args.ll
@@ -110,43 +110,43 @@ entry:
   store %struct.SmallStruct_1b* %ss7, %struct.SmallStruct_1b** %ss7.addr, align 8
   store %struct.SmallStruct_1b* %ss8, %struct.SmallStruct_1b** %ss8.addr, align 8
   store %struct.SmallStruct_1b* %ss9, %struct.SmallStruct_1b** %ss9.addr, align 8
-  %0 = load %struct.SmallStruct_1b** %ss1.addr, align 8
-  %1 = load %struct.SmallStruct_1b** %ss2.addr, align 8
-  %2 = load %struct.SmallStruct_1b** %ss3.addr, align 8
-  %3 = load %struct.SmallStruct_1b** %ss4.addr, align 8
-  %4 = load %struct.SmallStruct_1b** %ss5.addr, align 8
-  %5 = load %struct.SmallStruct_1b** %ss6.addr, align 8
-  %6 = load %struct.SmallStruct_1b** %ss7.addr, align 8
-  %7 = load %struct.SmallStruct_1b** %ss8.addr, align 8
-  %8 = load %struct.SmallStruct_1b** %ss9.addr, align 8
+  %0 = load %struct.SmallStruct_1b*, %struct.SmallStruct_1b** %ss1.addr, align 8
+  %1 = load %struct.SmallStruct_1b*, %struct.SmallStruct_1b** %ss2.addr, align 8
+  %2 = load %struct.SmallStruct_1b*, %struct.SmallStruct_1b** %ss3.addr, align 8
+  %3 = load %struct.SmallStruct_1b*, %struct.SmallStruct_1b** %ss4.addr, align 8
+  %4 = load %struct.SmallStruct_1b*, %struct.SmallStruct_1b** %ss5.addr, align 8
+  %5 = load %struct.SmallStruct_1b*, %struct.SmallStruct_1b** %ss6.addr, align 8
+  %6 = load %struct.SmallStruct_1b*, %struct.SmallStruct_1b** %ss7.addr, align 8
+  %7 = load %struct.SmallStruct_1b*, %struct.SmallStruct_1b** %ss8.addr, align 8
+  %8 = load %struct.SmallStruct_1b*, %struct.SmallStruct_1b** %ss9.addr, align 8
   %9 = bitcast %struct.SmallStruct_1b* %0 to { i8 }*
-  %10 = getelementptr { i8 }* %9, i32 0, i32 0
-  %11 = load i8* %10, align 1
+  %10 = getelementptr { i8 }, { i8 }* %9, i32 0, i32 0
+  %11 = load i8, i8* %10, align 1
   %12 = bitcast %struct.SmallStruct_1b* %1 to { i8 }*
-  %13 = getelementptr { i8 }* %12, i32 0, i32 0
-  %14 = load i8* %13, align 1
+  %13 = getelementptr { i8 }, { i8 }* %12, i32 0, i32 0
+  %14 = load i8, i8* %13, align 1
   %15 = bitcast %struct.SmallStruct_1b* %2 to { i8 }*
-  %16 = getelementptr { i8 }* %15, i32 0, i32 0
-  %17 = load i8* %16, align 1
+  %16 = getelementptr { i8 }, { i8 }* %15, i32 0, i32 0
+  %17 = load i8, i8* %16, align 1
   %18 = bitcast %struct.SmallStruct_1b* %3 to { i8 }*
-  %19 = getelementptr { i8 }* %18, i32 0, i32 0
-  %20 = load i8* %19, align 1
+  %19 = getelementptr { i8 }, { i8 }* %18, i32 0, i32 0
+  %20 = load i8, i8* %19, align 1
   %21 = bitcast %struct.SmallStruct_1b* %4 to { i8 }*
-  %22 = getelementptr { i8 }* %21, i32 0, i32 0
-  %23 = load i8* %22, align 1
+  %22 = getelementptr { i8 }, { i8 }* %21, i32 0, i32 0
+  %23 = load i8, i8* %22, align 1
   %24 = bitcast %struct.SmallStruct_1b* %5 to { i8 }*
-  %25 = getelementptr { i8 }* %24, i32 0, i32 0
-  %26 = load i8* %25, align 1
+  %25 = getelementptr { i8 }, { i8 }* %24, i32 0, i32 0
+  %26 = load i8, i8* %25, align 1
   %27 = bitcast %struct.SmallStruct_1b* %6 to { i8 }*
-  %28 = getelementptr { i8 }* %27, i32 0, i32 0
-  %29 = load i8* %28, align 1
+  %28 = getelementptr { i8 }, { i8 }* %27, i32 0, i32 0
+  %29 = load i8, i8* %28, align 1
   %30 = bitcast %struct.SmallStruct_1b* %7 to { i8 }*
-  %31 = getelementptr { i8 }* %30, i32 0, i32 0
-  %32 = load i8* %31, align 1
+  %31 = getelementptr { i8 }, { i8 }* %30, i32 0, i32 0
+  %32 = load i8, i8* %31, align 1
   %33 = bitcast %struct.SmallStruct_1b* %8 to { i8 }*
-  %34 = getelementptr { i8 }* %33, i32 0, i32 0
-  %35 = load i8* %34, align 1
-  call void (i8*, ...)* @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8]* @.str, i32 0, i32 0), i8 inreg %11, i8 inreg %14, i8 inreg %17, i8 inreg %20, i8 inreg %23, i8 inreg %26, i8 inreg %29, i8 inreg %32, i8 inreg %35)
+  %34 = getelementptr { i8 }, { i8 }* %33, i32 0, i32 0
+  %35 = load i8, i8* %34, align 1
+  call void (i8*, ...) @varArgF_SmallStruct(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @.str, i32 0, i32 0), i8 inreg %11, i8 inreg %14, i8 inreg %17, i8 inreg %20, i8 inreg %23, i8 inreg %26, i8 inreg %29, i8 inreg %32, i8 inreg %35)
   ret void
  ; CHECK-LABEL: smallStruct_1b_x9:
  ; CHECK: dsll $[[R1:[0-9]+]], $[[R2:[0-9]+]], 56
diff --git a/test/CodeGen/Mips/cconv/arguments-varargs.ll b/test/CodeGen/Mips/cconv/arguments-varargs.ll
index adacda5bc420..af217c92dab8 100644
--- a/test/CodeGen/Mips/cconv/arguments-varargs.ll
+++ b/test/CodeGen/Mips/cconv/arguments-varargs.ll
@@ -1,14 +1,14 @@
 ; RUN: llc -mtriple=mips-linux -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=O32 --check-prefix=O32-BE %s
 ; RUN: llc -mtriple=mipsel-linux -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=O32 --check-prefix=O32-LE %s
 
-; RUN-TODO: llc -march=mips64 -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
-; RUN-TODO: llc -march=mips64el -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64 -relocation-model=static -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64el -relocation-model=static -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 
-; RUN: llc -mtriple=mips64-linux -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=NEW --check-prefix=N32 --check-prefix=NEW-BE %s
-; RUN: llc -mtriple=mips64el-linux -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=NEW --check-prefix=N32 --check-prefix=NEW-LE %s
+; RUN: llc -mtriple=mips64-linux -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=NEW --check-prefix=N32 --check-prefix=NEW-BE %s
+; RUN: llc -mtriple=mips64el-linux -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=NEW --check-prefix=N32 --check-prefix=NEW-LE %s
 
-; RUN: llc -march=mips64 -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=NEW --check-prefix=N64 --check-prefix=NEW-BE %s
-; RUN: llc -march=mips64el -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=NEW --check-prefix=N64 --check-prefix=NEW-LE %s
+; RUN: llc -march=mips64 -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=NEW --check-prefix=N64 --check-prefix=NEW-BE %s
+; RUN: llc -march=mips64el -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=NEW --check-prefix=N64 --check-prefix=NEW-LE %s
 
 @hwords = global [3 x i16] zeroinitializer, align 1
 @words  = global [3 x i32] zeroinitializer, align 1
@@ -119,12 +119,12 @@ entry:
 
   call void asm sideeffect "# ANCHOR1", ""()
   %arg1 = va_arg i8** %ap, i16
-  %e1 = getelementptr [3 x i16]* @hwords, i32 0, i32 1
+  %e1 = getelementptr [3 x i16], [3 x i16]* @hwords, i32 0, i32 1
   store volatile i16 %arg1, i16* %e1, align 2
 
   call void asm sideeffect "# ANCHOR2", ""()
   %arg2 = va_arg i8** %ap, i16
-  %e2 = getelementptr [3 x i16]* @hwords, i32 0, i32 2
+  %e2 = getelementptr [3 x i16], [3 x i16]* @hwords, i32 0, i32 2
   store volatile i16 %arg2, i16* %e2, align 2
 
   call void @llvm.va_end(i8* %ap2)
@@ -237,12 +237,12 @@ entry:
 
   call void asm sideeffect "# ANCHOR1", ""()
   %arg1 = va_arg i8** %ap, i32
-  %e1 = getelementptr [3 x i32]* @words, i32 0, i32 1
+  %e1 = getelementptr [3 x i32], [3 x i32]* @words, i32 0, i32 1
   store volatile i32 %arg1, i32* %e1, align 4
 
   call void asm sideeffect "# ANCHOR2", ""()
   %arg2 = va_arg i8** %ap, i32
-  %e2 = getelementptr [3 x i32]* @words, i32 0, i32 2
+  %e2 = getelementptr [3 x i32], [3 x i32]* @words, i32 0, i32 2
   store volatile i32 %arg2, i32* %e2, align 4
 
   call void @llvm.va_end(i8* %ap2)
@@ -364,12 +364,12 @@ entry:
 
   call void asm sideeffect "# ANCHOR1", ""()
   %arg1 = va_arg i8** %ap, i64
-  %e1 = getelementptr [3 x i64]* @dwords, i32 0, i32 1
+  %e1 = getelementptr [3 x i64], [3 x i64]* @dwords, i32 0, i32 1
   store volatile i64 %arg1, i64* %e1, align 8
 
   call void asm sideeffect "# ANCHOR2", ""()
   %arg2 = va_arg i8** %ap, i64
-  %e2 = getelementptr [3 x i64]* @dwords, i32 0, i32 2
+  %e2 = getelementptr [3 x i64], [3 x i64]* @dwords, i32 0, i32 2
   store volatile i64 %arg2, i64* %e2, align 8
 
   call void @llvm.va_end(i8* %ap2)
@@ -482,12 +482,12 @@ entry:
 
   call void asm sideeffect "# ANCHOR1", ""()
   %arg1 = va_arg i8** %ap, i16
-  %e1 = getelementptr [3 x i16]* @hwords, i32 0, i32 1
+  %e1 = getelementptr [3 x i16], [3 x i16]* @hwords, i32 0, i32 1
   store volatile i16 %arg1, i16* %e1, align 2
 
   call void asm sideeffect "# ANCHOR2", ""()
   %arg2 = va_arg i8** %ap, i16
-  %e2 = getelementptr [3 x i16]* @hwords, i32 0, i32 2
+  %e2 = getelementptr [3 x i16], [3 x i16]* @hwords, i32 0, i32 2
   store volatile i16 %arg2, i16* %e2, align 2
 
   call void @llvm.va_end(i8* %ap2)
@@ -600,12 +600,12 @@ entry:
 
   call void asm sideeffect "# ANCHOR1", ""()
   %arg1 = va_arg i8** %ap, i32
-  %e1 = getelementptr [3 x i32]* @words, i32 0, i32 1
+  %e1 = getelementptr [3 x i32], [3 x i32]* @words, i32 0, i32 1
   store volatile i32 %arg1, i32* %e1, align 4
 
   call void asm sideeffect "# ANCHOR2", ""()
   %arg2 = va_arg i8** %ap, i32
-  %e2 = getelementptr [3 x i32]* @words, i32 0, i32 2
+  %e2 = getelementptr [3 x i32], [3 x i32]* @words, i32 0, i32 2
   store volatile i32 %arg2, i32* %e2, align 4
 
   call void @llvm.va_end(i8* %ap2)
@@ -727,12 +727,12 @@ entry:
 
   call void asm sideeffect "# ANCHOR1", ""()
   %arg1 = va_arg i8** %ap, i64
-  %e1 = getelementptr [3 x i64]* @dwords, i32 0, i32 1
+  %e1 = getelementptr [3 x i64], [3 x i64]* @dwords, i32 0, i32 1
   store volatile i64 %arg1, i64* %e1, align 8
 
   call void asm sideeffect "# ANCHOR2", ""()
   %arg2 = va_arg i8** %ap, i64
-  %e2 = getelementptr [3 x i64]* @dwords, i32 0, i32 2
+  %e2 = getelementptr [3 x i64], [3 x i64]* @dwords, i32 0, i32 2
   store volatile i64 %arg2, i64* %e2, align 8
 
   call void @llvm.va_end(i8* %ap2)
@@ -844,12 +844,12 @@ entry:
 
   call void asm sideeffect "# ANCHOR1", ""()
   %arg1 = va_arg i8** %ap, i16
-  %e1 = getelementptr [3 x i16]* @hwords, i32 0, i32 1
+  %e1 = getelementptr [3 x i16], [3 x i16]* @hwords, i32 0, i32 1
   store volatile i16 %arg1, i16* %e1, align 2
 
   call void asm sideeffect "# ANCHOR2", ""()
   %arg2 = va_arg i8** %ap, i16
-  %e2 = getelementptr [3 x i16]* @hwords, i32 0, i32 2
+  %e2 = getelementptr [3 x i16], [3 x i16]* @hwords, i32 0, i32 2
   store volatile i16 %arg2, i16* %e2, align 2
 
   call void @llvm.va_end(i8* %ap2)
@@ -961,12 +961,12 @@ entry:
 
   call void asm sideeffect "# ANCHOR1", ""()
   %arg1 = va_arg i8** %ap, i32
-  %e1 = getelementptr [3 x i32]* @words, i32 0, i32 1
+  %e1 = getelementptr [3 x i32], [3 x i32]* @words, i32 0, i32 1
   store volatile i32 %arg1, i32* %e1, align 4
 
   call void asm sideeffect "# ANCHOR2", ""()
   %arg2 = va_arg i8** %ap, i32
-  %e2 = getelementptr [3 x i32]* @words, i32 0, i32 2
+  %e2 = getelementptr [3 x i32], [3 x i32]* @words, i32 0, i32 2
   store volatile i32 %arg2, i32* %e2, align 4
 
   call void @llvm.va_end(i8* %ap2)
@@ -1087,12 +1087,12 @@ entry:
 
   call void asm sideeffect "# ANCHOR1", ""()
   %arg1 = va_arg i8** %ap, i64
-  %e1 = getelementptr [3 x i64]* @dwords, i32 0, i32 1
+  %e1 = getelementptr [3 x i64], [3 x i64]* @dwords, i32 0, i32 1
   store volatile i64 %arg1, i64* %e1, align 8
 
   call void asm sideeffect "# ANCHOR2", ""()
   %arg2 = va_arg i8** %ap, i64
-  %e2 = getelementptr [3 x i64]* @dwords, i32 0, i32 2
+  %e2 = getelementptr [3 x i64], [3 x i64]* @dwords, i32 0, i32 2
   store volatile i64 %arg2, i64* %e2, align 8
 
   call void @llvm.va_end(i8* %ap2)
diff --git a/test/CodeGen/Mips/cconv/arguments.ll b/test/CodeGen/Mips/cconv/arguments.ll
index 43da6044408b..430705f8d418 100644
--- a/test/CodeGen/Mips/cconv/arguments.ll
+++ b/test/CodeGen/Mips/cconv/arguments.ll
@@ -1,14 +1,14 @@
 ; RUN: llc -march=mips -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
 ; RUN: llc -march=mipsel -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
 
-; RUN-TODO: llc -march=mips64 -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
-; RUN-TODO: llc -march=mips64el -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64 -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64el -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=O32 %s
 
-; RUN: llc -march=mips64 -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=NEW %s
-; RUN: llc -march=mips64el -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=NEW %s
+; RUN: llc -march=mips64 -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=NEW %s
+; RUN: llc -march=mips64el -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM32 --check-prefix=NEW %s
 
-; RUN: llc -march=mips64 -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=NEW %s
-; RUN: llc -march=mips64el -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=NEW %s
+; RUN: llc -march=mips64 -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=NEW %s
+; RUN: llc -march=mips64el -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=SYM64 --check-prefix=NEW %s
 
 ; Test the integer arguments for all ABI's and byte orders as specified by
 ; section 5 of MD00305 (MIPS ABIs Described).
@@ -28,25 +28,25 @@ define void @align_to_arg_slots(i8 signext %a, i8 signext %b, i8 signext %c,
                                 i8 signext %g, i8 signext %h, i8 signext %i,
                                 i8 signext %j) nounwind {
 entry:
-        %0 = getelementptr [11 x i8]* @bytes, i32 0, i32 1
+        %0 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 1
         store volatile i8 %a, i8* %0
-        %1 = getelementptr [11 x i8]* @bytes, i32 0, i32 2
+        %1 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 2
         store volatile i8 %b, i8* %1
-        %2 = getelementptr [11 x i8]* @bytes, i32 0, i32 3
+        %2 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 3
         store volatile i8 %c, i8* %2
-        %3 = getelementptr [11 x i8]* @bytes, i32 0, i32 4
+        %3 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 4
         store volatile i8 %d, i8* %3
-        %4 = getelementptr [11 x i8]* @bytes, i32 0, i32 5
+        %4 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 5
         store volatile i8 %e, i8* %4
-        %5 = getelementptr [11 x i8]* @bytes, i32 0, i32 6
+        %5 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 6
         store volatile i8 %f, i8* %5
-        %6 = getelementptr [11 x i8]* @bytes, i32 0, i32 7
+        %6 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 7
         store volatile i8 %g, i8* %6
-        %7 = getelementptr [11 x i8]* @bytes, i32 0, i32 8
+        %7 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 8
         store volatile i8 %h, i8* %7
-        %8 = getelementptr [11 x i8]* @bytes, i32 0, i32 9
+        %8 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 9
         store volatile i8 %i, i8* %8
-        %9 = getelementptr [11 x i8]* @bytes, i32 0, i32 10
+        %9 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 10
         store volatile i8 %j, i8* %9
         ret void
 }
@@ -95,23 +95,23 @@ define void @slot_skipping(i8 signext %a, i64 signext %b, i8 signext %c,
                            i8 signext %d, i8 signext %e, i8 signext %f,
                            i8 signext %g, i64 signext %i, i8 signext %j) nounwind {
 entry:
-        %0 = getelementptr [11 x i8]* @bytes, i32 0, i32 1
+        %0 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 1
         store volatile i8 %a, i8* %0
-        %1 = getelementptr [11 x i64]* @dwords, i32 0, i32 1
+        %1 = getelementptr [11 x i64], [11 x i64]* @dwords, i32 0, i32 1
         store volatile i64 %b, i64* %1
-        %2 = getelementptr [11 x i8]* @bytes, i32 0, i32 2
+        %2 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 2
         store volatile i8 %c, i8* %2
-        %3 = getelementptr [11 x i8]* @bytes, i32 0, i32 3
+        %3 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 3
         store volatile i8 %d, i8* %3
-        %4 = getelementptr [11 x i8]* @bytes, i32 0, i32 4
+        %4 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 4
         store volatile i8 %e, i8* %4
-        %5 = getelementptr [11 x i8]* @bytes, i32 0, i32 5
+        %5 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 5
         store volatile i8 %f, i8* %5
-        %6 = getelementptr [11 x i8]* @bytes, i32 0, i32 6
+        %6 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 6
         store volatile i8 %g, i8* %6
-        %7 = getelementptr [11 x i64]* @dwords, i32 0, i32 2
+        %7 = getelementptr [11 x i64], [11 x i64]* @dwords, i32 0, i32 2
         store volatile i64 %i, i64* %7
-        %8 = getelementptr [11 x i8]* @bytes, i32 0, i32 7
+        %8 = getelementptr [11 x i8], [11 x i8]* @bytes, i32 0, i32 7
         store volatile i8 %j, i8* %8
         ret void
 }
diff --git a/test/CodeGen/Mips/cconv/callee-saved-float.ll b/test/CodeGen/Mips/cconv/callee-saved-float.ll
index de4d9171aec4..c84f0f439c26 100644
--- a/test/CodeGen/Mips/cconv/callee-saved-float.ll
+++ b/test/CodeGen/Mips/cconv/callee-saved-float.ll
@@ -3,20 +3,20 @@
 ; RUN: llc -march=mips < %s | FileCheck --check-prefix=ALL --check-prefix=O32-INV %s
 ; RUN: llc -march=mipsel < %s | FileCheck --check-prefix=ALL --check-prefix=O32-INV %s
 
-; RUN-TODO: llc -march=mips64 -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
-; RUN-TODO: llc -march=mips64el -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
-; RUN-TODO: llc -march=mips64 -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=ALL-INV --check-prefix=O32-INV %s
-; RUN-TODO: llc -march=mips64el -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=ALL-INV --check-prefix=O32-INV %s
+; RUN-TODO: llc -march=mips64 -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64el -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64 -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=ALL-INV --check-prefix=O32-INV %s
+; RUN-TODO: llc -march=mips64el -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=ALL-INV --check-prefix=O32-INV %s
 
-; RUN: llc -march=mips64 -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
-; RUN: llc -march=mips64el -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
-; RUN: llc -march=mips64 -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=ALL-INV --check-prefix=N32-INV %s
-; RUN: llc -march=mips64el -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=ALL-INV --check-prefix=N32-INV %s
+; RUN: llc -march=mips64 -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -march=mips64el -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -march=mips64 -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=ALL-INV --check-prefix=N32-INV %s
+; RUN: llc -march=mips64el -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=ALL-INV --check-prefix=N32-INV %s
 
-; RUN: llc -march=mips64 -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
-; RUN: llc -march=mips64el -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
-; RUN: llc -march=mips64 -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=ALL-INV --check-prefix=N64-INV %s
-; RUN: llc -march=mips64el -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=ALL-INV --check-prefix=N64-INV %s
+; RUN: llc -march=mips64 -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -march=mips64el -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -march=mips64 -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=ALL-INV --check-prefix=N64-INV %s
+; RUN: llc -march=mips64el -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=ALL-INV --check-prefix=N64-INV %s
 
 ; Test the the callee-saved registers are callee-saved as specified by section
 ; 2 of the MIPSpro N32 Handbook and section 3 of the SYSV ABI spec.
diff --git a/test/CodeGen/Mips/cconv/callee-saved.ll b/test/CodeGen/Mips/cconv/callee-saved.ll
index 293e99f0c8e6..d0b1e64cdeea 100644
--- a/test/CodeGen/Mips/cconv/callee-saved.ll
+++ b/test/CodeGen/Mips/cconv/callee-saved.ll
@@ -3,20 +3,20 @@
 ; RUN: llc -march=mips < %s | FileCheck --check-prefix=ALL --check-prefix=O32-INV %s
 ; RUN: llc -march=mipsel < %s | FileCheck --check-prefix=ALL --check-prefix=O32-INV %s
 
-; RUN-TODO: llc -march=mips64 -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
-; RUN-TODO: llc -march=mips64el -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
-; RUN-TODO: llc -march=mips64 -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32-INV %s
-; RUN-TODO: llc -march=mips64el -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32-INV %s
+; RUN-TODO: llc -march=mips64 -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64el -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64 -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32-INV %s
+; RUN-TODO: llc -march=mips64el -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32-INV %s
 
-; RUN: llc -march=mips64 -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
-; RUN: llc -march=mips64el -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
-; RUN: llc -march=mips64 -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32-INV %s
-; RUN: llc -march=mips64el -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32-INV %s
+; RUN: llc -march=mips64 -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -march=mips64el -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -march=mips64 -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32-INV %s
+; RUN: llc -march=mips64el -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32-INV %s
 
-; RUN: llc -march=mips64 -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
-; RUN: llc -march=mips64el -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
-; RUN: llc -march=mips64 -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64-INV %s
-; RUN: llc -march=mips64el -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64-INV %s
+; RUN: llc -march=mips64 -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -march=mips64el -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -march=mips64 -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64-INV %s
+; RUN: llc -march=mips64el -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64-INV %s
 
 ; Test the the callee-saved registers are callee-saved as specified by section
 ; 2 of the MIPSpro N32 Handbook and section 3 of the SYSV ABI spec.
diff --git a/test/CodeGen/Mips/cconv/memory-layout.ll b/test/CodeGen/Mips/cconv/memory-layout.ll
index 0c3cc9ecedfe..33a68da157f6 100644
--- a/test/CodeGen/Mips/cconv/memory-layout.ll
+++ b/test/CodeGen/Mips/cconv/memory-layout.ll
@@ -1,14 +1,14 @@
 ; RUN: llc -march=mips < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 ; RUN: llc -march=mipsel < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 
-; RUN-TODO: llc -march=mips64 -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
-; RUN-TODO: llc -march=mips64el -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64 -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64el -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 
-; RUN: llc -march=mips64 -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
-; RUN: llc -march=mips64el -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -march=mips64 -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -march=mips64el -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
 
-; RUN: llc -march=mips64 -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
-; RUN: llc -march=mips64el -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -march=mips64 -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -march=mips64el -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
 
 ; Test the memory layout for all ABI's and byte orders as specified by section
 ; 4 of MD00305 (MIPS ABIs Described).
diff --git a/test/CodeGen/Mips/cconv/reserved-space.ll b/test/CodeGen/Mips/cconv/reserved-space.ll
index b36f89ecc115..23190c2790cc 100644
--- a/test/CodeGen/Mips/cconv/reserved-space.ll
+++ b/test/CodeGen/Mips/cconv/reserved-space.ll
@@ -1,14 +1,14 @@
 ; RUN: llc -march=mips < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 ; RUN: llc -march=mipsel < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 
-; RUN-TODO: llc -march=mips64 -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
-; RUN-TODO: llc -march=mips64el -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64 -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64el -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 
-; RUN: llc -march=mips64 -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
-; RUN: llc -march=mips64el -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -march=mips64 -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -march=mips64el -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
 
-; RUN: llc -march=mips64 -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
-; RUN: llc -march=mips64el -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -march=mips64 -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -march=mips64el -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
 
 ; Test that O32 correctly reserved space for the four arguments, even when
 ; there aren't any as per section 5 of MD00305 (MIPS ABIs Described).
diff --git a/test/CodeGen/Mips/cconv/return-float.ll b/test/CodeGen/Mips/cconv/return-float.ll
index d1a5e4f2fa9d..8eb8c411e081 100644
--- a/test/CodeGen/Mips/cconv/return-float.ll
+++ b/test/CodeGen/Mips/cconv/return-float.ll
@@ -1,14 +1,14 @@
-; RUN: llc -mtriple=mips-linux-gnu -soft-float -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
-; RUN: llc -mtriple=mipsel-linux-gnu -soft-float -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN: llc -mtriple=mips-linux-gnu -mattr=+soft-float -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN: llc -mtriple=mipsel-linux-gnu -mattr=+soft-float -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 
-; RUN-TODO: llc -mtriple=mips64-linux-gnu -soft-float -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
-; RUN-TODO: llc -mtriple=mips64el-linux-gnu -soft-float -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -mtriple=mips64-linux-gnu -mattr=+soft-float -relocation-model=static -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -mtriple=mips64el-linux-gnu -mattr=+soft-float -relocation-model=static -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 
-; RUN: llc -mtriple=mips64-linux-gnu -soft-float -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
-; RUN: llc -mtriple=mips64el-linux-gnu -soft-float -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -mtriple=mips64-linux-gnu -mattr=+soft-float -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -mtriple=mips64el-linux-gnu -mattr=+soft-float -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
 
-; RUN: llc -mtriple=mips64-linux-gnu -soft-float -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
-; RUN: llc -mtriple=mips64el-linux-gnu -soft-float -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -mtriple=mips64-linux-gnu -mattr=+soft-float -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -mtriple=mips64el-linux-gnu -mattr=+soft-float -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
 
 ; Test the float returns for all ABI's and byte orders as specified by
 ; section 5 of MD00305 (MIPS ABIs Described).
@@ -21,7 +21,7 @@
 
 define float @retfloat() nounwind {
 entry:
-        %0 = load volatile float* @float
+        %0 = load volatile float, float* @float
         ret float %0
 }
 
@@ -35,7 +35,7 @@ entry:
 
 define double @retdouble() nounwind {
 entry:
-        %0 = load volatile double* @double
+        %0 = load volatile double, double* @double
         ret double %0
 }
 
diff --git a/test/CodeGen/Mips/cconv/return-hard-float.ll b/test/CodeGen/Mips/cconv/return-hard-float.ll
index 123b499185a9..14853c8ca6f7 100644
--- a/test/CodeGen/Mips/cconv/return-hard-float.ll
+++ b/test/CodeGen/Mips/cconv/return-hard-float.ll
@@ -1,14 +1,14 @@
 ; RUN: llc -mtriple=mips-linux-gnu -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 ; RUN: llc -mtriple=mipsel-linux-gnu -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 
-; RUN-TODO: llc -mtriple=mips64-linux-gnu -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
-; RUN-TODO: llc -mtriple=mips64el-linux-gnu -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -mtriple=mips64-linux-gnu -relocation-model=static -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -mtriple=mips64el-linux-gnu -relocation-model=static -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 
-; RUN: llc -mtriple=mips64-linux-gnu -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
-; RUN: llc -mtriple=mips64el-linux-gnu -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -mtriple=mips64-linux-gnu -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -mtriple=mips64el-linux-gnu -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
 
-; RUN: llc -mtriple=mips64-linux-gnu -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
-; RUN: llc -mtriple=mips64el-linux-gnu -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -mtriple=mips64-linux-gnu -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -mtriple=mips64el-linux-gnu -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
 
 ; RUN: llc -mtriple=mips-linux-gnu -relocation-model=static -mattr=+o32,+fp64 < %s | FileCheck --check-prefix=ALL --check-prefix=032FP64 %s
 ; RUN: llc -mtriple=mipsel-linux-gnu -relocation-model=static -mattr=+o32,+fp64 < %s | FileCheck --check-prefix=ALL --check-prefix=032FP64 %s
@@ -24,7 +24,7 @@
 
 define float @retfloat() nounwind {
 entry:
-        %0 = load volatile float* @float
+        %0 = load volatile float, float* @float
         ret float %0
 }
 
@@ -38,7 +38,7 @@ entry:
 
 define double @retdouble() nounwind {
 entry:
-        %0 = load volatile double* @double
+        %0 = load volatile double, double* @double
         ret double %0
 }
 
@@ -50,7 +50,7 @@ entry:
 
 define { double, double } @retComplexDouble() #0 {
   %retval = alloca { double, double }, align 8
-  %1 = load { double, double }* %retval
+  %1 = load { double, double }, { double, double }* %retval
   ret { double, double } %1
 }
 
diff --git a/test/CodeGen/Mips/cconv/return-hard-fp128.ll b/test/CodeGen/Mips/cconv/return-hard-fp128.ll
index 0da59efddd6c..34e9647acddd 100644
--- a/test/CodeGen/Mips/cconv/return-hard-fp128.ll
+++ b/test/CodeGen/Mips/cconv/return-hard-fp128.ll
@@ -1,8 +1,8 @@
-; RUN: llc -march=mips64 -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
-; RUN: llc -march=mips64el -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -march=mips64 -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -march=mips64el -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
 
-; RUN: llc -march=mips64 -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
-; RUN: llc -march=mips64el -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -march=mips64 -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -march=mips64el -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
 
 ; Test the fp128 returns for N32/N64 and all byte orders as specified by
 ; section 5 of MD00305 (MIPS ABIs Described).
@@ -13,7 +13,7 @@
 
 define fp128 @retldouble() nounwind {
 entry:
-        %0 = load volatile fp128* @fp128
+        %0 = load volatile fp128, fp128* @fp128
         ret fp128 %0
 }
 
diff --git a/test/CodeGen/Mips/cconv/return-hard-struct-f128.ll b/test/CodeGen/Mips/cconv/return-hard-struct-f128.ll
index 2e8447710281..c4c8f10ca3b4 100644
--- a/test/CodeGen/Mips/cconv/return-hard-struct-f128.ll
+++ b/test/CodeGen/Mips/cconv/return-hard-struct-f128.ll
@@ -1,8 +1,8 @@
-; RUN: llc -mtriple=mips64-linux-gnu -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
-; RUN: llc -mtriple=mips64el-linux-gnu -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -mtriple=mips64-linux-gnu -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -mtriple=mips64el-linux-gnu -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
 
-; RUN: llc -mtriple=mips64-linux-gnu -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
-; RUN: llc -mtriple=mips64el-linux-gnu -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -mtriple=mips64-linux-gnu -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -mtriple=mips64el-linux-gnu -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
 
 ; Test return of {fp128} agrees with de-facto N32/N64 ABI.
 
@@ -10,7 +10,7 @@
 
 define inreg {fp128} @ret_struct_fp128() nounwind {
 entry:
-        %0 = load volatile {fp128}* @struct_fp128
+        %0 = load volatile {fp128}, {fp128}* @struct_fp128
         ret {fp128} %0
 }
 
diff --git a/test/CodeGen/Mips/cconv/return-struct.ll b/test/CodeGen/Mips/cconv/return-struct.ll
index 11a8cf032148..8decd04f089b 100644
--- a/test/CodeGen/Mips/cconv/return-struct.ll
+++ b/test/CodeGen/Mips/cconv/return-struct.ll
@@ -1,14 +1,14 @@
 ; RUN: llc -mtriple=mips-linux-gnu -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=O32 --check-prefix=O32-BE %s
 ; RUN: llc -mtriple=mipsel-linux-gnu -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=O32 --check-prefix=O32-LE %s
 
-; RUN-TODO: llc -mtriple=mips64-linux-gnu -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
-; RUN-TODO: llc -mtriple=mips64el-linux-gnu -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -mtriple=mips64-linux-gnu -relocation-model=static -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -mtriple=mips64el-linux-gnu -relocation-model=static -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 
-; RUN: llc -mtriple=mips64-linux-gnu -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 --check-prefix=N32-BE %s
-; RUN: llc -mtriple=mips64el-linux-gnu -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 --check-prefix=N32-LE %s
+; RUN: llc -mtriple=mips64-linux-gnu -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 --check-prefix=N32-BE %s
+; RUN: llc -mtriple=mips64el-linux-gnu -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 --check-prefix=N32-LE %s
 
-; RUN: llc -mtriple=mips64-linux-gnu -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 --check-prefix=N64-BE %s
-; RUN: llc -mtriple=mips64el-linux-gnu -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 --check-prefix=N64-LE %s
+; RUN: llc -mtriple=mips64-linux-gnu -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 --check-prefix=N64-BE %s
+; RUN: llc -mtriple=mips64el-linux-gnu -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 --check-prefix=N64-LE %s
 
 ; Test struct returns for all ABI's and byte orders.
 
@@ -22,7 +22,7 @@ declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture, i8* nocapture readonly, i
 
 define inreg {i8} @ret_struct_i8() nounwind {
 entry:
-        %0 = load volatile {i8}* @struct_byte
+        %0 = load volatile {i8}, {i8}* @struct_byte
         ret {i8} %0
 }
 
@@ -52,9 +52,9 @@ define inreg {i16} @ret_struct_i16() nounwind {
 entry:
         %retval = alloca {i8,i8}, align 1
         %0 = bitcast {i8,i8}* %retval to i8*
-        call void @llvm.memcpy.p0i8.p0i8.i64(i8* %0, i8* getelementptr inbounds ({i8,i8}* @struct_2byte, i32 0, i32 0), i64 2, i32 1, i1 false)
+        call void @llvm.memcpy.p0i8.p0i8.i64(i8* %0, i8* getelementptr inbounds ({i8,i8}, {i8,i8}* @struct_2byte, i32 0, i32 0), i64 2, i32 1, i1 false)
         %1 = bitcast {i8,i8}* %retval to {i16}*
-        %2 = load volatile {i16}* %1
+        %2 = load volatile {i16}, {i16}* %1
         ret {i16} %2
 }
 
@@ -91,7 +91,7 @@ entry:
 ; missed by the CCPromoteToType and the shift didn't happen.
 define inreg {i48} @ret_struct_3xi16() nounwind {
 entry:
-        %0 = load volatile i48* bitcast ({[3 x i16]}* @struct_3xi16 to i48*), align 2
+        %0 = load volatile i48, i48* bitcast ({[3 x i16]}* @struct_3xi16 to i48*), align 2
         %1 = insertvalue {i48} undef, i48 %0, 0
         ret {i48} %1
 }
@@ -174,7 +174,7 @@ entry:
 ; This time we let the backend lower the sret argument.
 define {[6 x i32]} @ret_struct_6xi32() {
 entry:
-        %0 = load volatile {[6 x i32]}* @struct_6xi32, align 2
+        %0 = load volatile {[6 x i32]}, {[6 x i32]}* @struct_6xi32, align 2
         ret {[6 x i32]} %0
 }
 
diff --git a/test/CodeGen/Mips/cconv/return.ll b/test/CodeGen/Mips/cconv/return.ll
index 63f9b5f45a18..a53767275434 100644
--- a/test/CodeGen/Mips/cconv/return.ll
+++ b/test/CodeGen/Mips/cconv/return.ll
@@ -1,14 +1,14 @@
 ; RUN: llc -mtriple=mips-linux-gnu -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 ; RUN: llc -mtriple=mipsel-linux-gnu -relocation-model=static < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 
-; RUN-TODO: llc -mtriple=mips64-linux-gnu -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
-; RUN-TODO: llc -mtriple=mips64el-linux-gnu -relocation-model=static -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -mtriple=mips64-linux-gnu -relocation-model=static -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -mtriple=mips64el-linux-gnu -relocation-model=static -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 
-; RUN: llc -mtriple=mips64-linux-gnu -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
-; RUN: llc -mtriple=mips64el-linux-gnu -relocation-model=static -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -mtriple=mips64-linux-gnu -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -mtriple=mips64el-linux-gnu -relocation-model=static -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
 
-; RUN: llc -mtriple=mips64-linux-gnu -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
-; RUN: llc -mtriple=mips64el-linux-gnu -relocation-model=static -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -mtriple=mips64-linux-gnu -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -mtriple=mips64el-linux-gnu -relocation-model=static -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
 
 ; Test the integer returns for all ABI's and byte orders as specified by
 ; section 5 of MD00305 (MIPS ABIs Described).
@@ -24,7 +24,7 @@
 
 define i8 @reti8() nounwind {
 entry:
-        %0 = load volatile i8* @byte
+        %0 = load volatile i8, i8* @byte
         ret i8 %0
 }
 
@@ -38,7 +38,7 @@ entry:
 
 define i32 @reti32() nounwind {
 entry:
-        %0 = load volatile i32* @word
+        %0 = load volatile i32, i32* @word
         ret i32 %0
 }
 
@@ -52,7 +52,7 @@ entry:
 
 define i64 @reti64() nounwind {
 entry:
-        %0 = load volatile i64* @dword
+        %0 = load volatile i64, i64* @dword
         ret i64 %0
 }
 
diff --git a/test/CodeGen/Mips/cconv/stack-alignment.ll b/test/CodeGen/Mips/cconv/stack-alignment.ll
index 834033bc8da5..f21bc3066f72 100644
--- a/test/CodeGen/Mips/cconv/stack-alignment.ll
+++ b/test/CodeGen/Mips/cconv/stack-alignment.ll
@@ -1,14 +1,14 @@
 ; RUN: llc -march=mips < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 ; RUN: llc -march=mipsel < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 
-; RUN-TODO: llc -march=mips64 -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
-; RUN-TODO: llc -march=mips64el -mattr=-n64,+o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64 -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
+; RUN-TODO: llc -march=mips64el -target-abi o32 < %s | FileCheck --check-prefix=ALL --check-prefix=O32 %s
 
-; RUN: llc -march=mips64 -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
-; RUN: llc -march=mips64el -mattr=-n64,+n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -march=mips64 -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
+; RUN: llc -march=mips64el -target-abi n32 < %s | FileCheck --check-prefix=ALL --check-prefix=N32 %s
 
-; RUN: llc -march=mips64 -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
-; RUN: llc -march=mips64el -mattr=-n64,+n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -march=mips64 -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
+; RUN: llc -march=mips64el -target-abi n64 < %s | FileCheck --check-prefix=ALL --check-prefix=N64 %s
 
 ; Test the stack alignment for all ABI's and byte orders as specified by
 ; section 5 of MD00305 (MIPS ABIs Described).
diff --git a/test/CodeGen/Mips/cfi_offset.ll b/test/CodeGen/Mips/cfi_offset.ll
index e23855bd65d2..97233328fd55 100644
--- a/test/CodeGen/Mips/cfi_offset.ll
+++ b/test/CodeGen/Mips/cfi_offset.ll
@@ -32,9 +32,9 @@ define void @bar() {
 ; CHECK:  .cfi_offset 31, -20
 ; CHECK:  .cfi_offset 16, -24
 
-    %val1 = load volatile double* @var
-    %val2 = load volatile double* @var
-    call void (...)* @foo() nounwind
+    %val1 = load volatile double, double* @var
+    %val2 = load volatile double, double* @var
+    call void (...) @foo() nounwind
     store volatile double %val1, double* @var
     store volatile double %val2, double* @var
     ret void
diff --git a/test/CodeGen/Mips/check-adde-redundant-moves.ll b/test/CodeGen/Mips/check-adde-redundant-moves.ll
index 527c21770263..7bc63a494ac7 100644
--- a/test/CodeGen/Mips/check-adde-redundant-moves.ll
+++ b/test/CodeGen/Mips/check-adde-redundant-moves.ll
@@ -4,12 +4,18 @@
 ; RUN:    -check-prefix=ALL -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP32
 ; RUN: llc < %s -march=mips64 -mcpu=mips3 | FileCheck %s -check-prefix=ALL
 ; RUN: llc < %s -march=mips64 -mcpu=mips4 | FileCheck %s -check-prefix=ALL
 ; RUN: llc < %s -march=mips64 -mcpu=mips64 | FileCheck %s -check-prefix=ALL
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r2 | FileCheck %s -check-prefix=ALL
+; RUN: llc < %s -march=mips64 -mcpu=mips64r3 | FileCheck %s -check-prefix=ALL
+; RUN: llc < %s -march=mips64 -mcpu=mips64r5 | FileCheck %s -check-prefix=ALL
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r6 | FileCheck %s -check-prefix=ALL
 
 define i64 @add_i64(i64 %a) {
diff --git a/test/CodeGen/Mips/ci2.ll b/test/CodeGen/Mips/ci2.ll
index e2068fdf14e1..4687748879ac 100644
--- a/test/CodeGen/Mips/ci2.ll
+++ b/test/CodeGen/Mips/ci2.ll
@@ -1,4 +1,4 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static -mips16-constant-islands   < %s | FileCheck %s -check-prefix=constisle
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static -mips16-constant-islands   < %s | FileCheck %s -check-prefix=constisle
 
 @i = common global i32 0, align 4
 @b = common global i32 0, align 4
@@ -8,7 +8,7 @@
 define void @foo() #0 {
 entry:
   store i32 305419896, i32* @i, align 4
-  %0 = load i32* @b, align 4
+  %0 = load i32, i32* @b, align 4
   %tobool = icmp ne i32 %0, 0
   br i1 %tobool, label %if.then, label %if.else
 
diff --git a/test/CodeGen/Mips/cmov.ll b/test/CodeGen/Mips/cmov.ll
index b12c2df97c19..a8008a2cb29f 100755
--- a/test/CodeGen/Mips/cmov.ll
+++ b/test/CodeGen/Mips/cmov.ll
@@ -41,8 +41,8 @@
 define i32* @cmov1(i32 signext %s) nounwind readonly {
 entry:
   %tobool = icmp ne i32 %s, 0
-  %tmp1 = load i32** @i3, align 4
-  %cond = select i1 %tobool, i32* getelementptr inbounds ([3 x i32]* @i1, i32 0, i32 0), i32* %tmp1
+  %tmp1 = load i32*, i32** @i3, align 4
+  %cond = select i1 %tobool, i32* getelementptr inbounds ([3 x i32], [3 x i32]* @i1, i32 0, i32 0), i32* %tmp1
   ret i32* %cond
 }
 
@@ -81,8 +81,8 @@ entry:
 define i32 @cmov2(i32 signext %s) nounwind readonly {
 entry:
   %tobool = icmp ne i32 %s, 0
-  %tmp1 = load i32* @c, align 4
-  %tmp2 = load i32* @d, align 4
+  %tmp1 = load i32, i32* @c, align 4
+  %tmp2 = load i32, i32* @d, align 4
   %cond = select i1 %tobool, i32 %tmp1, i32 %tmp2
   ret i32 %cond
 }
diff --git a/test/CodeGen/Mips/cmplarge.ll b/test/CodeGen/Mips/cmplarge.ll
index 2a3d30a95492..79019065a905 100644
--- a/test/CodeGen/Mips/cmplarge.ll
+++ b/test/CodeGen/Mips/cmplarge.ll
@@ -9,8 +9,8 @@ target triple = "mipsel--linux-gnu"
 
 define void @getSubImagesLuma(%struct.StorablePicture* nocapture %s) #0 {
 entry:
-  %size_y = getelementptr inbounds %struct.StorablePicture* %s, i32 0, i32 1
-  %0 = load i32* %size_y, align 4
+  %size_y = getelementptr inbounds %struct.StorablePicture, %struct.StorablePicture* %s, i32 0, i32 1
+  %0 = load i32, i32* %size_y, align 4
   %sub = add nsw i32 %0, -1
   %add5 = add nsw i32 %0, 20
   %cmp6 = icmp sgt i32 %add5, -20
@@ -20,7 +20,7 @@ for.body:                                         ; preds = %entry, %for.body
   %j.07 = phi i32 [ %inc, %for.body ], [ -20, %entry ]
   %call = tail call i32 bitcast (i32 (...)* @iClip3 to i32 (i32, i32, i32)*)(i32 0, i32 %sub, i32 %j.07) #2
   %inc = add nsw i32 %j.07, 1
-  %1 = load i32* %size_y, align 4
+  %1 = load i32, i32* %size_y, align 4
   %add = add nsw i32 %1, 20
   %cmp = icmp slt i32 %inc, %add
   br i1 %cmp, label %for.body, label %for.end
diff --git a/test/CodeGen/Mips/const1.ll b/test/CodeGen/Mips/const1.ll
index f32ce244cf40..1a5d58bd3f95 100644
--- a/test/CodeGen/Mips/const1.ll
+++ b/test/CodeGen/Mips/const1.ll
@@ -1,4 +1,4 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static -mips16-constant-islands < %s | FileCheck %s 
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static -mips16-constant-islands < %s | FileCheck %s 
 
 ; ModuleID = 'const1.c'
 target datalayout = "e-p:32:32:32-i1:8:8-i8:8:32-i16:16:32-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-n32-S64"
diff --git a/test/CodeGen/Mips/const4a.ll b/test/CodeGen/Mips/const4a.ll
index ac6795b2c833..c31e54a01036 100644
--- a/test/CodeGen/Mips/const4a.ll
+++ b/test/CodeGen/Mips/const4a.ll
@@ -1,4 +1,4 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands -mips-constant-islands-no-load-relaxation  < %s | FileCheck %s -check-prefix=no-load-relax
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands -mips-constant-islands-no-load-relaxation  < %s | FileCheck %s -check-prefix=no-load-relax
 
 ; ModuleID = 'const4.c'
 target datalayout = "E-p:32:32:32-i1:8:8-i8:8:32-i16:16:32-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-n32-S64"
@@ -14,7 +14,7 @@ target triple = "mips--linux-gnu"
 define void @t() #0 {
 entry:
   store i32 -559023410, i32* @i, align 4
-  %0 = load i32* @b, align 4
+  %0 = load i32, i32* @b, align 4
 ; no-load-relax:	lw	${{[0-9]+}}, $CPI0_1	# 16 bit inst
   %tobool = icmp ne i32 %0, 0
   br i1 %tobool, label %if.then, label %if.else
diff --git a/test/CodeGen/Mips/const6.ll b/test/CodeGen/Mips/const6.ll
index c26e02f2ebba..49e98ea78703 100644
--- a/test/CodeGen/Mips/const6.ll
+++ b/test/CodeGen/Mips/const6.ll
@@ -1,6 +1,6 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=load-relax
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=load-relax
 
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands -mips-constant-islands-no-load-relaxation  < %s | FileCheck %s -check-prefix=no-load-relax
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands -mips-constant-islands-no-load-relaxation  < %s | FileCheck %s -check-prefix=no-load-relax
 
 ; ModuleID = 'const6.c'
 target datalayout = "E-p:32:32:32-i1:8:8-i8:8:32-i16:16:32-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-n32-S64"
diff --git a/test/CodeGen/Mips/const6a.ll b/test/CodeGen/Mips/const6a.ll
index aff1357c3a8f..54a3f2234dc2 100644
--- a/test/CodeGen/Mips/const6a.ll
+++ b/test/CodeGen/Mips/const6a.ll
@@ -1,6 +1,6 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=load-relax1
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=load-relax1
 
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=load-relax
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=load-relax
 
 ; ModuleID = 'const6a.c'
 target datalayout = "E-p:32:32:32-i1:8:8-i8:8:32-i16:16:32-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-n32-S64"
diff --git a/test/CodeGen/Mips/ctlz.ll b/test/CodeGen/Mips/ctlz.ll
index 1f871664a6cf..96af1973b196 100644
--- a/test/CodeGen/Mips/ctlz.ll
+++ b/test/CodeGen/Mips/ctlz.ll
@@ -9,7 +9,7 @@ define i32 @main() #0 {
 entry:
   %retval = alloca i32, align 4
   store i32 0, i32* %retval
-  %0 = load i32* @x, align 4
+  %0 = load i32, i32* @x, align 4
   %1 = call i32 @llvm.ctlz.i32(i32 %0, i1 true)
   store i32 %1, i32* @y, align 4
   ret i32 0
diff --git a/test/CodeGen/Mips/dagcombine_crash.ll b/test/CodeGen/Mips/dagcombine_crash.ll
new file mode 100644
index 000000000000..6fcf2b4e68b4
--- /dev/null
+++ b/test/CodeGen/Mips/dagcombine_crash.ll
@@ -0,0 +1,25 @@
+; RUN: llc -o - %s | FileCheck %s
+; The selection DAG select(select()) normalisation crashed for different types
+; on the condition inputs.
+target datalayout = "E-m:m-p:32:32-i8:8:32-i16:16:32-i64:64-n32-S64"
+target triple = "mips--"
+
+; CHECK-LABEL: foobar
+; CHECK: sltiu ${{[0-9]*}}, ${{[0-9]*}}, 42
+; CHECK: sltiu ${{[0-9]*}}, ${{[0-9]*}}, 23
+; CHECK: and ${{[0-9]*}}, ${{[0-9]*}}, ${{[0-9]*}}
+; CHECK: sltu ${{[0-9]*}}, ${{[0-9]*}}, ${{[0-9]*}}
+; CHECK: addiu ${{[0-9]*}}, ${{[0-9]*}}, -1
+; CHECK: movn ${{[0-9]*}}, ${{[0-9]*}}, ${{[0-9]*}}
+; CHECK: jr $ra
+; CHECK: move ${{[0-9]*}}, ${{[0-9]*}}
+define i64 @foobar(i32 %arg) #0 {
+entry:
+  %cmp0 = icmp ult i32 %arg, 23
+  %cmp1 = icmp ult i32 %arg, 42
+  %and = and i1 %cmp0, %cmp1
+  %cmp2 = icmp ugt i32 %arg, 0
+  %sext = sext i1 %cmp1 to i64
+  %retval.0 = select i1 %and, i64 %sext, i64 0
+  ret i64 %retval.0
+}
diff --git a/test/CodeGen/Mips/delay-slot-fill-forward.ll b/test/CodeGen/Mips/delay-slot-fill-forward.ll
new file mode 100644
index 000000000000..7fc011da92db
--- /dev/null
+++ b/test/CodeGen/Mips/delay-slot-fill-forward.ll
@@ -0,0 +1,183 @@
+; RUN: llc < %s -march=mips -mcpu=mips32r2 -O2 \
+; RUN:     -disable-mips-df-forward-search=false \
+; RUN:     -disable-mips-df-succbb-search=false \
+; RUN:     -relocation-model=static | FileCheck %s
+
+; This test was generated with bugpoint from
+; MultiSource/Applications/JM/lencod/me_fullsearch.c
+
+%struct.SubImageContainer = type { i16****, [2 x i16****] }
+%struct.storable_picture = type { i32, i32, i32, i32, i32, i32,
+  [6 x [33 x i64]], [6 x [33 x i64]], [6 x [33 x i64]], [6 x [33 x i64]],
+  i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32,
+  i32, i32, i32, i32, i32, i16**, i16****, i16****, i16*****, i16***,
+  i8*, i8***, i64***, i64***, i16****, i8**, i8**, %struct.storable_picture*,
+  %struct.storable_picture*, %struct.storable_picture*,
+  i32, i32, i32, i32, i32, i32, i32 }
+
+@img_height = external global i16, align 2
+@width_pad = external global i32, align 4
+@height_pad = external global i32, align 4
+@mvbits = external global i32*, align 4
+@ref_pic1_sub = external global %struct.SubImageContainer, align 4
+@ref_pic2_sub = external global %struct.SubImageContainer, align 4
+@wbp_weight = external global i32****, align 4
+@weight1 = external global i16, align 2
+@weight2 = external global i16, align 2
+@offsetBi = external global i16, align 2
+@computeBiPred2 = external global [3 x i32 (i16*, i32, i32, i32, i32, i32, i32, i32)*], align 4
+@computeBiPred = external global i32 (i16*, i32, i32, i32, i32, i32, i32, i32)*, align 4
+@bipred2_access_method = external global i32, align 4
+@start_me_refinement_hp = external global i32, align 4
+
+declare i32 @foobar(i16*, i32 signext , i32 signext , i32 signext ,
+                    i32 signext , i32 signext , i32 signext , i32 signext ) #1
+
+define void @SubPelBlockSearchBiPred(i16* %orig_pic, i16 signext %ref,
+    i32 signext %pic_pix_x, i32 signext %pic_pix_y, i16 signext %pred_mv_y,
+    i16* nocapture %mv_x, i16* nocapture %mv_y, i16* nocapture readonly %s_mv_x,
+    i32 signext %search_pos2, i32 signext %min_mcost) #0 {
+; CHECK-LABEL: SubPelBlockSearchBiPred:
+entry:
+  %add40 = shl i32 %pic_pix_x, 2
+  %shl = add i32 %add40, 80
+  %add41 = shl i32 %pic_pix_y, 2
+  %0 = load i32, i32* @start_me_refinement_hp, align 4, !tbaa !1
+  %cond47 = select i1 undef, i32 1, i32 %search_pos2
+  %1 = load i16, i16* %s_mv_x, align 2, !tbaa !5
+  %conv48 = sext i16 %1 to i32
+  %add49 = add nsw i32 %conv48, %shl
+  %idxprom52 = sext i16 %ref to i32
+  %2 = load i32, i32* null, align 4, !tbaa !1
+  store i32 undef, i32* bitcast (%struct.SubImageContainer* @ref_pic1_sub to i32*), align 4, !tbaa !7
+  %3 = load i32, i32* undef, align 4, !tbaa !10
+  store i32 %3, i32* bitcast (%struct.SubImageContainer* @ref_pic2_sub to i32*), align 4, !tbaa !7
+  store i16 0, i16* @img_height, align 2, !tbaa !5
+  %size_x_pad = getelementptr inbounds %struct.storable_picture, %struct.storable_picture* null, i32 0, i32 22
+  %4 = load i32, i32* %size_x_pad, align 4, !tbaa !12
+  store i32 %4, i32* @width_pad, align 4, !tbaa !1
+  %5 = load i32, i32* undef, align 4, !tbaa !13
+  store i32 %5, i32* @height_pad, align 4, !tbaa !1
+  %6 = load i32****, i32***** @wbp_weight, align 4, !tbaa !14
+  %arrayidx75 = getelementptr inbounds i32***, i32**** %6, i32 undef
+  %7 = load i32***, i32**** %arrayidx75, align 4, !tbaa !14
+  %arrayidx76 = getelementptr inbounds i32**, i32*** %7, i32 %idxprom52
+  %8 = load i32**, i32*** %arrayidx76, align 4, !tbaa !14
+  %cond87.in671 = load i32*, i32** %8, align 4
+  %cond87672 = load i32, i32* %cond87.in671, align 4
+  %conv88673 = trunc i32 %cond87672 to i16
+  store i16 %conv88673, i16* @weight1, align 2, !tbaa !5
+  %cond105 = load i32, i32* undef, align 4
+  %conv106 = trunc i32 %cond105 to i16
+  store i16 %conv106, i16* @weight2, align 2, !tbaa !5
+  store i16 0, i16* @offsetBi, align 2, !tbaa !5
+  %storemerge655 = load i32, i32* bitcast (i32 (i16*, i32, i32, i32, i32, i32, i32, i32)** getelementptr inbounds ([3 x i32 (i16*, i32, i32, i32, i32, i32, i32, i32)*], [3 x i32 (i16*, i32, i32, i32, i32, i32, i32, i32)*]* @computeBiPred2, i32 0, i32 1) to i32*), align 4
+  store i32 %storemerge655, i32* bitcast (i32 (i16*, i32, i32, i32, i32, i32, i32, i32)** @computeBiPred to i32*), align 4, !tbaa !14
+  %9 = load i16, i16* %mv_x, align 2, !tbaa !5
+  %cmp270 = icmp sgt i32 undef, 1
+  %or.cond = and i1 %cmp270, false
+  br i1 %or.cond, label %land.lhs.true277, label %if.else289
+
+land.lhs.true277:                                 ; preds = %entry
+  %10 = load i16, i16* %mv_y, align 2, !tbaa !5
+  %conv278 = sext i16 %10 to i32
+  %add279 = add nsw i32 %conv278, 0
+  %cmp280 = icmp sgt i32 %add279, 1
+  %or.cond660 = and i1 %cmp280, undef
+  br i1 %or.cond660, label %if.end290, label %if.else289
+
+if.else289:                                       ; preds = %land.lhs.true277, %entry
+  br label %if.end290
+
+if.end290:                                        ; preds = %if.else289, %land.lhs.true277
+  %storemerge = phi i32 [ 1, %if.else289 ], [ 0, %land.lhs.true277 ]
+  store i32 %storemerge, i32* @bipred2_access_method, align 4, !tbaa !1
+  %cmp315698 = icmp slt i32 %0, %cond47
+  br i1 %cmp315698, label %for.body.lr.ph, label %if.end358
+
+for.body.lr.ph:                                   ; preds = %if.end290
+  %conv328 = sext i16 %pred_mv_y to i32
+  br label %for.body
+
+for.body:                                         ; preds = %for.inc, %for.body.lr.ph
+  %11 = phi i16 [ %9, %for.body.lr.ph ], [ %.pre, %for.inc ]
+  %min_mcost.addr.0701 = phi i32 [ %min_mcost, %for.body.lr.ph ], [ undef, %for.inc ]
+  %pos.0700 = phi i32 [ %0, %for.body.lr.ph ], [ undef, %for.inc ]
+  %best_pos.0699 = phi i32 [ 0, %for.body.lr.ph ], [ %best_pos.1, %for.inc ]
+  %conv317 = sext i16 %11 to i32
+  %add320 = add nsw i32 0, %conv317
+  %12 = load i16, i16* %mv_y, align 2, !tbaa !5
+  %conv321 = sext i16 %12 to i32
+  %add324 = add nsw i32 0, %conv321
+  %13 = load i32*, i32** @mvbits, align 4, !tbaa !14
+  %14 = load i32, i32* undef, align 4, !tbaa !1
+  %sub329 = sub nsw i32 %add324, %conv328
+  %arrayidx330 = getelementptr inbounds i32, i32* %13, i32 %sub329
+  %15 = load i32, i32* %arrayidx330, align 4, !tbaa !1
+  %add331 = add nsw i32 %15, %14
+  %mul = mul nsw i32 %add331, %2
+  %shr332 = ashr i32 %mul, 16
+  %cmp333 = icmp sgt i32 %min_mcost.addr.0701, %shr332
+  br i1 %cmp333, label %if.end336, label %for.inc
+
+if.end336:                                        ; preds = %for.body
+  ; CHECK:      jalr  $25
+  ; CHECK-NOT:  move  $ra, {{.*}}
+  ; CHECK:      j     $BB{{.*}}
+  %add337 = add nsw i32 %add320, %shl
+  %add338 = add nsw i32 %add324, 0
+  %call340 = tail call i32 undef(i16* %orig_pic, i32 signext undef, i32 signext
+                                 undef, i32 signext 0, i32 signext %add49,
+                                 i32 signext undef, i32 signext %add337,
+                                 i32 signext %add338) #1
+  %cmp342 = icmp slt i32 0, %min_mcost.addr.0701
+  %pos.0.best_pos.0 = select i1 %cmp342, i32 %pos.0700, i32 %best_pos.0699
+  br label %for.inc
+
+for.inc:                                          ; preds = %if.end336, %for.body
+  %best_pos.1 = phi i32 [ %best_pos.0699, %for.body ], [ %pos.0.best_pos.0, %if.end336 ]
+  %.pre = load i16, i16* %mv_x, align 2, !tbaa !5
+  br label %for.body
+
+if.end358:                                        ; preds = %if.end290
+  %.min_mcost.addr.0 = select i1 false, i32 2147483647, i32 %min_mcost
+  br i1 undef, label %for.body415.lr.ph, label %if.end461
+
+for.body415.lr.ph:                                ; preds = %if.end358
+  %16 = load i16, i16* %mv_y, align 2, !tbaa !5
+  %conv420 = sext i16 %16 to i32
+  %add423 = add nsw i32 0, %conv420
+  %cmp433 = icmp sgt i32 %.min_mcost.addr.0, 0
+  br i1 %cmp433, label %if.end436, label %if.end461
+
+if.end436:                                        ; preds = %for.body415.lr.ph
+  %add438 = add nsw i32 %add423, 0
+  %call440 = tail call i32 @foobar(i16* %orig_pic, i32 signext undef, i32 signext undef,
+                                 i32 signext 0, i32 signext %add49, i32 signext undef,
+                                 i32 signext undef, i32 signext %add438) #1
+  br label %if.end461
+
+if.end461:                                        ; preds = %if.end436, %for.body415.lr.ph, %if.end358
+  ret void
+}
+
+attributes #0 = { nounwind "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="mips32r2" "target-features"="+mips32r2,+nooddspreg,+fpxx" "unsafe-fp-math"="false" "use-soft-float"="false" }
+attributes #1 = { nounwind }
+
+!llvm.ident = !{!0}
+
+!0 = !{!"clang version 3.7.0 (trunk 236218) (llvm/trunk 236237)"}
+!1 = !{!2, !2, i64 0}
+!2 = !{!"int", !3, i64 0}
+!3 = !{!"omnipotent char", !4, i64 0}
+!4 = !{!"Simple C/C++ TBAA"}
+!5 = !{!6, !6, i64 0}
+!6 = !{!"short", !3, i64 0}
+!7 = !{!8, !9, i64 0}
+!8 = !{!"", !9, i64 0, !3, i64 4}
+!9 = !{!"any pointer", !3, i64 0}
+!10 = !{!11, !9, i64 6440}
+!11 = !{!"storable_picture", !3, i64 0, !2, i64 4, !2, i64 8, !2, i64 12, !2, i64 16, !2, i64 20, !3, i64 24, !3, i64 1608, !3, i64 3192, !3, i64 4776, !2, i64 6360, !2, i64 6364, !2, i64 6368, !2, i64 6372, !2, i64 6376, !2, i64 6380, !2, i64 6384, !2, i64 6388, !2, i64 6392, !2, i64 6396, !2, i64 6400, !2, i64 6404, !2, i64 6408, !2, i64 6412, !2, i64 6416, !2, i64 6420, !2, i64 6424, !2, i64 6428, !2, i64 6432, !9, i64 6436, !9, i64 6440, !9, i64 6444, !9, i64 6448, !9, i64 6452, !9, i64 6456, !9, i64 6460, !9, i64 6464, !9, i64 6468, !9, i64 6472, !9, i64 6476, !9, i64 6480, !9, i64 6484, !9, i64 6488, !9, i64 6492, !2, i64 6496, !2, i64 6500, !2, i64 6504, !2, i64 6508, !2, i64 6512, !2, i64 6516, !2, i64 6520}
+!12 = !{!11, !2, i64 6408}
+!13 = !{!11, !2, i64 6412}
+!14 = !{!9, !9, i64 0}
diff --git a/test/CodeGen/Mips/delay-slot-kill.ll b/test/CodeGen/Mips/delay-slot-kill.ll
new file mode 100644
index 000000000000..57b630303c26
--- /dev/null
+++ b/test/CodeGen/Mips/delay-slot-kill.ll
@@ -0,0 +1,14 @@
+; RUN: llc < %s -march=mips64 -mcpu=mips3 | FileCheck %s
+
+; Currently, the following IR assembly generates a KILL instruction between
+; the bitwise-and instruction and the return instruction. We verify that the
+; delay slot filler ignores such KILL instructions by filling the slot of the
+; return instruction properly.
+define signext i32 @f1(i32 signext %a, i32 signext %b) {
+entry:
+  ; CHECK:          jr      $ra
+  ; CHECK-NEXT:     and     $2, $4, $5
+
+  %r = and i32 %a, %b
+  ret i32 %r
+}
diff --git a/test/CodeGen/Mips/disable-tail-merge.ll b/test/CodeGen/Mips/disable-tail-merge.ll
index b4c093aa8528..9396db7be7f6 100644
--- a/test/CodeGen/Mips/disable-tail-merge.ll
+++ b/test/CodeGen/Mips/disable-tail-merge.ll
@@ -9,20 +9,20 @@
 define i32 @test1(i32 %a) {
 entry:
   %tobool = icmp eq i32 %a, 0
-  %0 = load i32* @g0, align 4
+  %0 = load i32, i32* @g0, align 4
   br i1 %tobool, label %if.else, label %if.then
 
 if.then:
   %add = add nsw i32 %0, 1
   store i32 %add, i32* @g0, align 4
-  %1 = load i32* @g1, align 4
+  %1 = load i32, i32* @g1, align 4
   %add1 = add nsw i32 %1, 23
   br label %if.end
 
 if.else:
   %add2 = add nsw i32 %0, 11
   store i32 %add2, i32* @g0, align 4
-  %2 = load i32* @g1, align 4
+  %2 = load i32, i32* @g1, align 4
   %add3 = add nsw i32 %2, 23
   br label %if.end
 
diff --git a/test/CodeGen/Mips/div.ll b/test/CodeGen/Mips/div.ll
index 00e2c1927459..731841c554fa 100644
--- a/test/CodeGen/Mips/div.ll
+++ b/test/CodeGen/Mips/div.ll
@@ -6,8 +6,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @iiii, align 4
-  %1 = load i32* @jjjj, align 4
+  %0 = load i32, i32* @iiii, align 4
+  %1 = load i32, i32* @jjjj, align 4
   %div = sdiv i32 %0, %1
 ; 16:	div	$zero, ${{[0-9]+}}, ${{[0-9]+}}
 ; 16: 	mflo	${{[0-9]+}}
diff --git a/test/CodeGen/Mips/div_rem.ll b/test/CodeGen/Mips/div_rem.ll
index 950192eee169..e64529cee841 100644
--- a/test/CodeGen/Mips/div_rem.ll
+++ b/test/CodeGen/Mips/div_rem.ll
@@ -7,8 +7,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @iiii, align 4
-  %1 = load i32* @jjjj, align 4
+  %0 = load i32, i32* @iiii, align 4
+  %1 = load i32, i32* @jjjj, align 4
   %div = sdiv i32 %0, %1
   store i32 %div, i32* @kkkk, align 4
   %rem = srem i32 %0, %1
diff --git a/test/CodeGen/Mips/divrem.ll b/test/CodeGen/Mips/divrem.ll
index a9cfe0fa1523..918db053f5b6 100644
--- a/test/CodeGen/Mips/divrem.ll
+++ b/test/CodeGen/Mips/divrem.ll
@@ -220,8 +220,8 @@ entry:
 ; FIXME: It's not clear what this is supposed to test.
 define i32 @killFlags() {
 entry:
-  %0 = load i32* @g0, align 4
-  %1 = load i32* @g1, align 4
+  %0 = load i32, i32* @g0, align 4
+  %1 = load i32, i32* @g1, align 4
   %div = sdiv i32 %0, %1
   ret i32 %div
 }
diff --git a/test/CodeGen/Mips/divu.ll b/test/CodeGen/Mips/divu.ll
index b96a439390ca..5bc765a71eb9 100644
--- a/test/CodeGen/Mips/divu.ll
+++ b/test/CodeGen/Mips/divu.ll
@@ -6,8 +6,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @iiii, align 4
-  %1 = load i32* @jjjj, align 4
+  %0 = load i32, i32* @iiii, align 4
+  %1 = load i32, i32* @jjjj, align 4
   %div = udiv i32 %0, %1
 ; 16:	divu	$zero, ${{[0-9]+}}, ${{[0-9]+}}
 ; 16: 	mflo	${{[0-9]+}}
diff --git a/test/CodeGen/Mips/divu_remu.ll b/test/CodeGen/Mips/divu_remu.ll
index a6c1563ac195..a079440b913f 100644
--- a/test/CodeGen/Mips/divu_remu.ll
+++ b/test/CodeGen/Mips/divu_remu.ll
@@ -8,8 +8,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @iiii, align 4
-  %1 = load i32* @jjjj, align 4
+  %0 = load i32, i32* @iiii, align 4
+  %1 = load i32, i32* @jjjj, align 4
   %div = udiv i32 %0, %1
   store i32 %div, i32* @kkkk, align 4
   %rem = urem i32 %0, %1
diff --git a/test/CodeGen/Mips/dsp-patterns.ll b/test/CodeGen/Mips/dsp-patterns.ll
index f5bb3abed90e..837c0d8bfc52 100644
--- a/test/CodeGen/Mips/dsp-patterns.ll
+++ b/test/CodeGen/Mips/dsp-patterns.ll
@@ -6,8 +6,8 @@
 
 define zeroext i8 @test_lbux(i8* nocapture %b, i32 %i) {
 entry:
-  %add.ptr = getelementptr inbounds i8* %b, i32 %i
-  %0 = load i8* %add.ptr, align 1
+  %add.ptr = getelementptr inbounds i8, i8* %b, i32 %i
+  %0 = load i8, i8* %add.ptr, align 1
   ret i8 %0
 }
 
@@ -16,8 +16,8 @@ entry:
 
 define signext i16 @test_lhx(i16* nocapture %b, i32 %i) {
 entry:
-  %add.ptr = getelementptr inbounds i16* %b, i32 %i
-  %0 = load i16* %add.ptr, align 2
+  %add.ptr = getelementptr inbounds i16, i16* %b, i32 %i
+  %0 = load i16, i16* %add.ptr, align 2
   ret i16 %0
 }
 
@@ -26,8 +26,8 @@ entry:
 
 define i32 @test_lwx(i32* nocapture %b, i32 %i) {
 entry:
-  %add.ptr = getelementptr inbounds i32* %b, i32 %i
-  %0 = load i32* %add.ptr, align 4
+  %add.ptr = getelementptr inbounds i32, i32* %b, i32 %i
+  %0 = load i32, i32* %add.ptr, align 4
   ret i32 %0
 }
 
diff --git a/test/CodeGen/Mips/dsp-vec-load-store.ll b/test/CodeGen/Mips/dsp-vec-load-store.ll
index 7e4a8fedaa8c..f9251807d000 100644
--- a/test/CodeGen/Mips/dsp-vec-load-store.ll
+++ b/test/CodeGen/Mips/dsp-vec-load-store.ll
@@ -5,7 +5,7 @@
 
 define void @extend_load_trunc_store_v2i8() {
 entry:
-  %0 = load <2 x i8>* @g1, align 2
+  %0 = load <2 x i8>, <2 x i8>* @g1, align 2
   store <2 x i8> %0, <2 x i8>* @g0, align 2
   ret void
 }
diff --git a/test/CodeGen/Mips/eh-return32.ll b/test/CodeGen/Mips/eh-return32.ll
index 748050c4d34b..542c5bf4462e 100644
--- a/test/CodeGen/Mips/eh-return32.ll
+++ b/test/CodeGen/Mips/eh-return32.ll
@@ -7,7 +7,7 @@ declare void @foo(...)
 
 define i8* @f1(i32 %offset, i8* %handler) {
 entry:
-  call void (...)* @foo()
+  call void (...) @foo()
   call void @llvm.eh.return.i32(i32 %offset, i8* %handler)
   unreachable
 
diff --git a/test/CodeGen/Mips/eh-return64.ll b/test/CodeGen/Mips/eh-return64.ll
index 74a43231598c..2f8203d77c84 100644
--- a/test/CodeGen/Mips/eh-return64.ll
+++ b/test/CodeGen/Mips/eh-return64.ll
@@ -8,7 +8,7 @@ declare void @foo(...)
 
 define void @f1(i64 %offset, i8* %handler) {
 entry:
-  call void (...)* @foo()
+  call void (...) @foo()
   call void @llvm.eh.return.i64(i64 %offset, i8* %handler)
   unreachable
 
diff --git a/test/CodeGen/Mips/eh.ll b/test/CodeGen/Mips/eh.ll
index fc9e2ef21a8b..fcbd99ef737b 100644
--- a/test/CodeGen/Mips/eh.ll
+++ b/test/CodeGen/Mips/eh.ll
@@ -27,6 +27,7 @@ lpad:                                             ; preds = %entry
 ; CHECK-EL:  bne $5
 
   %exn.val = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
+           cleanup
            catch i8* bitcast (i8** @_ZTId to i8*)
   %exn = extractvalue { i8*, i32 } %exn.val, 0
   %sel = extractvalue { i8*, i32 } %exn.val, 1
@@ -37,7 +38,7 @@ lpad:                                             ; preds = %entry
 catch:                                            ; preds = %lpad
   %3 = tail call i8* @__cxa_begin_catch(i8* %exn) nounwind
   %4 = bitcast i8* %3 to double*
-  %exn.scalar = load double* %4, align 8
+  %exn.scalar = load double, double* %4, align 8
   %add = fadd double %exn.scalar, %i2
   store double %add, double* @g1, align 8
   tail call void @__cxa_end_catch() nounwind
diff --git a/test/CodeGen/Mips/ehframe-indirect.ll b/test/CodeGen/Mips/ehframe-indirect.ll
index b4efb40b6422..f124881a472f 100644
--- a/test/CodeGen/Mips/ehframe-indirect.ll
+++ b/test/CodeGen/Mips/ehframe-indirect.ll
@@ -1,16 +1,18 @@
-; RUN: llc -mtriple=mipsel-linux-gnu < %s | FileCheck  -check-prefix=CHECK32 %s
-; RUN: llc -mtriple=mipsel-linux-android < %s | FileCheck -check-prefix=CHECK32 %s
-; RUN: llc -mtriple=mips64el-linux-gnu < %s | FileCheck  -check-prefix=CHECK64 %s
-; RUN: llc -mtriple=mips64el-linux-android < %s | FileCheck -check-prefix=CHECK64 %s
+; RUN: llc -mtriple=mipsel-linux-gnu < %s | FileCheck -check-prefix=ALL -check-prefix=O32 %s
+; RUN: llc -mtriple=mipsel-linux-android < %s | FileCheck -check-prefix=ALL -check-prefix=O32 %s
+; RUN: llc -mtriple=mips64el-linux-gnu -target-abi=n32 < %s | FileCheck -check-prefix=ALL -check-prefix=N32 %s
+; RUN: llc -mtriple=mips64el-linux-android -target-abi=n32 < %s | FileCheck -check-prefix=ALL -check-prefix=N32 %s
+; RUN: llc -mtriple=mips64el-linux-gnu < %s | FileCheck -check-prefix=ALL -check-prefix=N64 %s
+; RUN: llc -mtriple=mips64el-linux-android < %s | FileCheck -check-prefix=ALL -check-prefix=N64 %s
 
 define i32 @main() {
-; CHECK: .cfi_startproc
-; CHECK: .cfi_personality 128, DW.ref.__gxx_personality_v0
+; ALL: .cfi_startproc
+; ALL: .cfi_personality 128, DW.ref.__gxx_personality_v0
 
 entry:
   invoke void @foo() to label %cont unwind label %lpad
-; CHECK: foo
-; CHECK: jalr
+; ALL: foo
+; ALL: jalr
 
 lpad:
   %0 = landingpad { i8*, i32 } personality i8*
@@ -20,20 +22,23 @@ lpad:
 cont:
   ret i32 0
 }
-; CHECK: .cfi_endproc
+; ALL: .cfi_endproc
 
 declare i32 @__gxx_personality_v0(...)
 
 declare void @foo()
 
-; CHECK: .hidden DW.ref.__gxx_personality_v0
-; CHECK: .weak DW.ref.__gxx_personality_v0
-; CHECK: .section .data.DW.ref.__gxx_personality_v0,"aGw",@progbits,DW.ref.__gxx_personality_v0,comdat
-; CHECK32: .align 2
-; CHECK64: .align 3
-; CHECK: .type DW.ref.__gxx_personality_v0,@object
-; CHECK32: .size DW.ref.__gxx_personality_v0, 4
-; CHECK64: .size DW.ref.__gxx_personality_v0, 8
-; CHECK: DW.ref.__gxx_personality_v0:
-; CHECK32: .4byte __gxx_personality_v0
-; CHECK64: .8byte __gxx_personality_v0
+; ALL: .hidden DW.ref.__gxx_personality_v0
+; ALL: .weak DW.ref.__gxx_personality_v0
+; ALL: .section .data.DW.ref.__gxx_personality_v0,"aGw",@progbits,DW.ref.__gxx_personality_v0,comdat
+; O32: .align 2
+; N32: .align 2
+; N64: .align 3
+; ALL: .type DW.ref.__gxx_personality_v0,@object
+; O32: .size DW.ref.__gxx_personality_v0, 4
+; N32: .size DW.ref.__gxx_personality_v0, 4
+; N64: .size DW.ref.__gxx_personality_v0, 8
+; ALL: DW.ref.__gxx_personality_v0:
+; O32: .4byte __gxx_personality_v0
+; N32: .4byte __gxx_personality_v0
+; N64: .8byte __gxx_personality_v0
diff --git a/test/CodeGen/Mips/emergency-spill-slot-near-fp.ll b/test/CodeGen/Mips/emergency-spill-slot-near-fp.ll
new file mode 100644
index 000000000000..3dc1cde77095
--- /dev/null
+++ b/test/CodeGen/Mips/emergency-spill-slot-near-fp.ll
@@ -0,0 +1,32 @@
+; Check that register scavenging spill slot is close to $fp.
+; RUN: llc -march=mipsel -O0 < %s | FileCheck %s
+
+; CHECK: sw ${{.*}}, 4($fp)
+; CHECK: lw ${{.*}}, 4($fp)
+
+define i32 @main(i32 signext %argc, i8** %argv) "no-frame-pointer-elim"="true" {
+entry:
+  %retval = alloca i32, align 4
+  %argc.addr = alloca i32, align 4
+  %argv.addr = alloca i8**, align 4
+  %v0 = alloca <16 x i8>, align 16
+  %.compoundliteral = alloca <16 x i8>, align 16
+  %v1 = alloca <16 x i8>, align 16
+  %.compoundliteral1 = alloca <16 x i8>, align 16
+  %unused_variable = alloca [16384 x i32], align 4
+  %result = alloca <16 x i8>, align 16
+  store i32 0, i32* %retval
+  store i32 %argc, i32* %argc.addr, align 4
+  store i8** %argv, i8*** %argv.addr, align 4
+  store <16 x i8> <i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 16>, <16 x i8>* %.compoundliteral
+  %0 = load <16 x i8>, <16 x i8>* %.compoundliteral
+  store <16 x i8> %0, <16 x i8>* %v0, align 16
+  store <16 x i8> zeroinitializer, <16 x i8>* %.compoundliteral1
+  %1 = load <16 x i8>, <16 x i8>* %.compoundliteral1
+  store <16 x i8> %1, <16 x i8>* %v1, align 16
+  %2 = load <16 x i8>, <16 x i8>* %v0, align 16
+  %3 = load <16 x i8>, <16 x i8>* %v1, align 16
+  %mul = mul <16 x i8> %2, %3
+  store <16 x i8> %mul, <16 x i8>* %result, align 16
+  ret i32 0
+}
diff --git a/test/CodeGen/Mips/emit-big-cst.ll b/test/CodeGen/Mips/emit-big-cst.ll
index a168743859a3..9bc96c89307d 100644
--- a/test/CodeGen/Mips/emit-big-cst.ll
+++ b/test/CodeGen/Mips/emit-big-cst.ll
@@ -10,7 +10,7 @@
 
 define void @accessBig(i64* %storage) {
   %addr = bitcast i64* %storage to i82*
-  %bigLoadedCst = load volatile i82* @bigCst
+  %bigLoadedCst = load volatile i82, i82* @bigCst
   %tmp = add i82 %bigLoadedCst, 1
   store i82 %tmp, i82* %addr
   ret void
diff --git a/test/CodeGen/Mips/ex2.ll b/test/CodeGen/Mips/ex2.ll
index 6d024c209c26..7547fdf81e35 100644
--- a/test/CodeGen/Mips/ex2.ll
+++ b/test/CodeGen/Mips/ex2.ll
@@ -17,12 +17,12 @@ entry:
   store i32 0, i32* %retval
   %exception = call i8* @__cxa_allocate_exception(i32 4) nounwind
   %0 = bitcast i8* %exception to i8**
-  store i8* getelementptr inbounds ([6 x i8]* @.str, i32 0, i32 0), i8** %0
+  store i8* getelementptr inbounds ([6 x i8], [6 x i8]* @.str, i32 0, i32 0), i8** %0
   call void @__cxa_throw(i8* %exception, i8* bitcast (i8** @_ZTIPKc to i8*), i8* null) noreturn
   unreachable
 
 return:                                           ; No predecessors!
-  %1 = load i32* %retval
+  %1 = load i32, i32* %retval
   ret i32 %1
 }
 
diff --git a/test/CodeGen/Mips/extins.ll b/test/CodeGen/Mips/extins.ll
index efaeeea96a5e..6604f89b1843 100644
--- a/test/CodeGen/Mips/extins.ll
+++ b/test/CodeGen/Mips/extins.ll
@@ -16,7 +16,7 @@ entry:
 ; 16-NOT: ins ${{[0-9]+}}
   %and = shl i32 %s, 5
   %shl = and i32 %and, 16352
-  %tmp3 = load i32* %d, align 4
+  %tmp3 = load i32, i32* %d, align 4
   %and5 = and i32 %tmp3, -16353
   %or = or i32 %and5, %shl
   store i32 %or, i32* %d, align 4
diff --git a/test/CodeGen/Mips/f16abs.ll b/test/CodeGen/Mips/f16abs.ll
index 0fba9c4fd08a..838983274e9b 100644
--- a/test/CodeGen/Mips/f16abs.ll
+++ b/test/CodeGen/Mips/f16abs.ll
@@ -11,12 +11,12 @@
 ; Function Attrs: nounwind optsize
 define i32 @main() #0 {
 entry:
-  %0 = load double* @y, align 8
+  %0 = load double, double* @y, align 8
   %call = tail call double @fabs(double %0) #2
   store double %call, double* @x, align 8
 ; static-NOT: 	.ent	__call_stub_fp_fabs
 ; static-NOT: 	jal fabs
-  %1 = load float* @y1, align 4
+  %1 = load float, float* @y1, align 4
   %call2 = tail call float @fabsf(float %1) #2
   store float %call2, float* @x1, align 4
 ; static-NOT: 	.ent	__call_stub_fp_fabsf
diff --git a/test/CodeGen/Mips/fastcc.ll b/test/CodeGen/Mips/fastcc.ll
index 6b022c5e36d9..299e0d696cbb 100644
--- a/test/CodeGen/Mips/fastcc.ll
+++ b/test/CodeGen/Mips/fastcc.ll
@@ -108,23 +108,23 @@ entry:
 ; CHECK-NACL-NOT: lw  $15
 ; CHECK-NACL-NOT: lw  $24
 
-  %0 = load i32* @gi0, align 4
-  %1 = load i32* @gi1, align 4
-  %2 = load i32* @gi2, align 4
-  %3 = load i32* @gi3, align 4
-  %4 = load i32* @gi4, align 4
-  %5 = load i32* @gi5, align 4
-  %6 = load i32* @gi6, align 4
-  %7 = load i32* @gi7, align 4
-  %8 = load i32* @gi8, align 4
-  %9 = load i32* @gi9, align 4
-  %10 = load i32* @gi10, align 4
-  %11 = load i32* @gi11, align 4
-  %12 = load i32* @gi12, align 4
-  %13 = load i32* @gi13, align 4
-  %14 = load i32* @gi14, align 4
-  %15 = load i32* @gi15, align 4
-  %16 = load i32* @gi16, align 4
+  %0 = load i32, i32* @gi0, align 4
+  %1 = load i32, i32* @gi1, align 4
+  %2 = load i32, i32* @gi2, align 4
+  %3 = load i32, i32* @gi3, align 4
+  %4 = load i32, i32* @gi4, align 4
+  %5 = load i32, i32* @gi5, align 4
+  %6 = load i32, i32* @gi6, align 4
+  %7 = load i32, i32* @gi7, align 4
+  %8 = load i32, i32* @gi8, align 4
+  %9 = load i32, i32* @gi9, align 4
+  %10 = load i32, i32* @gi10, align 4
+  %11 = load i32, i32* @gi11, align 4
+  %12 = load i32, i32* @gi12, align 4
+  %13 = load i32, i32* @gi13, align 4
+  %14 = load i32, i32* @gi14, align 4
+  %15 = load i32, i32* @gi15, align 4
+  %16 = load i32, i32* @gi16, align 4
   tail call fastcc void @callee0(i32 %0, i32 %1, i32 %2, i32 %3, i32 %4, i32 %5, i32 %6, i32 %7, i32 %8, i32 %9, i32 %10, i32 %11, i32 %12, i32 %13, i32 %14, i32 %15, i32 %16)
   ret void
 }
@@ -196,27 +196,27 @@ entry:
 ; CHECK: lwc1  $f1
 ; CHECK: lwc1  $f0
 
-  %0 = load float* @gfa0, align 4
-  %1 = load float* @gfa1, align 4
-  %2 = load float* @gfa2, align 4
-  %3 = load float* @gfa3, align 4
-  %4 = load float* @gfa4, align 4
-  %5 = load float* @gfa5, align 4
-  %6 = load float* @gfa6, align 4
-  %7 = load float* @gfa7, align 4
-  %8 = load float* @gfa8, align 4
-  %9 = load float* @gfa9, align 4
-  %10 = load float* @gfa10, align 4
-  %11 = load float* @gfa11, align 4
-  %12 = load float* @gfa12, align 4
-  %13 = load float* @gfa13, align 4
-  %14 = load float* @gfa14, align 4
-  %15 = load float* @gfa15, align 4
-  %16 = load float* @gfa16, align 4
-  %17 = load float* @gfa17, align 4
-  %18 = load float* @gfa18, align 4
-  %19 = load float* @gfa19, align 4
-  %20 = load float* @gfa20, align 4
+  %0 = load float, float* @gfa0, align 4
+  %1 = load float, float* @gfa1, align 4
+  %2 = load float, float* @gfa2, align 4
+  %3 = load float, float* @gfa3, align 4
+  %4 = load float, float* @gfa4, align 4
+  %5 = load float, float* @gfa5, align 4
+  %6 = load float, float* @gfa6, align 4
+  %7 = load float, float* @gfa7, align 4
+  %8 = load float, float* @gfa8, align 4
+  %9 = load float, float* @gfa9, align 4
+  %10 = load float, float* @gfa10, align 4
+  %11 = load float, float* @gfa11, align 4
+  %12 = load float, float* @gfa12, align 4
+  %13 = load float, float* @gfa13, align 4
+  %14 = load float, float* @gfa14, align 4
+  %15 = load float, float* @gfa15, align 4
+  %16 = load float, float* @gfa16, align 4
+  %17 = load float, float* @gfa17, align 4
+  %18 = load float, float* @gfa18, align 4
+  %19 = load float, float* @gfa19, align 4
+  %20 = load float, float* @gfa20, align 4
   tail call fastcc void @callee1(float %0, float %1, float %2, float %3, float %4, float %5, float %6, float %7, float %8, float %9, float %10, float %11, float %12, float %13, float %14, float %15, float %16, float %17, float %18, float %19, float %20)
   ret void
 }
@@ -292,17 +292,17 @@ entry:
 ; NOODDSPREG-DAG:    lwc1    $[[F0:f[0-9]*[02468]]], 40($[[R0]])
 ; NOODDSPREG-DAG:    swc1    $[[F0]], 0($sp)
 
-  %0 = load float* getelementptr ([11 x float]* @fa, i32 0, i32 0), align 4
-  %1 = load float* getelementptr ([11 x float]* @fa, i32 0, i32 1), align 4
-  %2 = load float* getelementptr ([11 x float]* @fa, i32 0, i32 2), align 4
-  %3 = load float* getelementptr ([11 x float]* @fa, i32 0, i32 3), align 4
-  %4 = load float* getelementptr ([11 x float]* @fa, i32 0, i32 4), align 4
-  %5 = load float* getelementptr ([11 x float]* @fa, i32 0, i32 5), align 4
-  %6 = load float* getelementptr ([11 x float]* @fa, i32 0, i32 6), align 4
-  %7 = load float* getelementptr ([11 x float]* @fa, i32 0, i32 7), align 4
-  %8 = load float* getelementptr ([11 x float]* @fa, i32 0, i32 8), align 4
-  %9 = load float* getelementptr ([11 x float]* @fa, i32 0, i32 9), align 4
-  %10 = load float* getelementptr ([11 x float]* @fa, i32 0, i32 10), align 4
+  %0 = load float, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 0), align 4
+  %1 = load float, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 1), align 4
+  %2 = load float, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 2), align 4
+  %3 = load float, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 3), align 4
+  %4 = load float, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 4), align 4
+  %5 = load float, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 5), align 4
+  %6 = load float, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 6), align 4
+  %7 = load float, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 7), align 4
+  %8 = load float, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 8), align 4
+  %9 = load float, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 9), align 4
+  %10 = load float, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 10), align 4
   tail call fastcc void @callee2(float %0, float %1, float %2, float %3,
                                  float %4, float %5, float %6, float %7,
                                  float %8, float %9, float %10)
@@ -336,17 +336,17 @@ entry:
 ; NOODDSPREG-DAG:    lwc1    $[[F0:f[0-9]*[02468]]], [[OFFSET]]($sp)
 ; NOODDSPREG-DAG:    swc1    $[[F0]], 40($[[R0]])
 
-  store float %a0, float* getelementptr ([11 x float]* @fa, i32 0, i32 0), align 4
-  store float %a1, float* getelementptr ([11 x float]* @fa, i32 0, i32 1), align 4
-  store float %a2, float* getelementptr ([11 x float]* @fa, i32 0, i32 2), align 4
-  store float %a3, float* getelementptr ([11 x float]* @fa, i32 0, i32 3), align 4
-  store float %a4, float* getelementptr ([11 x float]* @fa, i32 0, i32 4), align 4
-  store float %a5, float* getelementptr ([11 x float]* @fa, i32 0, i32 5), align 4
-  store float %a6, float* getelementptr ([11 x float]* @fa, i32 0, i32 6), align 4
-  store float %a7, float* getelementptr ([11 x float]* @fa, i32 0, i32 7), align 4
-  store float %a8, float* getelementptr ([11 x float]* @fa, i32 0, i32 8), align 4
-  store float %a9, float* getelementptr ([11 x float]* @fa, i32 0, i32 9), align 4
-  store float %a10, float* getelementptr ([11 x float]* @fa, i32 0, i32 10), align 4
+  store float %a0, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 0), align 4
+  store float %a1, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 1), align 4
+  store float %a2, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 2), align 4
+  store float %a3, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 3), align 4
+  store float %a4, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 4), align 4
+  store float %a5, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 5), align 4
+  store float %a6, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 6), align 4
+  store float %a7, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 7), align 4
+  store float %a8, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 8), align 4
+  store float %a9, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 9), align 4
+  store float %a10, float* getelementptr ([11 x float], [11 x float]* @fa, i32 0, i32 10), align 4
   ret void
 }
 
@@ -373,17 +373,17 @@ entry:
 ; FP64-NOODDSPREG-DAG:    ldc1    $[[F0:f[0-9]*[02468]]], 80($[[R0]])
 ; FP64-NOODDSPREG-DAG:    sdc1    $[[F0]], 0($sp)
 
-  %0 = load double* getelementptr ([11 x double]* @da, i32 0, i32 0), align 8
-  %1 = load double* getelementptr ([11 x double]* @da, i32 0, i32 1), align 8
-  %2 = load double* getelementptr ([11 x double]* @da, i32 0, i32 2), align 8
-  %3 = load double* getelementptr ([11 x double]* @da, i32 0, i32 3), align 8
-  %4 = load double* getelementptr ([11 x double]* @da, i32 0, i32 4), align 8
-  %5 = load double* getelementptr ([11 x double]* @da, i32 0, i32 5), align 8
-  %6 = load double* getelementptr ([11 x double]* @da, i32 0, i32 6), align 8
-  %7 = load double* getelementptr ([11 x double]* @da, i32 0, i32 7), align 8
-  %8 = load double* getelementptr ([11 x double]* @da, i32 0, i32 8), align 8
-  %9 = load double* getelementptr ([11 x double]* @da, i32 0, i32 9), align 8
-  %10 = load double* getelementptr ([11 x double]* @da, i32 0, i32 10), align 8
+  %0 = load double, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 0), align 8
+  %1 = load double, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 1), align 8
+  %2 = load double, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 2), align 8
+  %3 = load double, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 3), align 8
+  %4 = load double, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 4), align 8
+  %5 = load double, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 5), align 8
+  %6 = load double, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 6), align 8
+  %7 = load double, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 7), align 8
+  %8 = load double, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 8), align 8
+  %9 = load double, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 9), align 8
+  %10 = load double, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 10), align 8
   tail call fastcc void @callee3(double %0, double %1, double %2, double %3,
                                  double %4, double %5, double %6, double %7,
                                  double %8, double %9, double %10)
@@ -417,16 +417,16 @@ entry:
 ; FP64-NOODDSPREG-DAG:    ldc1    $[[F0:f[0-9]*[02468]]], [[OFFSET]]($sp)
 ; FP64-NOODDSPREG-DAG:    sdc1    $[[F0]], 80($[[R0]])
 
-  store double %a0, double* getelementptr ([11 x double]* @da, i32 0, i32 0), align 8
-  store double %a1, double* getelementptr ([11 x double]* @da, i32 0, i32 1), align 8
-  store double %a2, double* getelementptr ([11 x double]* @da, i32 0, i32 2), align 8
-  store double %a3, double* getelementptr ([11 x double]* @da, i32 0, i32 3), align 8
-  store double %a4, double* getelementptr ([11 x double]* @da, i32 0, i32 4), align 8
-  store double %a5, double* getelementptr ([11 x double]* @da, i32 0, i32 5), align 8
-  store double %a6, double* getelementptr ([11 x double]* @da, i32 0, i32 6), align 8
-  store double %a7, double* getelementptr ([11 x double]* @da, i32 0, i32 7), align 8
-  store double %a8, double* getelementptr ([11 x double]* @da, i32 0, i32 8), align 8
-  store double %a9, double* getelementptr ([11 x double]* @da, i32 0, i32 9), align 8
-  store double %a10, double* getelementptr ([11 x double]* @da, i32 0, i32 10), align 8
+  store double %a0, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 0), align 8
+  store double %a1, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 1), align 8
+  store double %a2, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 2), align 8
+  store double %a3, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 3), align 8
+  store double %a4, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 4), align 8
+  store double %a5, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 5), align 8
+  store double %a6, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 6), align 8
+  store double %a7, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 7), align 8
+  store double %a8, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 8), align 8
+  store double %a9, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 9), align 8
+  store double %a10, double* getelementptr ([11 x double], [11 x double]* @da, i32 0, i32 10), align 8
   ret void
 }
diff --git a/test/CodeGen/Mips/fcopysign-f32-f64.ll b/test/CodeGen/Mips/fcopysign-f32-f64.ll
index 148a780fb930..860bc79956fc 100644
--- a/test/CodeGen/Mips/fcopysign-f32-f64.ll
+++ b/test/CodeGen/Mips/fcopysign-f32-f64.ll
@@ -1,6 +1,6 @@
-; RUN: llc  < %s -march=mips64el -mcpu=mips4 -mattr=n64 | FileCheck %s -check-prefix=64
-; RUN: llc  < %s -march=mips64el -mcpu=mips64 -mattr=n64 | FileCheck %s -check-prefix=64
-; RUN: llc  < %s -march=mips64el -mcpu=mips64r2 -mattr=n64 | FileCheck %s -check-prefix=64R2
+; RUN: llc  < %s -march=mips64el -mcpu=mips4 -target-abi=n64 | FileCheck %s -check-prefix=64
+; RUN: llc  < %s -march=mips64el -mcpu=mips64 -target-abi=n64 | FileCheck %s -check-prefix=64
+; RUN: llc  < %s -march=mips64el -mcpu=mips64r2 -target-abi=n64 | FileCheck %s -check-prefix=64R2
 
 declare double @copysign(double, double) nounwind readnone
 
diff --git a/test/CodeGen/Mips/fcopysign.ll b/test/CodeGen/Mips/fcopysign.ll
index 3a9d9c73b279..6928f2fe507f 100644
--- a/test/CodeGen/Mips/fcopysign.ll
+++ b/test/CodeGen/Mips/fcopysign.ll
@@ -1,8 +1,8 @@
 ; RUN: llc  < %s -march=mipsel -mcpu=mips32 | FileCheck %s -check-prefix=32
 ; RUN: llc  < %s -march=mipsel -mcpu=mips32r2 | FileCheck %s -check-prefix=32R2
-; RUN: llc  < %s -march=mips64el -mcpu=mips4 -mattr=n64 | FileCheck %s -check-prefix=64
-; RUN: llc  < %s -march=mips64el -mcpu=mips64 -mattr=n64 | FileCheck %s -check-prefix=64
-; RUN: llc  < %s -march=mips64el -mcpu=mips64r2 -mattr=n64 | FileCheck %s -check-prefix=64R2
+; RUN: llc  < %s -march=mips64el -mcpu=mips4 -target-abi=n64 | FileCheck %s -check-prefix=64
+; RUN: llc  < %s -march=mips64el -mcpu=mips64 -target-abi=n64 | FileCheck %s -check-prefix=64
+; RUN: llc  < %s -march=mips64el -mcpu=mips64r2 -target-abi=n64 | FileCheck %s -check-prefix=64R2
 
 define double @func0(double %d0, double %d1) nounwind readnone {
 entry:
diff --git a/test/CodeGen/Mips/fixdfsf.ll b/test/CodeGen/Mips/fixdfsf.ll
index 4271ac222edb..869579922d51 100644
--- a/test/CodeGen/Mips/fixdfsf.ll
+++ b/test/CodeGen/Mips/fixdfsf.ll
@@ -7,7 +7,7 @@
 ; Function Attrs: nounwind optsize
 define void @foo()  {
 entry:
-  %0 = load double* @x, align 8
+  %0 = load double, double* @x, align 8
   %conv = fptoui double %0 to i32
   store i32 %conv, i32* @y, align 4
 ; pic1:	lw	${{[0-9]+}}, %call16(__fixunsdfsi)(${{[0-9]+}})
diff --git a/test/CodeGen/Mips/fmadd1.ll b/test/CodeGen/Mips/fmadd1.ll
index f0667eec3b33..99d99fada1cf 100644
--- a/test/CodeGen/Mips/fmadd1.ll
+++ b/test/CodeGen/Mips/fmadd1.ll
@@ -8,15 +8,15 @@
 ; RUN: llc < %s -march=mipsel   -mcpu=mips32              -enable-no-nans-fp-math | FileCheck %s -check-prefix=ALL -check-prefix=32   -check-prefix=32-NONAN
 ; RUN: llc < %s -march=mipsel   -mcpu=mips32r2            -enable-no-nans-fp-math | FileCheck %s -check-prefix=ALL -check-prefix=32R2 -check-prefix=32R2-NONAN
 ; RUN: llc < %s -march=mipsel   -mcpu=mips32r6            -enable-no-nans-fp-math | FileCheck %s -check-prefix=ALL -check-prefix=32R6 -check-prefix=32R6-NONAN
-; RUN: llc < %s -march=mips64el -mcpu=mips64   -mattr=n64 -enable-no-nans-fp-math | FileCheck %s -check-prefix=ALL -check-prefix=64   -check-prefix=64-NONAN
-; RUN: llc < %s -march=mips64el -mcpu=mips64r2 -mattr=n64 -enable-no-nans-fp-math | FileCheck %s -check-prefix=ALL -check-prefix=64R2 -check-prefix=64R2-NONAN
-; RUN: llc < %s -march=mips64el -mcpu=mips64r6 -mattr=n64 -enable-no-nans-fp-math | FileCheck %s -check-prefix=ALL -check-prefix=64R6 -check-prefix=64R6-NONAN
+; RUN: llc < %s -march=mips64el -mcpu=mips64   -target-abi=n64 -enable-no-nans-fp-math | FileCheck %s -check-prefix=ALL -check-prefix=64   -check-prefix=64-NONAN
+; RUN: llc < %s -march=mips64el -mcpu=mips64r2 -target-abi=n64 -enable-no-nans-fp-math | FileCheck %s -check-prefix=ALL -check-prefix=64R2 -check-prefix=64R2-NONAN
+; RUN: llc < %s -march=mips64el -mcpu=mips64r6 -target-abi=n64 -enable-no-nans-fp-math | FileCheck %s -check-prefix=ALL -check-prefix=64R6 -check-prefix=64R6-NONAN
 ; RUN: llc < %s -march=mipsel   -mcpu=mips32              | FileCheck %s -check-prefix=ALL -check-prefix=32 -check-prefix=32-NAN
 ; RUN: llc < %s -march=mipsel   -mcpu=mips32r2            | FileCheck %s -check-prefix=ALL -check-prefix=32R2 -check-prefix=32R2-NAN
 ; RUN: llc < %s -march=mipsel   -mcpu=mips32r6            | FileCheck %s -check-prefix=ALL -check-prefix=32R6 -check-prefix=32R6-NAN
-; RUN: llc < %s -march=mips64el -mcpu=mips64   -mattr=n64 | FileCheck %s -check-prefix=ALL -check-prefix=64   -check-prefix=64-NAN
-; RUN: llc < %s -march=mips64el -mcpu=mips64r2 -mattr=n64 | FileCheck %s -check-prefix=ALL -check-prefix=64R2 -check-prefix=64R2-NAN
-; RUN: llc < %s -march=mips64el -mcpu=mips64r6 -mattr=n64 | FileCheck %s -check-prefix=ALL -check-prefix=64R6 -check-prefix=64R6-NAN
+; RUN: llc < %s -march=mips64el -mcpu=mips64   -target-abi=n64 | FileCheck %s -check-prefix=ALL -check-prefix=64   -check-prefix=64-NAN
+; RUN: llc < %s -march=mips64el -mcpu=mips64r2 -target-abi=n64 | FileCheck %s -check-prefix=ALL -check-prefix=64R2 -check-prefix=64R2-NAN
+; RUN: llc < %s -march=mips64el -mcpu=mips64r6 -target-abi=n64 | FileCheck %s -check-prefix=ALL -check-prefix=64R6 -check-prefix=64R6-NAN
 
 define float @FOO0float(float %a, float %b, float %c) nounwind readnone {
 entry:
diff --git a/test/CodeGen/Mips/fp-indexed-ls.ll b/test/CodeGen/Mips/fp-indexed-ls.ll
index 787e131f6ec5..219ca99d3f94 100644
--- a/test/CodeGen/Mips/fp-indexed-ls.ll
+++ b/test/CodeGen/Mips/fp-indexed-ls.ll
@@ -1,10 +1,10 @@
 ; RUN: llc -march=mipsel   -mcpu=mips32   < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS32R1
 ; RUN: llc -march=mipsel   -mcpu=mips32r2 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS32R2
 ; RUN: llc -march=mipsel   -mcpu=mips32r6 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS32R6
-; RUN: llc -march=mips64el -mcpu=mips4    -mattr=n64 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS4
-; RUN: llc -march=mips64el -mcpu=mips64   -mattr=n64 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS4
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=n64 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS4
-; RUN: llc -march=mips64el -mcpu=mips64r6 -mattr=n64 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS64R6
+; RUN: llc -march=mips64el -mcpu=mips4    -target-abi=n64 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS4
+; RUN: llc -march=mips64el -mcpu=mips64   -target-abi=n64 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS4
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi=n64 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS4
+; RUN: llc -march=mips64el -mcpu=mips64r6 -target-abi=n64 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS64R6
 
 ; Check that [ls][dwu]xc1 are not emitted for nacl.
 ; RUN: llc -mtriple=mipsel-none-nacl-gnu -mcpu=mips32r2 < %s | FileCheck %s -check-prefix=CHECK-NACL
@@ -45,8 +45,8 @@ entry:
 
 ; CHECK-NACL-NOT: lwxc1
 
-  %arrayidx = getelementptr inbounds float* %b, i32 %o
-  %0 = load float* %arrayidx, align 4
+  %arrayidx = getelementptr inbounds float, float* %b, i32 %o
+  %0 = load float, float* %arrayidx, align 4
   ret float %0
 }
 
@@ -76,8 +76,8 @@ entry:
 
 ; CHECK-NACL-NOT: ldxc1
 
-  %arrayidx = getelementptr inbounds double* %b, i32 %o
-  %0 = load double* %arrayidx, align 8
+  %arrayidx = getelementptr inbounds double, double* %b, i32 %o
+  %0 = load double, double* %arrayidx, align 8
   ret double %0
 }
 
@@ -100,8 +100,8 @@ entry:
 ; luxc1 was removed in MIPS64r6
 ; MIPS64R6-NOT:  luxc1
 
-  %arrayidx1 = getelementptr inbounds [4 x %struct.S]* @s, i32 0, i32 %b, i32 0, i32 %c
-  %0 = load float* %arrayidx1, align 1
+  %arrayidx1 = getelementptr inbounds [4 x %struct.S], [4 x %struct.S]* @s, i32 0, i32 %b, i32 0, i32 %c
+  %0 = load float, float* %arrayidx1, align 1
   ret float %0
 }
 
@@ -129,8 +129,8 @@ entry:
 
 ; CHECK-NACL-NOT: swxc1
 
-  %0 = load float* @gf, align 4
-  %arrayidx = getelementptr inbounds float* %b, i32 %o
+  %0 = load float, float* @gf, align 4
+  %arrayidx = getelementptr inbounds float, float* %b, i32 %o
   store float %0, float* %arrayidx, align 4
   ret void
 }
@@ -159,8 +159,8 @@ entry:
 
 ; CHECK-NACL-NOT: sdxc1
 
-  %0 = load double* @gd, align 8
-  %arrayidx = getelementptr inbounds double* %b, i32 %o
+  %0 = load double, double* @gd, align 8
+  %arrayidx = getelementptr inbounds double, double* %b, i32 %o
   store double %0, double* %arrayidx, align 8
   ret void
 }
@@ -179,8 +179,8 @@ entry:
 
 ; MIPS64R6-NOT:  suxc1
 
-  %0 = load float* @gf, align 4
-  %arrayidx1 = getelementptr inbounds [4 x %struct.S]* @s, i32 0, i32 %b, i32 0, i32 %c
+  %0 = load float, float* @gf, align 4
+  %arrayidx1 = getelementptr inbounds [4 x %struct.S], [4 x %struct.S]* @s, i32 0, i32 %b, i32 0, i32 %c
   store float %0, float* %arrayidx1, align 1
   ret void
 }
@@ -199,8 +199,8 @@ entry:
 
 ; MIPS64R6-NOT:  luxc1
 
-  %arrayidx1 = getelementptr inbounds [4 x %struct.S2]* @s2, i32 0, i32 %b, i32 0, i32 %c
-  %0 = load double* %arrayidx1, align 1
+  %arrayidx1 = getelementptr inbounds [4 x %struct.S2], [4 x %struct.S2]* @s2, i32 0, i32 %b, i32 0, i32 %c
+  %0 = load double, double* %arrayidx1, align 1
   ret double %0
 }
 
@@ -218,8 +218,8 @@ entry:
 
 ; MIPS64R6-NOT:  suxc1
 
-  %0 = load double* @gd, align 8
-  %arrayidx1 = getelementptr inbounds [4 x %struct.S2]* @s2, i32 0, i32 %b, i32 0, i32 %c
+  %0 = load double, double* @gd, align 8
+  %arrayidx1 = getelementptr inbounds [4 x %struct.S2], [4 x %struct.S2]* @s2, i32 0, i32 %b, i32 0, i32 %c
   store double %0, double* %arrayidx1, align 1
   ret void
 }
@@ -238,7 +238,7 @@ entry:
 
 ; MIPS64R6-NOT:  luxc1
 
-  %0 = load float* getelementptr inbounds (%struct.S3* @s3, i32 0, i32 1), align 1
+  %0 = load float, float* getelementptr inbounds (%struct.S3, %struct.S3* @s3, i32 0, i32 1), align 1
   ret float %0
 }
 
@@ -256,7 +256,7 @@ entry:
 
 ; MIPS64R6-NOT:  suxc1
 
-  store float %f, float* getelementptr inbounds (%struct.S3* @s3, i32 0, i32 1), align 1
+  store float %f, float* getelementptr inbounds (%struct.S3, %struct.S3* @s3, i32 0, i32 1), align 1
   ret void
 }
 
diff --git a/test/CodeGen/Mips/fp-spill-reload.ll b/test/CodeGen/Mips/fp-spill-reload.ll
index f9887a55827f..4a53ad8c8e13 100644
--- a/test/CodeGen/Mips/fp-spill-reload.ll
+++ b/test/CodeGen/Mips/fp-spill-reload.ll
@@ -5,27 +5,27 @@ define void @foo0(i32* nocapture %b) nounwind {
 entry:
 ; CHECK: sw  $fp
 ; CHECK: lw  $fp
-  %0 = load i32* %b, align 4
-  %arrayidx.1 = getelementptr inbounds i32* %b, i32 1
-  %1 = load i32* %arrayidx.1, align 4
+  %0 = load i32, i32* %b, align 4
+  %arrayidx.1 = getelementptr inbounds i32, i32* %b, i32 1
+  %1 = load i32, i32* %arrayidx.1, align 4
   %add.1 = add nsw i32 %1, 1
-  %arrayidx.2 = getelementptr inbounds i32* %b, i32 2
-  %2 = load i32* %arrayidx.2, align 4
+  %arrayidx.2 = getelementptr inbounds i32, i32* %b, i32 2
+  %2 = load i32, i32* %arrayidx.2, align 4
   %add.2 = add nsw i32 %2, 2
-  %arrayidx.3 = getelementptr inbounds i32* %b, i32 3
-  %3 = load i32* %arrayidx.3, align 4
+  %arrayidx.3 = getelementptr inbounds i32, i32* %b, i32 3
+  %3 = load i32, i32* %arrayidx.3, align 4
   %add.3 = add nsw i32 %3, 3
-  %arrayidx.4 = getelementptr inbounds i32* %b, i32 4
-  %4 = load i32* %arrayidx.4, align 4
+  %arrayidx.4 = getelementptr inbounds i32, i32* %b, i32 4
+  %4 = load i32, i32* %arrayidx.4, align 4
   %add.4 = add nsw i32 %4, 4
-  %arrayidx.5 = getelementptr inbounds i32* %b, i32 5
-  %5 = load i32* %arrayidx.5, align 4
+  %arrayidx.5 = getelementptr inbounds i32, i32* %b, i32 5
+  %5 = load i32, i32* %arrayidx.5, align 4
   %add.5 = add nsw i32 %5, 5
-  %arrayidx.6 = getelementptr inbounds i32* %b, i32 6
-  %6 = load i32* %arrayidx.6, align 4
+  %arrayidx.6 = getelementptr inbounds i32, i32* %b, i32 6
+  %6 = load i32, i32* %arrayidx.6, align 4
   %add.6 = add nsw i32 %6, 6
-  %arrayidx.7 = getelementptr inbounds i32* %b, i32 7
-  %7 = load i32* %arrayidx.7, align 4
+  %arrayidx.7 = getelementptr inbounds i32, i32* %b, i32 7
+  %7 = load i32, i32* %arrayidx.7, align 4
   %add.7 = add nsw i32 %7, 7
   call void @foo2(i32 %0, i32 %add.1, i32 %add.2, i32 %add.3, i32 %add.4, i32 %add.5, i32 %add.6, i32 %add.7) nounwind
   call void bitcast (void (...)* @foo1 to void ()*)() nounwind
diff --git a/test/CodeGen/Mips/fp16-promote.ll b/test/CodeGen/Mips/fp16-promote.ll
new file mode 100644
index 000000000000..2ac46e028072
--- /dev/null
+++ b/test/CodeGen/Mips/fp16-promote.ll
@@ -0,0 +1,98 @@
+; RUN: llc -asm-verbose=false -mtriple=mipsel-linux-gnueabi < %s | FileCheck %s -check-prefix=CHECK-LIBCALL
+
+; CHECK-LIBCALL-LABEL: test_fadd:
+; CHECK-LIBCALL: %call16(__gnu_h2f_ieee)
+; CHECK-LIBCALL: %call16(__gnu_h2f_ieee)
+; CHECK-LIBCALL-DAG: add.s
+; CHECK-LIBCALL-DAG: %call16(__gnu_f2h_ieee)
+define void @test_fadd(half* %p, half* %q) #0 {
+  %a = load half, half* %p, align 2
+  %b = load half, half* %q, align 2
+  %r = fadd half %a, %b
+  store half %r, half* %p
+  ret void
+}
+
+; CHECK-LIBCALL-LABEL: test_fpext_float:
+; CHECK-LIBCALL: %call16(__gnu_h2f_ieee)
+define float @test_fpext_float(half* %p) {
+  %a = load half, half* %p, align 2
+  %r = fpext half %a to float
+  ret float %r
+}
+
+; CHECK-LIBCALL-LABEL: test_fpext_double:
+; CHECK-LIBCALL: %call16(__gnu_h2f_ieee)
+; CHECK-LIBCALL: cvt.d.s
+define double @test_fpext_double(half* %p) {
+  %a = load half, half* %p, align 2
+  %r = fpext half %a to double
+  ret double %r
+}
+
+; CHECK-LIBCALL-LABEL: test_fptrunc_float:
+; CHECK-LIBCALL: %call16(__gnu_f2h_ieee)
+define void @test_fptrunc_float(float %f, half* %p) #0 {
+  %a = fptrunc float %f to half
+  store half %a, half* %p
+  ret void
+}
+
+; CHECK-LIBCALL-LABEL: test_fptrunc_double:
+; CHECK-LIBCALL: %call16(__truncdfhf2)
+define void @test_fptrunc_double(double %d, half* %p) #0 {
+  %a = fptrunc double %d to half
+  store half %a, half* %p
+  ret void
+}
+
+; CHECK-LIBCALL-LABEL: test_vec_fpext_float:
+; CHECK-LIBCALL: %call16(__gnu_h2f_ieee)
+; CHECK-LIBCALL: %call16(__gnu_h2f_ieee)
+; CHECK-LIBCALL: %call16(__gnu_h2f_ieee)
+; CHECK-LIBCALL: %call16(__gnu_h2f_ieee)
+define <4 x float> @test_vec_fpext_float(<4 x half>* %p) #0 {
+  %a = load <4 x half>, <4 x half>* %p, align 8
+  %b = fpext <4 x half> %a to <4 x float>
+  ret <4 x float> %b
+}
+
+; This test is not robust against variations in instruction scheduling.
+; See the discussion in http://reviews.llvm.org/D8804
+; CHECK-LIBCALL-LABEL: test_vec_fpext_double:
+; CHECK-LIBCALL: %call16(__gnu_h2f_ieee)
+; CHECK-LIBCALL: %call16(__gnu_h2f_ieee)
+; CHECK-LIBCALL: %call16(__gnu_h2f_ieee)
+; CHECK-LIBCALL: cvt.d.s
+; CHECK-LIBCALL: cvt.d.s
+; CHECK-LIBCALL: cvt.d.s
+; CHECK-LIBCALL: %call16(__gnu_h2f_ieee)
+; CHECK-LIBCALL: cvt.d.s
+define <4 x double> @test_vec_fpext_double(<4 x half>* %p) #0 {
+  %a = load <4 x half>, <4 x half>* %p, align 8
+  %b = fpext <4 x half> %a to <4 x double>
+  ret <4 x double> %b
+}
+
+; CHECK-LIBCALL-LABEL: test_vec_fptrunc_float:
+; CHECK-LIBCALL: %call16(__gnu_f2h_ieee)
+; CHECK-LIBCALL: %call16(__gnu_f2h_ieee)
+; CHECK-LIBCALL: %call16(__gnu_f2h_ieee)
+; CHECK-LIBCALL: %call16(__gnu_f2h_ieee)
+define void @test_vec_fptrunc_float(<4 x float> %a, <4 x half>* %p) #0 {
+  %b = fptrunc <4 x float> %a to <4 x half>
+  store <4 x half> %b, <4 x half>* %p, align 8
+  ret void
+}
+
+; CHECK-LIBCALL-LABEL: test_vec_fptrunc_double:
+; CHECK-LIBCALL: %call16(__truncdfhf2)
+; CHECK-LIBCALL: %call16(__truncdfhf2)
+; CHECK-LIBCALL: %call16(__truncdfhf2)
+; CHECK-LIBCALL: %call16(__truncdfhf2)
+define void @test_vec_fptrunc_double(<4 x double> %a, <4 x half>* %p) #0 {
+  %b = fptrunc <4 x double> %a to <4 x half>
+  store <4 x half> %b, <4 x half>* %p, align 8
+  ret void
+}
+
diff --git a/test/CodeGen/Mips/fp16instrinsmc.ll b/test/CodeGen/Mips/fp16instrinsmc.ll
index 84d3814ee8b8..797be2668d40 100644
--- a/test/CodeGen/Mips/fp16instrinsmc.ll
+++ b/test/CodeGen/Mips/fp16instrinsmc.ll
@@ -23,8 +23,8 @@ define void @foo1() #0 {
 ; fmask: .set	reorder
 ; fmask: .end	foo1
 entry:
-  %0 = load float* @x, align 4
-  %1 = load float* @one, align 4
+  %0 = load float, float* @x, align 4
+  %1 = load float, float* @one, align 4
   %call = call float @copysignf(float %0, float %1) #2
   store float %call, float* @y, align 4
   ret void
@@ -39,8 +39,8 @@ define void @foo2() #0 {
 ; fmask:	save	{{.*}}
 ; fmask:	.end	foo2
 entry:
-  %0 = load float* @x, align 4
-  %1 = load float* @negone, align 4
+  %0 = load float, float* @x, align 4
+  %1 = load float, float* @negone, align 4
   %call = call float @copysignf(float %0, float %1) #2
   store float %call, float* @y, align 4
   ret void
@@ -57,8 +57,8 @@ entry:
 ; fmask: .set	macro
 ; fmask: .set	reorder
 ; fmask: .end	foo3
-  %0 = load double* @xd, align 8
-  %1 = load float* @oned, align 4
+  %0 = load double, double* @xd, align 8
+  %1 = load float, float* @oned, align 4
   %conv = fpext float %1 to double
   %call = call double @copysign(double %0, double %conv) #2
   store double %call, double* @yd, align 8
@@ -74,8 +74,8 @@ entry:
 ; fmask:	.ent	foo4
 ; fmask:	save	{{.*}}
 ; fmask:	.end	foo4
-  %0 = load double* @xd, align 8
-  %1 = load double* @negoned, align 8
+  %0 = load double, double* @xd, align 8
+  %1 = load double, double* @negoned, align 8
   %call = call double @copysign(double %0, double %1) #2
   store double %call, double* @yd, align 8
   ret void
@@ -84,7 +84,7 @@ entry:
 ; Function Attrs: nounwind
 define void @foo5() #0 {
 entry:
-  %0 = load float* @xn, align 4
+  %0 = load float, float* @xn, align 4
   %call = call float @fabsf(float %0) #2
   store float %call, float* @y, align 4
   ret void
@@ -96,7 +96,7 @@ declare float @fabsf(float) #1
 ; Function Attrs: nounwind
 define void @foo6() #0 {
 entry:
-  %0 = load double* @xdn, align 8
+  %0 = load double, double* @xdn, align 8
   %call = call double @fabs(double %0) #2
   store double %call, double* @yd, align 8
   ret void
@@ -108,7 +108,7 @@ declare double @fabs(double) #1
 ; Function Attrs: nounwind
 define void @foo7() #0 {
 entry:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   %call = call float @sinf(float %0) #3
 ;pic:	lw	${{[0-9]+}}, %call16(sinf)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_sf_1)(${{[0-9]+}})
@@ -122,7 +122,7 @@ declare float @sinf(float) #0
 ; Function Attrs: nounwind
 define void @foo8() #0 {
 entry:
-  %0 = load double* @xd, align 8
+  %0 = load double, double* @xd, align 8
   %call = call double @sin(double %0) #3
 ;pic:	lw	${{[0-9]+}}, %call16(sin)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_df_2)(${{[0-9]+}})
@@ -136,7 +136,7 @@ declare double @sin(double) #0
 ; Function Attrs: nounwind
 define void @foo9() #0 {
 entry:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   %call = call float @cosf(float %0) #3
 ;pic:	lw	${{[0-9]+}}, %call16(cosf)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_sf_1)(${{[0-9]+}})
@@ -150,7 +150,7 @@ declare float @cosf(float) #0
 ; Function Attrs: nounwind
 define void @foo10() #0 {
 entry:
-  %0 = load double* @xd, align 8
+  %0 = load double, double* @xd, align 8
   %call = call double @cos(double %0) #3
 ;pic:	lw	${{[0-9]+}}, %call16(cos)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_df_2)(${{[0-9]+}})
@@ -164,7 +164,7 @@ declare double @cos(double) #0
 ; Function Attrs: nounwind
 define void @foo11() #0 {
 entry:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   %call = call float @sqrtf(float %0) #3
 ;pic:	lw	${{[0-9]+}}, %call16(sqrtf)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_sf_1)(${{[0-9]+}})
@@ -178,7 +178,7 @@ declare float @sqrtf(float) #0
 ; Function Attrs: nounwind
 define void @foo12() #0 {
 entry:
-  %0 = load double* @xd, align 8
+  %0 = load double, double* @xd, align 8
   %call = call double @sqrt(double %0) #3
 ;pic:	lw	${{[0-9]+}}, %call16(sqrt)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_df_2)(${{[0-9]+}})
@@ -192,7 +192,7 @@ declare double @sqrt(double) #0
 ; Function Attrs: nounwind
 define void @foo13() #0 {
 entry:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   %call = call float @floorf(float %0) #2
 ;pic:	lw	${{[0-9]+}}, %call16(floorf)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_sf_1)(${{[0-9]+}})
@@ -206,7 +206,7 @@ declare float @floorf(float) #1
 ; Function Attrs: nounwind
 define void @foo14() #0 {
 entry:
-  %0 = load double* @xd, align 8
+  %0 = load double, double* @xd, align 8
   %call = call double @floor(double %0) #2
 ;pic:	lw	${{[0-9]+}}, %call16(floor)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_df_2)(${{[0-9]+}})
@@ -220,7 +220,7 @@ declare double @floor(double) #1
 ; Function Attrs: nounwind
 define void @foo15() #0 {
 entry:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   %call = call float @nearbyintf(float %0) #2
 ;pic:	lw	${{[0-9]+}}, %call16(nearbyintf)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_sf_1)(${{[0-9]+}})
@@ -234,7 +234,7 @@ declare float @nearbyintf(float) #1
 ; Function Attrs: nounwind
 define void @foo16() #0 {
 entry:
-  %0 = load double* @xd, align 8
+  %0 = load double, double* @xd, align 8
   %call = call double @nearbyint(double %0) #2
 ;pic:	lw	${{[0-9]+}}, %call16(nearbyint)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_df_2)(${{[0-9]+}})
@@ -248,7 +248,7 @@ declare double @nearbyint(double) #1
 ; Function Attrs: nounwind
 define void @foo17() #0 {
 entry:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   %call = call float @ceilf(float %0) #2
 ;pic:	lw	${{[0-9]+}}, %call16(ceilf)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_sf_1)(${{[0-9]+}})
@@ -262,7 +262,7 @@ declare float @ceilf(float) #1
 ; Function Attrs: nounwind
 define void @foo18() #0 {
 entry:
-  %0 = load double* @xd, align 8
+  %0 = load double, double* @xd, align 8
   %call = call double @ceil(double %0) #2
 ;pic:	lw	${{[0-9]+}}, %call16(ceil)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_df_2)(${{[0-9]+}})
@@ -276,7 +276,7 @@ declare double @ceil(double) #1
 ; Function Attrs: nounwind
 define void @foo19() #0 {
 entry:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   %call = call float @rintf(float %0) #2
 ;pic:	lw	${{[0-9]+}}, %call16(rintf)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_sf_1)(${{[0-9]+}})
@@ -290,7 +290,7 @@ declare float @rintf(float) #1
 ; Function Attrs: nounwind
 define void @foo20() #0 {
 entry:
-  %0 = load double* @xd, align 8
+  %0 = load double, double* @xd, align 8
   %call = call double @rint(double %0) #2
 ;pic:	lw	${{[0-9]+}}, %call16(rint)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_df_2)(${{[0-9]+}})
@@ -304,7 +304,7 @@ declare double @rint(double) #1
 ; Function Attrs: nounwind
 define void @foo21() #0 {
 entry:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   %call = call float @truncf(float %0) #2
 ;pic:	lw	${{[0-9]+}}, %call16(truncf)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_sf_1)(${{[0-9]+}})
@@ -318,7 +318,7 @@ declare float @truncf(float) #1
 ; Function Attrs: nounwind
 define void @foo22() #0 {
 entry:
-  %0 = load double* @xd, align 8
+  %0 = load double, double* @xd, align 8
   %call = call double @trunc(double %0) #2
 ;pic:	lw	${{[0-9]+}}, %call16(trunc)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_df_2)(${{[0-9]+}})
@@ -332,7 +332,7 @@ declare double @trunc(double) #1
 ; Function Attrs: nounwind
 define void @foo23() #0 {
 entry:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   %call = call float @log2f(float %0) #3
 ;pic:	lw	${{[0-9]+}}, %call16(log2f)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_sf_1)(${{[0-9]+}})
@@ -346,7 +346,7 @@ declare float @log2f(float) #0
 ; Function Attrs: nounwind
 define void @foo24() #0 {
 entry:
-  %0 = load double* @xd, align 8
+  %0 = load double, double* @xd, align 8
   %call = call double @log2(double %0) #3
 ;pic:	lw	${{[0-9]+}}, %call16(log2)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_df_2)(${{[0-9]+}})
@@ -360,7 +360,7 @@ declare double @log2(double) #0
 ; Function Attrs: nounwind
 define void @foo25() #0 {
 entry:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   %call = call float @exp2f(float %0) #3
 ;pic:	lw	${{[0-9]+}}, %call16(exp2f)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_sf_1)(${{[0-9]+}})
@@ -374,7 +374,7 @@ declare float @exp2f(float) #0
 ; Function Attrs: nounwind
 define void @foo26() #0 {
 entry:
-  %0 = load double* @xd, align 8
+  %0 = load double, double* @xd, align 8
   %call = call double @exp2(double %0) #3
 ;pic:	lw	${{[0-9]+}}, %call16(exp2)(${{[0-9]+}})
 ;pic:	lw	${{[0-9]+}}, %got(__mips16_call_stub_df_2)(${{[0-9]+}})
diff --git a/test/CodeGen/Mips/fp16static.ll b/test/CodeGen/Mips/fp16static.ll
index beb063db15ca..4e5059ed39e9 100644
--- a/test/CodeGen/Mips/fp16static.ll
+++ b/test/CodeGen/Mips/fp16static.ll
@@ -4,8 +4,8 @@
 
 define void @foo() nounwind {
 entry:
-  %0 = load float* @x, align 4
-  %1 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
+  %1 = load float, float* @x, align 4
   %mul = fmul float %0, %1
   store float %mul, float* @x, align 4
 ; CHECK-STATIC16: jal	__mips16_mulsf3
diff --git a/test/CodeGen/Mips/fpbr.ll b/test/CodeGen/Mips/fpbr.ll
index 311b83015a56..27d7094376e6 100644
--- a/test/CodeGen/Mips/fpbr.ll
+++ b/test/CodeGen/Mips/fpbr.ll
@@ -24,11 +24,11 @@ entry:
   br i1 %cmp, label %if.then, label %if.else
 
 if.then:                                          ; preds = %entry
-  tail call void (...)* @g0() nounwind
+  tail call void (...) @g0() nounwind
   br label %if.end
 
 if.else:                                          ; preds = %entry
-  tail call void (...)* @g1() nounwind
+  tail call void (...) @g1() nounwind
   br label %if.end
 
 if.end:                                           ; preds = %if.else, %if.then
@@ -57,11 +57,11 @@ entry:
   br i1 %cmp, label %if.then, label %if.else
 
 if.then:                                          ; preds = %entry
-  tail call void (...)* @g0() nounwind
+  tail call void (...) @g0() nounwind
   br label %if.end
 
 if.else:                                          ; preds = %entry
-  tail call void (...)* @g1() nounwind
+  tail call void (...) @g1() nounwind
   br label %if.end
 
 if.end:                                           ; preds = %if.else, %if.then
@@ -86,11 +86,11 @@ entry:
   br i1 %cmp, label %if.else, label %if.then
 
 if.then:                                          ; preds = %entry
-  tail call void (...)* @g0() nounwind
+  tail call void (...) @g0() nounwind
   br label %if.end
 
 if.else:                                          ; preds = %entry
-  tail call void (...)* @g1() nounwind
+  tail call void (...) @g1() nounwind
   br label %if.end
 
 if.end:                                           ; preds = %if.else, %if.then
@@ -116,11 +116,11 @@ entry:
   br i1 %cmp, label %if.then, label %if.else
 
 if.then:                                          ; preds = %entry
-  tail call void (...)* @g0() nounwind
+  tail call void (...) @g0() nounwind
   br label %if.end
 
 if.else:                                          ; preds = %entry
-  tail call void (...)* @g1() nounwind
+  tail call void (...) @g1() nounwind
   br label %if.end
 
 if.end:                                           ; preds = %if.else, %if.then
@@ -145,11 +145,11 @@ entry:
   br i1 %cmp, label %if.then, label %if.else
 
 if.then:                                          ; preds = %entry
-  tail call void (...)* @g0() nounwind
+  tail call void (...) @g0() nounwind
   br label %if.end
 
 if.else:                                          ; preds = %entry
-  tail call void (...)* @g1() nounwind
+  tail call void (...) @g1() nounwind
   br label %if.end
 
 if.end:                                           ; preds = %if.else, %if.then
@@ -174,11 +174,11 @@ entry:
   br i1 %cmp, label %if.else, label %if.then
 
 if.then:                                          ; preds = %entry
-  tail call void (...)* @g0() nounwind
+  tail call void (...) @g0() nounwind
   br label %if.end
 
 if.else:                                          ; preds = %entry
-  tail call void (...)* @g1() nounwind
+  tail call void (...) @g1() nounwind
   br label %if.end
 
 if.end:                                           ; preds = %if.else, %if.then
diff --git a/test/CodeGen/Mips/fpneeded.ll b/test/CodeGen/Mips/fpneeded.ll
index fdd8e8f707ef..a89e2a593a4d 100644
--- a/test/CodeGen/Mips/fpneeded.ll
+++ b/test/CodeGen/Mips/fpneeded.ll
@@ -76,8 +76,8 @@ entry:
 define void @foo1() #0 {
 entry:
   store float 1.000000e+00, float* @zz, align 4
-  %0 = load float* @y, align 4
-  %1 = load float* @x, align 4
+  %0 = load float, float* @y, align 4
+  %1 = load float, float* @x, align 4
   %add = fadd float %0, %1
   store float %add, float* @z, align 4
   ret void
@@ -96,7 +96,7 @@ entry:
 
 define void @foo2() #0 {
 entry:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   call void @vf(float %0)
   ret void
 }
diff --git a/test/CodeGen/Mips/fpnotneeded.ll b/test/CodeGen/Mips/fpnotneeded.ll
index e12d7baacdbb..02b8e8a345db 100644
--- a/test/CodeGen/Mips/fpnotneeded.ll
+++ b/test/CodeGen/Mips/fpnotneeded.ll
@@ -19,7 +19,7 @@ entry:
 
 define i32 @iv() #0 {
 entry:
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   ret i32 %0
 }
 
diff --git a/test/CodeGen/Mips/fpxx.ll b/test/CodeGen/Mips/fpxx.ll
index 7e2ed22e2d80..5b42ecec53e8 100644
--- a/test/CodeGen/Mips/fpxx.ll
+++ b/test/CodeGen/Mips/fpxx.ll
@@ -10,11 +10,11 @@
 ; RUN: llc -march=mips64 -mcpu=mips64 < %s | FileCheck %s -check-prefix=ALL -check-prefix=64-NOFPXX
 ; RUN: not llc -march=mips64 -mcpu=mips64 -mattr=fpxx < %s 2>&1 | FileCheck %s -check-prefix=64-FPXX
 
-; RUN-TODO: llc -march=mips64 -mcpu=mips4 -mattr=-n64,+o32 < %s | FileCheck %s -check-prefix=ALL -check-prefix=4-O32-NOFPXX
-; RUN-TODO: llc -march=mips64 -mcpu=mips4 -mattr=-n64,+o32 -mattr=fpxx < %s | FileCheck %s -check-prefix=ALL -check-prefix=4-O32-FPXX
+; RUN-TODO: llc -march=mips64 -mcpu=mips4 -target-abi o32 < %s | FileCheck %s -check-prefix=ALL -check-prefix=4-O32-NOFPXX
+; RUN-TODO: llc -march=mips64 -mcpu=mips4 -target-abi o32 -mattr=fpxx < %s | FileCheck %s -check-prefix=ALL -check-prefix=4-O32-FPXX
 
-; RUN-TODO: llc -march=mips64 -mcpu=mips64 -mattr=-n64,+o32 < %s | FileCheck %s -check-prefix=ALL -check-prefix=64-O32-NOFPXX
-; RUN-TODO: llc -march=mips64 -mcpu=mips64 -mattr=-n64,+o32 -mattr=fpxx < %s | FileCheck %s -check-prefix=ALL -check-prefix=64-O32-FPXX
+; RUN-TODO: llc -march=mips64 -mcpu=mips64 -target-abi o32 < %s | FileCheck %s -check-prefix=ALL -check-prefix=64-O32-NOFPXX
+; RUN-TODO: llc -march=mips64 -mcpu=mips64 -target-abi o32 -mattr=fpxx < %s | FileCheck %s -check-prefix=ALL -check-prefix=64-O32-FPXX
 
 declare double @dbl();
 
diff --git a/test/CodeGen/Mips/global-address.ll b/test/CodeGen/Mips/global-address.ll
index 0785cfcc0515..ecf5e563a577 100644
--- a/test/CodeGen/Mips/global-address.ll
+++ b/test/CodeGen/Mips/global-address.ll
@@ -1,9 +1,9 @@
 ; RUN: llc -march=mipsel -relocation-model=pic < %s | FileCheck %s -check-prefix=PIC-O32
 ; RUN: llc -march=mipsel -relocation-model=static -mtriple=mipsel-linux-gnu < %s | FileCheck %s -check-prefix=STATIC-O32
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=-n64,n32 -relocation-model=pic < %s | FileCheck %s -check-prefix=PIC-N32
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=-n64,n32 -relocation-model=static  -mtriple=mipsel-linux-gnu < %s | FileCheck %s -check-prefix=STATIC-N32
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=-n64,n64 -relocation-model=pic < %s | FileCheck %s -check-prefix=PIC-N64
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=-n64,n64 -relocation-model=static < %s | FileCheck %s -check-prefix=STATIC-N64
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi n32 -relocation-model=pic < %s | FileCheck %s -check-prefix=PIC-N32
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi n32 -relocation-model=static  -mtriple=mipsel-linux-gnu < %s | FileCheck %s -check-prefix=STATIC-N32
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi n64 -relocation-model=pic < %s | FileCheck %s -check-prefix=PIC-N64
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi n64 -relocation-model=static < %s | FileCheck %s -check-prefix=STATIC-N64
 
 @s1 = internal unnamed_addr global i32 8, align 4
 @g1 = external global i32
@@ -33,9 +33,9 @@ entry:
 ; STATIC-N64: lw  ${{[0-9]+}}, %got_ofst(s1)($[[R1]])
 ; STATIC-N64: ld  ${{[0-9]+}}, %got_disp(g1)
 
-  %0 = load i32* @s1, align 4
+  %0 = load i32, i32* @s1, align 4
   tail call void @foo1(i32 %0) nounwind
-  %1 = load i32* @g1, align 4
+  %1 = load i32, i32* @g1, align 4
   store i32 %1, i32* @s1, align 4
   %add = add nsw i32 %1, 2
   store i32 %add, i32* @g1, align 4
diff --git a/test/CodeGen/Mips/gpreg-lazy-binding.ll b/test/CodeGen/Mips/gpreg-lazy-binding.ll
index 3a636d82533f..800a74f5358f 100644
--- a/test/CodeGen/Mips/gpreg-lazy-binding.ll
+++ b/test/CodeGen/Mips/gpreg-lazy-binding.ll
@@ -19,7 +19,7 @@ declare void @externalFunc()
 
 define internal fastcc void @internalFunc() nounwind noinline {
 entry:
-  %0 = load i32* @g, align 4
+  %0 = load i32, i32* @g, align 4
   %inc = add nsw i32 %0, 1
   store i32 %inc, i32* @g, align 4
   ret void
diff --git a/test/CodeGen/Mips/gprestore.ll b/test/CodeGen/Mips/gprestore.ll
index cbcf0c93491c..b1c2ad1d2756 100644
--- a/test/CodeGen/Mips/gprestore.ll
+++ b/test/CodeGen/Mips/gprestore.ll
@@ -17,11 +17,11 @@ entry:
 ; CHECK: jalr
 ; CHECK-NOT: got({{.*}})($gp)
 ; CHECK: lw $gp
-  tail call void (...)* @f1() nounwind
-  %tmp = load i32* @p, align 4
+  tail call void (...) @f1() nounwind
+  %tmp = load i32, i32* @p, align 4
   tail call void @f2(i32 %tmp) nounwind
-  %tmp1 = load i32* @q, align 4
-  %tmp2 = load i32* @r, align 4
+  %tmp1 = load i32, i32* @q, align 4
+  %tmp2 = load i32, i32* @r, align 4
   tail call void @f3(i32 %tmp1, i32 %tmp2) nounwind
   ret void
 }
diff --git a/test/CodeGen/Mips/helloworld.ll b/test/CodeGen/Mips/helloworld.ll
index 36f4ad6b55c0..a0dbdf3afd47 100644
--- a/test/CodeGen/Mips/helloworld.ll
+++ b/test/CodeGen/Mips/helloworld.ll
@@ -12,7 +12,7 @@
 
 define i32 @main() nounwind {
 entry:
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([13 x i8]* @.str, i32 0, i32 0))
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([13 x i8], [13 x i8]* @.str, i32 0, i32 0))
   ret i32 0
 
 ; SR: 	.set	mips16
diff --git a/test/CodeGen/Mips/hf16_1.ll b/test/CodeGen/Mips/hf16_1.ll
index 9879cd523af3..103fd2d7fd63 100644
--- a/test/CodeGen/Mips/hf16_1.ll
+++ b/test/CodeGen/Mips/hf16_1.ll
@@ -11,96 +11,96 @@
 
 define void @foo() nounwind {
 entry:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   call void @v_sf(float %0)
-  %1 = load double* @xd, align 8
+  %1 = load double, double* @xd, align 8
   call void @v_df(double %1)
-  %2 = load float* @x, align 4
-  %3 = load float* @y, align 4
+  %2 = load float, float* @x, align 4
+  %3 = load float, float* @y, align 4
   call void @v_sf_sf(float %2, float %3)
-  %4 = load double* @xd, align 8
-  %5 = load float* @x, align 4
+  %4 = load double, double* @xd, align 8
+  %5 = load float, float* @x, align 4
   call void @v_df_sf(double %4, float %5)
-  %6 = load double* @xd, align 8
-  %7 = load double* @yd, align 8
+  %6 = load double, double* @xd, align 8
+  %7 = load double, double* @yd, align 8
   call void @v_df_df(double %6, double %7)
   %call = call float @sf_v()
-  %8 = load float* @x, align 4
+  %8 = load float, float* @x, align 4
   %call1 = call float @sf_sf(float %8)
-  %9 = load double* @xd, align 8
+  %9 = load double, double* @xd, align 8
   %call2 = call float @sf_df(double %9)
-  %10 = load float* @x, align 4
-  %11 = load float* @y, align 4
+  %10 = load float, float* @x, align 4
+  %11 = load float, float* @y, align 4
   %call3 = call float @sf_sf_sf(float %10, float %11)
-  %12 = load double* @xd, align 8
-  %13 = load float* @x, align 4
+  %12 = load double, double* @xd, align 8
+  %13 = load float, float* @x, align 4
   %call4 = call float @sf_df_sf(double %12, float %13)
-  %14 = load double* @xd, align 8
-  %15 = load double* @yd, align 8
+  %14 = load double, double* @xd, align 8
+  %15 = load double, double* @yd, align 8
   %call5 = call float @sf_df_df(double %14, double %15)
   %call6 = call double @df_v()
-  %16 = load float* @x, align 4
+  %16 = load float, float* @x, align 4
   %call7 = call double @df_sf(float %16)
-  %17 = load double* @xd, align 8
+  %17 = load double, double* @xd, align 8
   %call8 = call double @df_df(double %17)
-  %18 = load float* @x, align 4
-  %19 = load float* @y, align 4
+  %18 = load float, float* @x, align 4
+  %19 = load float, float* @y, align 4
   %call9 = call double @df_sf_sf(float %18, float %19)
-  %20 = load double* @xd, align 8
-  %21 = load float* @x, align 4
+  %20 = load double, double* @xd, align 8
+  %21 = load float, float* @x, align 4
   %call10 = call double @df_df_sf(double %20, float %21)
-  %22 = load double* @xd, align 8
-  %23 = load double* @yd, align 8
+  %22 = load double, double* @xd, align 8
+  %23 = load double, double* @yd, align 8
   %call11 = call double @df_df_df(double %22, double %23)
   %call12 = call { float, float } @sc_v()
   %24 = extractvalue { float, float } %call12, 0
   %25 = extractvalue { float, float } %call12, 1
-  %26 = load float* @x, align 4
+  %26 = load float, float* @x, align 4
   %call13 = call { float, float } @sc_sf(float %26)
   %27 = extractvalue { float, float } %call13, 0
   %28 = extractvalue { float, float } %call13, 1
-  %29 = load double* @xd, align 8
+  %29 = load double, double* @xd, align 8
   %call14 = call { float, float } @sc_df(double %29)
   %30 = extractvalue { float, float } %call14, 0
   %31 = extractvalue { float, float } %call14, 1
-  %32 = load float* @x, align 4
-  %33 = load float* @y, align 4
+  %32 = load float, float* @x, align 4
+  %33 = load float, float* @y, align 4
   %call15 = call { float, float } @sc_sf_sf(float %32, float %33)
   %34 = extractvalue { float, float } %call15, 0
   %35 = extractvalue { float, float } %call15, 1
-  %36 = load double* @xd, align 8
-  %37 = load float* @x, align 4
+  %36 = load double, double* @xd, align 8
+  %37 = load float, float* @x, align 4
   %call16 = call { float, float } @sc_df_sf(double %36, float %37)
   %38 = extractvalue { float, float } %call16, 0
   %39 = extractvalue { float, float } %call16, 1
-  %40 = load double* @xd, align 8
-  %41 = load double* @yd, align 8
+  %40 = load double, double* @xd, align 8
+  %41 = load double, double* @yd, align 8
   %call17 = call { float, float } @sc_df_df(double %40, double %41)
   %42 = extractvalue { float, float } %call17, 0
   %43 = extractvalue { float, float } %call17, 1
   %call18 = call { double, double } @dc_v()
   %44 = extractvalue { double, double } %call18, 0
   %45 = extractvalue { double, double } %call18, 1
-  %46 = load float* @x, align 4
+  %46 = load float, float* @x, align 4
   %call19 = call { double, double } @dc_sf(float %46)
   %47 = extractvalue { double, double } %call19, 0
   %48 = extractvalue { double, double } %call19, 1
-  %49 = load double* @xd, align 8
+  %49 = load double, double* @xd, align 8
   %call20 = call { double, double } @dc_df(double %49)
   %50 = extractvalue { double, double } %call20, 0
   %51 = extractvalue { double, double } %call20, 1
-  %52 = load float* @x, align 4
-  %53 = load float* @y, align 4
+  %52 = load float, float* @x, align 4
+  %53 = load float, float* @y, align 4
   %call21 = call { double, double } @dc_sf_sf(float %52, float %53)
   %54 = extractvalue { double, double } %call21, 0
   %55 = extractvalue { double, double } %call21, 1
-  %56 = load double* @xd, align 8
-  %57 = load float* @x, align 4
+  %56 = load double, double* @xd, align 8
+  %57 = load float, float* @x, align 4
   %call22 = call { double, double } @dc_df_sf(double %56, float %57)
   %58 = extractvalue { double, double } %call22, 0
   %59 = extractvalue { double, double } %call22, 1
-  %60 = load double* @xd, align 8
-  %61 = load double* @yd, align 8
+  %60 = load double, double* @xd, align 8
+  %61 = load double, double* @yd, align 8
   %call23 = call { double, double } @dc_df_df(double %60, double %61)
   %62 = extractvalue { double, double } %call23, 0
   %63 = extractvalue { double, double } %call23, 1
diff --git a/test/CodeGen/Mips/hf16call32.ll b/test/CodeGen/Mips/hf16call32.ll
index aec9c71c485b..3b3f8f799111 100644
--- a/test/CodeGen/Mips/hf16call32.ll
+++ b/test/CodeGen/Mips/hf16call32.ll
@@ -33,30 +33,30 @@ entry:
   store float 1.000000e+00, float* @y, align 4
   store double 1.000000e+00, double* @xd, align 8
   store double 1.000000e+00, double* @yd, align 8
-  store float 1.000000e+00, float* getelementptr inbounds ({ float, float }* @xy, i32 0, i32 0)
-  store float 0.000000e+00, float* getelementptr inbounds ({ float, float }* @xy, i32 0, i32 1)
-  store double 1.000000e+00, double* getelementptr inbounds ({ double, double }* @xyd, i32 0, i32 0)
-  store double 0.000000e+00, double* getelementptr inbounds ({ double, double }* @xyd, i32 0, i32 1)
+  store float 1.000000e+00, float* getelementptr inbounds ({ float, float }, { float, float }* @xy, i32 0, i32 0)
+  store float 0.000000e+00, float* getelementptr inbounds ({ float, float }, { float, float }* @xy, i32 0, i32 1)
+  store double 1.000000e+00, double* getelementptr inbounds ({ double, double }, { double, double }* @xyd, i32 0, i32 0)
+  store double 0.000000e+00, double* getelementptr inbounds ({ double, double }, { double, double }* @xyd, i32 0, i32 1)
   store float 1.000000e+00, float* @ret_sf, align 4
   store double 1.000000e+00, double* @ret_df, align 8
-  store float 1.000000e+00, float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 0)
-  store float 0.000000e+00, float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 1)
-  store double 1.000000e+00, double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 0)
-  store double 0.000000e+00, double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 1)
+  store float 1.000000e+00, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 0)
+  store float 0.000000e+00, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 1)
+  store double 1.000000e+00, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 0)
+  store double 0.000000e+00, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 1)
   store float 0.000000e+00, float* @lx, align 4
   store float 0.000000e+00, float* @ly, align 4
   store double 0.000000e+00, double* @lxd, align 8
   store double 0.000000e+00, double* @lyd, align 8
-  store float 0.000000e+00, float* getelementptr inbounds ({ float, float }* @lxy, i32 0, i32 0)
-  store float 0.000000e+00, float* getelementptr inbounds ({ float, float }* @lxy, i32 0, i32 1)
-  store double 0.000000e+00, double* getelementptr inbounds ({ double, double }* @lxyd, i32 0, i32 0)
-  store double 0.000000e+00, double* getelementptr inbounds ({ double, double }* @lxyd, i32 0, i32 1)
+  store float 0.000000e+00, float* getelementptr inbounds ({ float, float }, { float, float }* @lxy, i32 0, i32 0)
+  store float 0.000000e+00, float* getelementptr inbounds ({ float, float }, { float, float }* @lxy, i32 0, i32 1)
+  store double 0.000000e+00, double* getelementptr inbounds ({ double, double }, { double, double }* @lxyd, i32 0, i32 0)
+  store double 0.000000e+00, double* getelementptr inbounds ({ double, double }, { double, double }* @lxyd, i32 0, i32 1)
   store float 0.000000e+00, float* @lret_sf, align 4
   store double 0.000000e+00, double* @lret_df, align 8
-  store float 0.000000e+00, float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 0)
-  store float 0.000000e+00, float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 1)
-  store double 0.000000e+00, double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 0)
-  store double 0.000000e+00, double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 1)
+  store float 0.000000e+00, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 0)
+  store float 0.000000e+00, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 1)
+  store double 0.000000e+00, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 0)
+  store double 0.000000e+00, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 1)
   ret void
 }
 
@@ -67,686 +67,686 @@ entry:
   store i32 0, i32* %retval
   call void @clear()
   store float 1.500000e+00, float* @lx, align 4
-  %0 = load float* @lx, align 4
+  %0 = load float, float* @lx, align 4
   call void @v_sf(float %0)
-  %1 = load float* @x, align 4
+  %1 = load float, float* @x, align 4
   %conv = fpext float %1 to double
-  %2 = load float* @lx, align 4
+  %2 = load float, float* @lx, align 4
   %conv1 = fpext float %2 to double
-  %3 = load float* @x, align 4
-  %4 = load float* @lx, align 4
+  %3 = load float, float* @x, align 4
+  %4 = load float, float* @lx, align 4
   %cmp = fcmp oeq float %3, %4
   %conv2 = zext i1 %cmp to i32
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([10 x i8]* @.str, i32 0, i32 0), double %conv, double %conv1, i32 %conv2)
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([10 x i8], [10 x i8]* @.str, i32 0, i32 0), double %conv, double %conv1, i32 %conv2)
   call void @clear()
   store double 0x41678C29C0000000, double* @lxd, align 8
-  %5 = load double* @lxd, align 8
+  %5 = load double, double* @lxd, align 8
   call void @v_df(double %5)
-  %6 = load double* @xd, align 8
-  %7 = load double* @lxd, align 8
-  %8 = load double* @xd, align 8
-  %9 = load double* @lxd, align 8
+  %6 = load double, double* @xd, align 8
+  %7 = load double, double* @lxd, align 8
+  %8 = load double, double* @xd, align 8
+  %9 = load double, double* @lxd, align 8
   %cmp3 = fcmp oeq double %8, %9
   %conv4 = zext i1 %cmp3 to i32
-  %call5 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([10 x i8]* @.str, i32 0, i32 0), double %6, double %7, i32 %conv4)
+  %call5 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([10 x i8], [10 x i8]* @.str, i32 0, i32 0), double %6, double %7, i32 %conv4)
   call void @clear()
   store float 9.000000e+00, float* @lx, align 4
   store float 1.000000e+01, float* @ly, align 4
-  %10 = load float* @lx, align 4
-  %11 = load float* @ly, align 4
+  %10 = load float, float* @lx, align 4
+  %11 = load float, float* @ly, align 4
   call void @v_sf_sf(float %10, float %11)
-  %12 = load float* @x, align 4
+  %12 = load float, float* @x, align 4
   %conv6 = fpext float %12 to double
-  %13 = load float* @lx, align 4
+  %13 = load float, float* @lx, align 4
   %conv7 = fpext float %13 to double
-  %14 = load float* @y, align 4
+  %14 = load float, float* @y, align 4
   %conv8 = fpext float %14 to double
-  %15 = load float* @ly, align 4
+  %15 = load float, float* @ly, align 4
   %conv9 = fpext float %15 to double
-  %16 = load float* @x, align 4
-  %17 = load float* @lx, align 4
+  %16 = load float, float* @x, align 4
+  %17 = load float, float* @lx, align 4
   %cmp10 = fcmp oeq float %16, %17
   br i1 %cmp10, label %land.rhs, label %land.end
 
 land.rhs:                                         ; preds = %entry
-  %18 = load float* @y, align 4
-  %19 = load float* @ly, align 4
+  %18 = load float, float* @y, align 4
+  %19 = load float, float* @ly, align 4
   %cmp12 = fcmp oeq float %18, %19
   br label %land.end
 
 land.end:                                         ; preds = %land.rhs, %entry
   %20 = phi i1 [ false, %entry ], [ %cmp12, %land.rhs ]
   %land.ext = zext i1 %20 to i32
-  %call14 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([16 x i8]* @.str1, i32 0, i32 0), double %conv6, double %conv7, double %conv8, double %conv9, i32 %land.ext)
+  %call14 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([16 x i8], [16 x i8]* @.str1, i32 0, i32 0), double %conv6, double %conv7, double %conv8, double %conv9, i32 %land.ext)
   call void @clear()
   store float 0x3FFE666660000000, float* @lx, align 4
   store double 0x4007E613249FF279, double* @lyd, align 8
-  %21 = load float* @lx, align 4
-  %22 = load double* @lyd, align 8
+  %21 = load float, float* @lx, align 4
+  %22 = load double, double* @lyd, align 8
   call void @v_sf_df(float %21, double %22)
-  %23 = load float* @x, align 4
+  %23 = load float, float* @x, align 4
   %conv15 = fpext float %23 to double
-  %24 = load float* @lx, align 4
+  %24 = load float, float* @lx, align 4
   %conv16 = fpext float %24 to double
-  %25 = load double* @yd, align 8
-  %26 = load double* @lyd, align 8
-  %27 = load float* @x, align 4
-  %28 = load float* @lx, align 4
+  %25 = load double, double* @yd, align 8
+  %26 = load double, double* @lyd, align 8
+  %27 = load float, float* @x, align 4
+  %28 = load float, float* @lx, align 4
   %cmp17 = fcmp oeq float %27, %28
   %conv18 = zext i1 %cmp17 to i32
-  %29 = load double* @yd, align 8
-  %30 = load double* @lyd, align 8
+  %29 = load double, double* @yd, align 8
+  %30 = load double, double* @lyd, align 8
   %cmp19 = fcmp oeq double %29, %30
   %conv20 = zext i1 %cmp19 to i32
   %and = and i32 %conv18, %conv20
-  %call21 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([16 x i8]* @.str1, i32 0, i32 0), double %conv15, double %conv16, double %25, double %26, i32 %and)
+  %call21 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([16 x i8], [16 x i8]* @.str1, i32 0, i32 0), double %conv15, double %conv16, double %25, double %26, i32 %and)
   call void @clear()
   store double 0x4194E54F94000000, double* @lxd, align 8
   store float 7.600000e+01, float* @ly, align 4
-  %31 = load double* @lxd, align 8
-  %32 = load float* @ly, align 4
+  %31 = load double, double* @lxd, align 8
+  %32 = load float, float* @ly, align 4
   call void @v_df_sf(double %31, float %32)
-  %33 = load double* @xd, align 8
-  %34 = load double* @lxd, align 8
-  %35 = load float* @y, align 4
+  %33 = load double, double* @xd, align 8
+  %34 = load double, double* @lxd, align 8
+  %35 = load float, float* @y, align 4
   %conv22 = fpext float %35 to double
-  %36 = load float* @ly, align 4
+  %36 = load float, float* @ly, align 4
   %conv23 = fpext float %36 to double
-  %37 = load double* @xd, align 8
-  %38 = load double* @lxd, align 8
+  %37 = load double, double* @xd, align 8
+  %38 = load double, double* @lxd, align 8
   %cmp24 = fcmp oeq double %37, %38
   %conv25 = zext i1 %cmp24 to i32
-  %39 = load float* @y, align 4
-  %40 = load float* @ly, align 4
+  %39 = load float, float* @y, align 4
+  %40 = load float, float* @ly, align 4
   %cmp26 = fcmp oeq float %39, %40
   %conv27 = zext i1 %cmp26 to i32
   %and28 = and i32 %conv25, %conv27
-  %call29 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([16 x i8]* @.str1, i32 0, i32 0), double %33, double %34, double %conv22, double %conv23, i32 %and28)
+  %call29 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([16 x i8], [16 x i8]* @.str1, i32 0, i32 0), double %33, double %34, double %conv22, double %conv23, i32 %and28)
   call void @clear()
   store double 7.365198e+07, double* @lxd, align 8
   store double 0x416536CD80000000, double* @lyd, align 8
-  %41 = load double* @lxd, align 8
-  %42 = load double* @lyd, align 8
+  %41 = load double, double* @lxd, align 8
+  %42 = load double, double* @lyd, align 8
   call void @v_df_df(double %41, double %42)
-  %43 = load double* @xd, align 8
-  %44 = load double* @lxd, align 8
-  %45 = load double* @yd, align 8
-  %46 = load double* @lyd, align 8
-  %47 = load double* @xd, align 8
-  %48 = load double* @lxd, align 8
+  %43 = load double, double* @xd, align 8
+  %44 = load double, double* @lxd, align 8
+  %45 = load double, double* @yd, align 8
+  %46 = load double, double* @lyd, align 8
+  %47 = load double, double* @xd, align 8
+  %48 = load double, double* @lxd, align 8
   %cmp30 = fcmp oeq double %47, %48
   %conv31 = zext i1 %cmp30 to i32
-  %49 = load double* @yd, align 8
-  %50 = load double* @lyd, align 8
+  %49 = load double, double* @yd, align 8
+  %50 = load double, double* @lyd, align 8
   %cmp32 = fcmp oeq double %49, %50
   %conv33 = zext i1 %cmp32 to i32
   %and34 = and i32 %conv31, %conv33
-  %call35 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([16 x i8]* @.str1, i32 0, i32 0), double %43, double %44, double %45, double %46, i32 %and34)
+  %call35 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([16 x i8], [16 x i8]* @.str1, i32 0, i32 0), double %43, double %44, double %45, double %46, i32 %and34)
   call void @clear()
   store float 0x4016666660000000, float* @ret_sf, align 4
   %call36 = call float @sf_v()
   store float %call36, float* @lret_sf, align 4
-  %51 = load float* @ret_sf, align 4
+  %51 = load float, float* @ret_sf, align 4
   %conv37 = fpext float %51 to double
-  %52 = load float* @lret_sf, align 4
+  %52 = load float, float* @lret_sf, align 4
   %conv38 = fpext float %52 to double
-  %53 = load float* @ret_sf, align 4
-  %54 = load float* @lret_sf, align 4
+  %53 = load float, float* @ret_sf, align 4
+  %54 = load float, float* @lret_sf, align 4
   %cmp39 = fcmp oeq float %53, %54
   %conv40 = zext i1 %cmp39 to i32
-  %call41 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([10 x i8]* @.str, i32 0, i32 0), double %conv37, double %conv38, i32 %conv40)
+  %call41 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([10 x i8], [10 x i8]* @.str, i32 0, i32 0), double %conv37, double %conv38, i32 %conv40)
   call void @clear()
   store float 4.587300e+06, float* @ret_sf, align 4
   store float 3.420000e+02, float* @lx, align 4
-  %55 = load float* @lx, align 4
+  %55 = load float, float* @lx, align 4
   %call42 = call float @sf_sf(float %55)
   store float %call42, float* @lret_sf, align 4
-  %56 = load float* @ret_sf, align 4
+  %56 = load float, float* @ret_sf, align 4
   %conv43 = fpext float %56 to double
-  %57 = load float* @lret_sf, align 4
+  %57 = load float, float* @lret_sf, align 4
   %conv44 = fpext float %57 to double
-  %58 = load float* @x, align 4
+  %58 = load float, float* @x, align 4
   %conv45 = fpext float %58 to double
-  %59 = load float* @lx, align 4
+  %59 = load float, float* @lx, align 4
   %conv46 = fpext float %59 to double
-  %60 = load float* @ret_sf, align 4
-  %61 = load float* @lret_sf, align 4
+  %60 = load float, float* @ret_sf, align 4
+  %61 = load float, float* @lret_sf, align 4
   %cmp47 = fcmp oeq float %60, %61
   %conv48 = zext i1 %cmp47 to i32
-  %62 = load float* @x, align 4
-  %63 = load float* @lx, align 4
+  %62 = load float, float* @x, align 4
+  %63 = load float, float* @lx, align 4
   %cmp49 = fcmp oeq float %62, %63
   %conv50 = zext i1 %cmp49 to i32
   %and51 = and i32 %conv48, %conv50
-  %call52 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([16 x i8]* @.str1, i32 0, i32 0), double %conv43, double %conv44, double %conv45, double %conv46, i32 %and51)
+  %call52 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([16 x i8], [16 x i8]* @.str1, i32 0, i32 0), double %conv43, double %conv44, double %conv45, double %conv46, i32 %and51)
   call void @clear()
   store float 4.445910e+06, float* @ret_sf, align 4
   store double 0x419A7DB294000000, double* @lxd, align 8
-  %64 = load double* @lxd, align 8
+  %64 = load double, double* @lxd, align 8
   %call53 = call float @sf_df(double %64)
   store float %call53, float* @lret_sf, align 4
-  %65 = load float* @ret_sf, align 4
+  %65 = load float, float* @ret_sf, align 4
   %conv54 = fpext float %65 to double
-  %66 = load float* @lret_sf, align 4
+  %66 = load float, float* @lret_sf, align 4
   %conv55 = fpext float %66 to double
-  %67 = load double* @xd, align 8
-  %68 = load double* @lxd, align 8
-  %69 = load float* @ret_sf, align 4
-  %70 = load float* @lret_sf, align 4
+  %67 = load double, double* @xd, align 8
+  %68 = load double, double* @lxd, align 8
+  %69 = load float, float* @ret_sf, align 4
+  %70 = load float, float* @lret_sf, align 4
   %cmp56 = fcmp oeq float %69, %70
   %conv57 = zext i1 %cmp56 to i32
-  %71 = load double* @xd, align 8
-  %72 = load double* @lxd, align 8
+  %71 = load double, double* @xd, align 8
+  %72 = load double, double* @lxd, align 8
   %cmp58 = fcmp oeq double %71, %72
   %conv59 = zext i1 %cmp58 to i32
   %and60 = and i32 %conv57, %conv59
-  %call61 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([16 x i8]* @.str1, i32 0, i32 0), double %conv54, double %conv55, double %67, double %68, i32 %and60)
+  %call61 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([16 x i8], [16 x i8]* @.str1, i32 0, i32 0), double %conv54, double %conv55, double %67, double %68, i32 %and60)
   call void @clear()
   store float 0x3FFF4BC6A0000000, float* @ret_sf, align 4
   store float 4.445500e+03, float* @lx, align 4
   store float 0x4068ACCCC0000000, float* @ly, align 4
-  %73 = load float* @lx, align 4
-  %74 = load float* @ly, align 4
+  %73 = load float, float* @lx, align 4
+  %74 = load float, float* @ly, align 4
   %call62 = call float @sf_sf_sf(float %73, float %74)
   store float %call62, float* @lret_sf, align 4
-  %75 = load float* @ret_sf, align 4
+  %75 = load float, float* @ret_sf, align 4
   %conv63 = fpext float %75 to double
-  %76 = load float* @lret_sf, align 4
+  %76 = load float, float* @lret_sf, align 4
   %conv64 = fpext float %76 to double
-  %77 = load float* @x, align 4
+  %77 = load float, float* @x, align 4
   %conv65 = fpext float %77 to double
-  %78 = load float* @lx, align 4
+  %78 = load float, float* @lx, align 4
   %conv66 = fpext float %78 to double
-  %79 = load float* @y, align 4
+  %79 = load float, float* @y, align 4
   %conv67 = fpext float %79 to double
-  %80 = load float* @ly, align 4
+  %80 = load float, float* @ly, align 4
   %conv68 = fpext float %80 to double
-  %81 = load float* @ret_sf, align 4
-  %82 = load float* @lret_sf, align 4
+  %81 = load float, float* @ret_sf, align 4
+  %82 = load float, float* @lret_sf, align 4
   %cmp69 = fcmp oeq float %81, %82
   br i1 %cmp69, label %land.lhs.true, label %land.end76
 
 land.lhs.true:                                    ; preds = %land.end
-  %83 = load float* @x, align 4
-  %84 = load float* @lx, align 4
+  %83 = load float, float* @x, align 4
+  %84 = load float, float* @lx, align 4
   %cmp71 = fcmp oeq float %83, %84
   br i1 %cmp71, label %land.rhs73, label %land.end76
 
 land.rhs73:                                       ; preds = %land.lhs.true
-  %85 = load float* @y, align 4
-  %86 = load float* @ly, align 4
+  %85 = load float, float* @y, align 4
+  %86 = load float, float* @ly, align 4
   %cmp74 = fcmp oeq float %85, %86
   br label %land.end76
 
 land.end76:                                       ; preds = %land.rhs73, %land.lhs.true, %land.end
   %87 = phi i1 [ false, %land.lhs.true ], [ false, %land.end ], [ %cmp74, %land.rhs73 ]
   %land.ext77 = zext i1 %87 to i32
-  %call78 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([22 x i8]* @.str2, i32 0, i32 0), double %conv63, double %conv64, double %conv65, double %conv66, double %conv67, double %conv68, i32 %land.ext77)
+  %call78 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([22 x i8], [22 x i8]* @.str2, i32 0, i32 0), double %conv63, double %conv64, double %conv65, double %conv66, double %conv67, double %conv68, i32 %land.ext77)
   call void @clear()
   store float 9.991300e+04, float* @ret_sf, align 4
   store float 1.114500e+04, float* @lx, align 4
   store double 9.994445e+07, double* @lyd, align 8
-  %88 = load float* @lx, align 4
-  %89 = load double* @lyd, align 8
+  %88 = load float, float* @lx, align 4
+  %89 = load double, double* @lyd, align 8
   %call79 = call float @sf_sf_df(float %88, double %89)
   store float %call79, float* @lret_sf, align 4
-  %90 = load float* @ret_sf, align 4
+  %90 = load float, float* @ret_sf, align 4
   %conv80 = fpext float %90 to double
-  %91 = load float* @lret_sf, align 4
+  %91 = load float, float* @lret_sf, align 4
   %conv81 = fpext float %91 to double
-  %92 = load float* @x, align 4
+  %92 = load float, float* @x, align 4
   %conv82 = fpext float %92 to double
-  %93 = load float* @lx, align 4
+  %93 = load float, float* @lx, align 4
   %conv83 = fpext float %93 to double
-  %94 = load double* @yd, align 8
-  %95 = load double* @lyd, align 8
-  %96 = load float* @ret_sf, align 4
-  %97 = load float* @lret_sf, align 4
+  %94 = load double, double* @yd, align 8
+  %95 = load double, double* @lyd, align 8
+  %96 = load float, float* @ret_sf, align 4
+  %97 = load float, float* @lret_sf, align 4
   %cmp84 = fcmp oeq float %96, %97
   br i1 %cmp84, label %land.lhs.true86, label %land.end92
 
 land.lhs.true86:                                  ; preds = %land.end76
-  %98 = load float* @x, align 4
-  %99 = load float* @lx, align 4
+  %98 = load float, float* @x, align 4
+  %99 = load float, float* @lx, align 4
   %cmp87 = fcmp oeq float %98, %99
   br i1 %cmp87, label %land.rhs89, label %land.end92
 
 land.rhs89:                                       ; preds = %land.lhs.true86
-  %100 = load double* @yd, align 8
-  %101 = load double* @lyd, align 8
+  %100 = load double, double* @yd, align 8
+  %101 = load double, double* @lyd, align 8
   %cmp90 = fcmp oeq double %100, %101
   br label %land.end92
 
 land.end92:                                       ; preds = %land.rhs89, %land.lhs.true86, %land.end76
   %102 = phi i1 [ false, %land.lhs.true86 ], [ false, %land.end76 ], [ %cmp90, %land.rhs89 ]
   %land.ext93 = zext i1 %102 to i32
-  %call94 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([22 x i8]* @.str2, i32 0, i32 0), double %conv80, double %conv81, double %conv82, double %conv83, double %94, double %95, i32 %land.ext93)
+  %call94 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([22 x i8], [22 x i8]* @.str2, i32 0, i32 0), double %conv80, double %conv81, double %conv82, double %conv83, double %94, double %95, i32 %land.ext93)
   call void @clear()
   store float 0x417CCC7A00000000, float* @ret_sf, align 4
   store double 0x4172034530000000, double* @lxd, align 8
   store float 4.456200e+04, float* @ly, align 4
-  %103 = load double* @lxd, align 8
-  %104 = load float* @ly, align 4
+  %103 = load double, double* @lxd, align 8
+  %104 = load float, float* @ly, align 4
   %call95 = call float @sf_df_sf(double %103, float %104)
   store float %call95, float* @lret_sf, align 4
-  %105 = load float* @ret_sf, align 4
+  %105 = load float, float* @ret_sf, align 4
   %conv96 = fpext float %105 to double
-  %106 = load float* @lret_sf, align 4
+  %106 = load float, float* @lret_sf, align 4
   %conv97 = fpext float %106 to double
-  %107 = load double* @xd, align 8
-  %108 = load double* @lxd, align 8
-  %109 = load float* @y, align 4
+  %107 = load double, double* @xd, align 8
+  %108 = load double, double* @lxd, align 8
+  %109 = load float, float* @y, align 4
   %conv98 = fpext float %109 to double
-  %110 = load float* @ly, align 4
+  %110 = load float, float* @ly, align 4
   %conv99 = fpext float %110 to double
-  %111 = load float* @ret_sf, align 4
-  %112 = load float* @lret_sf, align 4
+  %111 = load float, float* @ret_sf, align 4
+  %112 = load float, float* @lret_sf, align 4
   %cmp100 = fcmp oeq float %111, %112
   br i1 %cmp100, label %land.lhs.true102, label %land.end108
 
 land.lhs.true102:                                 ; preds = %land.end92
-  %113 = load double* @xd, align 8
-  %114 = load double* @lxd, align 8
+  %113 = load double, double* @xd, align 8
+  %114 = load double, double* @lxd, align 8
   %cmp103 = fcmp oeq double %113, %114
   br i1 %cmp103, label %land.rhs105, label %land.end108
 
 land.rhs105:                                      ; preds = %land.lhs.true102
-  %115 = load float* @y, align 4
-  %116 = load float* @ly, align 4
+  %115 = load float, float* @y, align 4
+  %116 = load float, float* @ly, align 4
   %cmp106 = fcmp oeq float %115, %116
   br label %land.end108
 
 land.end108:                                      ; preds = %land.rhs105, %land.lhs.true102, %land.end92
   %117 = phi i1 [ false, %land.lhs.true102 ], [ false, %land.end92 ], [ %cmp106, %land.rhs105 ]
   %land.ext109 = zext i1 %117 to i32
-  %call110 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([22 x i8]* @.str2, i32 0, i32 0), double %conv96, double %conv97, double %107, double %108, double %conv98, double %conv99, i32 %land.ext109)
+  %call110 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([22 x i8], [22 x i8]* @.str2, i32 0, i32 0), double %conv96, double %conv97, double %107, double %108, double %conv98, double %conv99, i32 %land.ext109)
   call void @clear()
   store float 3.987721e+06, float* @ret_sf, align 4
   store double 0x3FF1F49F6DDDC2D8, double* @lxd, align 8
   store double 0x409129F306A2B170, double* @lyd, align 8
-  %118 = load double* @lxd, align 8
-  %119 = load double* @lyd, align 8
+  %118 = load double, double* @lxd, align 8
+  %119 = load double, double* @lyd, align 8
   %call111 = call float @sf_df_df(double %118, double %119)
   store float %call111, float* @lret_sf, align 4
-  %120 = load float* @ret_sf, align 4
+  %120 = load float, float* @ret_sf, align 4
   %conv112 = fpext float %120 to double
-  %121 = load float* @lret_sf, align 4
+  %121 = load float, float* @lret_sf, align 4
   %conv113 = fpext float %121 to double
-  %122 = load double* @xd, align 8
-  %123 = load double* @lxd, align 8
-  %124 = load double* @yd, align 8
-  %125 = load double* @lyd, align 8
-  %126 = load float* @ret_sf, align 4
-  %127 = load float* @lret_sf, align 4
+  %122 = load double, double* @xd, align 8
+  %123 = load double, double* @lxd, align 8
+  %124 = load double, double* @yd, align 8
+  %125 = load double, double* @lyd, align 8
+  %126 = load float, float* @ret_sf, align 4
+  %127 = load float, float* @lret_sf, align 4
   %cmp114 = fcmp oeq float %126, %127
   br i1 %cmp114, label %land.lhs.true116, label %land.end122
 
 land.lhs.true116:                                 ; preds = %land.end108
-  %128 = load double* @xd, align 8
-  %129 = load double* @lxd, align 8
+  %128 = load double, double* @xd, align 8
+  %129 = load double, double* @lxd, align 8
   %cmp117 = fcmp oeq double %128, %129
   br i1 %cmp117, label %land.rhs119, label %land.end122
 
 land.rhs119:                                      ; preds = %land.lhs.true116
-  %130 = load double* @yd, align 8
-  %131 = load double* @lyd, align 8
+  %130 = load double, double* @yd, align 8
+  %131 = load double, double* @lyd, align 8
   %cmp120 = fcmp oeq double %130, %131
   br label %land.end122
 
 land.end122:                                      ; preds = %land.rhs119, %land.lhs.true116, %land.end108
   %132 = phi i1 [ false, %land.lhs.true116 ], [ false, %land.end108 ], [ %cmp120, %land.rhs119 ]
   %land.ext123 = zext i1 %132 to i32
-  %call124 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([22 x i8]* @.str2, i32 0, i32 0), double %conv112, double %conv113, double %122, double %123, double %124, double %125, i32 %land.ext123)
+  %call124 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([22 x i8], [22 x i8]* @.str2, i32 0, i32 0), double %conv112, double %conv113, double %122, double %123, double %124, double %125, i32 %land.ext123)
   call void @clear()
   store double 1.561234e+01, double* @ret_df, align 8
   %call125 = call double @df_v()
   store double %call125, double* @lret_df, align 8
-  %133 = load double* @ret_df, align 8
-  %134 = load double* @lret_df, align 8
-  %135 = load double* @ret_df, align 8
-  %136 = load double* @lret_df, align 8
+  %133 = load double, double* @ret_df, align 8
+  %134 = load double, double* @lret_df, align 8
+  %135 = load double, double* @ret_df, align 8
+  %136 = load double, double* @lret_df, align 8
   %cmp126 = fcmp oeq double %135, %136
   %conv127 = zext i1 %cmp126 to i32
-  %call128 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([10 x i8]* @.str, i32 0, i32 0), double %133, double %134, i32 %conv127)
+  %call128 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([10 x i8], [10 x i8]* @.str, i32 0, i32 0), double %133, double %134, i32 %conv127)
   call void @clear()
   store double 1.345873e+01, double* @ret_df, align 8
   store float 3.434520e+05, float* @lx, align 4
-  %137 = load float* @lx, align 4
+  %137 = load float, float* @lx, align 4
   %call129 = call double @df_sf(float %137)
   store double %call129, double* @lret_df, align 8
-  %138 = load double* @ret_df, align 8
-  %139 = load double* @lret_df, align 8
-  %140 = load float* @x, align 4
+  %138 = load double, double* @ret_df, align 8
+  %139 = load double, double* @lret_df, align 8
+  %140 = load float, float* @x, align 4
   %conv130 = fpext float %140 to double
-  %141 = load float* @lx, align 4
+  %141 = load float, float* @lx, align 4
   %conv131 = fpext float %141 to double
-  %142 = load double* @ret_df, align 8
-  %143 = load double* @lret_df, align 8
+  %142 = load double, double* @ret_df, align 8
+  %143 = load double, double* @lret_df, align 8
   %cmp132 = fcmp oeq double %142, %143
   %conv133 = zext i1 %cmp132 to i32
-  %144 = load float* @x, align 4
-  %145 = load float* @lx, align 4
+  %144 = load float, float* @x, align 4
+  %145 = load float, float* @lx, align 4
   %cmp134 = fcmp oeq float %144, %145
   %conv135 = zext i1 %cmp134 to i32
   %and136 = and i32 %conv133, %conv135
-  %call137 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([16 x i8]* @.str1, i32 0, i32 0), double %138, double %139, double %conv130, double %conv131, i32 %and136)
+  %call137 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([16 x i8], [16 x i8]* @.str1, i32 0, i32 0), double %138, double %139, double %conv130, double %conv131, i32 %and136)
   call void @clear()
   store double 0x4084F3AB7AA25D8D, double* @ret_df, align 8
   store double 0x4114F671D2F1A9FC, double* @lxd, align 8
-  %146 = load double* @lxd, align 8
+  %146 = load double, double* @lxd, align 8
   %call138 = call double @df_df(double %146)
   store double %call138, double* @lret_df, align 8
-  %147 = load double* @ret_df, align 8
-  %148 = load double* @lret_df, align 8
-  %149 = load double* @xd, align 8
-  %150 = load double* @lxd, align 8
-  %151 = load double* @ret_df, align 8
-  %152 = load double* @lret_df, align 8
+  %147 = load double, double* @ret_df, align 8
+  %148 = load double, double* @lret_df, align 8
+  %149 = load double, double* @xd, align 8
+  %150 = load double, double* @lxd, align 8
+  %151 = load double, double* @ret_df, align 8
+  %152 = load double, double* @lret_df, align 8
   %cmp139 = fcmp oeq double %151, %152
   %conv140 = zext i1 %cmp139 to i32
-  %153 = load double* @xd, align 8
-  %154 = load double* @lxd, align 8
+  %153 = load double, double* @xd, align 8
+  %154 = load double, double* @lxd, align 8
   %cmp141 = fcmp oeq double %153, %154
   %conv142 = zext i1 %cmp141 to i32
   %and143 = and i32 %conv140, %conv142
-  %call144 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([16 x i8]* @.str1, i32 0, i32 0), double %147, double %148, double %149, double %150, i32 %and143)
+  %call144 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([16 x i8], [16 x i8]* @.str1, i32 0, i32 0), double %147, double %148, double %149, double %150, i32 %and143)
   call void @clear()
   store double 6.781956e+03, double* @ret_df, align 8
   store float 4.445500e+03, float* @lx, align 4
   store float 0x4068ACCCC0000000, float* @ly, align 4
-  %155 = load float* @lx, align 4
-  %156 = load float* @ly, align 4
+  %155 = load float, float* @lx, align 4
+  %156 = load float, float* @ly, align 4
   %call145 = call double @df_sf_sf(float %155, float %156)
   store double %call145, double* @lret_df, align 8
-  %157 = load double* @ret_df, align 8
-  %158 = load double* @lret_df, align 8
-  %159 = load float* @x, align 4
+  %157 = load double, double* @ret_df, align 8
+  %158 = load double, double* @lret_df, align 8
+  %159 = load float, float* @x, align 4
   %conv146 = fpext float %159 to double
-  %160 = load float* @lx, align 4
+  %160 = load float, float* @lx, align 4
   %conv147 = fpext float %160 to double
-  %161 = load float* @y, align 4
+  %161 = load float, float* @y, align 4
   %conv148 = fpext float %161 to double
-  %162 = load float* @ly, align 4
+  %162 = load float, float* @ly, align 4
   %conv149 = fpext float %162 to double
-  %163 = load double* @ret_df, align 8
-  %164 = load double* @lret_df, align 8
+  %163 = load double, double* @ret_df, align 8
+  %164 = load double, double* @lret_df, align 8
   %cmp150 = fcmp oeq double %163, %164
   br i1 %cmp150, label %land.lhs.true152, label %land.end158
 
 land.lhs.true152:                                 ; preds = %land.end122
-  %165 = load float* @x, align 4
-  %166 = load float* @lx, align 4
+  %165 = load float, float* @x, align 4
+  %166 = load float, float* @lx, align 4
   %cmp153 = fcmp oeq float %165, %166
   br i1 %cmp153, label %land.rhs155, label %land.end158
 
 land.rhs155:                                      ; preds = %land.lhs.true152
-  %167 = load float* @y, align 4
-  %168 = load float* @ly, align 4
+  %167 = load float, float* @y, align 4
+  %168 = load float, float* @ly, align 4
   %cmp156 = fcmp oeq float %167, %168
   br label %land.end158
 
 land.end158:                                      ; preds = %land.rhs155, %land.lhs.true152, %land.end122
   %169 = phi i1 [ false, %land.lhs.true152 ], [ false, %land.end122 ], [ %cmp156, %land.rhs155 ]
   %land.ext159 = zext i1 %169 to i32
-  %call160 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([22 x i8]* @.str2, i32 0, i32 0), double %157, double %158, double %conv146, double %conv147, double %conv148, double %conv149, i32 %land.ext159)
+  %call160 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([22 x i8], [22 x i8]* @.str2, i32 0, i32 0), double %157, double %158, double %conv146, double %conv147, double %conv148, double %conv149, i32 %land.ext159)
   call void @clear()
   store double 1.889130e+05, double* @ret_df, align 8
   store float 9.111450e+05, float* @lx, align 4
   store double 0x4185320A58000000, double* @lyd, align 8
-  %170 = load float* @lx, align 4
-  %171 = load double* @lyd, align 8
+  %170 = load float, float* @lx, align 4
+  %171 = load double, double* @lyd, align 8
   %call161 = call double @df_sf_df(float %170, double %171)
   store double %call161, double* @lret_df, align 8
-  %172 = load double* @ret_df, align 8
-  %173 = load double* @lret_df, align 8
-  %174 = load float* @x, align 4
+  %172 = load double, double* @ret_df, align 8
+  %173 = load double, double* @lret_df, align 8
+  %174 = load float, float* @x, align 4
   %conv162 = fpext float %174 to double
-  %175 = load float* @lx, align 4
+  %175 = load float, float* @lx, align 4
   %conv163 = fpext float %175 to double
-  %176 = load double* @yd, align 8
-  %177 = load double* @lyd, align 8
-  %178 = load double* @ret_df, align 8
-  %179 = load double* @lret_df, align 8
+  %176 = load double, double* @yd, align 8
+  %177 = load double, double* @lyd, align 8
+  %178 = load double, double* @ret_df, align 8
+  %179 = load double, double* @lret_df, align 8
   %cmp164 = fcmp oeq double %178, %179
   br i1 %cmp164, label %land.lhs.true166, label %land.end172
 
 land.lhs.true166:                                 ; preds = %land.end158
-  %180 = load float* @x, align 4
-  %181 = load float* @lx, align 4
+  %180 = load float, float* @x, align 4
+  %181 = load float, float* @lx, align 4
   %cmp167 = fcmp oeq float %180, %181
   br i1 %cmp167, label %land.rhs169, label %land.end172
 
 land.rhs169:                                      ; preds = %land.lhs.true166
-  %182 = load double* @yd, align 8
-  %183 = load double* @lyd, align 8
+  %182 = load double, double* @yd, align 8
+  %183 = load double, double* @lyd, align 8
   %cmp170 = fcmp oeq double %182, %183
   br label %land.end172
 
 land.end172:                                      ; preds = %land.rhs169, %land.lhs.true166, %land.end158
   %184 = phi i1 [ false, %land.lhs.true166 ], [ false, %land.end158 ], [ %cmp170, %land.rhs169 ]
   %land.ext173 = zext i1 %184 to i32
-  %call174 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([22 x i8]* @.str2, i32 0, i32 0), double %172, double %173, double %conv162, double %conv163, double %176, double %177, i32 %land.ext173)
+  %call174 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([22 x i8], [22 x i8]* @.str2, i32 0, i32 0), double %172, double %173, double %conv162, double %conv163, double %176, double %177, i32 %land.ext173)
   call void @clear()
   store double 0x418B2DB900000000, double* @ret_df, align 8
   store double 0x41B1EF2ED3000000, double* @lxd, align 8
   store float 1.244562e+06, float* @ly, align 4
-  %185 = load double* @lxd, align 8
-  %186 = load float* @ly, align 4
+  %185 = load double, double* @lxd, align 8
+  %186 = load float, float* @ly, align 4
   %call175 = call double @df_df_sf(double %185, float %186)
   store double %call175, double* @lret_df, align 8
-  %187 = load double* @ret_df, align 8
-  %188 = load double* @lret_df, align 8
-  %189 = load double* @xd, align 8
-  %190 = load double* @lxd, align 8
-  %191 = load float* @y, align 4
+  %187 = load double, double* @ret_df, align 8
+  %188 = load double, double* @lret_df, align 8
+  %189 = load double, double* @xd, align 8
+  %190 = load double, double* @lxd, align 8
+  %191 = load float, float* @y, align 4
   %conv176 = fpext float %191 to double
-  %192 = load float* @ly, align 4
+  %192 = load float, float* @ly, align 4
   %conv177 = fpext float %192 to double
-  %193 = load double* @ret_df, align 8
-  %194 = load double* @lret_df, align 8
+  %193 = load double, double* @ret_df, align 8
+  %194 = load double, double* @lret_df, align 8
   %cmp178 = fcmp oeq double %193, %194
   br i1 %cmp178, label %land.lhs.true180, label %land.end186
 
 land.lhs.true180:                                 ; preds = %land.end172
-  %195 = load double* @xd, align 8
-  %196 = load double* @lxd, align 8
+  %195 = load double, double* @xd, align 8
+  %196 = load double, double* @lxd, align 8
   %cmp181 = fcmp oeq double %195, %196
   br i1 %cmp181, label %land.rhs183, label %land.end186
 
 land.rhs183:                                      ; preds = %land.lhs.true180
-  %197 = load float* @y, align 4
-  %198 = load float* @ly, align 4
+  %197 = load float, float* @y, align 4
+  %198 = load float, float* @ly, align 4
   %cmp184 = fcmp oeq float %197, %198
   br label %land.end186
 
 land.end186:                                      ; preds = %land.rhs183, %land.lhs.true180, %land.end172
   %199 = phi i1 [ false, %land.lhs.true180 ], [ false, %land.end172 ], [ %cmp184, %land.rhs183 ]
   %land.ext187 = zext i1 %199 to i32
-  %call188 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([22 x i8]* @.str2, i32 0, i32 0), double %187, double %188, double %189, double %190, double %conv176, double %conv177, i32 %land.ext187)
+  %call188 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([22 x i8], [22 x i8]* @.str2, i32 0, i32 0), double %187, double %188, double %189, double %190, double %conv176, double %conv177, i32 %land.ext187)
   call void @clear()
   store double 3.987721e+06, double* @ret_df, align 8
   store double 5.223560e+00, double* @lxd, align 8
   store double 0x40B7D37CC1A8AC5C, double* @lyd, align 8
-  %200 = load double* @lxd, align 8
-  %201 = load double* @lyd, align 8
+  %200 = load double, double* @lxd, align 8
+  %201 = load double, double* @lyd, align 8
   %call189 = call double @df_df_df(double %200, double %201)
   store double %call189, double* @lret_df, align 8
-  %202 = load double* @ret_df, align 8
-  %203 = load double* @lret_df, align 8
-  %204 = load double* @xd, align 8
-  %205 = load double* @lxd, align 8
-  %206 = load double* @yd, align 8
-  %207 = load double* @lyd, align 8
-  %208 = load double* @ret_df, align 8
-  %209 = load double* @lret_df, align 8
+  %202 = load double, double* @ret_df, align 8
+  %203 = load double, double* @lret_df, align 8
+  %204 = load double, double* @xd, align 8
+  %205 = load double, double* @lxd, align 8
+  %206 = load double, double* @yd, align 8
+  %207 = load double, double* @lyd, align 8
+  %208 = load double, double* @ret_df, align 8
+  %209 = load double, double* @lret_df, align 8
   %cmp190 = fcmp oeq double %208, %209
   br i1 %cmp190, label %land.lhs.true192, label %land.end198
 
 land.lhs.true192:                                 ; preds = %land.end186
-  %210 = load double* @xd, align 8
-  %211 = load double* @lxd, align 8
+  %210 = load double, double* @xd, align 8
+  %211 = load double, double* @lxd, align 8
   %cmp193 = fcmp oeq double %210, %211
   br i1 %cmp193, label %land.rhs195, label %land.end198
 
 land.rhs195:                                      ; preds = %land.lhs.true192
-  %212 = load double* @yd, align 8
-  %213 = load double* @lyd, align 8
+  %212 = load double, double* @yd, align 8
+  %213 = load double, double* @lyd, align 8
   %cmp196 = fcmp oeq double %212, %213
   br label %land.end198
 
 land.end198:                                      ; preds = %land.rhs195, %land.lhs.true192, %land.end186
   %214 = phi i1 [ false, %land.lhs.true192 ], [ false, %land.end186 ], [ %cmp196, %land.rhs195 ]
   %land.ext199 = zext i1 %214 to i32
-  %call200 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([22 x i8]* @.str2, i32 0, i32 0), double %202, double %203, double %204, double %205, double %206, double %207, i32 %land.ext199)
+  %call200 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([22 x i8], [22 x i8]* @.str2, i32 0, i32 0), double %202, double %203, double %204, double %205, double %206, double %207, i32 %land.ext199)
   call void @clear()
-  store float 4.500000e+00, float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 0)
-  store float 7.000000e+00, float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 1)
+  store float 4.500000e+00, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 0)
+  store float 7.000000e+00, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 1)
   %call201 = call { float, float } @sc_v()
   %215 = extractvalue { float, float } %call201, 0
   %216 = extractvalue { float, float } %call201, 1
-  store float %215, float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 0)
-  store float %216, float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 1)
-  %ret_sc.real = load float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 0)
-  %ret_sc.imag = load float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 1)
+  store float %215, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 0)
+  store float %216, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 1)
+  %ret_sc.real = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 0)
+  %ret_sc.imag = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 1)
   %conv202 = fpext float %ret_sc.real to double
   %conv203 = fpext float %ret_sc.imag to double
-  %ret_sc.real204 = load float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 0)
-  %ret_sc.imag205 = load float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 1)
+  %ret_sc.real204 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 0)
+  %ret_sc.imag205 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 1)
   %conv206 = fpext float %ret_sc.real204 to double
   %conv207 = fpext float %ret_sc.imag205 to double
-  %lret_sc.real = load float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 0)
-  %lret_sc.imag = load float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 1)
+  %lret_sc.real = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 0)
+  %lret_sc.imag = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 1)
   %conv208 = fpext float %lret_sc.real to double
   %conv209 = fpext float %lret_sc.imag to double
-  %lret_sc.real210 = load float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 0)
-  %lret_sc.imag211 = load float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 1)
+  %lret_sc.real210 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 0)
+  %lret_sc.imag211 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 1)
   %conv212 = fpext float %lret_sc.real210 to double
   %conv213 = fpext float %lret_sc.imag211 to double
-  %ret_sc.real214 = load float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 0)
-  %ret_sc.imag215 = load float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 1)
-  %lret_sc.real216 = load float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 0)
-  %lret_sc.imag217 = load float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 1)
+  %ret_sc.real214 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 0)
+  %ret_sc.imag215 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 1)
+  %lret_sc.real216 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 0)
+  %lret_sc.imag217 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 1)
   %cmp.r = fcmp oeq float %ret_sc.real214, %lret_sc.real216
   %cmp.i = fcmp oeq float %ret_sc.imag215, %lret_sc.imag217
   %and.ri = and i1 %cmp.r, %cmp.i
   %conv218 = zext i1 %and.ri to i32
-  %call219 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([18 x i8]* @.str3, i32 0, i32 0), double %conv202, double %conv207, double %conv208, double %conv213, i32 %conv218)
+  %call219 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([18 x i8], [18 x i8]* @.str3, i32 0, i32 0), double %conv202, double %conv207, double %conv208, double %conv213, i32 %conv218)
   call void @clear()
   store float 0x3FF7A99300000000, float* @lx, align 4
-  store float 4.500000e+00, float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 0)
-  store float 7.000000e+00, float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 1)
-  %217 = load float* @lx, align 4
+  store float 4.500000e+00, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 0)
+  store float 7.000000e+00, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 1)
+  %217 = load float, float* @lx, align 4
   %call220 = call { float, float } @sc_sf(float %217)
   %218 = extractvalue { float, float } %call220, 0
   %219 = extractvalue { float, float } %call220, 1
-  store float %218, float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 0)
-  store float %219, float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 1)
-  %ret_sc.real221 = load float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 0)
-  %ret_sc.imag222 = load float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 1)
+  store float %218, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 0)
+  store float %219, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 1)
+  %ret_sc.real221 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 0)
+  %ret_sc.imag222 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 1)
   %conv223 = fpext float %ret_sc.real221 to double
   %conv224 = fpext float %ret_sc.imag222 to double
-  %ret_sc.real225 = load float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 0)
-  %ret_sc.imag226 = load float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 1)
+  %ret_sc.real225 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 0)
+  %ret_sc.imag226 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 1)
   %conv227 = fpext float %ret_sc.real225 to double
   %conv228 = fpext float %ret_sc.imag226 to double
-  %lret_sc.real229 = load float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 0)
-  %lret_sc.imag230 = load float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 1)
+  %lret_sc.real229 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 0)
+  %lret_sc.imag230 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 1)
   %conv231 = fpext float %lret_sc.real229 to double
   %conv232 = fpext float %lret_sc.imag230 to double
-  %lret_sc.real233 = load float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 0)
-  %lret_sc.imag234 = load float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 1)
+  %lret_sc.real233 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 0)
+  %lret_sc.imag234 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 1)
   %conv235 = fpext float %lret_sc.real233 to double
   %conv236 = fpext float %lret_sc.imag234 to double
-  %220 = load float* @x, align 4
+  %220 = load float, float* @x, align 4
   %conv237 = fpext float %220 to double
-  %221 = load float* @lx, align 4
+  %221 = load float, float* @lx, align 4
   %conv238 = fpext float %221 to double
-  %ret_sc.real239 = load float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 0)
-  %ret_sc.imag240 = load float* getelementptr inbounds ({ float, float }* @ret_sc, i32 0, i32 1)
-  %lret_sc.real241 = load float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 0)
-  %lret_sc.imag242 = load float* getelementptr inbounds ({ float, float }* @lret_sc, i32 0, i32 1)
+  %ret_sc.real239 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 0)
+  %ret_sc.imag240 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @ret_sc, i32 0, i32 1)
+  %lret_sc.real241 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 0)
+  %lret_sc.imag242 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @lret_sc, i32 0, i32 1)
   %cmp.r243 = fcmp oeq float %ret_sc.real239, %lret_sc.real241
   %cmp.i244 = fcmp oeq float %ret_sc.imag240, %lret_sc.imag242
   %and.ri245 = and i1 %cmp.r243, %cmp.i244
   br i1 %and.ri245, label %land.rhs247, label %land.end250
 
 land.rhs247:                                      ; preds = %land.end198
-  %222 = load float* @x, align 4
-  %223 = load float* @lx, align 4
+  %222 = load float, float* @x, align 4
+  %223 = load float, float* @lx, align 4
   %cmp248 = fcmp oeq float %222, %223
   br label %land.end250
 
 land.end250:                                      ; preds = %land.rhs247, %land.end198
   %224 = phi i1 [ false, %land.end198 ], [ %cmp248, %land.rhs247 ]
   %land.ext251 = zext i1 %224 to i32
-  %call252 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([24 x i8]* @.str4, i32 0, i32 0), double %conv223, double %conv228, double %conv231, double %conv236, double %conv237, double %conv238, i32 %land.ext251)
+  %call252 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([24 x i8], [24 x i8]* @.str4, i32 0, i32 0), double %conv223, double %conv228, double %conv231, double %conv236, double %conv237, double %conv238, i32 %land.ext251)
   call void @clear()
-  store double 1.234500e+03, double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 0)
-  store double 7.677000e+03, double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 1)
+  store double 1.234500e+03, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 0)
+  store double 7.677000e+03, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 1)
   %call253 = call { double, double } @dc_v()
   %225 = extractvalue { double, double } %call253, 0
   %226 = extractvalue { double, double } %call253, 1
-  store double %225, double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 0)
-  store double %226, double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 1)
-  %ret_dc.real = load double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 0)
-  %ret_dc.imag = load double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 1)
-  %ret_dc.real254 = load double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 0)
-  %ret_dc.imag255 = load double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 1)
-  %lret_dc.real = load double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 0)
-  %lret_dc.imag = load double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 1)
-  %lret_dc.real256 = load double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 0)
-  %lret_dc.imag257 = load double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 1)
-  %ret_dc.real258 = load double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 0)
-  %ret_dc.imag259 = load double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 1)
-  %lret_dc.real260 = load double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 0)
-  %lret_dc.imag261 = load double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 1)
+  store double %225, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 0)
+  store double %226, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 1)
+  %ret_dc.real = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 0)
+  %ret_dc.imag = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 1)
+  %ret_dc.real254 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 0)
+  %ret_dc.imag255 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 1)
+  %lret_dc.real = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 0)
+  %lret_dc.imag = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 1)
+  %lret_dc.real256 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 0)
+  %lret_dc.imag257 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 1)
+  %ret_dc.real258 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 0)
+  %ret_dc.imag259 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 1)
+  %lret_dc.real260 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 0)
+  %lret_dc.imag261 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 1)
   %cmp.r262 = fcmp oeq double %ret_dc.real258, %lret_dc.real260
   %cmp.i263 = fcmp oeq double %ret_dc.imag259, %lret_dc.imag261
   %and.ri264 = and i1 %cmp.r262, %cmp.i263
   %conv265 = zext i1 %and.ri264 to i32
-  %call266 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([18 x i8]* @.str3, i32 0, i32 0), double %ret_dc.real, double %ret_dc.imag255, double %lret_dc.real, double %lret_dc.imag257, i32 %conv265)
+  %call266 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([18 x i8], [18 x i8]* @.str3, i32 0, i32 0), double %ret_dc.real, double %ret_dc.imag255, double %lret_dc.real, double %lret_dc.imag257, i32 %conv265)
   call void @clear()
   store double 0x40AAF6F532617C1C, double* @lxd, align 8
-  store double 4.444500e+03, double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 0)
-  store double 7.888000e+03, double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 1)
-  %227 = load float* @lx, align 4
+  store double 4.444500e+03, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 0)
+  store double 7.888000e+03, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 1)
+  %227 = load float, float* @lx, align 4
   %call267 = call { double, double } @dc_sf(float %227)
   %228 = extractvalue { double, double } %call267, 0
   %229 = extractvalue { double, double } %call267, 1
-  store double %228, double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 0)
-  store double %229, double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 1)
-  %ret_dc.real268 = load double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 0)
-  %ret_dc.imag269 = load double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 1)
-  %ret_dc.real270 = load double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 0)
-  %ret_dc.imag271 = load double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 1)
-  %lret_dc.real272 = load double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 0)
-  %lret_dc.imag273 = load double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 1)
-  %lret_dc.real274 = load double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 0)
-  %lret_dc.imag275 = load double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 1)
-  %230 = load float* @x, align 4
+  store double %228, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 0)
+  store double %229, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 1)
+  %ret_dc.real268 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 0)
+  %ret_dc.imag269 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 1)
+  %ret_dc.real270 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 0)
+  %ret_dc.imag271 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 1)
+  %lret_dc.real272 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 0)
+  %lret_dc.imag273 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 1)
+  %lret_dc.real274 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 0)
+  %lret_dc.imag275 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 1)
+  %230 = load float, float* @x, align 4
   %conv276 = fpext float %230 to double
-  %231 = load float* @lx, align 4
+  %231 = load float, float* @lx, align 4
   %conv277 = fpext float %231 to double
-  %ret_dc.real278 = load double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 0)
-  %ret_dc.imag279 = load double* getelementptr inbounds ({ double, double }* @ret_dc, i32 0, i32 1)
-  %lret_dc.real280 = load double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 0)
-  %lret_dc.imag281 = load double* getelementptr inbounds ({ double, double }* @lret_dc, i32 0, i32 1)
+  %ret_dc.real278 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 0)
+  %ret_dc.imag279 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @ret_dc, i32 0, i32 1)
+  %lret_dc.real280 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 0)
+  %lret_dc.imag281 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @lret_dc, i32 0, i32 1)
   %cmp.r282 = fcmp oeq double %ret_dc.real278, %lret_dc.real280
   %cmp.i283 = fcmp oeq double %ret_dc.imag279, %lret_dc.imag281
   %and.ri284 = and i1 %cmp.r282, %cmp.i283
   br i1 %and.ri284, label %land.rhs286, label %land.end289
 
 land.rhs286:                                      ; preds = %land.end250
-  %232 = load float* @x, align 4
-  %233 = load float* @lx, align 4
+  %232 = load float, float* @x, align 4
+  %233 = load float, float* @lx, align 4
   %cmp287 = fcmp oeq float %232, %233
   br label %land.end289
 
 land.end289:                                      ; preds = %land.rhs286, %land.end250
   %234 = phi i1 [ false, %land.end250 ], [ %cmp287, %land.rhs286 ]
   %land.ext290 = zext i1 %234 to i32
-  %call291 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([24 x i8]* @.str4, i32 0, i32 0), double %ret_dc.real268, double %ret_dc.imag271, double %lret_dc.real272, double %lret_dc.imag275, double %conv276, double %conv277, i32 %land.ext290)
-  %235 = load i32* %retval
+  %call291 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([24 x i8], [24 x i8]* @.str4, i32 0, i32 0), double %ret_dc.real268, double %ret_dc.imag271, double %lret_dc.real272, double %lret_dc.imag275, double %conv276, double %conv277, i32 %land.ext290)
+  %235 = load i32, i32* %retval
   ret i32 %235
 }
 
diff --git a/test/CodeGen/Mips/hf16call32_body.ll b/test/CodeGen/Mips/hf16call32_body.ll
index adac31460c44..d06256cc564f 100644
--- a/test/CodeGen/Mips/hf16call32_body.ll
+++ b/test/CodeGen/Mips/hf16call32_body.ll
@@ -14,7 +14,7 @@ define void @v_sf(float %p) #0 {
 entry:
   %p.addr = alloca float, align 4
   store float %p, float* %p.addr, align 4
-  %0 = load float* %p.addr, align 4
+  %0 = load float, float* %p.addr, align 4
   store float %0, float* @x, align 4
   ret void
 }
@@ -33,7 +33,7 @@ define void @v_df(double %p) #0 {
 entry:
   %p.addr = alloca double, align 8
   store double %p, double* %p.addr, align 8
-  %0 = load double* %p.addr, align 8
+  %0 = load double, double* %p.addr, align 8
   store double %0, double* @xd, align 8
   ret void
 }
@@ -54,9 +54,9 @@ entry:
   %p2.addr = alloca float, align 4
   store float %p1, float* %p1.addr, align 4
   store float %p2, float* %p2.addr, align 4
-  %0 = load float* %p1.addr, align 4
+  %0 = load float, float* %p1.addr, align 4
   store float %0, float* @x, align 4
-  %1 = load float* %p2.addr, align 4
+  %1 = load float, float* %p2.addr, align 4
   store float %1, float* @y, align 4
   ret void
 }
@@ -77,9 +77,9 @@ entry:
   %p2.addr = alloca double, align 8
   store float %p1, float* %p1.addr, align 4
   store double %p2, double* %p2.addr, align 8
-  %0 = load float* %p1.addr, align 4
+  %0 = load float, float* %p1.addr, align 4
   store float %0, float* @x, align 4
-  %1 = load double* %p2.addr, align 8
+  %1 = load double, double* %p2.addr, align 8
   store double %1, double* @yd, align 8
   ret void
 }
@@ -101,9 +101,9 @@ entry:
   %p2.addr = alloca float, align 4
   store double %p1, double* %p1.addr, align 8
   store float %p2, float* %p2.addr, align 4
-  %0 = load double* %p1.addr, align 8
+  %0 = load double, double* %p1.addr, align 8
   store double %0, double* @xd, align 8
-  %1 = load float* %p2.addr, align 4
+  %1 = load float, float* %p2.addr, align 4
   store float %1, float* @y, align 4
   ret void
 }
@@ -125,9 +125,9 @@ entry:
   %p2.addr = alloca double, align 8
   store double %p1, double* %p1.addr, align 8
   store double %p2, double* %p2.addr, align 8
-  %0 = load double* %p1.addr, align 8
+  %0 = load double, double* %p1.addr, align 8
   store double %0, double* @xd, align 8
-  %1 = load double* %p2.addr, align 8
+  %1 = load double, double* %p2.addr, align 8
   store double %1, double* @yd, align 8
   ret void
 }
@@ -146,7 +146,7 @@ entry:
 ; Function Attrs: nounwind
 define float @sf_v() #0 {
 entry:
-  %0 = load float* @ret_sf, align 4
+  %0 = load float, float* @ret_sf, align 4
   ret float %0
 }
 
@@ -155,9 +155,9 @@ define float @sf_sf(float %p) #0 {
 entry:
   %p.addr = alloca float, align 4
   store float %p, float* %p.addr, align 4
-  %0 = load float* %p.addr, align 4
+  %0 = load float, float* %p.addr, align 4
   store float %0, float* @x, align 4
-  %1 = load float* @ret_sf, align 4
+  %1 = load float, float* @ret_sf, align 4
   ret float %1
 }
 
@@ -176,9 +176,9 @@ define float @sf_df(double %p) #0 {
 entry:
   %p.addr = alloca double, align 8
   store double %p, double* %p.addr, align 8
-  %0 = load double* %p.addr, align 8
+  %0 = load double, double* %p.addr, align 8
   store double %0, double* @xd, align 8
-  %1 = load float* @ret_sf, align 4
+  %1 = load float, float* @ret_sf, align 4
   ret float %1
 }
 
@@ -198,11 +198,11 @@ entry:
   %p2.addr = alloca float, align 4
   store float %p1, float* %p1.addr, align 4
   store float %p2, float* %p2.addr, align 4
-  %0 = load float* %p1.addr, align 4
+  %0 = load float, float* %p1.addr, align 4
   store float %0, float* @x, align 4
-  %1 = load float* %p2.addr, align 4
+  %1 = load float, float* %p2.addr, align 4
   store float %1, float* @y, align 4
-  %2 = load float* @ret_sf, align 4
+  %2 = load float, float* @ret_sf, align 4
   ret float %2
 }
 
@@ -222,11 +222,11 @@ entry:
   %p2.addr = alloca double, align 8
   store float %p1, float* %p1.addr, align 4
   store double %p2, double* %p2.addr, align 8
-  %0 = load float* %p1.addr, align 4
+  %0 = load float, float* %p1.addr, align 4
   store float %0, float* @x, align 4
-  %1 = load double* %p2.addr, align 8
+  %1 = load double, double* %p2.addr, align 8
   store double %1, double* @yd, align 8
-  %2 = load float* @ret_sf, align 4
+  %2 = load float, float* @ret_sf, align 4
   ret float %2
 }
 
@@ -247,11 +247,11 @@ entry:
   %p2.addr = alloca float, align 4
   store double %p1, double* %p1.addr, align 8
   store float %p2, float* %p2.addr, align 4
-  %0 = load double* %p1.addr, align 8
+  %0 = load double, double* %p1.addr, align 8
   store double %0, double* @xd, align 8
-  %1 = load float* %p2.addr, align 4
+  %1 = load float, float* %p2.addr, align 4
   store float %1, float* @y, align 4
-  %2 = load float* @ret_sf, align 4
+  %2 = load float, float* @ret_sf, align 4
   ret float %2
 }
 
@@ -272,11 +272,11 @@ entry:
   %p2.addr = alloca double, align 8
   store double %p1, double* %p1.addr, align 8
   store double %p2, double* %p2.addr, align 8
-  %0 = load double* %p1.addr, align 8
+  %0 = load double, double* %p1.addr, align 8
   store double %0, double* @xd, align 8
-  %1 = load double* %p2.addr, align 8
+  %1 = load double, double* %p2.addr, align 8
   store double %1, double* @yd, align 8
-  %2 = load float* @ret_sf, align 4
+  %2 = load float, float* @ret_sf, align 4
   ret float %2
 }
 
diff --git a/test/CodeGen/Mips/hf1_body.ll b/test/CodeGen/Mips/hf1_body.ll
index 5acfe86373d9..71a1b960c5b3 100644
--- a/test/CodeGen/Mips/hf1_body.ll
+++ b/test/CodeGen/Mips/hf1_body.ll
@@ -7,7 +7,7 @@ define void @v_sf(float %p) #0 {
 entry:
   %p.addr = alloca float, align 4
   store float %p, float* %p.addr, align 4
-  %0 = load float* %p.addr, align 4
+  %0 = load float, float* %p.addr, align 4
   store float %0, float* @x, align 4
   ret void
 }
diff --git a/test/CodeGen/Mips/hfptrcall.ll b/test/CodeGen/Mips/hfptrcall.ll
index 683952d0e4ec..c9f1fe973796 100644
--- a/test/CodeGen/Mips/hfptrcall.ll
+++ b/test/CodeGen/Mips/hfptrcall.ll
@@ -34,11 +34,11 @@ entry:
 define { float, float } @scv() #0 {
 entry:
   %retval = alloca { float, float }, align 4
-  %real = getelementptr inbounds { float, float }* %retval, i32 0, i32 0
-  %imag = getelementptr inbounds { float, float }* %retval, i32 0, i32 1
+  %real = getelementptr inbounds { float, float }, { float, float }* %retval, i32 0, i32 0
+  %imag = getelementptr inbounds { float, float }, { float, float }* %retval, i32 0, i32 1
   store float 5.000000e+00, float* %real
   store float 9.900000e+01, float* %imag
-  %0 = load { float, float }* %retval
+  %0 = load { float, float }, { float, float }* %retval
   ret { float, float } %0
 }
 
@@ -50,11 +50,11 @@ entry:
 define { double, double } @dcv() #0 {
 entry:
   %retval = alloca { double, double }, align 8
-  %real = getelementptr inbounds { double, double }* %retval, i32 0, i32 0
-  %imag = getelementptr inbounds { double, double }* %retval, i32 0, i32 1
+  %real = getelementptr inbounds { double, double }, { double, double }* %retval, i32 0, i32 0
+  %imag = getelementptr inbounds { double, double }, { double, double }* %retval, i32 0, i32 1
   store double 0x416BC8B0A0000000, double* %real
   store double 0x41CDCCB763800000, double* %imag
-  %0 = load { double, double }* %retval
+  %0 = load { double, double }, { double, double }* %retval
   ret { double, double } %0
 }
 
@@ -65,43 +65,43 @@ entry:
 ; Function Attrs: nounwind
 define i32 @main() #0 {
 entry:
-  %0 = load float ()** @ptrsv, align 4
+  %0 = load float ()*, float ()** @ptrsv, align 4
   %call = call float %0()
   store float %call, float* @x, align 4
-  %1 = load float* @x, align 4
+  %1 = load float, float* @x, align 4
   %conv = fpext float %1 to double
-  %call1 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([4 x i8]* @.str, i32 0, i32 0), double %conv)
-  %2 = load double ()** @ptrdv, align 4
+  %call1 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([4 x i8], [4 x i8]* @.str, i32 0, i32 0), double %conv)
+  %2 = load double ()*, double ()** @ptrdv, align 4
   %call2 = call double %2()
   store double %call2, double* @xd, align 8
-  %3 = load double* @xd, align 8
-  %call3 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([4 x i8]* @.str, i32 0, i32 0), double %3)
-  %4 = load { float, float } ()** @ptrscv, align 4
+  %3 = load double, double* @xd, align 8
+  %call3 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([4 x i8], [4 x i8]* @.str, i32 0, i32 0), double %3)
+  %4 = load { float, float } ()*, { float, float } ()** @ptrscv, align 4
   %call4 = call { float, float } %4()
   %5 = extractvalue { float, float } %call4, 0
   %6 = extractvalue { float, float } %call4, 1
-  store float %5, float* getelementptr inbounds ({ float, float }* @xy, i32 0, i32 0)
-  store float %6, float* getelementptr inbounds ({ float, float }* @xy, i32 0, i32 1)
-  %xy.real = load float* getelementptr inbounds ({ float, float }* @xy, i32 0, i32 0)
-  %xy.imag = load float* getelementptr inbounds ({ float, float }* @xy, i32 0, i32 1)
+  store float %5, float* getelementptr inbounds ({ float, float }, { float, float }* @xy, i32 0, i32 0)
+  store float %6, float* getelementptr inbounds ({ float, float }, { float, float }* @xy, i32 0, i32 1)
+  %xy.real = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @xy, i32 0, i32 0)
+  %xy.imag = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @xy, i32 0, i32 1)
   %conv5 = fpext float %xy.real to double
   %conv6 = fpext float %xy.imag to double
-  %xy.real7 = load float* getelementptr inbounds ({ float, float }* @xy, i32 0, i32 0)
-  %xy.imag8 = load float* getelementptr inbounds ({ float, float }* @xy, i32 0, i32 1)
+  %xy.real7 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @xy, i32 0, i32 0)
+  %xy.imag8 = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @xy, i32 0, i32 1)
   %conv9 = fpext float %xy.real7 to double
   %conv10 = fpext float %xy.imag8 to double
-  %call11 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([10 x i8]* @.str1, i32 0, i32 0), double %conv5, double %conv10)
-  %7 = load { double, double } ()** @ptrdcv, align 4
+  %call11 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([10 x i8], [10 x i8]* @.str1, i32 0, i32 0), double %conv5, double %conv10)
+  %7 = load { double, double } ()*, { double, double } ()** @ptrdcv, align 4
   %call12 = call { double, double } %7()
   %8 = extractvalue { double, double } %call12, 0
   %9 = extractvalue { double, double } %call12, 1
-  store double %8, double* getelementptr inbounds ({ double, double }* @xyd, i32 0, i32 0)
-  store double %9, double* getelementptr inbounds ({ double, double }* @xyd, i32 0, i32 1)
-  %xyd.real = load double* getelementptr inbounds ({ double, double }* @xyd, i32 0, i32 0)
-  %xyd.imag = load double* getelementptr inbounds ({ double, double }* @xyd, i32 0, i32 1)
-  %xyd.real13 = load double* getelementptr inbounds ({ double, double }* @xyd, i32 0, i32 0)
-  %xyd.imag14 = load double* getelementptr inbounds ({ double, double }* @xyd, i32 0, i32 1)
-  %call15 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([10 x i8]* @.str1, i32 0, i32 0), double %xyd.real, double %xyd.imag14)
+  store double %8, double* getelementptr inbounds ({ double, double }, { double, double }* @xyd, i32 0, i32 0)
+  store double %9, double* getelementptr inbounds ({ double, double }, { double, double }* @xyd, i32 0, i32 1)
+  %xyd.real = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @xyd, i32 0, i32 0)
+  %xyd.imag = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @xyd, i32 0, i32 1)
+  %xyd.real13 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @xyd, i32 0, i32 0)
+  %xyd.imag14 = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @xyd, i32 0, i32 1)
+  %call15 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([10 x i8], [10 x i8]* @.str1, i32 0, i32 0), double %xyd.real, double %xyd.imag14)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/i32k.ll b/test/CodeGen/Mips/i32k.ll
index 73f1302beec0..ba9cf7342308 100644
--- a/test/CodeGen/Mips/i32k.ll
+++ b/test/CodeGen/Mips/i32k.ll
@@ -4,14 +4,14 @@
 
 define i32 @main() nounwind {
 entry:
-  %call = tail call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([4 x i8]* @.str, i32 0, i32 0), i32 1075344593) nounwind
+  %call = tail call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([4 x i8], [4 x i8]* @.str, i32 0, i32 0), i32 1075344593) nounwind
 ; 16:	lw	${{[0-9]+}}, 1f
 ; 16:	b	2f
 ; 16:	.align	2
 ; 16: 1: 	.word	1075344593
 ; 16: 2:
 
-  %call1 = tail call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([4 x i8]* @.str, i32 0, i32 0), i32 -1075344593) nounwind
+  %call1 = tail call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([4 x i8], [4 x i8]* @.str, i32 0, i32 0), i32 -1075344593) nounwind
 
 ; 16:	lw	${{[0-9]+}}, 1f
 ; 16:	b	2f
diff --git a/test/CodeGen/Mips/inlineasm-assembler-directives.ll b/test/CodeGen/Mips/inlineasm-assembler-directives.ll
index e4a6d1e26c69..88ceed4114c2 100644
--- a/test/CodeGen/Mips/inlineasm-assembler-directives.ll
+++ b/test/CodeGen/Mips/inlineasm-assembler-directives.ll
@@ -16,7 +16,7 @@ entry:
   %a = alloca i32, align 4
   %b = alloca i32, align 4
   store i32 20, i32* %a, align 4
-  %0 = load i32* %a, align 4
+  %0 = load i32, i32* %a, align 4
   %1 = call i32 asm sideeffect "addi $$9, $1, 8\0A\09subi $0, $$9, 6", "=r,r,~{$1}"(i32 %0)
   store i32 %1, i32* %b, align 4
   ret void
diff --git a/test/CodeGen/Mips/inlineasm-cnstrnt-reg64.ll b/test/CodeGen/Mips/inlineasm-cnstrnt-reg64.ll
index a7ba762b1064..acce63203812 100644
--- a/test/CodeGen/Mips/inlineasm-cnstrnt-reg64.ll
+++ b/test/CodeGen/Mips/inlineasm-cnstrnt-reg64.ll
@@ -3,7 +3,7 @@
 ; The target is 64 bit.
 ;
 ;
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=n64 < %s | FileCheck %s
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi=n64 < %s | FileCheck %s
 
 
 define i32 @main() nounwind {
diff --git a/test/CodeGen/Mips/inlineasm-operand-code.ll b/test/CodeGen/Mips/inlineasm-operand-code.ll
index 3d9dec76fb37..b9415ee90cdb 100644
--- a/test/CodeGen/Mips/inlineasm-operand-code.ll
+++ b/test/CodeGen/Mips/inlineasm-operand-code.ll
@@ -125,7 +125,7 @@ entry:
 ;CHECK_BIG_32:       #APP
 ;CHECK_BIG_32:       or ${{[0-9]+}},$[[SECOND]],${{[0-9]+}}
 ;CHECK_BIG_32:       #NO_APP
-  %bosco = load i64* getelementptr inbounds (%union.u_tag* @uval, i32 0, i32 0), align 8
+  %bosco = load i64, i64* getelementptr inbounds (%union.u_tag, %union.u_tag* @uval, i32 0, i32 0), align 8
   %trunc1 = trunc i64 %bosco to i32
   tail call i32 asm sideeffect "or $0,${1:D},$2", "=r,r,r"(i64 %bosco, i32 %trunc1) nounwind
   ret i32 0
@@ -149,7 +149,7 @@ entry:
 ;CHECK_BIG_32:       #APP
 ;CHECK_BIG_32:       or ${{[0-9]+}},$[[SECOND]],${{[0-9]+}}
 ;CHECK_BIG_32:       #NO_APP
-  %bosco = load i64* getelementptr inbounds (%union.u_tag* @uval, i32 0, i32 0), align 8
+  %bosco = load i64, i64* getelementptr inbounds (%union.u_tag, %union.u_tag* @uval, i32 0, i32 0), align 8
   %trunc1 = trunc i64 %bosco to i32
   tail call i32 asm sideeffect "or $0,${1:L},$2", "=r,r,r"(i64 %bosco, i32 %trunc1) nounwind
   ret i32 0
@@ -173,7 +173,7 @@ entry:
 ;CHECK_BIG_32:       #APP
 ;CHECK_BIG_32:       or ${{[0-9]+}},$[[FIRST]],${{[0-9]+}}
 ;CHECK_BIG_32:       #NO_APP
-  %bosco = load i64* getelementptr inbounds (%union.u_tag* @uval, i32 0, i32 0), align 8
+  %bosco = load i64, i64* getelementptr inbounds (%union.u_tag, %union.u_tag* @uval, i32 0, i32 0), align 8
   %trunc1 = trunc i64 %bosco to i32
   tail call i32 asm sideeffect "or $0,${1:M},$2", "=r,r,r"(i64 %bosco, i32 %trunc1) nounwind
   ret i32 0
diff --git a/test/CodeGen/Mips/inlineasm64.ll b/test/CodeGen/Mips/inlineasm64.ll
index dbce3c394e96..82abdf82a3ed 100644
--- a/test/CodeGen/Mips/inlineasm64.ll
+++ b/test/CodeGen/Mips/inlineasm64.ll
@@ -1,4 +1,4 @@
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=n64 < %s | FileCheck %s
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi=n64 < %s | FileCheck %s
 
 @gl2 = external global i64
 @gl1 = external global i64
@@ -8,8 +8,8 @@ define void @foo1() nounwind {
 entry:
 ; CHECK: foo1
 ; CHECK: daddu
-  %0 = load i64* @gl1, align 8
-  %1 = load i64* @gl0, align 8
+  %0 = load i64, i64* @gl1, align 8
+  %1 = load i64, i64* @gl0, align 8
   %2 = tail call i64 asm "daddu $0, $1, $2", "=r,r,r"(i64 %0, i64 %1) nounwind
   store i64 %2, i64* @gl2, align 8
   ret void
diff --git a/test/CodeGen/Mips/inlineasm_constraint.ll b/test/CodeGen/Mips/inlineasm_constraint.ll
index 76b73dc276ae..868433e0941f 100644
--- a/test/CodeGen/Mips/inlineasm_constraint.ll
+++ b/test/CodeGen/Mips/inlineasm_constraint.ll
@@ -51,14 +51,5 @@ entry:
 ; CHECK: #NO_APP	
   tail call i32 asm sideeffect "addiu $0,$1,$2", "=r,r,P"(i32 7, i32 65535) nounwind
 
-; Now R Which takes the address of c
-  %c = alloca i32, align 4
-  store i32 -4469539, i32* %c, align 4
-  %8 = call i32 asm sideeffect "lw $0, 1 + $1\0A\09lw $0, 2 + $1\0A\09", "=r,*R"(i32* %c) #1
-; CHECK: #APP
-; CHECK: lw ${{[0-9]+}}, 1 + 0(${{[0-9]+}})
-; CHECK: lw ${{[0-9]+}}, 2 + 0(${{[0-9]+}})
-; CHECK: #NO_APP	
-
   ret i32 0
 }
diff --git a/test/CodeGen/Mips/inlineasm_constraint_R.ll b/test/CodeGen/Mips/inlineasm_constraint_R.ll
new file mode 100644
index 000000000000..c4105ae6b22c
--- /dev/null
+++ b/test/CodeGen/Mips/inlineasm_constraint_R.ll
@@ -0,0 +1,60 @@
+; RUN: llc -march=mipsel < %s | FileCheck %s
+
+@data = global [8193 x i32] zeroinitializer
+
+define void @R(i32 *%p) nounwind {
+entry:
+  ; CHECK-LABEL: R:
+
+  call void asm sideeffect "lw $$1, $0", "*R,~{$1}"(i32* getelementptr inbounds ([8193 x i32], [8193 x i32]* @data, i32 0, i32 0))
+
+  ; CHECK: lw $[[BASEPTR:[0-9]+]], %got(data)(
+  ; CHECK: #APP
+  ; CHECK: lw $1, 0($[[BASEPTR]])
+  ; CHECK: #NO_APP
+
+  ret void
+}
+
+define void @R_offset_4(i32 *%p) nounwind {
+entry:
+  ; CHECK-LABEL: R_offset_4:
+
+  call void asm sideeffect "lw $$1, $0", "*R,~{$1}"(i32* getelementptr inbounds ([8193 x i32], [8193 x i32]* @data, i32 0, i32 1))
+
+  ; CHECK: lw $[[BASEPTR:[0-9]+]], %got(data)(
+  ; CHECK: #APP
+  ; CHECK: lw $1, 4($[[BASEPTR]])
+  ; CHECK: #NO_APP
+
+  ret void
+}
+
+define void @R_offset_254(i32 *%p) nounwind {
+entry:
+  ; CHECK-LABEL: R_offset_254:
+
+  call void asm sideeffect "lw $$1, $0", "*R,~{$1}"(i32* getelementptr inbounds ([8193 x i32], [8193 x i32]* @data, i32 0, i32 63))
+
+  ; CHECK-DAG: lw $[[BASEPTR:[0-9]+]], %got(data)(
+  ; CHECK: #APP
+  ; CHECK: lw $1, 252($[[BASEPTR]])
+  ; CHECK: #NO_APP
+
+  ret void
+}
+
+define void @R_offset_256(i32 *%p) nounwind {
+entry:
+  ; CHECK-LABEL: R_offset_256:
+
+  call void asm sideeffect "lw $$1, $0", "*R,~{$1}"(i32* getelementptr inbounds ([8193 x i32], [8193 x i32]* @data, i32 0, i32 64))
+
+  ; CHECK-DAG: lw $[[BASEPTR:[0-9]+]], %got(data)(
+  ; CHECK: addiu $[[BASEPTR2:[0-9]+]], $[[BASEPTR]], 256
+  ; CHECK: #APP
+  ; CHECK: lw $1, 0($[[BASEPTR2]])
+  ; CHECK: #NO_APP
+
+  ret void
+}
diff --git a/test/CodeGen/Mips/inlineasm_constraint_ZC.ll b/test/CodeGen/Mips/inlineasm_constraint_ZC.ll
new file mode 100644
index 000000000000..c1746a67564f
--- /dev/null
+++ b/test/CodeGen/Mips/inlineasm_constraint_ZC.ll
@@ -0,0 +1,167 @@
+; RUN: llc -march=mipsel -mcpu=mips32r6 < %s | FileCheck %s -check-prefix=ALL -check-prefix=09BIT
+; RUN: llc -march=mipsel -mattr=+micromips < %s | FileCheck %s -check-prefix=ALL -check-prefix=12BIT
+; RUN: llc -march=mipsel < %s | FileCheck %s -check-prefix=ALL -check-prefix=16BIT
+
+@data = global [8193 x i32] zeroinitializer
+
+define void @ZC(i32 *%p) nounwind {
+entry:
+  ; ALL-LABEL: ZC:
+
+  call void asm sideeffect "lw $$1, $0", "*^ZC,~{$1}"(i32* getelementptr inbounds ([8193 x i32], [8193 x i32]* @data, i32 0, i32 0))
+
+  ; ALL: lw $[[BASEPTR:[0-9]+]], %got(data)(
+  ; ALL: #APP
+  ; ALL: lw $1, 0($[[BASEPTR]])
+  ; ALL: #NO_APP
+
+  ret void
+}
+
+define void @ZC_offset_n4(i32 *%p) nounwind {
+entry:
+  ; ALL-LABEL: ZC_offset_n4:
+
+  call void asm sideeffect "lw $$1, $0", "*^ZC,~{$1}"(i32* getelementptr inbounds ([8193 x i32], [8193 x i32]* @data, i32 0, i32 -1))
+
+  ; ALL: lw $[[BASEPTR:[0-9]+]], %got(data)(
+  ; ALL: #APP
+  ; ALL: lw $1, -4($[[BASEPTR]])
+  ; ALL: #NO_APP
+
+  ret void
+}
+
+define void @ZC_offset_4(i32 *%p) nounwind {
+entry:
+  ; ALL-LABEL: ZC_offset_4:
+
+  call void asm sideeffect "lw $$1, $0", "*^ZC,~{$1}"(i32* getelementptr inbounds ([8193 x i32], [8193 x i32]* @data, i32 0, i32 1))
+
+  ; ALL: lw $[[BASEPTR:[0-9]+]], %got(data)(
+  ; ALL: #APP
+  ; ALL: lw $1, 4($[[BASEPTR]])
+  ; ALL: #NO_APP
+
+  ret void
+}
+
+define void @ZC_offset_252(i32 *%p) nounwind {
+entry:
+  ; ALL-LABEL: ZC_offset_252:
+
+  call void asm sideeffect "lw $$1, $0", "*^ZC,~{$1}"(i32* getelementptr inbounds ([8193 x i32], [8193 x i32]* @data, i32 0, i32 63))
+
+  ; ALL: lw $[[BASEPTR:[0-9]+]], %got(data)(
+  ; ALL: #APP
+  ; ALL: lw $1, 252($[[BASEPTR]])
+  ; ALL: #NO_APP
+
+  ret void
+}
+
+define void @ZC_offset_256(i32 *%p) nounwind {
+entry:
+  ; ALL-LABEL: ZC_offset_256:
+
+  call void asm sideeffect "lw $$1, $0", "*^ZC,~{$1}"(i32* getelementptr inbounds ([8193 x i32], [8193 x i32]* @data, i32 0, i32 64))
+
+  ; ALL: lw $[[BASEPTR:[0-9]+]], %got(data)(
+
+  ; 09BIT: addiu $[[BASEPTR2:[0-9]+]], $[[BASEPTR]], 256
+
+  ; ALL: #APP
+
+  ; 09BIT: lw $1, 0($[[BASEPTR2]])
+  ; 12BIT: lw $1, 256($[[BASEPTR]])
+  ; 16BIT: lw $1, 256($[[BASEPTR]])
+
+  ; ALL: #NO_APP
+
+  ret void
+}
+
+define void @ZC_offset_2044(i32 *%p) nounwind {
+entry:
+  ; ALL-LABEL: ZC_offset_2044:
+
+  call void asm sideeffect "lw $$1, $0", "*^ZC,~{$1}"(i32* getelementptr inbounds ([8193 x i32], [8193 x i32]* @data, i32 0, i32 511))
+
+  ; ALL: lw $[[BASEPTR:[0-9]+]], %got(data)(
+
+  ; 09BIT: addiu $[[BASEPTR2:[0-9]+]], $[[BASEPTR]], 2044
+
+  ; ALL: #APP
+
+  ; 09BIT: lw $1, 0($[[BASEPTR2]])
+  ; 12BIT: lw $1, 2044($[[BASEPTR]])
+  ; 16BIT: lw $1, 2044($[[BASEPTR]])
+
+  ; ALL: #NO_APP
+
+  ret void
+}
+
+define void @ZC_offset_2048(i32 *%p) nounwind {
+entry:
+  ; ALL-LABEL: ZC_offset_2048:
+
+  call void asm sideeffect "lw $$1, $0", "*^ZC,~{$1}"(i32* getelementptr inbounds ([8193 x i32], [8193 x i32]* @data, i32 0, i32 512))
+
+  ; ALL: lw $[[BASEPTR:[0-9]+]], %got(data)(
+
+  ; 09BIT: addiu $[[BASEPTR2:[0-9]+]], $[[BASEPTR]], 2048
+  ; 12BIT: addiu $[[BASEPTR2:[0-9]+]], $[[BASEPTR]], 2048
+
+  ; ALL: #APP
+
+  ; 09BIT: lw $1, 0($[[BASEPTR2]])
+  ; 12BIT: lw $1, 0($[[BASEPTR2]])
+  ; 16BIT: lw $1, 2048($[[BASEPTR]])
+
+  ; ALL: #NO_APP
+
+  ret void
+}
+
+define void @ZC_offset_32764(i32 *%p) nounwind {
+entry:
+  ; ALL-LABEL: ZC_offset_32764:
+
+  call void asm sideeffect "lw $$1, $0", "*^ZC,~{$1}"(i32* getelementptr inbounds ([8193 x i32], [8193 x i32]* @data, i32 0, i32 8191))
+
+  ; ALL-DAG: lw $[[BASEPTR:[0-9]+]], %got(data)(
+
+  ; 09BIT: addiu $[[BASEPTR2:[0-9]+]], $[[BASEPTR]], 32764
+  ; 12BIT: addiu $[[BASEPTR2:[0-9]+]], $[[BASEPTR]], 32764
+
+  ; ALL: #APP
+
+  ; 09BIT: lw $1, 0($[[BASEPTR2]])
+  ; 12BIT: lw $1, 0($[[BASEPTR2]])
+  ; 16BIT: lw $1, 32764($[[BASEPTR]])
+
+  ; ALL: #NO_APP
+
+  ret void
+}
+
+define void @ZC_offset_32768(i32 *%p) nounwind {
+entry:
+  ; ALL-LABEL: ZC_offset_32768:
+
+  call void asm sideeffect "lw $$1, $0", "*^ZC,~{$1}"(i32* getelementptr inbounds ([8193 x i32], [8193 x i32]* @data, i32 0, i32 8192))
+
+  ; ALL-DAG: lw $[[BASEPTR:[0-9]+]], %got(data)(
+  ; ALL-DAG: ori $[[T0:[0-9]+]], $zero, 32768
+
+  ; 09BIT: addu $[[BASEPTR2:[0-9]+]], $[[BASEPTR]], $[[T0]]
+  ; 12BIT: addu16 $[[BASEPTR2:[0-9]+]], $[[BASEPTR]], $[[T0]]
+  ; 16BIT: addu $[[BASEPTR2:[0-9]+]], $[[BASEPTR]], $[[T0]]
+
+  ; ALL: #APP
+  ; ALL: lw $1, 0($[[BASEPTR2]])
+  ; ALL: #NO_APP
+
+  ret void
+}
diff --git a/test/CodeGen/Mips/inlineasm_constraint_m.ll b/test/CodeGen/Mips/inlineasm_constraint_m.ll
new file mode 100644
index 000000000000..00053ad3c105
--- /dev/null
+++ b/test/CodeGen/Mips/inlineasm_constraint_m.ll
@@ -0,0 +1,61 @@
+; RUN: llc -march=mipsel < %s | FileCheck %s
+
+@data = global [8193 x i32] zeroinitializer
+
+define void @m(i32 *%p) nounwind {
+entry:
+  ; CHECK-LABEL: m:
+
+  call void asm sideeffect "lw $$1, $0", "*m,~{$1}"(i32* getelementptr inbounds ([8193 x i32], [8193 x i32]* @data, i32 0, i32 0))
+
+  ; CHECK: lw $[[BASEPTR:[0-9]+]], %got(data)(
+  ; CHECK: #APP
+  ; CHECK: lw $1, 0($[[BASEPTR]])
+  ; CHECK: #NO_APP
+
+  ret void
+}
+
+define void @m_offset_4(i32 *%p) nounwind {
+entry:
+  ; CHECK-LABEL: m_offset_4:
+
+  call void asm sideeffect "lw $$1, $0", "*m,~{$1}"(i32* getelementptr inbounds ([8193 x i32], [8193 x i32]* @data, i32 0, i32 1))
+
+  ; CHECK: lw $[[BASEPTR:[0-9]+]], %got(data)(
+  ; CHECK: #APP
+  ; CHECK: lw $1, 4($[[BASEPTR]])
+  ; CHECK: #NO_APP
+
+  ret void
+}
+
+define void @m_offset_32764(i32 *%p) nounwind {
+entry:
+  ; CHECK-LABEL: m_offset_32764:
+
+  call void asm sideeffect "lw $$1, $0", "*m,~{$1}"(i32* getelementptr inbounds ([8193 x i32], [8193 x i32]* @data, i32 0, i32 8191))
+
+  ; CHECK-DAG: lw $[[BASEPTR:[0-9]+]], %got(data)(
+  ; CHECK: #APP
+  ; CHECK: lw $1, 32764($[[BASEPTR]])
+  ; CHECK: #NO_APP
+
+  ret void
+}
+
+define void @m_offset_32768(i32 *%p) nounwind {
+entry:
+  ; CHECK-LABEL: m_offset_32768:
+
+  call void asm sideeffect "lw $$1, $0", "*m,~{$1}"(i32* getelementptr inbounds ([8193 x i32], [8193 x i32]* @data, i32 0, i32 8192))
+
+  ; CHECK-DAG: lw $[[BASEPTR:[0-9]+]], %got(data)(
+  ; CHECK-DAG: ori $[[T0:[0-9]+]], $zero, 32768
+  ; CHECK: addu $[[BASEPTR2:[0-9]+]], $[[BASEPTR]], $[[T0]]
+  ; CHECK: #APP
+  ; CHECK: lw $1, 0($[[BASEPTR2]])
+  ; CHECK: #NO_APP
+
+  ret void
+}
diff --git a/test/CodeGen/Mips/inlineasmmemop.ll b/test/CodeGen/Mips/inlineasmmemop.ll
index 5518520c5491..9e9b6cd089ea 100644
--- a/test/CodeGen/Mips/inlineasmmemop.ll
+++ b/test/CodeGen/Mips/inlineasmmemop.ll
@@ -6,14 +6,13 @@
 define i32 @f1(i32 %x) nounwind {
 entry:
 ; CHECK-LABEL: f1:
-; CHECK: addiu $[[T0:[0-9]+]], $sp
 ; CHECK: #APP
-; CHECK: sw $4, 0($[[T0]])
+; CHECK: sw $4, [[OFFSET:[0-9]+]]($sp)
 ; CHECK: #NO_APP
+; CHECK: lw  $[[T1:[0-9]+]], %got(g1)
 ; CHECK: #APP
-; CHECK: lw $[[T3:[0-9]+]], 0($[[T0]])
+; CHECK: lw $[[T3:[0-9]+]], [[OFFSET]]($sp)
 ; CHECK: #NO_APP
-; CHECK: lw  $[[T1:[0-9]+]], %got(g1)
 ; CHECK: sw  $[[T3]], 0($[[T1]])
 
   %l1 = alloca i32, align 4
@@ -27,13 +26,13 @@ entry:
 ; "D": Second word of a double word. This works for any memory element
 ; double or single.
 ; CHECK: #APP
-; CHECK: lw ${{[0-9]+}},4(${{[0-9]+}});
+; CHECK: lw ${{[0-9]+}}, 16(${{[0-9]+}});
 ; CHECK: #NO_APP
 
 ; No "D": First word of a double word. This works for any memory element
 ; double or single.
 ; CHECK: #APP
-; CHECK: lw ${{[0-9]+}},0(${{[0-9]+}});
+; CHECK: lw ${{[0-9]+}}, 12(${{[0-9]+}});
 ; CHECK: #NO_APP
 
 @b = common global [20 x i32] zeroinitializer, align 4
@@ -41,8 +40,8 @@ entry:
 define void @main() {
 entry:
 ; Second word:
-  tail call void asm sideeffect "    lw    $0,${1:D};", "r,*m,~{$11}"(i32 undef, i32* getelementptr inbounds ([20 x i32]* @b, i32 0, i32 3))
+  tail call void asm sideeffect "    lw    $0, ${1:D};", "r,*m,~{$11}"(i32 undef, i32* getelementptr inbounds ([20 x i32], [20 x i32]* @b, i32 0, i32 3))
 ; First word. Notice, no 'D':
-  tail call void asm sideeffect "    lw    $0,${1};", "r,*m,~{$11}"(i32 undef, i32* getelementptr inbounds ([20 x i32]* @b, i32 0, i32 3))
+  tail call void asm sideeffect "    lw    $0, ${1};", "r,*m,~{$11}"(i32 undef, i32* getelementptr inbounds ([20 x i32], [20 x i32]* @b, i32 0, i32 3))
   ret void
 }
diff --git a/test/CodeGen/Mips/insn-zero-size-bb.ll b/test/CodeGen/Mips/insn-zero-size-bb.ll
new file mode 100644
index 000000000000..9739c6f17fab
--- /dev/null
+++ b/test/CodeGen/Mips/insn-zero-size-bb.ll
@@ -0,0 +1,27 @@
+; RUN: llc < %s -march=mips -mcpu=mips32 | FileCheck %s
+; RUN: llc < %s -march=mips -mcpu=mips32r3 -mattr=+micromips | FileCheck %s
+; RUN: llc < %s -march=mips -mcpu=mips16 | FileCheck %s
+
+; Verify that we emit the .insn directive for zero-sized (empty) basic blocks.
+; This only really matters for microMIPS and MIPS16.
+
+declare i32 @foo(...)
+declare void @bar()
+
+define void @main() {
+entry:
+  invoke void @bar() #0
+          to label %unreachable unwind label %return
+
+unreachable:
+; CHECK:          ${{.*}}: # %unreachable
+; CHECK-NEXT:         .insn
+  unreachable
+
+return:
+  %0 = landingpad { i8*, i32 } personality i8* bitcast (i32 (...)* @foo to i8*)
+          catch i8* null
+  ret void
+}
+
+attributes #0 = { noreturn }
diff --git a/test/CodeGen/Mips/internalfunc.ll b/test/CodeGen/Mips/internalfunc.ll
index 863375ad4d4a..2b4a0397f45f 100644
--- a/test/CodeGen/Mips/internalfunc.ll
+++ b/test/CodeGen/Mips/internalfunc.ll
@@ -20,8 +20,8 @@ entry:
   br i1 %tobool, label %if.end, label %if.then
 
 if.then:                                          ; preds = %entry
-  %tmp1 = load void (...)** @caller.sf1, align 4
-  tail call void (...)* %tmp1() nounwind
+  %tmp1 = load void (...)*, void (...)** @caller.sf1, align 4
+  tail call void (...) %tmp1() nounwind
   br label %if.end
 
 if.end:                                           ; preds = %entry, %if.then
@@ -30,7 +30,7 @@ if.end:                                           ; preds = %entry, %if.then
 ; CHECK: lw  $[[R3:[0-9]+]], %got(caller.sf1)
 ; CHECK: sw  ${{[0-9]+}}, %lo(caller.sf1)($[[R3]])
   %tobool3 = icmp ne i32 %a0, 0
-  %tmp4 = load void (...)** @gf1, align 4
+  %tmp4 = load void (...)*, void (...)** @gf1, align 4
   %cond = select i1 %tobool3, void (...)* %tmp4, void (...)* bitcast (void ()* @sf2 to void (...)*)
   store void (...)* %cond, void (...)** @caller.sf1, align 4
   ret void
@@ -38,7 +38,7 @@ if.end:                                           ; preds = %entry, %if.then
 
 define internal void @sf2() nounwind {
 entry:
-  %call = tail call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([3 x i8]* @.str, i32 0, i32 0)) nounwind
+  %call = tail call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @.str, i32 0, i32 0)) nounwind
   ret void
 }
 
@@ -46,7 +46,7 @@ declare i32 @printf(i8* nocapture, ...) nounwind
 
 define internal fastcc void @f2() nounwind noinline {
 entry:
-  %call = tail call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([3 x i8]* @.str, i32 0, i32 0)) nounwind
+  %call = tail call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @.str, i32 0, i32 0)) nounwind
   ret void
 }
 
diff --git a/test/CodeGen/Mips/jtstat.ll b/test/CodeGen/Mips/jtstat.ll
index 01afc080c2ed..35f71cf2dc85 100644
--- a/test/CodeGen/Mips/jtstat.ll
+++ b/test/CodeGen/Mips/jtstat.ll
@@ -8,7 +8,7 @@ define void @test(i32 %i) nounwind {
 entry:
   %i.addr = alloca i32, align 4
   store i32 %i, i32* %i.addr, align 4
-  %0 = load i32* %i.addr, align 4
+  %0 = load i32, i32* %i.addr, align 4
   switch i32 %0, label %sw.epilog [
     i32 115, label %sw.bb
     i32 105, label %sw.bb1
diff --git a/test/CodeGen/Mips/l3mc.ll b/test/CodeGen/Mips/l3mc.ll
index 3bfb389ba05d..c1bff11595c9 100644
--- a/test/CodeGen/Mips/l3mc.ll
+++ b/test/CodeGen/Mips/l3mc.ll
@@ -1,22 +1,22 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck  %s -check-prefix=__call_stub_fp___fixunsdfsi 
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck  %s -check-prefix=__call_stub_fp___fixunsdfsi 
 
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___floatdidf 
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___floatdidf 
 
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___floatdisf 
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___floatdisf 
 
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___floatundidf
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___floatundidf
 
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___fixsfdi 
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___fixsfdi 
 
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___fixunsdfdi 
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___fixunsdfdi 
 
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___fixdfdi
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___fixdfdi
 
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___fixunssfsi 
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___fixunssfsi 
 
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___fixunssfdi 
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___fixunssfdi 
 
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___floatundisf 
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static  < %s | FileCheck %s -check-prefix=__call_stub_fp___floatundisf 
 
 @ll1 = global i64 0, align 8
 @ll2 = global i64 0, align 8
@@ -42,28 +42,28 @@
 ; Function Attrs: nounwind
 define void @_Z3foov() #0 {
 entry:
-  %0 = load double* @d1, align 8
+  %0 = load double, double* @d1, align 8
   %conv = fptosi double %0 to i64
   store i64 %conv, i64* @ll1, align 8
-  %1 = load double* @d2, align 8
+  %1 = load double, double* @d2, align 8
   %conv1 = fptoui double %1 to i64
   store i64 %conv1, i64* @ull1, align 8
-  %2 = load float* @f1, align 4
+  %2 = load float, float* @f1, align 4
   %conv2 = fptosi float %2 to i64
   store i64 %conv2, i64* @ll2, align 8
-  %3 = load float* @f2, align 4
+  %3 = load float, float* @f2, align 4
   %conv3 = fptoui float %3 to i64
   store i64 %conv3, i64* @ull2, align 8
-  %4 = load double* @d3, align 8
+  %4 = load double, double* @d3, align 8
   %conv4 = fptosi double %4 to i32
   store i32 %conv4, i32* @l1, align 4
-  %5 = load double* @d4, align 8
+  %5 = load double, double* @d4, align 8
   %conv5 = fptoui double %5 to i32
   store i32 %conv5, i32* @ul1, align 4
-  %6 = load float* @f3, align 4
+  %6 = load float, float* @f3, align 4
   %conv6 = fptosi float %6 to i32
   store i32 %conv6, i32* @l2, align 4
-  %7 = load float* @f4, align 4
+  %7 = load float, float* @f4, align 4
   %conv7 = fptoui float %7 to i32
   store i32 %conv7, i32* @ul2, align 4
   ret void
@@ -72,28 +72,28 @@ entry:
 ; Function Attrs: nounwind
 define void @_Z3goov() #0 {
 entry:
-  %0 = load i64* @ll1, align 8
+  %0 = load i64, i64* @ll1, align 8
   %conv = sitofp i64 %0 to double
   store double %conv, double* @d1, align 8
-  %1 = load i64* @ull1, align 8
+  %1 = load i64, i64* @ull1, align 8
   %conv1 = uitofp i64 %1 to double
   store double %conv1, double* @d2, align 8
-  %2 = load i64* @ll2, align 8
+  %2 = load i64, i64* @ll2, align 8
   %conv2 = sitofp i64 %2 to float
   store float %conv2, float* @f1, align 4
-  %3 = load i64* @ull2, align 8
+  %3 = load i64, i64* @ull2, align 8
   %conv3 = uitofp i64 %3 to float
   store float %conv3, float* @f2, align 4
-  %4 = load i32* @l1, align 4
+  %4 = load i32, i32* @l1, align 4
   %conv4 = sitofp i32 %4 to double
   store double %conv4, double* @d3, align 8
-  %5 = load i32* @ul1, align 4
+  %5 = load i32, i32* @ul1, align 4
   %conv5 = uitofp i32 %5 to double
   store double %conv5, double* @d4, align 8
-  %6 = load i32* @l2, align 4
+  %6 = load i32, i32* @l2, align 4
   %conv6 = sitofp i32 %6 to float
   store float %conv6, float* @f3, align 4
-  %7 = load i32* @ul2, align 4
+  %7 = load i32, i32* @ul2, align 4
   %conv7 = uitofp i32 %7 to float
   store float %conv7, float* @f4, align 4
   ret void
diff --git a/test/CodeGen/Mips/largeimm1.ll b/test/CodeGen/Mips/largeimm1.ll
index 1c0f69c59011..06c4d6bd9603 100644
--- a/test/CodeGen/Mips/largeimm1.ll
+++ b/test/CodeGen/Mips/largeimm1.ll
@@ -5,7 +5,7 @@
 define void @f() nounwind {
 entry:
   %a1 = alloca [1073741824 x i8], align 1
-  %arrayidx = getelementptr inbounds [1073741824 x i8]* %a1, i32 0, i32 1048676
+  %arrayidx = getelementptr inbounds [1073741824 x i8], [1073741824 x i8]* %a1, i32 0, i32 1048676
   call void @f2(i8* %arrayidx) nounwind
   ret void
 }
diff --git a/test/CodeGen/Mips/largeimmprinting.ll b/test/CodeGen/Mips/largeimmprinting.ll
index 0e9c91fb46df..a53a953a7883 100644
--- a/test/CodeGen/Mips/largeimmprinting.ll
+++ b/test/CodeGen/Mips/largeimmprinting.ll
@@ -1,7 +1,7 @@
 ; RUN: llc -march=mipsel < %s | FileCheck %s -check-prefix=32
-; RUN: llc -march=mips64el -mcpu=mips4 -mattr=n64 < %s | \
+; RUN: llc -march=mips64el -mcpu=mips4 -target-abi=n64 < %s | \
 ; RUN:     FileCheck %s -check-prefix=64
-; RUN: llc -march=mips64el -mcpu=mips64 -mattr=n64 < %s | \
+; RUN: llc -march=mips64el -mcpu=mips64 -target-abi=n64 < %s | \
 ; RUN:     FileCheck %s -check-prefix=64
 
 %struct.S1 = type { [65536 x i8] }
@@ -27,8 +27,8 @@ entry:
 ; 64:  sd  $ra, 24($[[R1]])
 
   %agg.tmp = alloca %struct.S1, align 1
-  %tmp = getelementptr inbounds %struct.S1* %agg.tmp, i32 0, i32 0, i32 0
-  call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp, i8* getelementptr inbounds (%struct.S1* @s1, i32 0, i32 0, i32 0), i32 65536, i32 1, i1 false)
+  %tmp = getelementptr inbounds %struct.S1, %struct.S1* %agg.tmp, i32 0, i32 0, i32 0
+  call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp, i8* getelementptr inbounds (%struct.S1, %struct.S1* @s1, i32 0, i32 0, i32 0), i32 65536, i32 1, i1 false)
   call void @f2(%struct.S1* byval %agg.tmp) nounwind
   ret void
 }
diff --git a/test/CodeGen/Mips/lb1.ll b/test/CodeGen/Mips/lb1.ll
index aac2767a4e40..21648d7572a5 100644
--- a/test/CodeGen/Mips/lb1.ll
+++ b/test/CodeGen/Mips/lb1.ll
@@ -6,12 +6,12 @@
 define i32 @main() nounwind {
 entry:
   %i = alloca i32, align 4
-  %0 = load i8* @c, align 1
+  %0 = load i8, i8* @c, align 1
 ; 16:	lb	${{[0-9]+}}, 0(${{[0-9]+}})
   %conv = sext i8 %0 to i32
   store i32 %conv, i32* %i, align 4
-  %1 = load i32* %i, align 4
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i32 0, i32 0), i32 %1)
+  %1 = load i32, i32* %i, align 4
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([5 x i8], [5 x i8]* @.str, i32 0, i32 0), i32 %1)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/lbu1.ll b/test/CodeGen/Mips/lbu1.ll
index 63e0cca1684d..28ca27132467 100644
--- a/test/CodeGen/Mips/lbu1.ll
+++ b/test/CodeGen/Mips/lbu1.ll
@@ -6,13 +6,13 @@
 define i32 @main() nounwind {
 entry:
   %i = alloca i32, align 4
-  %0 = load i8* @c, align 1
+  %0 = load i8, i8* @c, align 1
   %conv = zext i8 %0 to i32
 ; 16:	lbu	${{[0-9]+}}, 0(${{[0-9]+}})
   store i32 %conv, i32* %i, align 4
-  %1 = load i8* @c, align 1
+  %1 = load i8, i8* @c, align 1
   %conv1 = zext i8 %1 to i32
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i32 0, i32 0), i32 %conv1)
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([5 x i8], [5 x i8]* @.str, i32 0, i32 0), i32 %conv1)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/lcb2.ll b/test/CodeGen/Mips/lcb2.ll
index 59b96e64e95e..a6f4968e6d23 100644
--- a/test/CodeGen/Mips/lcb2.ll
+++ b/test/CodeGen/Mips/lcb2.ll
@@ -1,6 +1,6 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static -mips16-constant-islands=true   < %s | FileCheck %s -check-prefix=lcb
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static -mips16-constant-islands=true   < %s | FileCheck %s -check-prefix=lcb
 
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static -mips16-constant-islands=true   < %s | FileCheck %s -check-prefix=lcbn
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static -mips16-constant-islands=true   < %s | FileCheck %s -check-prefix=lcbn
 
 @i = global i32 0, align 4
 @j = common global i32 0, align 4
@@ -9,7 +9,7 @@
 ; Function Attrs: nounwind optsize
 define i32 @bnez() #0 {
 entry:
-  %0 = load i32* @i, align 4, !tbaa !1
+  %0 = load i32, i32* @i, align 4, !tbaa !1
   %cmp = icmp eq i32 %0, 0
   br i1 %cmp, label %if.then, label %if.end
 
@@ -31,7 +31,7 @@ if.end:                                           ; preds = %if.then, %entry
 ; Function Attrs: nounwind optsize
 define i32 @beqz() #0 {
 entry:
-  %0 = load i32* @i, align 4, !tbaa !1
+  %0 = load i32, i32* @i, align 4, !tbaa !1
   %cmp = icmp eq i32 %0, 0
   br i1 %cmp, label %if.then, label %if.else
 
@@ -60,8 +60,8 @@ if.end:                                           ; preds = %if.else, %if.then
 ; Function Attrs: nounwind optsize
 define void @bteqz() #0 {
 entry:
-  %0 = load i32* @i, align 4, !tbaa !1
-  %1 = load i32* @j, align 4, !tbaa !1
+  %0 = load i32, i32* @i, align 4, !tbaa !1
+  %1 = load i32, i32* @j, align 4, !tbaa !1
   %cmp = icmp eq i32 %0, %1
   br i1 %cmp, label %if.then, label %if.else
 
@@ -90,15 +90,15 @@ if.end:                                           ; preds = %if.else, %if.then
 ; Function Attrs: nounwind optsize
 define void @btz() #0 {
 entry:
-  %0 = load i32* @i, align 4, !tbaa !1
-  %1 = load i32* @j, align 4, !tbaa !1
+  %0 = load i32, i32* @i, align 4, !tbaa !1
+  %1 = load i32, i32* @j, align 4, !tbaa !1
   %cmp1 = icmp sgt i32 %0, %1
   br i1 %cmp1, label %if.then, label %if.end
 
 if.then:                                          ; preds = %entry, %if.then
   tail call void asm sideeffect ".space 60000", ""() #1, !srcloc !10
-  %2 = load i32* @i, align 4, !tbaa !1
-  %3 = load i32* @j, align 4, !tbaa !1
+  %2 = load i32, i32* @i, align 4, !tbaa !1
+  %3 = load i32, i32* @j, align 4, !tbaa !1
   %cmp = icmp sgt i32 %2, %3
   br i1 %cmp, label %if.then, label %if.end
 
diff --git a/test/CodeGen/Mips/lcb3c.ll b/test/CodeGen/Mips/lcb3c.ll
index eb8329145421..4c6f2c036a0b 100644
--- a/test/CodeGen/Mips/lcb3c.ll
+++ b/test/CodeGen/Mips/lcb3c.ll
@@ -1,4 +1,4 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static -O0    < %s | FileCheck %s -check-prefix=lcb
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static -O0    < %s | FileCheck %s -check-prefix=lcb
 
 @i = global i32 0, align 4
 @j = common global i32 0, align 4
@@ -7,7 +7,7 @@
 ; Function Attrs: nounwind
 define i32 @s() #0 {
 entry:
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %cmp = icmp eq i32 %0, 0
   br i1 %cmp, label %if.then, label %if.else
 
@@ -30,7 +30,7 @@ if.end:                                           ; preds = %if.else, %if.then
 ; Function Attrs: nounwind
 define i32 @b() #0 {
 entry:
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %cmp = icmp eq i32 %0, 0
   br i1 %cmp, label %if.then, label %if.else
 
diff --git a/test/CodeGen/Mips/lcb4a.ll b/test/CodeGen/Mips/lcb4a.ll
index fbcadd2552f8..9e97b5bf1433 100644
--- a/test/CodeGen/Mips/lcb4a.ll
+++ b/test/CodeGen/Mips/lcb4a.ll
@@ -1,4 +1,4 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static     < %s | FileCheck %s -check-prefix=ci
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static     < %s | FileCheck %s -check-prefix=ci
 
 @i = global i32 0, align 4
 @j = common global i32 0, align 4
@@ -7,7 +7,7 @@
 ; Function Attrs: nounwind optsize
 define i32 @foo() #0 {
 entry:
-  %0 = load i32* @i, align 4, !tbaa !1
+  %0 = load i32, i32* @i, align 4, !tbaa !1
   %cmp = icmp eq i32 %0, 0
   br i1 %cmp, label %if.then, label %if.else
 
@@ -32,7 +32,7 @@ if.end:                                           ; preds = %if.else, %if.then
 ; Function Attrs: nounwind optsize
 define i32 @goo() #0 {
 entry:
-  %0 = load i32* @i, align 4, !tbaa !1
+  %0 = load i32, i32* @i, align 4, !tbaa !1
   %cmp = icmp eq i32 %0, 0
   br i1 %cmp, label %if.then, label %if.else
 
diff --git a/test/CodeGen/Mips/lcb5.ll b/test/CodeGen/Mips/lcb5.ll
index b2a8d1d33ef6..41878d5f8817 100644
--- a/test/CodeGen/Mips/lcb5.ll
+++ b/test/CodeGen/Mips/lcb5.ll
@@ -1,4 +1,4 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static     < %s | FileCheck %s -check-prefix=ci
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static     < %s | FileCheck %s -check-prefix=ci
 
 @i = global i32 0, align 4
 @j = common global i32 0, align 4
@@ -7,7 +7,7 @@
 ; Function Attrs: nounwind optsize
 define i32 @x0() #0 {
 entry:
-  %0 = load i32* @i, align 4, !tbaa !1
+  %0 = load i32, i32* @i, align 4, !tbaa !1
   %cmp = icmp eq i32 %0, 0
   br i1 %cmp, label %if.then, label %if.else
 
@@ -33,7 +33,7 @@ if.end:                                           ; preds = %if.else, %if.then
 ; Function Attrs: nounwind optsize
 define i32 @x1() #0 {
 entry:
-  %0 = load i32* @i, align 4, !tbaa !1
+  %0 = load i32, i32* @i, align 4, !tbaa !1
   %cmp = icmp eq i32 %0, 0
   br i1 %cmp, label %if.then, label %if.else
 
@@ -61,7 +61,7 @@ if.end:                                           ; preds = %if.else, %if.then
 ; Function Attrs: nounwind optsize
 define i32 @y0() #0 {
 entry:
-  %0 = load i32* @i, align 4, !tbaa !1
+  %0 = load i32, i32* @i, align 4, !tbaa !1
   %cmp = icmp eq i32 %0, 0
   br i1 %cmp, label %if.then, label %if.else
 
@@ -86,7 +86,7 @@ if.end:                                           ; preds = %if.else, %if.then
 ; Function Attrs: nounwind optsize
 define i32 @y1() #0 {
 entry:
-  %0 = load i32* @i, align 4, !tbaa !1
+  %0 = load i32, i32* @i, align 4, !tbaa !1
   %cmp = icmp eq i32 %0, 0
   br i1 %cmp, label %if.then, label %if.else
 
@@ -114,8 +114,8 @@ if.end:                                           ; preds = %if.else, %if.then
 ; Function Attrs: nounwind optsize
 define void @z0() #0 {
 entry:
-  %0 = load i32* @i, align 4, !tbaa !1
-  %1 = load i32* @j, align 4, !tbaa !1
+  %0 = load i32, i32* @i, align 4, !tbaa !1
+  %1 = load i32, i32* @j, align 4, !tbaa !1
   %cmp = icmp eq i32 %0, %1
   br i1 %cmp, label %if.then, label %if.else
 
@@ -140,8 +140,8 @@ if.end:                                           ; preds = %if.else, %if.then
 ; Function Attrs: nounwind optsize
 define void @z1() #0 {
 entry:
-  %0 = load i32* @i, align 4, !tbaa !1
-  %1 = load i32* @j, align 4, !tbaa !1
+  %0 = load i32, i32* @i, align 4, !tbaa !1
+  %1 = load i32, i32* @j, align 4, !tbaa !1
   %cmp = icmp eq i32 %0, %1
   br i1 %cmp, label %if.then, label %if.else
 
@@ -169,15 +169,15 @@ if.end:                                           ; preds = %if.else, %if.then
 ; Function Attrs: nounwind optsize
 define void @z3() #0 {
 entry:
-  %0 = load i32* @i, align 4, !tbaa !1
-  %1 = load i32* @j, align 4, !tbaa !1
+  %0 = load i32, i32* @i, align 4, !tbaa !1
+  %1 = load i32, i32* @j, align 4, !tbaa !1
   %cmp1 = icmp sgt i32 %0, %1
   br i1 %cmp1, label %if.then, label %if.end
 
 if.then:                                          ; preds = %entry, %if.then
   tail call void asm sideeffect ".space 10000", ""() #1, !srcloc !17
-  %2 = load i32* @i, align 4, !tbaa !1
-  %3 = load i32* @j, align 4, !tbaa !1
+  %2 = load i32, i32* @i, align 4, !tbaa !1
+  %3 = load i32, i32* @j, align 4, !tbaa !1
   %cmp = icmp sgt i32 %2, %3
   br i1 %cmp, label %if.then, label %if.end
 
@@ -192,15 +192,15 @@ if.end:                                           ; preds = %if.then, %entry
 ; Function Attrs: nounwind optsize
 define void @z4() #0 {
 entry:
-  %0 = load i32* @i, align 4, !tbaa !1
-  %1 = load i32* @j, align 4, !tbaa !1
+  %0 = load i32, i32* @i, align 4, !tbaa !1
+  %1 = load i32, i32* @j, align 4, !tbaa !1
   %cmp1 = icmp sgt i32 %0, %1
   br i1 %cmp1, label %if.then, label %if.end
 
 if.then:                                          ; preds = %entry, %if.then
   tail call void asm sideeffect ".space 10000000", ""() #1, !srcloc !18
-  %2 = load i32* @i, align 4, !tbaa !1
-  %3 = load i32* @j, align 4, !tbaa !1
+  %2 = load i32, i32* @i, align 4, !tbaa !1
+  %3 = load i32, i32* @j, align 4, !tbaa !1
   %cmp = icmp sgt i32 %2, %3
   br i1 %cmp, label %if.then, label %if.end
 
diff --git a/test/CodeGen/Mips/lh1.ll b/test/CodeGen/Mips/lh1.ll
index 1f95b0903466..31967e5a5379 100644
--- a/test/CodeGen/Mips/lh1.ll
+++ b/test/CodeGen/Mips/lh1.ll
@@ -6,12 +6,12 @@
 define i32 @main() nounwind {
 entry:
   %i = alloca i32, align 4
-  %0 = load i16* @s, align 2
+  %0 = load i16, i16* @s, align 2
   %conv = sext i16 %0 to i32
 ; 16:	lh	${{[0-9]+}}, 0(${{[0-9]+}})
   store i32 %conv, i32* %i, align 4
-  %1 = load i32* %i, align 4
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i32 0, i32 0), i32 %1)
+  %1 = load i32, i32* %i, align 4
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([5 x i8], [5 x i8]* @.str, i32 0, i32 0), i32 %1)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/lhu1.ll b/test/CodeGen/Mips/lhu1.ll
index 0cfcede669e0..413da46d4a31 100644
--- a/test/CodeGen/Mips/lhu1.ll
+++ b/test/CodeGen/Mips/lhu1.ll
@@ -7,12 +7,12 @@
 define i32 @main() nounwind {
 entry:
   %i = alloca i32, align 4
-  %0 = load i16* @s, align 2
+  %0 = load i16, i16* @s, align 2
   %conv = zext i16 %0 to i32
 ; 16:	lhu	${{[0-9]+}}, 0(${{[0-9]+}})
   store i32 %conv, i32* %i, align 4
-  %1 = load i32* %i, align 4
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i32 0, i32 0), i32 %1)
+  %1 = load i32, i32* %i, align 4
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([5 x i8], [5 x i8]* @.str, i32 0, i32 0), i32 %1)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/llcarry.ll b/test/CodeGen/Mips/llcarry.ll
index 7763daec3b32..f4120ecec175 100644
--- a/test/CodeGen/Mips/llcarry.ll
+++ b/test/CodeGen/Mips/llcarry.ll
@@ -9,8 +9,8 @@
 
 define void @test1() nounwind {
 entry:
-  %0 = load i64* @i, align 8
-  %1 = load i64* @j, align 8
+  %0 = load i64, i64* @i, align 8
+  %1 = load i64, i64* @j, align 8
   %add = add nsw i64 %1, %0
   store i64 %add, i64* @k, align 8
 ; 16:	addu	${{[0-9]+}}, ${{[0-9]+}}, ${{[0-9]+}}
@@ -23,8 +23,8 @@ entry:
 
 define void @test2() nounwind {
 entry:
-  %0 = load i64* @i, align 8
-  %1 = load i64* @j, align 8
+  %0 = load i64, i64* @i, align 8
+  %1 = load i64, i64* @j, align 8
   %sub = sub nsw i64 %0, %1
 ; 16:	subu	${{[0-9]+}}, ${{[0-9]+}}, ${{[0-9]+}}
 ; 16:	sltu	${{[0-9]+}}, ${{[0-9]+}}
@@ -37,7 +37,7 @@ entry:
 
 define void @test3() nounwind {
 entry:
-  %0 = load i64* @ii, align 8
+  %0 = load i64, i64* @ii, align 8
   %add = add nsw i64 %0, 15
 ; 16:	addiu	${{[0-9]+}}, 15
 ; 16:	sltu	${{[0-9]+}}, ${{[0-9]+}}
diff --git a/test/CodeGen/Mips/llvm-ir/add.ll b/test/CodeGen/Mips/llvm-ir/add.ll
index 83774eda634f..6cccc7df19f9 100644
--- a/test/CodeGen/Mips/llvm-ir/add.ll
+++ b/test/CodeGen/Mips/llvm-ir/add.ll
@@ -4,6 +4,10 @@
 ; RUN:    -check-prefix=ALL -check-prefix=NOT-R2-R6 -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=R2-R6 -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=R2-R6 -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=R2-R6 -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=R2-R6 -check-prefix=GP32
 ; RUN: llc < %s -march=mips64 -mcpu=mips3 | FileCheck %s \
@@ -14,6 +18,10 @@
 ; RUN:    -check-prefix=ALL -check-prefix=NOT-R2-R6 -check-prefix=GP64
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=R2-R6 -check-prefix=GP64
+; RUN: llc < %s -march=mips64 -mcpu=mips64r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=R2-R6 -check-prefix=GP64
+; RUN: llc < %s -march=mips64 -mcpu=mips64r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=R2-R6 -check-prefix=GP64
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=R2-R6 -check-prefix=GP64
 
diff --git a/test/CodeGen/Mips/llvm-ir/and.ll b/test/CodeGen/Mips/llvm-ir/and.ll
index 09d0ef9238af..8ebcfe4a3f64 100644
--- a/test/CodeGen/Mips/llvm-ir/and.ll
+++ b/test/CodeGen/Mips/llvm-ir/and.ll
@@ -4,6 +4,10 @@
 ; RUN:    -check-prefix=ALL -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP32
 ; RUN: llc < %s -march=mips64 -mcpu=mips3 | FileCheck %s \
@@ -14,6 +18,10 @@
 ; RUN:    -check-prefix=ALL -check-prefix=GP64
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64
+; RUN: llc < %s -march=mips64 -mcpu=mips64r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP64
+; RUN: llc < %s -march=mips64 -mcpu=mips64r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP64
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64
 
@@ -51,10 +59,7 @@ define signext i32 @and_i32(i32 signext %a, i32 signext %b) {
 entry:
 ; ALL-LABEL: and_i32:
 
-  ; GP32:         and     $2, $4, $5
-
-  ; GP64:         and     $[[T0:[0-9]+]], $4, $5
-  ; GP64:         sll     $2, $[[T0]], 0
+  ; ALL:          and     $2, $4, $5
 
   %r = and i32 %a, %b
   ret i32 %r
diff --git a/test/CodeGen/Mips/llvm-ir/ashr.ll b/test/CodeGen/Mips/llvm-ir/ashr.ll
index 415998929aa0..cad4a39d7743 100644
--- a/test/CodeGen/Mips/llvm-ir/ashr.ll
+++ b/test/CodeGen/Mips/llvm-ir/ashr.ll
@@ -1,30 +1,42 @@
 ; RUN: llc < %s -march=mips -mcpu=mips2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP32 \
-; RUN:    -check-prefix=M2 -check-prefix=NOT-R2-R6
+; RUN:    -check-prefix=M2
 ; RUN: llc < %s -march=mips -mcpu=mips32 | FileCheck %s \
-; RUN:    -check-prefix=ALL -check-prefix=GP32 -check-prefix=NOT-R2-R6 \
-; RUN:    -check-prefix=32R1-R2
+; RUN:    -check-prefix=ALL -check-prefix=GP32 \
+; RUN:    -check-prefix=32R1-R5
 ; RUN: llc < %s -march=mips -mcpu=mips32r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP32 \
-; RUN:    -check-prefix=32R1-R2 -check-prefix=R2-R6
+; RUN:    -check-prefix=32R1-R5
+; RUN: llc < %s -march=mips -mcpu=mips32r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP32 \
+; RUN:    -check-prefix=32R1-R5
+; RUN: llc < %s -march=mips -mcpu=mips32r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP32 \
+; RUN:    -check-prefix=32R1-R5
 ; RUN: llc < %s -march=mips -mcpu=mips32r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP32 \
-; RUN:    -check-prefix=32R6 -check-prefix=R2-R6
+; RUN:    -check-prefix=32R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips3 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64 \
-; RUN:    -check-prefix=M3 -check-prefix=NOT-R2-R6
+; RUN:    -check-prefix=M3
 ; RUN: llc < %s -march=mips64 -mcpu=mips4 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64 \
-; RUN:    -check-prefix=GP64-NOT-R6 -check-prefix=NOT-R2-R6
+; RUN:    -check-prefix=GP64-NOT-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64 \
-; RUN:    -check-prefix=GP64-NOT-R6 -check-prefix=NOT-R2-R6
+; RUN:    -check-prefix=GP64-NOT-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64 \
-; RUN:    -check-prefix=GP64-NOT-R6 -check-prefix R2-R6
+; RUN:    -check-prefix=GP64-NOT-R6
+; RUN: llc < %s -march=mips64 -mcpu=mips64r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP64 \
+; RUN:    -check-prefix=GP64-NOT-R6
+; RUN: llc < %s -march=mips64 -mcpu=mips64r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP64 \
+; RUN:    -check-prefix=GP64-NOT-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64 \
-; RUN:    -check-prefix=64R6 -check-prefix=R2-R6
+; RUN:    -check-prefix=64R6
 
 define signext i1 @ashr_i1(i1 signext %a, i1 signext %b) {
 entry:
@@ -91,17 +103,17 @@ entry:
   ; M2:         jr        $ra
   ; M2:         nop
 
-  ; 32R1-R2:    srlv      $[[T0:[0-9]+]], $5, $7
-  ; 32R1-R2:    not       $[[T1:[0-9]+]], $7
-  ; 32R1-R2:    sll       $[[T2:[0-9]+]], $4, 1
-  ; 32R1-R2:    sllv      $[[T3:[0-9]+]], $[[T2]], $[[T1]]
-  ; 32R1-R2:    or        $3, $[[T3]], $[[T0]]
-  ; 32R1-R2:    srav      $[[T4:[0-9]+]], $4, $7
-  ; 32R1-R2:    andi      $[[T5:[0-9]+]], $7, 32
-  ; 32R1-R2:    movn      $3, $[[T4]], $[[T5]]
-  ; 32R1-R2:    sra       $4, $4, 31
-  ; 32R1-R2:    jr        $ra
-  ; 32R1-R2:    movn      $2, $4, $[[T5]]
+  ; 32R1-R5:    srlv      $[[T0:[0-9]+]], $5, $7
+  ; 32R1-R5:    not       $[[T1:[0-9]+]], $7
+  ; 32R1-R5:    sll       $[[T2:[0-9]+]], $4, 1
+  ; 32R1-R5:    sllv      $[[T3:[0-9]+]], $[[T2]], $[[T1]]
+  ; 32R1-R5:    or        $3, $[[T3]], $[[T0]]
+  ; 32R1-R5:    srav      $[[T4:[0-9]+]], $4, $7
+  ; 32R1-R5:    andi      $[[T5:[0-9]+]], $7, 32
+  ; 32R1-R5:    movn      $3, $[[T4]], $[[T5]]
+  ; 32R1-R5:    sra       $4, $4, 31
+  ; 32R1-R5:    jr        $ra
+  ; 32R1-R5:    movn      $2, $4, $[[T5]]
 
   ; 32R6:       srav      $[[T0:[0-9]+]], $4, $7
   ; 32R6:       andi      $[[T1:[0-9]+]], $7, 32
@@ -119,9 +131,7 @@ entry:
   ; 32R6:       jr        $ra
   ; 32R6:       or        $3, $[[T0]], $[[T11]]
 
-  ; FIXME: The sll instruction below is redundant.
-  ; GP64:       sll       $[[T0:[0-9]+]], $5, 0
-  ; GP64:       dsrav     $2, $4, $[[T0]]
+  ; GP64:       dsrav     $2, $4, $5
 
   %r = ashr i64 %a, %b
   ret i64 %r
@@ -134,11 +144,11 @@ entry:
   ; GP32:           lw        $25, %call16(__ashrti3)($gp)
 
   ; M3:             sll       $[[T0:[0-9]+]], $7, 0
-  ; M3:             dsrav     $[[T1:[0-9]+]], $4, $[[T0]]
+  ; M3:             dsrav     $[[T1:[0-9]+]], $4, $7
   ; M3:             andi      $[[T2:[0-9]+]], $[[T0]], 64
   ; M3:             bnez      $[[T3:[0-9]+]], $[[BB0:BB[0-9_]+]]
   ; M3:             move      $3, $[[T1]]
-  ; M3:             dsrlv     $[[T4:[0-9]+]], $5, $[[T0]]
+  ; M3:             dsrlv     $[[T4:[0-9]+]], $5, $7
   ; M3:             dsll      $[[T5:[0-9]+]], $4, 1
   ; M3:             not       $[[T6:[0-9]+]], $[[T0]]
   ; M3:             dsllv     $[[T7:[0-9]+]], $[[T5]], $[[T6]]
@@ -151,35 +161,34 @@ entry:
   ; M3:             jr        $ra
   ; M3:             nop
 
-  ; GP64-NOT-R6:    sll       $[[T0:[0-9]+]], $7, 0
-  ; GP64-NOT-R6:    dsrlv     $[[T1:[0-9]+]], $5, $[[T0]]
-  ; GP64-NOT-R6:    dsll      $[[T2:[0-9]+]], $4, 1
-  ; GP64-NOT-R6:    not       $[[T3:[0-9]+]], $[[T0]]
-  ; GP64-NOT-R6:    dsllv     $[[T4:[0-9]+]], $[[T2]], $[[T3]]
-  ; GP64-NOT-R6:    or        $3, $[[T4]], $[[T1]]
-  ; GP64-NOT-R6:    dsrav     $2, $4, $[[T0]]
-  ; GP64-NOT-R6:    andi      $[[T5:[0-9]+]], $[[T0]], 64
-
+  ; GP64-NOT-R6:    dsrlv     $[[T0:[0-9]+]], $5, $7
+  ; GP64-NOT-R6:    dsll      $[[T1:[0-9]+]], $4, 1
+  ; GP64-NOT-R6:    sll       $[[T2:[0-9]+]], $7, 0
+  ; GP64-NOT-R6:    not       $[[T3:[0-9]+]], $[[T2]]
+  ; GP64-NOT-R6:    dsllv     $[[T4:[0-9]+]], $[[T1]], $[[T3]]
+  ; GP64-NOT-R6:    or        $3, $[[T4]], $[[T0]]
+  ; GP64-NOT-R6:    dsrav     $2, $4, $7
+  ; GP64-NOT-R6:    andi      $[[T5:[0-9]+]], $[[T2]], 64
   ; GP64-NOT-R6:    movn      $3, $2, $[[T5]]
   ; GP64-NOT-R6:    dsra      $[[T6:[0-9]+]], $4, 63
   ; GP64-NOT-R6:    jr        $ra
   ; GP64-NOT-R6:    movn      $2, $[[T6]], $[[T5]]
 
-  ; 64R6:           sll       $[[T0:[0-9]+]], $7, 0
-  ; 64R6:           dsrav     $[[T1:[0-9]+]], $4, $[[T0]]
-  ; 64R6:           andi      $[[T2:[0-9]+]], $[[T0]], 64
+  ; 64R6:           dsrav     $[[T0:[0-9]+]], $4, $7
+  ; 64R6:           sll       $[[T1:[0-9]+]], $7, 0
+  ; 64R6:           andi      $[[T2:[0-9]+]], $[[T1]], 64
   ; 64R6:           sll       $[[T3:[0-9]+]], $[[T2]], 0
-  ; 64R6:           seleqz    $[[T4:[0-9]+]], $[[T1]], $[[T3]]
+  ; 64R6:           seleqz    $[[T4:[0-9]+]], $[[T0]], $[[T3]]
   ; 64R6:           dsra      $[[T5:[0-9]+]], $4, 63
   ; 64R6:           selnez    $[[T6:[0-9]+]], $[[T5]], $[[T3]]
   ; 64R6:           or        $2, $[[T6]], $[[T4]]
-  ; 64R6:           dsrlv     $[[T7:[0-9]+]], $5, $[[T0]]
+  ; 64R6:           dsrlv     $[[T7:[0-9]+]], $5, $7
   ; 64R6:           dsll      $[[T8:[0-9]+]], $4, 1
-  ; 64R6:           not       $[[T9:[0-9]+]], $[[T0]]
+  ; 64R6:           not       $[[T9:[0-9]+]], $[[T1]]
   ; 64R6:           dsllv     $[[T10:[0-9]+]], $[[T8]], $[[T9]]
   ; 64R6:           or        $[[T11:[0-9]+]], $[[T10]], $[[T7]]
   ; 64R6:           seleqz    $[[T12:[0-9]+]], $[[T11]], $[[T3]]
-  ; 64R6:           selnez    $[[T13:[0-9]+]], $[[T1]], $[[T3]]
+  ; 64R6:           selnez    $[[T13:[0-9]+]], $[[T0]], $[[T3]]
   ; 64R6:           jr        $ra
   ; 64R6:           or        $3, $[[T13]], $[[T12]]
 
diff --git a/test/CodeGen/Mips/llvm-ir/call.ll b/test/CodeGen/Mips/llvm-ir/call.ll
index 4cbf43cae28e..112ab8ee8c7f 100644
--- a/test/CodeGen/Mips/llvm-ir/call.ll
+++ b/test/CodeGen/Mips/llvm-ir/call.ll
@@ -3,10 +3,14 @@
 ; FIXME: We should remove the need for -enable-mips-tail-calls
 ; RUN: llc -march=mips   -mcpu=mips32   -enable-mips-tail-calls < %s | FileCheck %s -check-prefix=ALL -check-prefix=O32
 ; RUN: llc -march=mips   -mcpu=mips32r2 -enable-mips-tail-calls < %s | FileCheck %s -check-prefix=ALL -check-prefix=O32
+; RUN: llc -march=mips   -mcpu=mips32r3 -enable-mips-tail-calls < %s | FileCheck %s -check-prefix=ALL -check-prefix=O32
+; RUN: llc -march=mips   -mcpu=mips32r5 -enable-mips-tail-calls < %s | FileCheck %s -check-prefix=ALL -check-prefix=O32
 ; RUN: llc -march=mips   -mcpu=mips32r6 -enable-mips-tail-calls < %s | FileCheck %s -check-prefix=ALL -check-prefix=O32
 ; RUN: llc -march=mips64 -mcpu=mips4    -enable-mips-tail-calls < %s | FileCheck %s -check-prefix=ALL -check-prefix=N64
 ; RUN: llc -march=mips64 -mcpu=mips64   -enable-mips-tail-calls < %s | FileCheck %s -check-prefix=ALL -check-prefix=N64
 ; RUN: llc -march=mips64 -mcpu=mips64r2 -enable-mips-tail-calls < %s | FileCheck %s -check-prefix=ALL -check-prefix=N64
+; RUN: llc -march=mips64 -mcpu=mips64r3 -enable-mips-tail-calls < %s | FileCheck %s -check-prefix=ALL -check-prefix=N64
+; RUN: llc -march=mips64 -mcpu=mips64r5 -enable-mips-tail-calls < %s | FileCheck %s -check-prefix=ALL -check-prefix=N64
 ; RUN: llc -march=mips64 -mcpu=mips64r6 -enable-mips-tail-calls < %s | FileCheck %s -check-prefix=ALL -check-prefix=N64
 
 declare void @extern_void_void()
diff --git a/test/CodeGen/Mips/llvm-ir/indirectbr.ll b/test/CodeGen/Mips/llvm-ir/indirectbr.ll
index d8fd78774553..debfeb35b213 100644
--- a/test/CodeGen/Mips/llvm-ir/indirectbr.ll
+++ b/test/CodeGen/Mips/llvm-ir/indirectbr.ll
@@ -2,10 +2,14 @@
 
 ; RUN: llc -march=mips   -mcpu=mips32   -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=NOT-R6
 ; RUN: llc -march=mips   -mcpu=mips32r2 -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=NOT-R6
+; RUN: llc -march=mips   -mcpu=mips32r3 -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=NOT-R6
+; RUN: llc -march=mips   -mcpu=mips32r5 -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=NOT-R6
 ; RUN: llc -march=mips   -mcpu=mips32r6 -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=R6
 ; RUN: llc -march=mips64 -mcpu=mips4    -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=NOT-R6
 ; RUN: llc -march=mips64 -mcpu=mips64   -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=NOT-R6
 ; RUN: llc -march=mips64 -mcpu=mips64r2 -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=NOT-R6
+; RUN: llc -march=mips64 -mcpu=mips64r3 -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=NOT-R6
+; RUN: llc -march=mips64 -mcpu=mips64r5 -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=NOT-R6
 ; RUN: llc -march=mips64 -mcpu=mips64r6 -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=R6
 
 define i32 @br(i8 *%addr) {
diff --git a/test/CodeGen/Mips/llvm-ir/lshr.ll b/test/CodeGen/Mips/llvm-ir/lshr.ll
index 59f4330dde6c..3a7029fa5b7a 100644
--- a/test/CodeGen/Mips/llvm-ir/lshr.ll
+++ b/test/CodeGen/Mips/llvm-ir/lshr.ll
@@ -1,30 +1,42 @@
 ; RUN: llc < %s -march=mips -mcpu=mips2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP32 \
-; RUN:    -check-prefix=M2 -check-prefix=NOT-R2-R6
+; RUN:    -check-prefix=M2
 ; RUN: llc < %s -march=mips -mcpu=mips32 | FileCheck %s \
-; RUN:    -check-prefix=ALL -check-prefix=GP32 -check-prefix=NOT-R2-R6 \
-; RUN:    -check-prefix=32R1-R2
+; RUN:    -check-prefix=ALL -check-prefix=GP32 \
+; RUN:    -check-prefix=32R1-R5
 ; RUN: llc < %s -march=mips -mcpu=mips32r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP32 \
-; RUN:    -check-prefix=32R1-R2 -check-prefix=R2-R6
+; RUN:    -check-prefix=32R1-R5
+; RUN: llc < %s -march=mips -mcpu=mips32r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP32 \
+; RUN:    -check-prefix=32R1-R5
+; RUN: llc < %s -march=mips -mcpu=mips32r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP32 \
+; RUN:    -check-prefix=32R1-R5
 ; RUN: llc < %s -march=mips -mcpu=mips32r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP32 \
-; RUN:    -check-prefix=32R6 -check-prefix=R2-R6
+; RUN:    -check-prefix=32R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips3 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64 \
-; RUN:    -check-prefix=M3 -check-prefix=NOT-R2-R6
+; RUN:    -check-prefix=M3
 ; RUN: llc < %s -march=mips64 -mcpu=mips4 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64 \
-; RUN:    -check-prefix=GP64-NOT-R6 -check-prefix=NOT-R2-R6
+; RUN:    -check-prefix=GP64-NOT-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64 \
-; RUN:    -check-prefix=GP64-NOT-R6 -check-prefix=NOT-R2-R6
+; RUN:    -check-prefix=GP64-NOT-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64 \
-; RUN:    -check-prefix=GP64-NOT-R6 -check-prefix R2-R6
+; RUN:    -check-prefix=GP64-NOT-R6
+; RUN: llc < %s -march=mips64 -mcpu=mips64r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP64 \
+; RUN:    -check-prefix=GP64-NOT-R6
+; RUN: llc < %s -march=mips64 -mcpu=mips64r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP64 \
+; RUN:    -check-prefix=GP64-NOT-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64 \
-; RUN:    -check-prefix=64R6 -check-prefix=R2-R6
+; RUN:    -check-prefix=64R6
 
 define signext i1 @lshr_i1(i1 signext %a, i1 signext %b) {
 entry:
@@ -89,16 +101,16 @@ entry:
   ; M2:         jr        $ra
   ; M2:         nop
 
-  ; 32R1-R2:    srlv      $[[T0:[0-9]+]], $5, $7
-  ; 32R1-R2:    not       $[[T1:[0-9]+]], $7
-  ; 32R1-R2:    sll       $[[T2:[0-9]+]], $4, 1
-  ; 32R1-R2:    sllv      $[[T3:[0-9]+]], $[[T2]], $[[T1]]
-  ; 32R1-R2:    or        $3, $[[T3]], $[[T0]]
-  ; 32R1-R2:    srlv      $[[T4:[0-9]+]], $4, $7
-  ; 32R1-R2:    andi      $[[T5:[0-9]+]], $7, 32
-  ; 32R1-R2:    movn      $3, $[[T4]], $[[T5]]
-  ; 32R1-R2:    jr        $ra
-  ; 32R1-R2:    movn      $2, $zero, $[[T5]]
+  ; 32R1-R5:    srlv      $[[T0:[0-9]+]], $5, $7
+  ; 32R1-R5:    not       $[[T1:[0-9]+]], $7
+  ; 32R1-R5:    sll       $[[T2:[0-9]+]], $4, 1
+  ; 32R1-R5:    sllv      $[[T3:[0-9]+]], $[[T2]], $[[T1]]
+  ; 32R1-R5:    or        $3, $[[T3]], $[[T0]]
+  ; 32R1-R5:    srlv      $[[T4:[0-9]+]], $4, $7
+  ; 32R1-R5:    andi      $[[T5:[0-9]+]], $7, 32
+  ; 32R1-R5:    movn      $3, $[[T4]], $[[T5]]
+  ; 32R1-R5:    jr        $ra
+  ; 32R1-R5:    movn      $2, $zero, $[[T5]]
 
   ; 32R6:       srlv      $[[T0:[0-9]+]], $5, $7
   ; 32R6:       not       $[[T1:[0-9]+]], $7
@@ -113,8 +125,7 @@ entry:
   ; 32R6:       jr        $ra
   ; 32R6:       seleqz    $2, $[[T7]], $[[T5]]
 
-  ; GP64:         sll     $[[T0:[0-9]+]], $5, 0
-  ; GP64:         dsrlv   $2, $4, $[[T0]]
+  ; GP64:         dsrlv   $2, $4, $5
 
   %r = lshr i64 %a, %b
   ret i64 %r
@@ -127,11 +138,11 @@ entry:
   ; GP32:         lw      $25, %call16(__lshrti3)($gp)
 
   ; M3:             sll       $[[T0:[0-9]+]], $7, 0
-  ; M3:             dsrlv     $[[T1:[0-9]+]], $4, $[[T0]]
+  ; M3:             dsrlv     $[[T1:[0-9]+]], $4, $7
   ; M3:             andi      $[[T2:[0-9]+]], $[[T0]], 64
   ; M3:             bnez      $[[T3:[0-9]+]], $[[BB0:BB[0-9_]+]]
   ; M3:             move      $3, $[[T1]]
-  ; M3:             dsrlv     $[[T4:[0-9]+]], $5, $[[T0]]
+  ; M3:             dsrlv     $[[T4:[0-9]+]], $5, $7
   ; M3:             dsll      $[[T5:[0-9]+]], $4, 1
   ; M3:             not       $[[T6:[0-9]+]], $[[T0]]
   ; M3:             dsllv     $[[T7:[0-9]+]], $[[T5]], $[[T6]]
@@ -144,32 +155,32 @@ entry:
   ; M3:             jr        $ra
   ; M3:             nop
 
-  ; GP64-NOT-R6:    sll       $[[T0:[0-9]+]], $7, 0
-  ; GP64-NOT-R6:    dsrlv     $[[T1:[0-9]+]], $5, $[[T0]]
-  ; GP64-NOT-R6:    dsll      $[[T2:[0-9]+]], $4, 1
-  ; GP64-NOT-R6:    not       $[[T3:[0-9]+]], $[[T0]]
-  ; GP64-NOT-R6:    dsllv     $[[T4:[0-9]+]], $[[T2]], $[[T3]]
-  ; GP64-NOT-R6:    or        $3, $[[T4]], $[[T1]]
-  ; GP64-NOT-R6:    dsrlv     $2, $4, $[[T0]]
-  ; GP64-NOT-R6:    andi      $[[T5:[0-9]+]], $[[T0]], 64
+  ; GP64-NOT-R6:    dsrlv     $[[T0:[0-9]+]], $5, $7
+  ; GP64-NOT-R6:    dsll      $[[T1:[0-9]+]], $4, 1
+  ; GP64-NOT-R6:    sll       $[[T2:[0-9]+]], $7, 0
+  ; GP64-NOT-R6:    not       $[[T3:[0-9]+]], $[[T2]]
+  ; GP64-NOT-R6:    dsllv     $[[T4:[0-9]+]], $[[T1]], $[[T3]]
+  ; GP64-NOT-R6:    or        $3, $[[T4]], $[[T0]]
+  ; GP64-NOT-R6:    dsrlv     $2, $4, $7
+  ; GP64-NOT-R6:    andi      $[[T5:[0-9]+]], $[[T2]], 64
   ; GP64-NOT-R6:    movn      $3, $2, $[[T5]]
   ; GP64-NOT-R6:    jr        $ra
   ; GP64-NOT-R6:    movn      $2, $zero, $1
 
-  ; 64R6:           sll       $[[T0:[0-9]+]], $7, 0
-  ; 64R6:           dsrlv     $[[T1:[0-9]+]], $5, $[[T0]]
-  ; 64R6:           dsll      $[[T2:[0-9]+]], $4, 1
-  ; 64R6:           not       $[[T3:[0-9]+]], $[[T0]]
-  ; 64R6:           dsllv     $[[T4:[0-9]+]], $[[T2]], $[[T3]]
-  ; 64R6:           or        $[[T5:[0-9]+]], $[[T4]], $[[T1]]
-  ; 64R6:           andi      $[[T6:[0-9]+]], $[[T0]], 64
+  ; 64R6:           dsrlv     $[[T0:[0-9]+]], $5, $7
+  ; 64R6:           dsll      $[[T1:[0-9]+]], $4, 1
+  ; 64R6:           sll       $[[T2:[0-9]+]], $7, 0
+  ; 64R6:           not       $[[T3:[0-9]+]], $[[T2]]
+  ; 64R6:           dsllv     $[[T4:[0-9]+]], $[[T1]], $[[T3]]
+  ; 64R6:           or        $[[T5:[0-9]+]], $[[T4]], $[[T0]]
+  ; 64R6:           andi      $[[T6:[0-9]+]], $[[T2]], 64
   ; 64R6:           sll       $[[T7:[0-9]+]], $[[T6]], 0
   ; 64R6:           seleqz    $[[T8:[0-9]+]], $[[T5]], $[[T7]]
-  ; 64R6:           dsrlv     $[[T9:[0-9]+]], $4, $[[T0]]
+  ; 64R6:           dsrlv     $[[T9:[0-9]+]], $4, $7
   ; 64R6:           selnez    $[[T10:[0-9]+]], $[[T9]], $[[T7]]
   ; 64R6:           or        $3, $[[T10]], $[[T8]]
   ; 64R6:           jr        $ra
-  ; 64R6:           seleqz    $2, $[[T0]], $[[T7]]
+  ; 64R6:           seleqz    $2, $[[T9]], $[[T7]]
 
   %r = lshr i128 %a, %b
   ret i128 %r
diff --git a/test/CodeGen/Mips/llvm-ir/mul.ll b/test/CodeGen/Mips/llvm-ir/mul.ll
index 5f7f338c7789..a7582805dd74 100644
--- a/test/CodeGen/Mips/llvm-ir/mul.ll
+++ b/test/CodeGen/Mips/llvm-ir/mul.ll
@@ -1,17 +1,25 @@
 ; RUN: llc < %s -march=mips -mcpu=mips2 | FileCheck %s -check-prefix=ALL \
 ; RUN:    -check-prefix=M2 -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32 | FileCheck %s -check-prefix=ALL \
-; RUN:    -check-prefix=32R1-R2 -check-prefix=GP32
+; RUN:    -check-prefix=32R1-R5 -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r2 | FileCheck %s -check-prefix=ALL \
-; RUN:    -check-prefix=32R1-R2 -check-prefix=32R2 -check-prefix=GP32
+; RUN:    -check-prefix=32R1-R5 -check-prefix=32R2-R5 -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r3 | FileCheck %s -check-prefix=ALL \
+; RUN:    -check-prefix=32R1-R5 -check-prefix=32R2-R5 -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r5 | FileCheck %s -check-prefix=ALL \
+; RUN:    -check-prefix=32R1-R5 -check-prefix=32R2-R5 -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r6 | FileCheck %s -check-prefix=ALL \
 ; RUN:    -check-prefix=32R6 -check-prefix=GP32
 ; RUN: llc < %s -march=mips64 -mcpu=mips4 | FileCheck %s -check-prefix=ALL \
 ; RUN:    -check-prefix=M4 -check-prefix=GP64-NOT-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64 | FileCheck %s -check-prefix=ALL \
-; RUN:    -check-prefix=64R1-R2 -check-prefix=GP64-NOT-R6
+; RUN:    -check-prefix=64R1-R5 -check-prefix=GP64-NOT-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r2 | FileCheck %s -check-prefix=ALL \
-; RUN:    -check-prefix=64R1-R2 -check-prefix=GP64 -check-prefix=GP64-NOT-R6
+; RUN:    -check-prefix=64R1-R5 -check-prefix=GP64 -check-prefix=GP64-NOT-R6
+; RUN: llc < %s -march=mips64 -mcpu=mips64r3 | FileCheck %s -check-prefix=ALL \
+; RUN:    -check-prefix=64R1-R5 -check-prefix=GP64 -check-prefix=GP64-NOT-R6
+; RUN: llc < %s -march=mips64 -mcpu=mips64r5 | FileCheck %s -check-prefix=ALL \
+; RUN:    -check-prefix=64R1-R5 -check-prefix=GP64 -check-prefix=GP64-NOT-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r6 | FileCheck %s -check-prefix=ALL \
 ; RUN:    -check-prefix=64R6
 
@@ -24,9 +32,9 @@ entry:
   ; M2:         sll     $[[T0]], $[[T0]], 31
   ; M2:         sra     $2, $[[T0]], 31
 
-  ; 32R1-R2:    mul     $[[T0:[0-9]+]], $4, $5
-  ; 32R1-R2:    sll     $[[T0]], $[[T0]], 31
-  ; 32R1-R2:    sra     $2, $[[T0]], 31
+  ; 32R1-R5:    mul     $[[T0:[0-9]+]], $4, $5
+  ; 32R1-R5:    sll     $[[T0]], $[[T0]], 31
+  ; 32R1-R5:    sra     $2, $[[T0]], 31
 
   ; 32R6:       mul     $[[T0:[0-9]+]], $4, $5
   ; 32R6:       sll     $[[T0]], $[[T0]], 31
@@ -37,9 +45,9 @@ entry:
   ; M4:         sll     $[[T0]], $[[T0]], 31
   ; M4:         sra     $2, $[[T0]], 31
 
-  ; 64R1-R2:    mul     $[[T0:[0-9]+]], $4, $5
-  ; 64R1-R2:    sll     $[[T0]], $[[T0]], 31
-  ; 64R1-R2:    sra     $2, $[[T0]], 31
+  ; 64R1-R5:    mul     $[[T0:[0-9]+]], $4, $5
+  ; 64R1-R5:    sll     $[[T0]], $[[T0]], 31
+  ; 64R1-R5:    sra     $2, $[[T0]], 31
 
   ; 64R6:       mul     $[[T0:[0-9]+]], $4, $5
   ; 64R6:       sll     $[[T0]], $[[T0]], 31
@@ -62,8 +70,8 @@ entry:
   ; 32R1:       sll     $[[T0]], $[[T0]], 24
   ; 32R1:       sra     $2, $[[T0]], 24
 
-  ; 32R2:       mul     $[[T0:[0-9]+]], $4, $5
-  ; 32R2:       seb     $2, $[[T0]]
+  ; 32R2-R5:    mul     $[[T0:[0-9]+]], $4, $5
+  ; 32R2-R5:    seb     $2, $[[T0]]
 
   ; 32R6:       mul     $[[T0:[0-9]+]], $4, $5
   ; 32R6:       seb     $2, $[[T0]]
@@ -99,8 +107,8 @@ entry:
   ; 32R1:       sll     $[[T0]], $[[T0]], 16
   ; 32R1:       sra     $2, $[[T0]], 16
 
-  ; 32R2:       mul     $[[T0:[0-9]+]], $4, $5
-  ; 32R2:       seh     $2, $[[T0]]
+  ; 32R2-R5:    mul     $[[T0:[0-9]+]], $4, $5
+  ; 32R2-R5:    seh     $2, $[[T0]]
 
   ; 32R6:       mul     $[[T0:[0-9]+]], $4, $5
   ; 32R6:       seh     $2, $[[T0]]
@@ -130,10 +138,10 @@ entry:
   ; M2:         mult    $4, $5
   ; M2:         mflo    $2
 
-  ; 32R1-R2:    mul     $2, $4, $5
+  ; 32R1-R5:    mul     $2, $4, $5
   ; 32R6:       mul     $2, $4, $5
 
-  ; 64R1-R2:    mul     $2, $4, $5
+  ; 64R1-R5:    mul     $2, $4, $5
   ; 64R6:       mul     $2, $4, $5
   %r = mul i32 %a, %b
   ret i32 %r
@@ -153,13 +161,13 @@ entry:
   ; M2:         addu    $[[T2:[0-9]+]], $4, $[[T1]]
   ; M2:         addu    $2, $[[T2]], $[[T0]]
 
-  ; 32R1-R2:    multu   $5, $7
-  ; 32R1-R2:    mflo    $3
-  ; 32R1-R2:    mfhi    $[[T0:[0-9]+]]
-  ; 32R1-R2:    mul     $[[T1:[0-9]+]], $4, $7
-  ; 32R1-R2:    mul     $[[T2:[0-9]+]], $5, $6
-  ; 32R1-R2:    addu    $[[T0]], $[[T0]], $[[T2:[0-9]+]]
-  ; 32R1-R2:    addu    $2, $[[T0]], $[[T1]]
+  ; 32R1-R5:    multu   $5, $7
+  ; 32R1-R5:    mflo    $3
+  ; 32R1-R5:    mfhi    $[[T0:[0-9]+]]
+  ; 32R1-R5:    mul     $[[T1:[0-9]+]], $4, $7
+  ; 32R1-R5:    mul     $[[T2:[0-9]+]], $5, $6
+  ; 32R1-R5:    addu    $[[T0]], $[[T0]], $[[T2:[0-9]+]]
+  ; 32R1-R5:    addu    $2, $[[T0]], $[[T1]]
 
   ; 32R6:       mul     $[[T0:[0-9]+]], $5, $6
   ; 32R6:       muhu    $[[T1:[0-9]+]], $5, $7
@@ -171,8 +179,8 @@ entry:
   ; M4:         dmult   $4, $5
   ; M4:         mflo    $2
 
-  ; 64R1-R2:    dmult   $4, $5
-  ; 64R1-R2:    mflo    $2
+  ; 64R1-R5:    dmult   $4, $5
+  ; 64R1-R5:    mflo    $2
 
   ; 64R6:       dmul    $2, $4, $5
 
diff --git a/test/CodeGen/Mips/llvm-ir/or.ll b/test/CodeGen/Mips/llvm-ir/or.ll
index 21d1d4fca2a3..6215e4036325 100644
--- a/test/CodeGen/Mips/llvm-ir/or.ll
+++ b/test/CodeGen/Mips/llvm-ir/or.ll
@@ -4,6 +4,10 @@
 ; RUN:    -check-prefix=ALL -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP32
 ; RUN: llc < %s -march=mips64 -mcpu=mips3 | FileCheck %s \
@@ -14,6 +18,10 @@
 ; RUN:    -check-prefix=ALL -check-prefix=GP64
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64
+; RUN: llc < %s -march=mips64 -mcpu=mips64r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP64
+; RUN: llc < %s -march=mips64 -mcpu=mips64r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP64
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64
 
@@ -51,11 +59,7 @@ define signext i32 @or_i32(i32 signext %a, i32 signext %b) {
 entry:
 ; ALL-LABEL: or_i32:
 
-  ; GP32:         or     $2, $4, $5
-
-  ; GP64:         or     $[[T0:[0-9]+]], $4, $5
-  ; FIXME: The sll instruction below is redundant.
-  ; GP64:         sll     $2, $[[T0]], 0
+  ; ALL:          or     $2, $4, $5
 
   %r = or i32 %a, %b
   ret i32 %r
diff --git a/test/CodeGen/Mips/llvm-ir/ret.ll b/test/CodeGen/Mips/llvm-ir/ret.ll
index 8f5b1159760c..0561c24219ce 100644
--- a/test/CodeGen/Mips/llvm-ir/ret.ll
+++ b/test/CodeGen/Mips/llvm-ir/ret.ll
@@ -9,10 +9,14 @@
 
 ; RUN: llc -march=mips   -mcpu=mips32   -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=GPR32 -check-prefix=NO-MTHC1 -check-prefix=NOT-R6
 ; RUN: llc -march=mips   -mcpu=mips32r2 -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=GPR32 -check-prefix=MTHC1 -check-prefix=NOT-R6
+; RUN: llc -march=mips   -mcpu=mips32r3 -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=GPR32 -check-prefix=MTHC1 -check-prefix=NOT-R6
+; RUN: llc -march=mips   -mcpu=mips32r5 -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=GPR32 -check-prefix=MTHC1 -check-prefix=NOT-R6
 ; RUN: llc -march=mips   -mcpu=mips32r6 -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=GPR32 -check-prefix=MTHC1 -check-prefix=R6
 ; RUN: llc -march=mips64 -mcpu=mips4    -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=GPR64 -check-prefix=DMTC1 -check-prefix=NOT-R6
 ; RUN: llc -march=mips64 -mcpu=mips64   -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=GPR64 -check-prefix=DMTC1 -check-prefix=NOT-R6
 ; RUN: llc -march=mips64 -mcpu=mips64r2 -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=GPR64 -check-prefix=DMTC1 -check-prefix=NOT-R6
+; RUN: llc -march=mips64 -mcpu=mips64r3 -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=GPR64 -check-prefix=DMTC1 -check-prefix=NOT-R6
+; RUN: llc -march=mips64 -mcpu=mips64r5 -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=GPR64 -check-prefix=DMTC1 -check-prefix=NOT-R6
 ; RUN: llc -march=mips64 -mcpu=mips64r6 -asm-show-inst < %s | FileCheck %s -check-prefix=ALL -check-prefix=GPR64 -check-prefix=DMTC1 -check-prefix=R6
 
 define void @ret_void() {
diff --git a/test/CodeGen/Mips/llvm-ir/sdiv.ll b/test/CodeGen/Mips/llvm-ir/sdiv.ll
index 54b7f70b1dac..929ee88bb7f7 100644
--- a/test/CodeGen/Mips/llvm-ir/sdiv.ll
+++ b/test/CodeGen/Mips/llvm-ir/sdiv.ll
@@ -3,7 +3,11 @@
 ; RUN: llc < %s -march=mips -mcpu=mips32 | FileCheck %s \
 ; RUN:    -check-prefix=NOT-R6 -check-prefix=NOT-R2-R6 -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r2 | FileCheck %s \
-; RUN:    -check-prefix=NOT-R6 -check-prefix=R2 -check-prefix=GP32
+; RUN:    -check-prefix=NOT-R6 -check-prefix=R2-R5 -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r3 | FileCheck %s \
+; RUN:    -check-prefix=NOT-R6 -check-prefix=R2-R5 -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r5 | FileCheck %s \
+; RUN:    -check-prefix=NOT-R6 -check-prefix=R2-R5 -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r6 | FileCheck %s \
 ; RUN:    -check-prefix=R6 -check-prefix=GP32
 ; RUN: llc < %s -march=mips64 -mcpu=mips3 | FileCheck %s \
@@ -13,7 +17,11 @@
 ; RUN: llc < %s -march=mips64 -mcpu=mips64 | FileCheck %s \
 ; RUN:    -check-prefix=NOT-R6 -check-prefix=NOT-R2-R6 -check-prefix=GP64-NOT-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r2 | FileCheck %s \
-; RUN:    -check-prefix=NOT-R6 -check-prefix=R2 -check-prefix=GP64-NOT-R6
+; RUN:    -check-prefix=NOT-R6 -check-prefix=R2-R5 -check-prefix=GP64-NOT-R6
+; RUN: llc < %s -march=mips64 -mcpu=mips64r3 | FileCheck %s \
+; RUN:    -check-prefix=NOT-R6 -check-prefix=R2-R5 -check-prefix=GP64-NOT-R6
+; RUN: llc < %s -march=mips64 -mcpu=mips64r5 | FileCheck %s \
+; RUN:    -check-prefix=NOT-R6 -check-prefix=R2-R5 -check-prefix=GP64-NOT-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r6 | FileCheck %s \
 ; RUN:    -check-prefix=R6 -check-prefix=64R6
 
@@ -49,11 +57,11 @@ entry:
   ; NOT-R2-R6:    sll     $[[T1:[0-9]+]], $[[T0]], 24
   ; NOT-R2-R6:    sra     $2, $[[T1]], 24
 
-  ; R2:           div     $zero, $4, $5
-  ; R2:           teq     $5, $zero, 7
-  ; R2:           mflo    $[[T0:[0-9]+]]
+  ; R2-R5:        div     $zero, $4, $5
+  ; R2-R5:        teq     $5, $zero, 7
+  ; R2-R5:        mflo    $[[T0:[0-9]+]]
   ; FIXME: This instruction is redundant.
-  ; R2:           seb     $2, $[[T0]]
+  ; R2-R5:        seb     $2, $[[T0]]
 
   ; R6:           div     $[[T0:[0-9]+]], $4, $5
   ; R6:           teq     $5, $zero, 7
@@ -75,11 +83,11 @@ entry:
   ; NOT-R2-R6:    sll     $[[T1:[0-9]+]], $[[T0]], 16
   ; NOT-R2-R6:    sra     $2, $[[T1]], 16
 
-  ; R2:           div     $zero, $4, $5
-  ; R2:           teq     $5, $zero, 7
-  ; R2:           mflo    $[[T0:[0-9]+]]
+  ; R2-R5:        div     $zero, $4, $5
+  ; R2-R5:        teq     $5, $zero, 7
+  ; R2-R5:        mflo    $[[T0:[0-9]+]]
   ; FIXME: This is instruction is redundant since div is signed.
-  ; R2:           seh     $2, $[[T0]]
+  ; R2-R5:        seh     $2, $[[T0]]
 
   ; R6:           div     $[[T0:[0-9]+]], $4, $5
   ; R6:           teq     $5, $zero, 7
diff --git a/test/CodeGen/Mips/llvm-ir/select.ll b/test/CodeGen/Mips/llvm-ir/select.ll
index 736bc579088d..f17670adca33 100644
--- a/test/CodeGen/Mips/llvm-ir/select.ll
+++ b/test/CodeGen/Mips/llvm-ir/select.ll
@@ -5,7 +5,13 @@
 ; RUN:    -check-prefix=CMOV-32 -check-prefix=CMOV-32R1
 ; RUN: llc < %s -march=mips -mcpu=mips32r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=CMOV \
-; RUN:    -check-prefix=CMOV-32 -check-prefix=CMOV-32R2
+; RUN:    -check-prefix=CMOV-32 -check-prefix=CMOV-32R2-R5
+; RUN: llc < %s -march=mips -mcpu=mips32r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=CMOV \
+; RUN:    -check-prefix=CMOV-32 -check-prefix=CMOV-32R2-R5
+; RUN: llc < %s -march=mips -mcpu=mips32r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=CMOV \
+; RUN:    -check-prefix=CMOV-32 -check-prefix=CMOV-32R2-R5
 ; RUN: llc < %s -march=mips -mcpu=mips32r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=SEL -check-prefix=SEL-32
 ; RUN: llc < %s -march=mips64 -mcpu=mips3 | FileCheck %s \
@@ -16,6 +22,10 @@
 ; RUN:    -check-prefix=ALL -check-prefix=CMOV -check-prefix=CMOV-64
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=CMOV -check-prefix=CMOV-64
+; RUN: llc < %s -march=mips64 -mcpu=mips64r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=CMOV -check-prefix=CMOV-64
+; RUN: llc < %s -march=mips64 -mcpu=mips64r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=CMOV -check-prefix=CMOV-64
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=SEL -check-prefix=SEL-64
 
@@ -232,12 +242,12 @@ entry:
   ; M2:         jr      $ra
   ; M2:         mtc1    $6, $f1
 
-  ; CMOV-32:    mtc1    $7, $[[F0:f[0-9]+]]
-  ; CMOV-32R1:  mtc1    $6, $f{{[0-9]+}}
-  ; CMOV-32R2   mthc1   $6, $[[F0]]
-  ; CMOV-32:    andi    $[[T0:[0-9]+]], $4, 1
-  ; CMOV-32:    ldc1    $f0, 16($sp)
-  ; CMOV-32:    movn.d  $f0, $[[F0]], $[[T0]]
+  ; CMOV-32:      mtc1    $7, $[[F0:f[0-9]+]]
+  ; CMOV-32R1:    mtc1    $6, $f{{[0-9]+}}
+  ; CMOV-32R2-R5: mthc1   $6, $[[F0]]
+  ; CMOV-32:      andi    $[[T0:[0-9]+]], $4, 1
+  ; CMOV-32:      ldc1    $f0, 16($sp)
+  ; CMOV-32:      movn.d  $f0, $[[F0]], $[[T0]]
 
   ; SEL-32:     mtc1    $7, $[[F0:f[0-9]+]]
   ; SEL-32:     mthc1   $6, $[[F0]]
diff --git a/test/CodeGen/Mips/llvm-ir/shl.ll b/test/CodeGen/Mips/llvm-ir/shl.ll
index fc5243cc97f2..bba34c47ea82 100644
--- a/test/CodeGen/Mips/llvm-ir/shl.ll
+++ b/test/CodeGen/Mips/llvm-ir/shl.ll
@@ -3,10 +3,16 @@
 ; RUN:    -check-prefix=M2 -check-prefix=NOT-R2-R6
 ; RUN: llc < %s -march=mips -mcpu=mips32 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP32 -check-prefix=NOT-R2-R6 \
-; RUN:    -check-prefix=32R1-R2
+; RUN:    -check-prefix=32R1-R5
 ; RUN: llc < %s -march=mips -mcpu=mips32r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP32 \
-; RUN:    -check-prefix=32R1-R2 -check-prefix=R2-R6
+; RUN:    -check-prefix=32R1-R5 -check-prefix=R2-R6
+; RUN: llc < %s -march=mips -mcpu=mips32r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP32 \
+; RUN:    -check-prefix=32R1-R5 -check-prefix=R2-R6
+; RUN: llc < %s -march=mips -mcpu=mips32r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP32 \
+; RUN:    -check-prefix=32R1-R5 -check-prefix=R2-R6
 ; RUN: llc < %s -march=mips -mcpu=mips32r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP32 \
 ; RUN:    -check-prefix=32R6 -check-prefix=R2-R6
@@ -22,6 +28,12 @@
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64 \
 ; RUN:    -check-prefix=GP64-NOT-R6 -check-prefix R2-R6
+; RUN: llc < %s -march=mips64 -mcpu=mips64r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP64 \
+; RUN:    -check-prefix=GP64-NOT-R6 -check-prefix R2-R6
+; RUN: llc < %s -march=mips64 -mcpu=mips64r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP64 \
+; RUN:    -check-prefix=GP64-NOT-R6 -check-prefix R2-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64 \
 ; RUN:    -check-prefix=64R6 -check-prefix=R2-R6
@@ -101,16 +113,16 @@ entry:
   ; M2:         jr        $ra
   ; M2:         nop
 
-  ; 32R1-R2:    sllv      $[[T0:[0-9]+]], $4, $7
-  ; 32R1-R2:    not       $[[T1:[0-9]+]], $7
-  ; 32R1-R2:    srl       $[[T2:[0-9]+]], $5, 1
-  ; 32R1-R2:    srlv      $[[T3:[0-9]+]], $[[T2]], $[[T1]]
-  ; 32R1-R2:    or        $2, $[[T0]], $[[T3]]
-  ; 32R1-R2:    sllv      $[[T4:[0-9]+]], $5, $7
-  ; 32R1-R2:    andi      $[[T5:[0-9]+]], $7, 32
-  ; 32R1-R2:    movn      $2, $[[T4]], $[[T5]]
-  ; 32R1-R2:    jr        $ra
-  ; 32R1-R2:    movn      $3, $zero, $[[T5]]
+  ; 32R1-R5:    sllv      $[[T0:[0-9]+]], $4, $7
+  ; 32R1-R5:    not       $[[T1:[0-9]+]], $7
+  ; 32R1-R5:    srl       $[[T2:[0-9]+]], $5, 1
+  ; 32R1-R5:    srlv      $[[T3:[0-9]+]], $[[T2]], $[[T1]]
+  ; 32R1-R5:    or        $2, $[[T0]], $[[T3]]
+  ; 32R1-R5:    sllv      $[[T4:[0-9]+]], $5, $7
+  ; 32R1-R5:    andi      $[[T5:[0-9]+]], $7, 32
+  ; 32R1-R5:    movn      $2, $[[T4]], $[[T5]]
+  ; 32R1-R5:    jr        $ra
+  ; 32R1-R5:    movn      $3, $zero, $[[T5]]
 
   ; 32R6:       sllv      $[[T0:[0-9]+]], $4, $7
   ; 32R6:       not       $[[T1:[0-9]+]], $7
@@ -125,8 +137,7 @@ entry:
   ; 32R6:       jr        $ra
   ; 32R6:       seleqz    $3, $[[T7]], $[[T5]]
 
-  ; GP64:       sll       $[[T0:[0-9]+]], $5, 0
-  ; GP64:       dsllv     $2, $4, $1
+  ; GP64:       dsllv     $2, $4, $5
 
   %r = shl i64 %a, %b
   ret i64 %r
@@ -139,11 +150,11 @@ entry:
   ; GP32:           lw        $25, %call16(__ashlti3)($gp)
 
   ; M3:             sll       $[[T0:[0-9]+]], $7, 0
-  ; M3:             dsllv     $[[T1:[0-9]+]], $5, $[[T0]]
+  ; M3:             dsllv     $[[T1:[0-9]+]], $5, $7
   ; M3:             andi      $[[T2:[0-9]+]], $[[T0]], 64
   ; M3:             bnez      $[[T3:[0-9]+]], $[[BB0:BB[0-9_]+]]
   ; M3:             move      $2, $[[T1]]
-  ; M3:             dsllv     $[[T4:[0-9]+]], $4, $[[T0]]
+  ; M3:             dsllv     $[[T4:[0-9]+]], $4, $7
   ; M3:             dsrl      $[[T5:[0-9]+]], $5, 1
   ; M3:             not       $[[T6:[0-9]+]], $[[T0]]
   ; M3:             dsrlv     $[[T7:[0-9]+]], $[[T5]], $[[T6]]
@@ -156,32 +167,32 @@ entry:
   ; M3:             jr        $ra
   ; M3:             nop
 
-  ; GP64-NOT-R6:    sll       $[[T0:[0-9]+]], $7, 0
-  ; GP64-NOT-R6:    dsllv     $[[T1:[0-9]+]], $4, $[[T0]]
-  ; GP64-NOT-R6:    dsrl      $[[T2:[0-9]+]], $5, 1
-  ; GP64-NOT-R6:    not       $[[T3:[0-9]+]], $[[T0]]
-  ; GP64-NOT-R6:    dsrlv     $[[T4:[0-9]+]], $[[T2]], $[[T3]]
-  ; GP64-NOT-R6:    or        $2, $[[T1]], $[[T4]]
-  ; GP64-NOT-R6:    dsllv     $3, $5, $[[T0]]
-  ; GP64-NOT-R6:    andi      $[[T5:[0-9]+]], $[[T0]], 64
+  ; GP64-NOT-R6:    dsllv     $[[T0:[0-9]+]], $4, $7
+  ; GP64-NOT-R6:    dsrl      $[[T1:[0-9]+]], $5, 1
+  ; GP64-NOT-R6:    sll       $[[T2:[0-9]+]], $7, 0
+  ; GP64-NOT-R6:    not       $[[T3:[0-9]+]], $[[T2]]
+  ; GP64-NOT-R6:    dsrlv     $[[T4:[0-9]+]], $[[T1]], $[[T3]]
+  ; GP64-NOT-R6:    or        $2, $[[T0]], $[[T4]]
+  ; GP64-NOT-R6:    dsllv     $3, $5, $7
+  ; GP64-NOT-R6:    andi      $[[T5:[0-9]+]], $[[T2]], 64
   ; GP64-NOT-R6:    movn      $2, $3, $[[T5]]
   ; GP64-NOT-R6:    jr        $ra
   ; GP64-NOT-R6:    movn      $3, $zero, $1
 
-  ; 64R6:           sll       $[[T0:[0-9]+]], $7, 0
-  ; 64R6:           dsllv     $[[T1:[0-9]+]], $4, $[[T0]]
-  ; 64R6:           dsrl      $[[T2:[0-9]+]], $5, 1
-  ; 64R6:           not       $[[T3:[0-9]+]], $[[T0]]
-  ; 64R6:           dsrlv     $[[T4:[0-9]+]], $[[T2]], $[[T3]]
-  ; 64R6:           or        $[[T5:[0-9]+]], $[[T1]], $[[T4]]
-  ; 64R6:           andi      $[[T6:[0-9]+]], $[[T0]], 64
+  ; 64R6:           dsllv     $[[T0:[0-9]+]], $4, $7
+  ; 64R6:           dsrl      $[[T1:[0-9]+]], $5, 1
+  ; 64R6:           sll       $[[T2:[0-9]+]], $7, 0
+  ; 64R6:           not       $[[T3:[0-9]+]], $[[T2]]
+  ; 64R6:           dsrlv     $[[T4:[0-9]+]], $[[T1]], $[[T3]]
+  ; 64R6:           or        $[[T5:[0-9]+]], $[[T0]], $[[T4]]
+  ; 64R6:           andi      $[[T6:[0-9]+]], $[[T2]], 64
   ; 64R6:           sll       $[[T7:[0-9]+]], $[[T6]], 0
   ; 64R6:           seleqz    $[[T8:[0-9]+]], $[[T5]], $[[T7]]
-  ; 64R6:           dsllv     $[[T9:[0-9]+]], $5, $[[T0]]
+  ; 64R6:           dsllv     $[[T9:[0-9]+]], $5, $7
   ; 64R6:           selnez    $[[T10:[0-9]+]], $[[T9]], $[[T7]]
   ; 64R6:           or        $2, $[[T10]], $[[T8]]
   ; 64R6:           jr        $ra
-  ; 64R6:           seleqz    $3, $[[T0]], $[[T7]]
+  ; 64R6:           seleqz    $3, $[[T9]], $[[T7]]
 
   %r = shl i128 %a, %b
   ret i128 %r
diff --git a/test/CodeGen/Mips/llvm-ir/srem.ll b/test/CodeGen/Mips/llvm-ir/srem.ll
index 1e949d24678b..ceb53ee7033a 100644
--- a/test/CodeGen/Mips/llvm-ir/srem.ll
+++ b/test/CodeGen/Mips/llvm-ir/srem.ll
@@ -3,7 +3,11 @@
 ; RUN: llc < %s -march=mips -mcpu=mips32 | FileCheck %s \
 ; RUN:  -check-prefix=GP32 -check-prefix=NOT-R6 -check-prefix=NOT-R2-R6
 ; RUN: llc < %s -march=mips -mcpu=mips32r2 | FileCheck %s -check-prefix=GP32 \
-; RUN:  -check-prefix=R2 -check-prefix=R2-R6 -check-prefix=NOT-R6
+; RUN:  -check-prefix=R2-R5 -check-prefix=R2-R6 -check-prefix=NOT-R6
+; RUN: llc < %s -march=mips -mcpu=mips32r3 | FileCheck %s -check-prefix=GP32 \
+; RUN:  -check-prefix=R2-R5 -check-prefix=R2-R6 -check-prefix=NOT-R6
+; RUN: llc < %s -march=mips -mcpu=mips32r5 | FileCheck %s -check-prefix=GP32 \
+; RUN:  -check-prefix=R2-R5 -check-prefix=R2-R6 -check-prefix=NOT-R6
 ; RUN: llc < %s -march=mips -mcpu=mips32r6 | FileCheck %s \
 ; RUN:   -check-prefix=GP32 -check-prefix=R6 -check-prefix=R2-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips3 | FileCheck %s \
@@ -13,7 +17,13 @@
 ; RUN: llc < %s -march=mips64 -mcpu=mips64 | FileCheck %s \
 ; RUN:  -check-prefix=GP64-NOT-R6 -check-prefix=NOT-R6 -check-prefix=NOT-R2-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r2 | FileCheck %s \
-; RUN:  -check-prefix=R2 -check-prefix=R2-R6 \
+; RUN:  -check-prefix=R2-R5 -check-prefix=R2-R6 \
+; RUN:  -check-prefix=GP64-NOT-R6 -check-prefix=NOT-R6
+; RUN: llc < %s -march=mips64 -mcpu=mips64r3 | FileCheck %s \
+; RUN:  -check-prefix=R2-R5 -check-prefix=R2-R6 \
+; RUN:  -check-prefix=GP64-NOT-R6 -check-prefix=NOT-R6
+; RUN: llc < %s -march=mips64 -mcpu=mips64r5 | FileCheck %s \
+; RUN:  -check-prefix=R2-R5 -check-prefix=R2-R6 \
 ; RUN:  -check-prefix=GP64-NOT-R6 -check-prefix=NOT-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r6 | FileCheck %s \
 ; RUN:  -check-prefix=64R6 -check-prefix=R6 -check-prefix=R2-R6
@@ -47,10 +57,10 @@ entry:
   ; NOT-R2-R6:    sll     $[[T1:[0-9]+]], $[[T0]], 24
   ; NOT-R2-R6:    sra     $2, $[[T1]], 24
 
-  ; R2:           div     $zero, $4, $5
-  ; R2:           teq     $5, $zero, 7
-  ; R2:           mfhi    $[[T0:[0-9]+]]
-  ; R2:           seb     $2, $[[T0]]
+  ; R2-R5:        div     $zero, $4, $5
+  ; R2-R5:        teq     $5, $zero, 7
+  ; R2-R5:        mfhi    $[[T0:[0-9]+]]
+  ; R2-R5:        seb     $2, $[[T0]]
 
   ; R6:           mod     $[[T0:[0-9]+]], $4, $5
   ; R6:           teq     $5, $zero, 7
@@ -70,10 +80,10 @@ entry:
   ; NOT-R2-R6:    sll     $[[T1:[0-9]+]], $[[T0]], 16
   ; NOT-R2-R6:    sra     $2, $[[T1]], 16
 
-  ; R2:           div     $zero, $4, $5
-  ; R2:           teq     $5, $zero, 7
-  ; R2:           mfhi    $[[T0:[0-9]+]]
-  ; R2:           seh     $2, $[[T1]]
+  ; R2-R5:        div     $zero, $4, $5
+  ; R2-R5:        teq     $5, $zero, 7
+  ; R2-R5:        mfhi    $[[T0:[0-9]+]]
+  ; R2-R5:        seh     $2, $[[T1]]
 
   ; R6:           mod     $[[T0:[0-9]+]], $4, $5
   ; R6:           teq     $5, $zero, 7
diff --git a/test/CodeGen/Mips/llvm-ir/sub.ll b/test/CodeGen/Mips/llvm-ir/sub.ll
index 6d592be38211..164975844d73 100644
--- a/test/CodeGen/Mips/llvm-ir/sub.ll
+++ b/test/CodeGen/Mips/llvm-ir/sub.ll
@@ -4,6 +4,10 @@
 ; RUN:    -check-prefix=ALL -check-prefix=NOT-R2-R6 -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=R2-R6 -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=R2-R6 -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=R2-R6 -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=R2-R6 -check-prefix=GP32
 ; RUN: llc < %s -march=mips64 -mcpu=mips3 | FileCheck %s \
@@ -14,6 +18,10 @@
 ; RUN:    -check-prefix=ALL -check-prefix=NOT-R2-R6 -check-prefix=GP64
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=R2-R6 -check-prefix=GP64
+; RUN: llc < %s -march=mips64 -mcpu=mips64r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=R2-R6 -check-prefix=GP64
+; RUN: llc < %s -march=mips64 -mcpu=mips64r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=R2-R6 -check-prefix=GP64
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=R2-R6 -check-prefix=GP64
 
diff --git a/test/CodeGen/Mips/llvm-ir/udiv.ll b/test/CodeGen/Mips/llvm-ir/udiv.ll
index 1f7aa0d5f4ce..a7cafe52d1ac 100644
--- a/test/CodeGen/Mips/llvm-ir/udiv.ll
+++ b/test/CodeGen/Mips/llvm-ir/udiv.ll
@@ -4,6 +4,10 @@
 ; RUN:    -check-prefix=NOT-R6 -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r2 | FileCheck %s \
 ; RUN:    -check-prefix=NOT-R6 -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r3 | FileCheck %s \
+; RUN:    -check-prefix=NOT-R6 -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r5 | FileCheck %s \
+; RUN:    -check-prefix=NOT-R6 -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r6 | FileCheck %s \
 ; RUN:    -check-prefix=R6 -check-prefix=GP32
 ; RUN: llc < %s -march=mips64 -mcpu=mips3 | FileCheck %s \
@@ -14,6 +18,10 @@
 ; RUN:    -check-prefix=NOT-R6 -check-prefix=GP64-NOT-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r2 | FileCheck %s \
 ; RUN:    -check-prefix=NOT-R6 -check-prefix=GP64-NOT-R6
+; RUN: llc < %s -march=mips64 -mcpu=mips64r3 | FileCheck %s \
+; RUN:    -check-prefix=NOT-R6 -check-prefix=GP64-NOT-R6
+; RUN: llc < %s -march=mips64 -mcpu=mips64r5 | FileCheck %s \
+; RUN:    -check-prefix=NOT-R6 -check-prefix=GP64-NOT-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r6 | FileCheck %s \
 ; RUN:    -check-prefix=R6 -check-prefix=64R6
 
diff --git a/test/CodeGen/Mips/llvm-ir/urem.ll b/test/CodeGen/Mips/llvm-ir/urem.ll
index 73235341a42f..d5a231c8dfca 100644
--- a/test/CodeGen/Mips/llvm-ir/urem.ll
+++ b/test/CodeGen/Mips/llvm-ir/urem.ll
@@ -3,7 +3,11 @@
 ; RUN: llc < %s -march=mips -mcpu=mips32 | FileCheck %s \
 ; RUN:  -check-prefix=GP32 -check-prefix=NOT-R6 -check-prefix=NOT-R2-R6
 ; RUN: llc < %s -march=mips -mcpu=mips32r2 | FileCheck %s -check-prefix=GP32 \
-; RUN:  -check-prefix=R2 -check-prefix=R2-R6 -check-prefix=NOT-R6
+; RUN:  -check-prefix=R2-R5 -check-prefix=R2-R6 -check-prefix=NOT-R6
+; RUN: llc < %s -march=mips -mcpu=mips32r3 | FileCheck %s -check-prefix=GP32 \
+; RUN:  -check-prefix=R2-R5 -check-prefix=R2-R6 -check-prefix=NOT-R6
+; RUN: llc < %s -march=mips -mcpu=mips32r5 | FileCheck %s -check-prefix=GP32 \
+; RUN:  -check-prefix=R2-R5 -check-prefix=R2-R6 -check-prefix=NOT-R6
 ; RUN: llc < %s -march=mips -mcpu=mips32r6 | FileCheck %s \
 ; RUN:   -check-prefix=GP32 -check-prefix=R6 -check-prefix=R2-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips3 | FileCheck %s \
@@ -13,7 +17,13 @@
 ; RUN: llc < %s -march=mips64 -mcpu=mips64 | FileCheck %s \
 ; RUN:  -check-prefix=GP64-NOT-R6 -check-prefix=NOT-R6 -check-prefix=NOT-R2-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r2 | FileCheck %s \
-; RUN:  -check-prefix=R2 -check-prefix=R2-R6 \
+; RUN:  -check-prefix=R2-R5 -check-prefix=R2-R6 \
+; RUN:  -check-prefix=GP64-NOT-R6 -check-prefix=NOT-R6
+; RUN: llc < %s -march=mips64 -mcpu=mips64r3 | FileCheck %s \
+; RUN:  -check-prefix=R2-R5 -check-prefix=R2-R6 \
+; RUN:  -check-prefix=GP64-NOT-R6 -check-prefix=NOT-R6
+; RUN: llc < %s -march=mips64 -mcpu=mips64r5 | FileCheck %s \
+; RUN:  -check-prefix=R2-R5 -check-prefix=R2-R6 \
 ; RUN:  -check-prefix=GP64-NOT-R6 -check-prefix=NOT-R6
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r6 | FileCheck %s \
 ; RUN:  -check-prefix=64R6 -check-prefix=R6 -check-prefix=R2-R6
@@ -53,12 +63,12 @@ entry:
   ; NOT-R2-R6:    sll     $[[T3:[0-9]+]], $[[T2]], 24
   ; NOT-R2-R6:    sra     $2, $[[T3]], 24
 
-  ; R2:           andi    $[[T0:[0-9]+]], $5, 255
-  ; R2:           andi    $[[T1:[0-9]+]], $4, 255
-  ; R2:           divu    $zero, $[[T1]], $[[T0]]
-  ; R2:           teq     $[[T0]], $zero, 7
-  ; R2:           mfhi    $[[T2:[0-9]+]]
-  ; R2:           seb     $2, $[[T2]]
+  ; R2-R5:        andi    $[[T0:[0-9]+]], $5, 255
+  ; R2-R5:        andi    $[[T1:[0-9]+]], $4, 255
+  ; R2-R5:        divu    $zero, $[[T1]], $[[T0]]
+  ; R2-R5:        teq     $[[T0]], $zero, 7
+  ; R2-R5:        mfhi    $[[T2:[0-9]+]]
+  ; R2-R5:        seb     $2, $[[T2]]
 
   ; R6:           andi    $[[T0:[0-9]+]], $5, 255
   ; R6:           andi    $[[T1:[0-9]+]], $4, 255
@@ -82,12 +92,12 @@ entry:
   ; NOT-R2-R6:    sll     $[[T3:[0-9]+]], $[[T2]], 16
   ; NOT-R2-R6:    sra     $2, $[[T3]], 16
 
-  ; R2:           andi    $[[T0:[0-9]+]], $5, 65535
-  ; R2:           andi    $[[T1:[0-9]+]], $4, 65535
-  ; R2:           divu    $zero, $[[T1]], $[[T0]]
-  ; R2:           teq     $[[T0]], $zero, 7
-  ; R2:           mfhi    $[[T3:[0-9]+]]
-  ; R2:           seh     $2, $[[T2]]
+  ; R2-R5:        andi    $[[T0:[0-9]+]], $5, 65535
+  ; R2-R5:        andi    $[[T1:[0-9]+]], $4, 65535
+  ; R2-R5:        divu    $zero, $[[T1]], $[[T0]]
+  ; R2-R5:        teq     $[[T0]], $zero, 7
+  ; R2-R5:        mfhi    $[[T3:[0-9]+]]
+  ; R2-R5:        seh     $2, $[[T2]]
 
   ; R6:           andi    $[[T0:[0-9]+]], $5, 65535
   ; R6:           andi    $[[T1:[0-9]+]], $4, 65535
diff --git a/test/CodeGen/Mips/llvm-ir/xor.ll b/test/CodeGen/Mips/llvm-ir/xor.ll
index 94dead1eff41..89af99981a3c 100644
--- a/test/CodeGen/Mips/llvm-ir/xor.ll
+++ b/test/CodeGen/Mips/llvm-ir/xor.ll
@@ -4,6 +4,10 @@
 ; RUN:    -check-prefix=ALL -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP32
+; RUN: llc < %s -march=mips -mcpu=mips32r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP32
 ; RUN: llc < %s -march=mips -mcpu=mips32r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP32
 ; RUN: llc < %s -march=mips64 -mcpu=mips3 | FileCheck %s \
@@ -14,6 +18,10 @@
 ; RUN:    -check-prefix=ALL -check-prefix=GP64
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r2 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64
+; RUN: llc < %s -march=mips64 -mcpu=mips64r3 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP64
+; RUN: llc < %s -march=mips64 -mcpu=mips64r5 | FileCheck %s \
+; RUN:    -check-prefix=ALL -check-prefix=GP64
 ; RUN: llc < %s -march=mips64 -mcpu=mips64r6 | FileCheck %s \
 ; RUN:    -check-prefix=ALL -check-prefix=GP64
 
@@ -51,10 +59,7 @@ define signext i32 @xor_i32(i32 signext %a, i32 signext %b) {
 entry:
 ; ALL-LABEL: xor_i32:
 
-  ; GP32:         xor     $2, $4, $5
-
-  ; GP64:         xor     $[[T0:[0-9]+]], $4, $5
-  ; GP64:         sll     $2, $[[T0]], 0
+  ; ALL:          xor     $2, $4, $5
 
   %r = xor i32 %a, %b
   ret i32 %r
diff --git a/test/CodeGen/Mips/load-store-left-right.ll b/test/CodeGen/Mips/load-store-left-right.ll
index f6d0e8debb36..a01d246ae460 100644
--- a/test/CodeGen/Mips/load-store-left-right.ll
+++ b/test/CodeGen/Mips/load-store-left-right.ll
@@ -4,14 +4,14 @@
 ; RUN: llc -march=mips     -mcpu=mips32r2            < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS32 -check-prefix=MIPS32-EB %s
 ; RUN: llc -march=mipsel   -mcpu=mips32r6            < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS32R6 -check-prefix=MIPS32R6-EL %s
 ; RUN: llc -march=mips     -mcpu=mips32r6            < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS32R6 -check-prefix=MIPS32R6-EB %s
-; RUN: llc -march=mips64el -mcpu=mips4    -mattr=n64 < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS64 -check-prefix=MIPS64-EL %s
-; RUN: llc -march=mips64   -mcpu=mips4    -mattr=n64 < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS64 -check-prefix=MIPS64-EB %s
-; RUN: llc -march=mips64el -mcpu=mips64   -mattr=n64 < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS64 -check-prefix=MIPS64-EL %s
-; RUN: llc -march=mips64   -mcpu=mips64   -mattr=n64 < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS64 -check-prefix=MIPS64-EB %s
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=n64 < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS64 -check-prefix=MIPS64-EL %s
-; RUN: llc -march=mips64   -mcpu=mips64r2 -mattr=n64 < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS64 -check-prefix=MIPS64-EB %s
-; RUN: llc -march=mips64el -mcpu=mips64r6 -mattr=n64 < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS64R6 -check-prefix=MIPS64R6-EL %s
-; RUN: llc -march=mips64   -mcpu=mips64r6 -mattr=n64 < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS64R6 -check-prefix=MIPS64R6-EB %s
+; RUN: llc -march=mips64el -mcpu=mips4    -target-abi=n64 < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS64 -check-prefix=MIPS64-EL %s
+; RUN: llc -march=mips64   -mcpu=mips4    -target-abi=n64 < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS64 -check-prefix=MIPS64-EB %s
+; RUN: llc -march=mips64el -mcpu=mips64   -target-abi=n64 < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS64 -check-prefix=MIPS64-EL %s
+; RUN: llc -march=mips64   -mcpu=mips64   -target-abi=n64 < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS64 -check-prefix=MIPS64-EB %s
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi=n64 < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS64 -check-prefix=MIPS64-EL %s
+; RUN: llc -march=mips64   -mcpu=mips64r2 -target-abi=n64 < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS64 -check-prefix=MIPS64-EB %s
+; RUN: llc -march=mips64el -mcpu=mips64r6 -target-abi=n64 < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS64R6 -check-prefix=MIPS64R6-EL %s
+; RUN: llc -march=mips64   -mcpu=mips64r6 -target-abi=n64 < %s | FileCheck -check-prefix=ALL -check-prefix=MIPS64R6 -check-prefix=MIPS64R6-EB %s
 
 %struct.SLL = type { i64 }
 %struct.SI = type { i32 }
@@ -43,7 +43,7 @@ entry:
 ; MIPS64R6:      ld $[[PTR:[0-9]+]], %got_disp(si)(
 ; MIPS64R6:      lw $2, 0($[[PTR]])
 
-  %0 = load i32* getelementptr inbounds (%struct.SI* @si, i32 0, i32 0), align 1
+  %0 = load i32, i32* getelementptr inbounds (%struct.SI, %struct.SI* @si, i32 0, i32 0), align 1
   ret i32 %0
 }
 
@@ -69,7 +69,7 @@ entry:
 ; MIPS64R6:      ld $[[PTR:[0-9]+]], %got_disp(si)(
 ; MIPS64R6:      sw $4, 0($[[PTR]])
 
-  store i32 %a, i32* getelementptr inbounds (%struct.SI* @si, i32 0, i32 0), align 1
+  store i32 %a, i32* getelementptr inbounds (%struct.SI, %struct.SI* @si, i32 0, i32 0), align 1
   ret void
 }
 
@@ -100,7 +100,7 @@ entry:
 ; MIPS64R6:      ld $[[PTR:[0-9]+]], %got_disp(sll)(
 ; MIPS64R6:      ld $2, 0($[[PTR]])
 
-  %0 = load i64* getelementptr inbounds (%struct.SLL* @sll, i64 0, i32 0), align 1
+  %0 = load i64, i64* getelementptr inbounds (%struct.SLL, %struct.SLL* @sll, i64 0, i32 0), align 1
   ret i64 %0
 }
 
@@ -129,7 +129,7 @@ entry:
 ; MIPS64R6:      ld $[[PTR:[0-9]+]], %got_disp(si)(
 ; MIPS64R6:      lw $2, 0($[[PTR]])
 
-  %0 = load i32* getelementptr inbounds (%struct.SI* @si, i64 0, i32 0), align 1
+  %0 = load i32, i32* getelementptr inbounds (%struct.SI, %struct.SI* @si, i64 0, i32 0), align 1
   %conv = sext i32 %0 to i64
   ret i64 %conv
 }
@@ -165,7 +165,7 @@ entry:
 ; MIPS64R6:      ld $[[PTR:[0-9]+]], %got_disp(sui)(
 ; MIPS64R6:      lwu $2, 0($[[PTR]])
 
-  %0 = load i32* getelementptr inbounds (%struct.SUI* @sui, i64 0, i32 0), align 1
+  %0 = load i32, i32* getelementptr inbounds (%struct.SUI, %struct.SUI* @sui, i64 0, i32 0), align 1
   %conv = zext i32 %0 to i64
   ret i64 %conv
 }
@@ -197,7 +197,7 @@ entry:
 ; MIPS64R6:      ld $[[PTR:[0-9]+]], %got_disp(sll)(
 ; MIPS64R6:      sd $4, 0($[[PTR]])
 
-  store i64 %a, i64* getelementptr inbounds (%struct.SLL* @sll, i64 0, i32 0), align 1
+  store i64 %a, i64* getelementptr inbounds (%struct.SLL, %struct.SLL* @sll, i64 0, i32 0), align 1
   ret void
 }
 
@@ -223,7 +223,7 @@ entry:
 ; MIPS64R6:      ld $[[PTR:[0-9]+]], %got_disp(si)(
 ; MIPS64R6:      sw $4, 0($[[PTR]])
 
-  store i32 %a, i32* getelementptr inbounds (%struct.SI* @si, i64 0, i32 0), align 1
+  store i32 %a, i32* getelementptr inbounds (%struct.SI, %struct.SI* @si, i64 0, i32 0), align 1
   ret void
 }
 
@@ -257,8 +257,8 @@ entry:
 ; ALL-DAG:       lbu $[[R1:[0-9]+]], 1($[[PTR]])
 ; ALL-DAG:       sb $[[R1]], 3($[[PTR]])
 
-  %0 = load %struct.S0* getelementptr inbounds (%struct.S0* @struct_s0, i32 0), align 1
-  store %struct.S0 %0, %struct.S0* getelementptr inbounds (%struct.S0* @struct_s0, i32 1), align 1
+  %0 = load %struct.S0, %struct.S0* getelementptr inbounds (%struct.S0, %struct.S0* @struct_s0, i32 0), align 1
+  store %struct.S0 %0, %struct.S0* getelementptr inbounds (%struct.S0, %struct.S0* @struct_s0, i32 1), align 1
   ret void
 }
 
@@ -300,8 +300,8 @@ entry:
 ; MIPS64R6-DAG:  lhu $[[R1:[0-9]+]], 2($[[PTR]])
 ; MIPS64R6-DAG:  sh $[[R1]], 6($[[PTR]])
 
-  %0 = load %struct.S1* getelementptr inbounds (%struct.S1* @struct_s1, i32 0), align 1
-  store %struct.S1 %0, %struct.S1* getelementptr inbounds (%struct.S1* @struct_s1, i32 1), align 1
+  %0 = load %struct.S1, %struct.S1* getelementptr inbounds (%struct.S1, %struct.S1* @struct_s1, i32 0), align 1
+  store %struct.S1 %0, %struct.S1* getelementptr inbounds (%struct.S1, %struct.S1* @struct_s1, i32 1), align 1
   ret void
 }
 
@@ -361,8 +361,8 @@ entry:
 ; MIPS64R6-DAG:  lw $[[R1:[0-9]+]], 4($[[PTR]])
 ; MIPS64R6-DAG:  sw $[[R1]],       12($[[PTR]])
 
-  %0 = load %struct.S2* getelementptr inbounds (%struct.S2* @struct_s2, i32 0), align 1
-  store %struct.S2 %0, %struct.S2* getelementptr inbounds (%struct.S2* @struct_s2, i32 1), align 1
+  %0 = load %struct.S2, %struct.S2* getelementptr inbounds (%struct.S2, %struct.S2* @struct_s2, i32 0), align 1
+  store %struct.S2 %0, %struct.S2* getelementptr inbounds (%struct.S2, %struct.S2* @struct_s2, i32 1), align 1
   ret void
 }
 
diff --git a/test/CodeGen/Mips/longbranch.ll b/test/CodeGen/Mips/longbranch.ll
index ad0235eb0af4..9f5b7417b859 100644
--- a/test/CodeGen/Mips/longbranch.ll
+++ b/test/CodeGen/Mips/longbranch.ll
@@ -1,9 +1,9 @@
 ; RUN: llc -march=mipsel < %s | FileCheck %s
 ; RUN: llc -march=mipsel -force-mips-long-branch -O3 < %s \
 ; RUN:   | FileCheck %s -check-prefix=O32
-; RUN: llc -march=mips64el -mcpu=mips4 -mattr=n64 -force-mips-long-branch -O3 \
+; RUN: llc -march=mips64el -mcpu=mips4 -target-abi=n64 -force-mips-long-branch -O3 \
 ; RUN:   < %s | FileCheck %s -check-prefix=N64
-; RUN: llc -march=mips64el -mcpu=mips64 -mattr=n64 -force-mips-long-branch -O3 \
+; RUN: llc -march=mips64el -mcpu=mips64 -target-abi=n64 -force-mips-long-branch -O3 \
 ; RUN:   < %s | FileCheck %s -check-prefix=N64
 ; RUN: llc -march=mipsel -mcpu=mips32r2 -mattr=micromips \
 ; RUN:   -force-mips-long-branch -O3 < %s | FileCheck %s -check-prefix=MICROMIPS
@@ -124,10 +124,9 @@ end:
 ; MICROMIPS:   $[[BB0]]:
 ; MICROMIPS:        lw      $[[R1:[0-9]+]], %got(x)($[[GP]])
 ; MICROMIPS:        li16    $[[R2:[0-9]+]], 1
-; MICROMIPS:        sw      $[[R2]], 0($[[R1]])
+; MICROMIPS:        sw16    $[[R2]], 0($[[R1]])
 ; MICROMIPS:   $[[BB2]]:
-; MICROMIPS:        jr      $ra
-; MICROMIPS:        nop
+; MICROMIPS:        jrc      $ra
 
 
 ; Check the NaCl version.  Check that sp change is not in the branch delay slot
diff --git a/test/CodeGen/Mips/machineverifier.ll b/test/CodeGen/Mips/machineverifier.ll
index c673fe557e6b..d496b833a6c6 100644
--- a/test/CodeGen/Mips/machineverifier.ll
+++ b/test/CodeGen/Mips/machineverifier.ll
@@ -6,7 +6,7 @@
 
 define void @foo() nounwind {
 entry:
-  %0 = load i32* @g, align 4
+  %0 = load i32, i32* @g, align 4
   %tobool = icmp eq i32 %0, 0
   br i1 %tobool, label %if.end, label %if.then
 
diff --git a/test/CodeGen/Mips/mbrsize4a.ll b/test/CodeGen/Mips/mbrsize4a.ll
index 15e1f47ce29e..264d2284afc9 100644
--- a/test/CodeGen/Mips/mbrsize4a.ll
+++ b/test/CodeGen/Mips/mbrsize4a.ll
@@ -1,4 +1,4 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=static -mips16-constant-islands   < %s | FileCheck %s -check-prefix=jal16
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=static -mips16-constant-islands   < %s | FileCheck %s -check-prefix=jal16
 
 @j = global i32 10, align 4
 @.str = private unnamed_addr constant [11 x i8] c"at bottom\0A\00", align 1
@@ -17,11 +17,11 @@ z:                                                ; preds = %y, %entry
   br label %y
 
 y:                                                ; preds = %z
-  %call1 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([11 x i8]* @.str, i32 0, i32 0))
+  %call1 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([11 x i8], [11 x i8]* @.str, i32 0, i32 0))
   br label %z
 
 return:                                           ; No predecessors!
-  %0 = load i32* %retval
+  %0 = load i32, i32* %retval
   ret i32 %0
 ; jal16: 	jal	$BB{{[0-9]+}}_{{[0-9]+}}
 }
diff --git a/test/CodeGen/Mips/memcpy.ll b/test/CodeGen/Mips/memcpy.ll
index 39764a936381..5c4ebb27dde4 100644
--- a/test/CodeGen/Mips/memcpy.ll
+++ b/test/CodeGen/Mips/memcpy.ll
@@ -8,9 +8,9 @@ define void @foo1(%struct.S1* %s1, i8 signext %n) nounwind {
 entry:
 ; CHECK-NOT: call16(memcpy
 
-  %arraydecay = getelementptr inbounds %struct.S1* %s1, i32 0, i32 1, i32 0
-  tail call void @llvm.memcpy.p0i8.p0i8.i32(i8* %arraydecay, i8* getelementptr inbounds ([31 x i8]* @.str, i32 0, i32 0), i32 31, i32 1, i1 false)
-  %arrayidx = getelementptr inbounds %struct.S1* %s1, i32 0, i32 1, i32 40
+  %arraydecay = getelementptr inbounds %struct.S1, %struct.S1* %s1, i32 0, i32 1, i32 0
+  tail call void @llvm.memcpy.p0i8.p0i8.i32(i8* %arraydecay, i8* getelementptr inbounds ([31 x i8], [31 x i8]* @.str, i32 0, i32 0), i32 31, i32 1, i1 false)
+  %arrayidx = getelementptr inbounds %struct.S1, %struct.S1* %s1, i32 0, i32 1, i32 40
   store i8 %n, i8* %arrayidx, align 1
   ret void
 }
diff --git a/test/CodeGen/Mips/micromips-addiu.ll b/test/CodeGen/Mips/micromips-addiu.ll
index c5bee34028c8..e0743c9c088b 100644
--- a/test/CodeGen/Mips/micromips-addiu.ll
+++ b/test/CodeGen/Mips/micromips-addiu.ll
@@ -8,20 +8,20 @@
 
 define i32 @main() nounwind {
 entry:
-  %0 = load i32* @x, align 4
+  %0 = load i32, i32* @x, align 4
   %addiu1 = add i32 %0, -7
-  %call1 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds
-                                  ([7 x i8]* @.str, i32 0, i32 0), i32 %addiu1)
+  %call1 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds
+                                  ([7 x i8], [7 x i8]* @.str, i32 0, i32 0), i32 %addiu1)
 
-  %1 = load i32* @y, align 4
+  %1 = load i32, i32* @y, align 4
   %addiu2 = add i32 %1, 55
-  %call2 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds
-                                  ([7 x i8]* @.str, i32 0, i32 0), i32 %addiu2)
+  %call2 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds
+                                  ([7 x i8], [7 x i8]* @.str, i32 0, i32 0), i32 %addiu2)
 
-  %2 = load i32* @z, align 4
+  %2 = load i32, i32* @z, align 4
   %addiu3 = add i32 %2, 24
-  %call3 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds
-                                  ([7 x i8]* @.str, i32 0, i32 0), i32 %addiu3)
+  %call3 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds
+                                  ([7 x i8], [7 x i8]* @.str, i32 0, i32 0), i32 %addiu3)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/micromips-addu16.ll b/test/CodeGen/Mips/micromips-addu16.ll
new file mode 100644
index 000000000000..3ecdf2488d2c
--- /dev/null
+++ b/test/CodeGen/Mips/micromips-addu16.ll
@@ -0,0 +1,18 @@
+; RUN: llc -march=mipsel -mcpu=mips32r2 -mattr=+micromips \
+; RUN:   -relocation-model=pic -O3 < %s | FileCheck %s
+
+define i32 @main() {
+entry:
+  %retval = alloca i32, align 4
+  %a = alloca i32, align 4
+  %b = alloca i32, align 4
+  %c = alloca i32, align 4
+  store i32 0, i32* %retval
+  %0 = load i32, i32* %b, align 4
+  %1 = load i32, i32* %c, align 4
+  %add = add nsw i32 %0, %1
+  store i32 %add, i32* %a, align 4
+  ret i32 0
+}
+
+; CHECK: addu16
diff --git a/test/CodeGen/Mips/micromips-and16.ll b/test/CodeGen/Mips/micromips-and16.ll
new file mode 100644
index 000000000000..d0a16ac28a09
--- /dev/null
+++ b/test/CodeGen/Mips/micromips-and16.ll
@@ -0,0 +1,18 @@
+; RUN: llc -march=mipsel -mcpu=mips32r2 -mattr=+micromips \
+; RUN:   -relocation-model=pic -O3 < %s | FileCheck %s
+
+define i32 @main() {
+entry:
+  %retval = alloca i32, align 4
+  %a = alloca i32, align 4
+  %b = alloca i32, align 4
+  %c = alloca i32, align 4
+  store i32 0, i32* %retval
+  %0 = load i32, i32* %b, align 4
+  %1 = load i32, i32* %c, align 4
+  %and = and i32 %0, %1
+  store i32 %and, i32* %a, align 4
+  ret i32 0
+}
+
+; CHECK: and16
diff --git a/test/CodeGen/Mips/micromips-andi.ll b/test/CodeGen/Mips/micromips-andi.ll
index b82d2b09eae4..cd7a794cd1be 100644
--- a/test/CodeGen/Mips/micromips-andi.ll
+++ b/test/CodeGen/Mips/micromips-andi.ll
@@ -7,15 +7,15 @@
 
 define i32 @main() nounwind {
 entry:
-  %0 = load i32* @x, align 4
+  %0 = load i32, i32* @x, align 4
   %and1 = and i32 %0, 4
-  %call1 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds
-                                  ([7 x i8]* @.str, i32 0, i32 0), i32 %and1)
+  %call1 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds
+                                  ([7 x i8], [7 x i8]* @.str, i32 0, i32 0), i32 %and1)
 
-  %1 = load i32* @y, align 4
+  %1 = load i32, i32* @y, align 4
   %and2 = and i32 %1, 5
-  %call2 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds
-                                  ([7 x i8]* @.str, i32 0, i32 0), i32 %and2)
+  %call2 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds
+                                  ([7 x i8], [7 x i8]* @.str, i32 0, i32 0), i32 %and2)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/micromips-compact-branches.ll b/test/CodeGen/Mips/micromips-compact-branches.ll
index 670f9a05064f..c689944d386b 100644
--- a/test/CodeGen/Mips/micromips-compact-branches.ll
+++ b/test/CodeGen/Mips/micromips-compact-branches.ll
@@ -4,7 +4,7 @@
 define void @main() nounwind uwtable {
 entry:
   %x = alloca i32, align 4
-  %0 = load i32* %x, align 4
+  %0 = load i32, i32* %x, align 4
   %cmp = icmp eq i32 %0, 0
   br i1 %cmp, label %if.then, label %if.end
 
diff --git a/test/CodeGen/Mips/micromips-compact-jump.ll b/test/CodeGen/Mips/micromips-compact-jump.ll
new file mode 100644
index 000000000000..70cff84e967f
--- /dev/null
+++ b/test/CodeGen/Mips/micromips-compact-jump.ll
@@ -0,0 +1,11 @@
+; RUN: llc -march=mipsel -mcpu=mips32r2 -mattr=+micromips \
+; RUN:   -disable-mips-delay-filler -O3 < %s | FileCheck %s
+
+define i32 @foo(i32 signext %a) #0 {
+entry:
+  ret i32 0
+}
+
+declare i32 @bar(i32 signext) #1
+
+; CHECK:      jrc
diff --git a/test/CodeGen/Mips/micromips-delay-slot-jr.ll b/test/CodeGen/Mips/micromips-delay-slot-jr.ll
index df593b35e2a6..c6636ff5b4c7 100644
--- a/test/CodeGen/Mips/micromips-delay-slot-jr.ll
+++ b/test/CodeGen/Mips/micromips-delay-slot-jr.ll
@@ -11,14 +11,14 @@ entry:
 
 L1:                                               ; preds = %entry, %L1
   %i.0 = phi i32 [ 0, %entry ], [ %inc, %L1 ]
-  %puts = tail call i32 @puts(i8* getelementptr inbounds ([2 x i8]* @str, i32 0, i32 0))
+  %puts = tail call i32 @puts(i8* getelementptr inbounds ([2 x i8], [2 x i8]* @str, i32 0, i32 0))
   %inc = add i32 %i.0, 1
-  %arrayidx = getelementptr inbounds [3 x i8*]* @main.L, i32 0, i32 %i.0
-  %0 = load i8** %arrayidx, align 4, !tbaa !1
+  %arrayidx = getelementptr inbounds [3 x i8*], [3 x i8*]* @main.L, i32 0, i32 %i.0
+  %0 = load i8*, i8** %arrayidx, align 4, !tbaa !1
   indirectbr i8* %0, [label %L1, label %L2]
 
 L2:                                               ; preds = %L1
-  %puts2 = tail call i32 @puts(i8* getelementptr inbounds ([2 x i8]* @str2, i32 0, i32 0))
+  %puts2 = tail call i32 @puts(i8* getelementptr inbounds ([2 x i8], [2 x i8]* @str2, i32 0, i32 0))
   ret i32 0
 }
 
@@ -29,20 +29,18 @@ declare i32 @puts(i8* nocapture readonly) #1
 !3 = !{!"omnipotent char", !4, i64 0}
 !4 = !{!"Simple C/C++ TBAA"}
 
-; CHECK:      jr
-; CHECK-NEXT: nop
+; CHECK:      jrc
 
 %struct.foostruct = type { [3 x float] }
 %struct.barstruct = type { %struct.foostruct, float }
 @bar_ary = common global [4 x %struct.barstruct] zeroinitializer, align 4
 define float* @spooky(i32 signext %i) #0 {
 
-  %safe = getelementptr inbounds [4 x %struct.barstruct]* @bar_ary, i32 0, i32 %i, i32 1
+  %safe = getelementptr inbounds [4 x %struct.barstruct], [4 x %struct.barstruct]* @bar_ary, i32 0, i32 %i, i32 1
   store float 1.420000e+02, float* %safe, align 4, !tbaa !1
   ret float* %safe
 }
 
 ; CHECK:      spooky:
-; CHECK:      jr $ra
-; CHECK-NEXT: nop
+; CHECK:      jrc $ra
 
diff --git a/test/CodeGen/Mips/micromips-delay-slot.ll b/test/CodeGen/Mips/micromips-delay-slot.ll
index b5f6c56235bc..ef6546232835 100644
--- a/test/CodeGen/Mips/micromips-delay-slot.ll
+++ b/test/CodeGen/Mips/micromips-delay-slot.ll
@@ -6,7 +6,7 @@ define i32 @foo(i32 signext %a) #0 {
 entry:
   %a.addr = alloca i32, align 4
   store i32 %a, i32* %a.addr, align 4
-  %0 = load i32* %a.addr, align 4
+  %0 = load i32, i32* %a.addr, align 4
   %shl = shl i32 %0, 2
   %call = call i32 @bar(i32 signext %shl)
   ret i32 %call
diff --git a/test/CodeGen/Mips/micromips-gp-rc.ll b/test/CodeGen/Mips/micromips-gp-rc.ll
new file mode 100644
index 000000000000..f139f7a8486d
--- /dev/null
+++ b/test/CodeGen/Mips/micromips-gp-rc.ll
@@ -0,0 +1,18 @@
+; RUN: llc -march=mipsel -mcpu=mips32r2 -mattr=+micromips \
+; RUN:   -relocation-model=pic -O3 < %s | FileCheck %s
+
+@g = external global i32
+
+; Function Attrs: noreturn nounwind
+define void @foo() #0 {
+entry:
+  %0 = load i32, i32* @g, align 4
+  tail call void @exit(i32 signext %0)
+  unreachable
+}
+
+; Function Attrs: noreturn
+declare void @exit(i32 signext)
+
+; CHECK: move $gp, ${{[0-9]+}}
+
diff --git a/test/CodeGen/Mips/micromips-jal.ll b/test/CodeGen/Mips/micromips-jal.ll
index fccc22919728..51832fe333dc 100644
--- a/test/CodeGen/Mips/micromips-jal.ll
+++ b/test/CodeGen/Mips/micromips-jal.ll
@@ -7,8 +7,8 @@ entry:
   %b.addr = alloca i32, align 4
   store i32 %a, i32* %a.addr, align 4
   store i32 %b, i32* %b.addr, align 4
-  %0 = load i32* %a.addr, align 4
-  %1 = load i32* %b.addr, align 4
+  %0 = load i32, i32* %a.addr, align 4
+  %1 = load i32, i32* %b.addr, align 4
   %add = add nsw i32 %0, %1
   ret i32 %add
 }
@@ -20,11 +20,11 @@ entry:
   %y = alloca i32, align 4
   %z = alloca i32, align 4
   store i32 0, i32* %retval
-  %0 = load i32* %y, align 4
-  %1 = load i32* %z, align 4
+  %0 = load i32, i32* %y, align 4
+  %1 = load i32, i32* %z, align 4
   %call = call i32 @sum(i32 %0, i32 %1)
   store i32 %call, i32* %x, align 4
-  %2 = load i32* %x, align 4
+  %2 = load i32, i32* %x, align 4
   ret i32 %2
 }
 
diff --git a/test/CodeGen/Mips/micromips-load-effective-address.ll b/test/CodeGen/Mips/micromips-load-effective-address.ll
index afba760f0e62..470458098212 100644
--- a/test/CodeGen/Mips/micromips-load-effective-address.ll
+++ b/test/CodeGen/Mips/micromips-load-effective-address.ll
@@ -7,10 +7,10 @@ entry:
   %y.addr = alloca i32*, align 8
   store i32* %x, i32** %x.addr, align 8
   store i32* %y, i32** %y.addr, align 8
-  %0 = load i32** %x.addr, align 8
-  %1 = load i32* %0, align 4
-  %2 = load i32** %y.addr, align 8
-  %3 = load i32* %2, align 4
+  %0 = load i32*, i32** %x.addr, align 8
+  %1 = load i32, i32* %0, align 4
+  %2 = load i32*, i32** %y.addr, align 8
+  %3 = load i32, i32* %2, align 4
   %add = add nsw i32 %1, %3
   ret i32 %add
 }
diff --git a/test/CodeGen/Mips/micromips-not16.ll b/test/CodeGen/Mips/micromips-not16.ll
new file mode 100644
index 000000000000..d31aefae6f0d
--- /dev/null
+++ b/test/CodeGen/Mips/micromips-not16.ll
@@ -0,0 +1,26 @@
+; RUN: llc -march=mipsel -mcpu=mips32r2 -mattr=+micromips \
+; RUN:   -relocation-model=pic -O3 < %s | FileCheck %s
+
+define i32 @main() {
+entry:
+  %retval = alloca i32, align 4
+  %x = alloca i64, align 8
+  store i32 0, i32* %retval
+  %0 = load i64, i64* %x, align 8
+  %cmp = icmp ne i64 %0, 9223372036854775807
+  br i1 %cmp, label %if.then, label %if.end
+
+if.then:
+  store i32 1, i32* %retval
+  br label %return
+
+if.end:
+  store i32 0, i32* %retval
+  br label %return
+
+return:
+  %1 = load i32, i32* %retval
+  ret i32 %1
+}
+
+; CHECK: not16
diff --git a/test/CodeGen/Mips/micromips-or16.ll b/test/CodeGen/Mips/micromips-or16.ll
new file mode 100644
index 000000000000..82ea9c687df4
--- /dev/null
+++ b/test/CodeGen/Mips/micromips-or16.ll
@@ -0,0 +1,18 @@
+; RUN: llc -march=mipsel -mcpu=mips32r2 -mattr=+micromips \
+; RUN:   -relocation-model=pic -O3 < %s | FileCheck %s
+
+define i32 @main() {
+entry:
+  %retval = alloca i32, align 4
+  %a = alloca i32, align 4
+  %b = alloca i32, align 4
+  %c = alloca i32, align 4
+  store i32 0, i32* %retval
+  %0 = load i32, i32* %b, align 4
+  %1 = load i32, i32* %c, align 4
+  %or = or i32 %0, %1
+  store i32 %or, i32* %a, align 4
+  ret i32 0
+}
+
+; CHECK: or16
diff --git a/test/CodeGen/Mips/micromips-rdhwr-directives.ll b/test/CodeGen/Mips/micromips-rdhwr-directives.ll
index af40a8796824..ebe4dddd0124 100644
--- a/test/CodeGen/Mips/micromips-rdhwr-directives.ll
+++ b/test/CodeGen/Mips/micromips-rdhwr-directives.ll
@@ -10,6 +10,6 @@ entry:
 ; CHECK: rdhwr
 ; CHECK: .set  pop
 
-  %0 = load i32* @a, align 4
+  %0 = load i32, i32* @a, align 4
   ret i32 %0
 }
diff --git a/test/CodeGen/Mips/micromips-shift.ll b/test/CodeGen/Mips/micromips-shift.ll
index 8215010bfc78..ed1bcbbf0831 100644
--- a/test/CodeGen/Mips/micromips-shift.ll
+++ b/test/CodeGen/Mips/micromips-shift.ll
@@ -8,11 +8,11 @@
 
 define i32 @shift_left() nounwind {
 entry:
-  %0 = load i32* @a, align 4
+  %0 = load i32, i32* @a, align 4
   %shl = shl i32 %0, 4
   store i32 %shl, i32* @b, align 4
 
-  %1 = load i32* @c, align 4
+  %1 = load i32, i32* @c, align 4
   %shl1 = shl i32 %1, 10
   store i32 %shl1, i32* @d, align 4
 
@@ -29,11 +29,11 @@ entry:
 
 define i32 @shift_right() nounwind {
 entry:
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %shr = lshr i32 %0, 4
   store i32 %shr, i32* @j, align 4
 
-  %1 = load i32* @m, align 4
+  %1 = load i32, i32* @m, align 4
   %shr1 = lshr i32 %1, 10
   store i32 %shr1, i32* @n, align 4
 
diff --git a/test/CodeGen/Mips/micromips-subu16.ll b/test/CodeGen/Mips/micromips-subu16.ll
new file mode 100644
index 000000000000..d415574f443a
--- /dev/null
+++ b/test/CodeGen/Mips/micromips-subu16.ll
@@ -0,0 +1,18 @@
+; RUN: llc -march=mipsel -mcpu=mips32r2 -mattr=+micromips \
+; RUN:   -relocation-model=pic -O3 < %s | FileCheck %s
+
+define i32 @main() {
+entry:
+  %retval = alloca i32, align 4
+  %a = alloca i32, align 4
+  %b = alloca i32, align 4
+  %c = alloca i32, align 4
+  store i32 0, i32* %retval
+  %0 = load i32, i32* %b, align 4
+  %1 = load i32, i32* %c, align 4
+  %sub = sub nsw i32 %0, %1
+  store i32 %sub, i32* %a, align 4
+  ret i32 0
+}
+
+; CHECK: subu16
diff --git a/test/CodeGen/Mips/micromips-sw-lw-16.ll b/test/CodeGen/Mips/micromips-sw-lw-16.ll
new file mode 100644
index 000000000000..358372649b5d
--- /dev/null
+++ b/test/CodeGen/Mips/micromips-sw-lw-16.ll
@@ -0,0 +1,27 @@
+; RUN: llc %s -march=mipsel -mattr=micromips -filetype=asm \
+; RUN: -relocation-model=pic -O3 -o - | FileCheck %s
+
+; Function Attrs: noinline nounwind
+define void @bar(i32* %p) #0 {
+entry:
+  %p.addr = alloca i32*, align 4
+  store i32* %p, i32** %p.addr, align 4
+  %0 = load i32*, i32** %p.addr, align 4
+  %1 = load i32, i32* %0, align 4
+  %add = add nsw i32 7, %1
+  %2 = load i32*, i32** %p.addr, align 4
+  store i32 %add, i32* %2, align 4
+  %3 = load i32*, i32** %p.addr, align 4
+  %add.ptr = getelementptr inbounds i32, i32* %3, i32 1
+  %4 = load i32, i32* %add.ptr, align 4
+  %add1 = add nsw i32 7, %4
+  %5 = load i32*, i32** %p.addr, align 4
+  %add.ptr2 = getelementptr inbounds i32, i32* %5, i32 1
+  store i32 %add1, i32* %add.ptr2, align 4
+  ret void
+}
+
+; CHECK: lw16 ${{[0-9]+}}, 0($4)
+; CHECK: sw16 ${{[0-9]+}}, 0($4)
+; CHECK: lw16 ${{[0-9]+}}, 4(${{[0-9]+}})
+; CHECK: sw16 ${{[0-9]+}}, 4(${{[0-9]+}})
diff --git a/test/CodeGen/Mips/micromips-xor16.ll b/test/CodeGen/Mips/micromips-xor16.ll
new file mode 100644
index 000000000000..53c75acd4d3b
--- /dev/null
+++ b/test/CodeGen/Mips/micromips-xor16.ll
@@ -0,0 +1,18 @@
+; RUN: llc -march=mipsel -mcpu=mips32r2 -mattr=+micromips \
+; RUN:   -relocation-model=pic -O3 < %s | FileCheck %s
+
+define i32 @main() {
+entry:
+  %retval = alloca i32, align 4
+  %a = alloca i32, align 4
+  %b = alloca i32, align 4
+  %c = alloca i32, align 4
+  store i32 0, i32* %retval
+  %0 = load i32, i32* %b, align 4
+  %1 = load i32, i32* %c, align 4
+  %xor = xor i32 %0, %1
+  store i32 %xor, i32* %a, align 4
+  ret i32 0
+}
+
+; CHECK: xor16
diff --git a/test/CodeGen/Mips/mips16_32_8.ll b/test/CodeGen/Mips/mips16_32_8.ll
index 2f5bc219cf35..5e03928a11f1 100644
--- a/test/CodeGen/Mips/mips16_32_8.ll
+++ b/test/CodeGen/Mips/mips16_32_8.ll
@@ -22,13 +22,13 @@ entry:
 define void @nofoo() #1 {
 entry:
   store i32 20, i32* @i, align 4
-  %0 = load float* @x, align 4
-  %1 = load float* @y, align 4
+  %0 = load float, float* @x, align 4
+  %1 = load float, float* @y, align 4
   %add = fadd float %0, %1
   store float %add, float* @f, align 4
-  %2 = load float* @f, align 4
+  %2 = load float, float* @f, align 4
   %conv = fpext float %2 to double
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([8 x i8]* @.str, i32 0, i32 0), double %conv)
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([8 x i8], [8 x i8]* @.str, i32 0, i32 0), double %conv)
   ret void
 }
 
@@ -48,11 +48,11 @@ declare i32 @printf(i8*, ...) #2
 define i32 @main() #3 {
 entry:
   call void @foo()
-  %0 = load i32* @i, align 4
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([11 x i8]* @.str1, i32 0, i32 0), i32 %0)
+  %0 = load i32, i32* @i, align 4
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([11 x i8], [11 x i8]* @.str1, i32 0, i32 0), i32 %0)
   call void @nofoo()
-  %1 = load i32* @i, align 4
-  %call1 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([13 x i8]* @.str2, i32 0, i32 0), i32 %1)
+  %1 = load i32, i32* @i, align 4
+  %call1 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([13 x i8], [13 x i8]* @.str2, i32 0, i32 0), i32 %1)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/mips16_fpret.ll b/test/CodeGen/Mips/mips16_fpret.ll
index fe87604d6107..0f09c4105a17 100644
--- a/test/CodeGen/Mips/mips16_fpret.ll
+++ b/test/CodeGen/Mips/mips16_fpret.ll
@@ -11,7 +11,7 @@
 
 define float @foox()  {
 entry:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   ret float %0
 ; 1: 	.ent	foox
 ; 1:	lw	$2, %lo(x)(${{[0-9]+}})
@@ -20,7 +20,7 @@ entry:
 
 define double @foodx()  {
 entry:
-  %0 = load double* @dx, align 8
+  %0 = load double, double* @dx, align 8
   ret double %0
 ; 1: 	.ent	foodx
 ; 1: 	lw	$2, %lo(dx)(${{[0-9]+}})
@@ -34,13 +34,13 @@ entry:
 define { float, float } @foocx()  {
 entry:
   %retval = alloca { float, float }, align 4
-  %cx.real = load float* getelementptr inbounds ({ float, float }* @cx, i32 0, i32 0)
-  %cx.imag = load float* getelementptr inbounds ({ float, float }* @cx, i32 0, i32 1)
-  %real = getelementptr inbounds { float, float }* %retval, i32 0, i32 0
-  %imag = getelementptr inbounds { float, float }* %retval, i32 0, i32 1
+  %cx.real = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @cx, i32 0, i32 0)
+  %cx.imag = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @cx, i32 0, i32 1)
+  %real = getelementptr inbounds { float, float }, { float, float }* %retval, i32 0, i32 0
+  %imag = getelementptr inbounds { float, float }, { float, float }* %retval, i32 0, i32 1
   store float %cx.real, float* %real
   store float %cx.imag, float* %imag
-  %0 = load { float, float }* %retval
+  %0 = load { float, float }, { float, float }* %retval
   ret { float, float } %0
 ; 1: 	.ent	foocx
 ; 1: 	lw	$2, %lo(cx)(${{[0-9]+}})
@@ -53,13 +53,13 @@ entry:
 define { double, double } @foodcx()  {
 entry:
   %retval = alloca { double, double }, align 8
-  %dcx.real = load double* getelementptr inbounds ({ double, double }* @dcx, i32 0, i32 0)
-  %dcx.imag = load double* getelementptr inbounds ({ double, double }* @dcx, i32 0, i32 1)
-  %real = getelementptr inbounds { double, double }* %retval, i32 0, i32 0
-  %imag = getelementptr inbounds { double, double }* %retval, i32 0, i32 1
+  %dcx.real = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @dcx, i32 0, i32 0)
+  %dcx.imag = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @dcx, i32 0, i32 1)
+  %real = getelementptr inbounds { double, double }, { double, double }* %retval, i32 0, i32 0
+  %imag = getelementptr inbounds { double, double }, { double, double }* %retval, i32 0, i32 1
   store double %dcx.real, double* %real
   store double %dcx.imag, double* %imag
-  %0 = load { double, double }* %retval
+  %0 = load { double, double }, { double, double }* %retval
   ret { double, double } %0
 ; 1: 	.ent	foodcx
 ; 1: 	lw	${{[0-9]}}, %lo(dcx)(${{[0-9]+}})
diff --git a/test/CodeGen/Mips/mips16ex.ll b/test/CodeGen/Mips/mips16ex.ll
index a1a99191595d..68b584604b27 100644
--- a/test/CodeGen/Mips/mips16ex.ll
+++ b/test/CodeGen/Mips/mips16ex.ll
@@ -1,8 +1,10 @@
 ; RUN: llc  -march=mipsel -mcpu=mips16 -relocation-model=pic -O3 < %s | FileCheck %s -check-prefix=16
 
-;16: .cfi_personality
+;16: main:
 ;16-NEXT: [[TMP:.*]]:
-;16-NEXT: $eh_func_begin0 = ([[TMP]])
+;16-NEXT: $func_begin0 = ([[TMP]])
+;16-NEXT: .cfi_startproc
+;16-NEXT: .cfi_personality
 @.str = private unnamed_addr constant [7 x i8] c"hello\0A\00", align 1
 @_ZTIi = external constant i8*
 @.str1 = private unnamed_addr constant [15 x i8] c"exception %i \0A\00", align 1
@@ -14,7 +16,7 @@ entry:
   %ehselector.slot = alloca i32
   %e = alloca i32, align 4
   store i32 0, i32* %retval
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([7 x i8]* @.str, i32 0, i32 0))
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([7 x i8], [7 x i8]* @.str, i32 0, i32 0))
   %exception = call i8* @__cxa_allocate_exception(i32 4) nounwind
   %0 = bitcast i8* %exception to i32*
   store i32 20, i32* %0
@@ -31,19 +33,19 @@ lpad:                                             ; preds = %entry
   br label %catch.dispatch
 
 catch.dispatch:                                   ; preds = %lpad
-  %sel = load i32* %ehselector.slot
+  %sel = load i32, i32* %ehselector.slot
   %4 = call i32 @llvm.eh.typeid.for(i8* bitcast (i8** @_ZTIi to i8*)) nounwind
   %matches = icmp eq i32 %sel, %4
   br i1 %matches, label %catch, label %eh.resume
 
 catch:                                            ; preds = %catch.dispatch
-  %exn = load i8** %exn.slot
+  %exn = load i8*, i8** %exn.slot
   %5 = call i8* @__cxa_begin_catch(i8* %exn) nounwind
   %6 = bitcast i8* %5 to i32*
-  %exn.scalar = load i32* %6
+  %exn.scalar = load i32, i32* %6
   store i32 %exn.scalar, i32* %e, align 4
-  %7 = load i32* %e, align 4
-  %call2 = invoke i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([15 x i8]* @.str1, i32 0, i32 0), i32 %7)
+  %7 = load i32, i32* %e, align 4
+  %call2 = invoke i32 (i8*, ...) @printf(i8* getelementptr inbounds ([15 x i8], [15 x i8]* @.str1, i32 0, i32 0), i32 %7)
           to label %invoke.cont unwind label %lpad1
 
 invoke.cont:                                      ; preds = %catch
@@ -64,8 +66,8 @@ lpad1:                                            ; preds = %catch
   br label %eh.resume
 
 eh.resume:                                        ; preds = %lpad1, %catch.dispatch
-  %exn3 = load i8** %exn.slot
-  %sel4 = load i32* %ehselector.slot
+  %exn3 = load i8*, i8** %exn.slot
+  %sel4 = load i32, i32* %ehselector.slot
   %lpad.val = insertvalue { i8*, i32 } undef, i8* %exn3, 0
   %lpad.val5 = insertvalue { i8*, i32 } %lpad.val, i32 %sel4, 1
   resume { i8*, i32 } %lpad.val5
diff --git a/test/CodeGen/Mips/mips16fpe.ll b/test/CodeGen/Mips/mips16fpe.ll
index 987980e080ff..f8b916da3a49 100644
--- a/test/CodeGen/Mips/mips16fpe.ll
+++ b/test/CodeGen/Mips/mips16fpe.ll
@@ -42,8 +42,8 @@
 define void @test_addsf3() nounwind {
 entry:
 ;16hf-LABEL: test_addsf3:
-  %0 = load float* @x, align 4
-  %1 = load float* @y, align 4
+  %0 = load float, float* @x, align 4
+  %1 = load float, float* @y, align 4
   %add = fadd float %0, %1
   store float %add, float* @addsf3_result, align 4
 ;16hf:  lw	${{[0-9]+}}, %call16(__mips16_addsf3)(${{[0-9]+}})
@@ -53,8 +53,8 @@ entry:
 define void @test_adddf3() nounwind {
 entry:
 ;16hf-LABEL: test_adddf3:
-  %0 = load double* @xd, align 8
-  %1 = load double* @yd, align 8
+  %0 = load double, double* @xd, align 8
+  %1 = load double, double* @yd, align 8
   %add = fadd double %0, %1
   store double %add, double* @adddf3_result, align 8
 ;16hf:  lw	${{[0-9]+}}, %call16(__mips16_adddf3)(${{[0-9]+}})
@@ -64,8 +64,8 @@ entry:
 define void @test_subsf3() nounwind {
 entry:
 ;16hf-LABEL: test_subsf3:
-  %0 = load float* @x, align 4
-  %1 = load float* @y, align 4
+  %0 = load float, float* @x, align 4
+  %1 = load float, float* @y, align 4
   %sub = fsub float %0, %1
   store float %sub, float* @subsf3_result, align 4
 ;16hf:  lw	${{[0-9]+}}, %call16(__mips16_subsf3)(${{[0-9]+}})
@@ -75,8 +75,8 @@ entry:
 define void @test_subdf3() nounwind {
 entry:
 ;16hf-LABEL: test_subdf3:
-  %0 = load double* @xd, align 8
-  %1 = load double* @yd, align 8
+  %0 = load double, double* @xd, align 8
+  %1 = load double, double* @yd, align 8
   %sub = fsub double %0, %1
   store double %sub, double* @subdf3_result, align 8
 ;16hf:  lw	${{[0-9]+}}, %call16(__mips16_subdf3)(${{[0-9]+}})
@@ -86,8 +86,8 @@ entry:
 define void @test_mulsf3() nounwind {
 entry:
 ;16hf-LABEL: test_mulsf3:
-  %0 = load float* @x, align 4
-  %1 = load float* @y, align 4
+  %0 = load float, float* @x, align 4
+  %1 = load float, float* @y, align 4
   %mul = fmul float %0, %1
   store float %mul, float* @mulsf3_result, align 4
 ;16hf:  lw	${{[0-9]+}}, %call16(__mips16_mulsf3)(${{[0-9]+}})
@@ -97,8 +97,8 @@ entry:
 define void @test_muldf3() nounwind {
 entry:
 ;16hf-LABEL: test_muldf3:
-  %0 = load double* @xd, align 8
-  %1 = load double* @yd, align 8
+  %0 = load double, double* @xd, align 8
+  %1 = load double, double* @yd, align 8
   %mul = fmul double %0, %1
   store double %mul, double* @muldf3_result, align 8
 ;16hf:  lw	${{[0-9]+}}, %call16(__mips16_muldf3)(${{[0-9]+}})
@@ -108,8 +108,8 @@ entry:
 define void @test_divsf3() nounwind {
 entry:
 ;16hf-LABEL: test_divsf3:
-  %0 = load float* @y, align 4
-  %1 = load float* @x, align 4
+  %0 = load float, float* @y, align 4
+  %1 = load float, float* @x, align 4
   %div = fdiv float %0, %1
   store float %div, float* @divsf3_result, align 4
 ;16hf:  lw	${{[0-9]+}}, %call16(__mips16_divsf3)(${{[0-9]+}})
@@ -119,9 +119,9 @@ entry:
 define void @test_divdf3() nounwind {
 entry:
 ;16hf-LABEL: test_divdf3:
-  %0 = load double* @yd, align 8
+  %0 = load double, double* @yd, align 8
   %mul = fmul double %0, 2.000000e+00
-  %1 = load double* @xd, align 8
+  %1 = load double, double* @xd, align 8
   %div = fdiv double %mul, %1
   store double %div, double* @divdf3_result, align 8
 ;16hf:  lw	${{[0-9]+}}, %call16(__mips16_divdf3)(${{[0-9]+}})
@@ -131,7 +131,7 @@ entry:
 define void @test_extendsfdf2() nounwind {
 entry:
 ;16hf-LABEL: test_extendsfdf2:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   %conv = fpext float %0 to double
   store double %conv, double* @extendsfdf2_result, align 8
 ;16hf:  lw	${{[0-9]+}}, %call16(__mips16_extendsfdf2)(${{[0-9]+}})
@@ -141,7 +141,7 @@ entry:
 define void @test_truncdfsf2() nounwind {
 entry:
 ;16hf-LABEL: test_truncdfsf2:
-  %0 = load double* @xd2, align 8
+  %0 = load double, double* @xd2, align 8
   %conv = fptrunc double %0 to float
   store float %conv, float* @truncdfsf2_result, align 4
 ;16hf:  lw	${{[0-9]+}}, %call16(__mips16_truncdfsf2)(${{[0-9]+}})
@@ -151,7 +151,7 @@ entry:
 define void @test_fix_truncsfsi() nounwind {
 entry:
 ;16hf-LABEL: test_fix_truncsfsi:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   %conv = fptosi float %0 to i32
   store i32 %conv, i32* @fix_truncsfsi_result, align 4
 ;16hf:  lw	${{[0-9]+}}, %call16(__mips16_fix_truncsfsi)(${{[0-9]+}})
@@ -161,7 +161,7 @@ entry:
 define void @test_fix_truncdfsi() nounwind {
 entry:
 ;16hf-LABEL: test_fix_truncdfsi:
-  %0 = load double* @xd, align 8
+  %0 = load double, double* @xd, align 8
   %conv = fptosi double %0 to i32
   store i32 %conv, i32* @fix_truncdfsi_result, align 4
 ;16hf:  lw	${{[0-9]+}}, %call16(__mips16_fix_truncdfsi)(${{[0-9]+}})
@@ -171,7 +171,7 @@ entry:
 define void @test_floatsisf() nounwind {
 entry:
 ;16hf-LABEL: test_floatsisf:
-  %0 = load i32* @si, align 4
+  %0 = load i32, i32* @si, align 4
   %conv = sitofp i32 %0 to float
   store float %conv, float* @floatsisf_result, align 4
 ;16hf:  lw	${{[0-9]+}}, %call16(__mips16_floatsisf)(${{[0-9]+}})
@@ -181,7 +181,7 @@ entry:
 define void @test_floatsidf() nounwind {
 entry:
 ;16hf-LABEL: test_floatsidf:
-  %0 = load i32* @si, align 4
+  %0 = load i32, i32* @si, align 4
   %conv = sitofp i32 %0 to double
   store double %conv, double* @floatsidf_result, align 8
 ;16hf:  lw	${{[0-9]+}}, %call16(__mips16_floatsidf)(${{[0-9]+}})
@@ -191,7 +191,7 @@ entry:
 define void @test_floatunsisf() nounwind {
 entry:
 ;16hf-LABEL: test_floatunsisf:
-  %0 = load i32* @ui, align 4
+  %0 = load i32, i32* @ui, align 4
   %conv = uitofp i32 %0 to float
   store float %conv, float* @floatunsisf_result, align 4
 ;16hf:  lw	${{[0-9]+}}, %call16(__mips16_floatunsisf)(${{[0-9]+}})
@@ -201,7 +201,7 @@ entry:
 define void @test_floatunsidf() nounwind {
 entry:
 ;16hf-LABEL: test_floatunsidf:
-  %0 = load i32* @ui, align 4
+  %0 = load i32, i32* @ui, align 4
   %conv = uitofp i32 %0 to double
   store double %conv, double* @floatunsidf_result, align 8
 ;16hf:  lw	${{[0-9]+}}, %call16(__mips16_floatunsidf)(${{[0-9]+}})
@@ -211,8 +211,8 @@ entry:
 define void @test_eqsf2() nounwind {
 entry:
 ;16hf-LABEL: test_eqsf2:
-  %0 = load float* @x, align 4
-  %1 = load float* @xx, align 4
+  %0 = load float, float* @x, align 4
+  %1 = load float, float* @xx, align 4
   %cmp = fcmp oeq float %0, %1
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @eqsf2_result, align 4
@@ -223,8 +223,8 @@ entry:
 define void @test_eqdf2() nounwind {
 entry:
 ;16hf-LABEL: test_eqdf2:
-  %0 = load double* @xd, align 8
-  %1 = load double* @xxd, align 8
+  %0 = load double, double* @xd, align 8
+  %1 = load double, double* @xxd, align 8
   %cmp = fcmp oeq double %0, %1
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @eqdf2_result, align 4
@@ -235,8 +235,8 @@ entry:
 define void @test_nesf2() nounwind {
 entry:
 ;16hf-LABEL: test_nesf2:
-  %0 = load float* @x, align 4
-  %1 = load float* @y, align 4
+  %0 = load float, float* @x, align 4
+  %1 = load float, float* @y, align 4
   %cmp = fcmp une float %0, %1
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @nesf2_result, align 4
@@ -247,8 +247,8 @@ entry:
 define void @test_nedf2() nounwind {
 entry:
 ;16hf-LABEL: test_nedf2:
-  %0 = load double* @xd, align 8
-  %1 = load double* @yd, align 8
+  %0 = load double, double* @xd, align 8
+  %1 = load double, double* @yd, align 8
   %cmp = fcmp une double %0, %1
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @nedf2_result, align 4
@@ -259,10 +259,10 @@ entry:
 define void @test_gesf2() nounwind {
 entry:
 ;16hf-LABEL: test_gesf2:
-  %0 = load float* @x, align 4
-  %1 = load float* @xx, align 4
+  %0 = load float, float* @x, align 4
+  %1 = load float, float* @xx, align 4
   %cmp = fcmp oge float %0, %1
-  %2 = load float* @y, align 4
+  %2 = load float, float* @y, align 4
   %cmp1 = fcmp oge float %2, %0
   %and3 = and i1 %cmp, %cmp1
   %and = zext i1 %and3 to i32
@@ -274,10 +274,10 @@ entry:
 define void @test_gedf2() nounwind {
 entry:
 ;16hf-LABEL: test_gedf2:
-  %0 = load double* @xd, align 8
-  %1 = load double* @xxd, align 8
+  %0 = load double, double* @xd, align 8
+  %1 = load double, double* @xxd, align 8
   %cmp = fcmp oge double %0, %1
-  %2 = load double* @yd, align 8
+  %2 = load double, double* @yd, align 8
   %cmp1 = fcmp oge double %2, %0
   %and3 = and i1 %cmp, %cmp1
   %and = zext i1 %and3 to i32
@@ -289,10 +289,10 @@ entry:
 define void @test_ltsf2() nounwind {
 entry:
 ;16hf-LABEL: test_ltsf2:
-  %0 = load float* @x, align 4
-  %1 = load float* @xx, align 4
+  %0 = load float, float* @x, align 4
+  %1 = load float, float* @xx, align 4
   %lnot = fcmp uge float %0, %1
-  %2 = load float* @y, align 4
+  %2 = load float, float* @y, align 4
   %cmp1 = fcmp olt float %0, %2
   %and2 = and i1 %lnot, %cmp1
   %and = zext i1 %and2 to i32
@@ -305,10 +305,10 @@ entry:
 define void @test_ltdf2() nounwind {
 entry:
 ;16hf-LABEL: test_ltdf2:
-  %0 = load double* @xd, align 8
-  %1 = load double* @xxd, align 8
+  %0 = load double, double* @xd, align 8
+  %1 = load double, double* @xxd, align 8
   %lnot = fcmp uge double %0, %1
-  %2 = load double* @yd, align 8
+  %2 = load double, double* @yd, align 8
   %cmp1 = fcmp olt double %0, %2
   %and2 = and i1 %lnot, %cmp1
   %and = zext i1 %and2 to i32
@@ -321,10 +321,10 @@ entry:
 define void @test_lesf2() nounwind {
 entry:
 ;16hf-LABEL: test_lesf2:
-  %0 = load float* @x, align 4
-  %1 = load float* @xx, align 4
+  %0 = load float, float* @x, align 4
+  %1 = load float, float* @xx, align 4
   %cmp = fcmp ole float %0, %1
-  %2 = load float* @y, align 4
+  %2 = load float, float* @y, align 4
   %cmp1 = fcmp ole float %0, %2
   %and3 = and i1 %cmp, %cmp1
   %and = zext i1 %and3 to i32
@@ -336,10 +336,10 @@ entry:
 define void @test_ledf2() nounwind {
 entry:
 ;16hf-LABEL: test_ledf2:
-  %0 = load double* @xd, align 8
-  %1 = load double* @xxd, align 8
+  %0 = load double, double* @xd, align 8
+  %1 = load double, double* @xxd, align 8
   %cmp = fcmp ole double %0, %1
-  %2 = load double* @yd, align 8
+  %2 = load double, double* @yd, align 8
   %cmp1 = fcmp ole double %0, %2
   %and3 = and i1 %cmp, %cmp1
   %and = zext i1 %and3 to i32
@@ -351,10 +351,10 @@ entry:
 define void @test_gtsf2() nounwind {
 entry:
 ;16hf-LABEL: test_gtsf2:
-  %0 = load float* @x, align 4
-  %1 = load float* @xx, align 4
+  %0 = load float, float* @x, align 4
+  %1 = load float, float* @xx, align 4
   %lnot = fcmp ule float %0, %1
-  %2 = load float* @y, align 4
+  %2 = load float, float* @y, align 4
   %cmp1 = fcmp ogt float %2, %0
   %and2 = and i1 %lnot, %cmp1
   %and = zext i1 %and2 to i32
@@ -366,10 +366,10 @@ entry:
 define void @test_gtdf2() nounwind {
 entry:
 ;16hf-LABEL: test_gtdf2:
-  %0 = load double* @xd, align 8
-  %1 = load double* @xxd, align 8
+  %0 = load double, double* @xd, align 8
+  %1 = load double, double* @xxd, align 8
   %lnot = fcmp ule double %0, %1
-  %2 = load double* @yd, align 8
+  %2 = load double, double* @yd, align 8
   %cmp1 = fcmp ogt double %2, %0
   %and2 = and i1 %lnot, %cmp1
   %and = zext i1 %and2 to i32
diff --git a/test/CodeGen/Mips/mips64-f128-call.ll b/test/CodeGen/Mips/mips64-f128-call.ll
index 455e540e5df1..9a093e6f9825 100644
--- a/test/CodeGen/Mips/mips64-f128-call.ll
+++ b/test/CodeGen/Mips/mips64-f128-call.ll
@@ -19,7 +19,7 @@ entry:
 
 define void @foo1() {
 entry:
-  %0 = load fp128* @gld0, align 16
+  %0 = load fp128, fp128* @gld0, align 16
   tail call void @foo2(fp128 %0)
   ret void
 }
@@ -38,7 +38,7 @@ define fp128 @foo3() {
 entry:
   %call = tail call fp128 @foo4()
   store fp128 %call, fp128* @gld0, align 16
-  %0 = load fp128* @gld1, align 16
+  %0 = load fp128, fp128* @gld1, align 16
   ret fp128 %0
 }
 
diff --git a/test/CodeGen/Mips/mips64-f128.ll b/test/CodeGen/Mips/mips64-f128.ll
index f0cbbd08d79a..d9c93810438f 100644
--- a/test/CodeGen/Mips/mips64-f128.ll
+++ b/test/CodeGen/Mips/mips64-f128.ll
@@ -1,10 +1,10 @@
-; RUN: llc -mtriple=mips64el-unknown-unknown -mcpu=mips4 -soft-float -O1 \
+; RUN: llc -mtriple=mips64el-unknown-unknown -mcpu=mips4 -mattr=+soft-float -O1 \
 ; RUN:     -disable-mips-delay-filler < %s | FileCheck %s -check-prefix=ALL -check-prefix=C_CC_FMT
-; RUN: llc -mtriple=mips64el-unknown-unknown -mcpu=mips64 -soft-float -O1 \
+; RUN: llc -mtriple=mips64el-unknown-unknown -mcpu=mips64 -mattr=+soft-float -O1 \
 ; RUN:     -disable-mips-delay-filler < %s | FileCheck %s -check-prefix=ALL -check-prefix=C_CC_FMT
-; RUN: llc -mtriple=mips64el-unknown-unknown -mcpu=mips64r2 -soft-float -O1 \
+; RUN: llc -mtriple=mips64el-unknown-unknown -mcpu=mips64r2 -mattr=+soft-float -O1 \
 ; RUN:     -disable-mips-delay-filler < %s | FileCheck %s -check-prefix=ALL -check-prefix=C_CC_FMT
-; RUN: llc -mtriple=mips64el-unknown-unknown -mcpu=mips64r6 -soft-float -O1 \
+; RUN: llc -mtriple=mips64el-unknown-unknown -mcpu=mips64r6 -mattr=+soft-float -O1 \
 ; RUN:     -disable-mips-delay-filler < %s | FileCheck %s -check-prefix=ALL -check-prefix=CMP_CC_FMT
 
 @gld0 = external global fp128
@@ -18,8 +18,8 @@
 
 define fp128 @addLD() {
 entry:
-  %0 = load fp128* @gld0, align 16
-  %1 = load fp128* @gld1, align 16
+  %0 = load fp128, fp128* @gld0, align 16
+  %1 = load fp128, fp128* @gld1, align 16
   %add = fadd fp128 %0, %1
   ret fp128 %add
 }
@@ -29,8 +29,8 @@ entry:
 
 define fp128 @subLD() {
 entry:
-  %0 = load fp128* @gld0, align 16
-  %1 = load fp128* @gld1, align 16
+  %0 = load fp128, fp128* @gld0, align 16
+  %1 = load fp128, fp128* @gld1, align 16
   %sub = fsub fp128 %0, %1
   ret fp128 %sub
 }
@@ -40,8 +40,8 @@ entry:
 
 define fp128 @mulLD() {
 entry:
-  %0 = load fp128* @gld0, align 16
-  %1 = load fp128* @gld1, align 16
+  %0 = load fp128, fp128* @gld0, align 16
+  %1 = load fp128, fp128* @gld1, align 16
   %mul = fmul fp128 %0, %1
   ret fp128 %mul
 }
@@ -51,8 +51,8 @@ entry:
 
 define fp128 @divLD() {
 entry:
-  %0 = load fp128* @gld0, align 16
-  %1 = load fp128* @gld1, align 16
+  %0 = load fp128, fp128* @gld0, align 16
+  %1 = load fp128, fp128* @gld1, align 16
   %div = fdiv fp128 %0, %1
   ret fp128 %div
 }
@@ -247,7 +247,7 @@ entry:
 
 define fp128 @libcall1_fabsl() {
 entry:
-  %0 = load fp128* @gld0, align 16
+  %0 = load fp128, fp128* @gld0, align 16
   %call = tail call fp128 @fabsl(fp128 %0) nounwind readnone
   ret fp128 %call
 }
@@ -259,7 +259,7 @@ declare fp128 @fabsl(fp128) #1
 
 define fp128 @libcall1_ceill() {
 entry:
-  %0 = load fp128* @gld0, align 16
+  %0 = load fp128, fp128* @gld0, align 16
   %call = tail call fp128 @ceill(fp128 %0) nounwind readnone
   ret fp128 %call
 }
@@ -271,7 +271,7 @@ declare fp128 @ceill(fp128) #1
 
 define fp128 @libcall1_sinl() {
 entry:
-  %0 = load fp128* @gld0, align 16
+  %0 = load fp128, fp128* @gld0, align 16
   %call = tail call fp128 @sinl(fp128 %0) nounwind
   ret fp128 %call
 }
@@ -283,7 +283,7 @@ declare fp128 @sinl(fp128) #2
 
 define fp128 @libcall1_cosl() {
 entry:
-  %0 = load fp128* @gld0, align 16
+  %0 = load fp128, fp128* @gld0, align 16
   %call = tail call fp128 @cosl(fp128 %0) nounwind
   ret fp128 %call
 }
@@ -295,7 +295,7 @@ declare fp128 @cosl(fp128) #2
 
 define fp128 @libcall1_expl() {
 entry:
-  %0 = load fp128* @gld0, align 16
+  %0 = load fp128, fp128* @gld0, align 16
   %call = tail call fp128 @expl(fp128 %0) nounwind
   ret fp128 %call
 }
@@ -307,7 +307,7 @@ declare fp128 @expl(fp128) #2
 
 define fp128 @libcall1_exp2l() {
 entry:
-  %0 = load fp128* @gld0, align 16
+  %0 = load fp128, fp128* @gld0, align 16
   %call = tail call fp128 @exp2l(fp128 %0) nounwind
   ret fp128 %call
 }
@@ -319,7 +319,7 @@ declare fp128 @exp2l(fp128) #2
 
 define fp128 @libcall1_logl() {
 entry:
-  %0 = load fp128* @gld0, align 16
+  %0 = load fp128, fp128* @gld0, align 16
   %call = tail call fp128 @logl(fp128 %0) nounwind
   ret fp128 %call
 }
@@ -331,7 +331,7 @@ declare fp128 @logl(fp128) #2
 
 define fp128 @libcall1_log2l() {
 entry:
-  %0 = load fp128* @gld0, align 16
+  %0 = load fp128, fp128* @gld0, align 16
   %call = tail call fp128 @log2l(fp128 %0) nounwind
   ret fp128 %call
 }
@@ -343,7 +343,7 @@ declare fp128 @log2l(fp128) #2
 
 define fp128 @libcall1_log10l() {
 entry:
-  %0 = load fp128* @gld0, align 16
+  %0 = load fp128, fp128* @gld0, align 16
   %call = tail call fp128 @log10l(fp128 %0) nounwind
   ret fp128 %call
 }
@@ -355,7 +355,7 @@ declare fp128 @log10l(fp128) #2
 
 define fp128 @libcall1_nearbyintl() {
 entry:
-  %0 = load fp128* @gld0, align 16
+  %0 = load fp128, fp128* @gld0, align 16
   %call = tail call fp128 @nearbyintl(fp128 %0) nounwind readnone
   ret fp128 %call
 }
@@ -367,7 +367,7 @@ declare fp128 @nearbyintl(fp128) #1
 
 define fp128 @libcall1_floorl() {
 entry:
-  %0 = load fp128* @gld0, align 16
+  %0 = load fp128, fp128* @gld0, align 16
   %call = tail call fp128 @floorl(fp128 %0) nounwind readnone
   ret fp128 %call
 }
@@ -379,7 +379,7 @@ declare fp128 @floorl(fp128) #1
 
 define fp128 @libcall1_sqrtl() {
 entry:
-  %0 = load fp128* @gld0, align 16
+  %0 = load fp128, fp128* @gld0, align 16
   %call = tail call fp128 @sqrtl(fp128 %0) nounwind
   ret fp128 %call
 }
@@ -391,7 +391,7 @@ declare fp128 @sqrtl(fp128) #2
 
 define fp128 @libcall1_rintl() {
 entry:
-  %0 = load fp128* @gld0, align 16
+  %0 = load fp128, fp128* @gld0, align 16
   %call = tail call fp128 @rintl(fp128 %0) nounwind readnone
   ret fp128 %call
 }
@@ -424,8 +424,8 @@ declare fp128 @llvm.powi.f128(fp128, i32) #3
 
 define fp128 @libcall2_copysignl() {
 entry:
-  %0 = load fp128* @gld0, align 16
-  %1 = load fp128* @gld1, align 16
+  %0 = load fp128, fp128* @gld0, align 16
+  %1 = load fp128, fp128* @gld1, align 16
   %call = tail call fp128 @copysignl(fp128 %0, fp128 %1) nounwind readnone
   ret fp128 %call
 }
@@ -437,8 +437,8 @@ declare fp128 @copysignl(fp128, fp128) #1
 
 define fp128 @libcall2_powl() {
 entry:
-  %0 = load fp128* @gld0, align 16
-  %1 = load fp128* @gld1, align 16
+  %0 = load fp128, fp128* @gld0, align 16
+  %1 = load fp128, fp128* @gld1, align 16
   %call = tail call fp128 @powl(fp128 %0, fp128 %1) nounwind
   ret fp128 %call
 }
@@ -450,8 +450,8 @@ declare fp128 @powl(fp128, fp128) #2
 
 define fp128 @libcall2_fmodl() {
 entry:
-  %0 = load fp128* @gld0, align 16
-  %1 = load fp128* @gld1, align 16
+  %0 = load fp128, fp128* @gld0, align 16
+  %1 = load fp128, fp128* @gld1, align 16
   %call = tail call fp128 @fmodl(fp128 %0, fp128 %1) nounwind
   ret fp128 %call
 }
@@ -463,9 +463,9 @@ declare fp128 @fmodl(fp128, fp128) #2
 
 define fp128 @libcall3_fmal() {
 entry:
-  %0 = load fp128* @gld0, align 16
-  %1 = load fp128* @gld2, align 16
-  %2 = load fp128* @gld1, align 16
+  %0 = load fp128, fp128* @gld0, align 16
+  %1 = load fp128, fp128* @gld2, align 16
+  %2 = load fp128, fp128* @gld1, align 16
   %3 = tail call fp128 @llvm.fma.f128(fp128 %0, fp128 %2, fp128 %1)
   ret fp128 %3
 }
@@ -539,7 +539,7 @@ entry:
 
 define fp128 @load_LD_LD() {
 entry:
-  %0 = load fp128* @gld1, align 16
+  %0 = load fp128, fp128* @gld1, align 16
   ret fp128 %0
 }
 
@@ -551,7 +551,7 @@ entry:
 
 define fp128 @load_LD_float() {
 entry:
-  %0 = load float* @gf1, align 4
+  %0 = load float, float* @gf1, align 4
   %conv = fpext float %0 to fp128
   ret fp128 %conv
 }
@@ -564,7 +564,7 @@ entry:
 
 define fp128 @load_LD_double() {
 entry:
-  %0 = load double* @gd1, align 8
+  %0 = load double, double* @gd1, align 8
   %conv = fpext double %0 to fp128
   ret fp128 %conv
 }
@@ -579,7 +579,7 @@ entry:
 
 define void @store_LD_LD() {
 entry:
-  %0 = load fp128* @gld1, align 16
+  %0 = load fp128, fp128* @gld1, align 16
   store fp128 %0, fp128* @gld0, align 16
   ret void
 }
@@ -595,7 +595,7 @@ entry:
 
 define void @store_LD_float() {
 entry:
-  %0 = load fp128* @gld1, align 16
+  %0 = load fp128, fp128* @gld1, align 16
   %conv = fptrunc fp128 %0 to float
   store float %conv, float* @gf1, align 4
   ret void
@@ -612,7 +612,7 @@ entry:
 
 define void @store_LD_double() {
 entry:
-  %0 = load fp128* @gld1, align 16
+  %0 = load fp128, fp128* @gld1, align 16
   %conv = fptrunc fp128 %0 to double
   store double %conv, double* @gd1, align 8
   ret void
diff --git a/test/CodeGen/Mips/mips64-libcall.ll b/test/CodeGen/Mips/mips64-libcall.ll
index 290baafd18b9..8512e9fcb72e 100644
--- a/test/CodeGen/Mips/mips64-libcall.ll
+++ b/test/CodeGen/Mips/mips64-libcall.ll
@@ -1,6 +1,6 @@
 ; RUN: llc -march=mips64el -mcpu=mips64r2 -O3 < %s |\
 ; RUN: FileCheck %s -check-prefix=HARD
-; RUN: llc -march=mips64el -mcpu=mips64r2 -soft-float < %s |\
+; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=+soft-float < %s |\
 ; RUN: FileCheck %s -check-prefix=SOFT
 
 ; Check that %add is not passed in an integer register.
diff --git a/test/CodeGen/Mips/mips64-sret.ll b/test/CodeGen/Mips/mips64-sret.ll
index ed494e965b7d..0559747f62cc 100644
--- a/test/CodeGen/Mips/mips64-sret.ll
+++ b/test/CodeGen/Mips/mips64-sret.ll
@@ -1,4 +1,4 @@
-; RUN: llc -march=mips64el -mcpu=mips64r2 -mattr=n64 < %s | FileCheck %s
+; RUN: llc -march=mips64el -mcpu=mips64r2 -target-abi=n64 < %s | FileCheck %s
 
 define void @foo(i32* noalias sret %agg.result) nounwind {
 entry:
diff --git a/test/CodeGen/Mips/mips64directive.ll b/test/CodeGen/Mips/mips64directive.ll
index 3d95f519bc64..b1052f77f5a8 100644
--- a/test/CodeGen/Mips/mips64directive.ll
+++ b/test/CodeGen/Mips/mips64directive.ll
@@ -1,12 +1,12 @@
-; RUN: llc  < %s -march=mips64el -mcpu=mips4 -mattr=n64 | FileCheck %s
-; RUN: llc  < %s -march=mips64el -mcpu=mips64 -mattr=n64 | FileCheck %s
+; RUN: llc  < %s -march=mips64el -mcpu=mips4 -target-abi=n64 | FileCheck %s
+; RUN: llc  < %s -march=mips64el -mcpu=mips64 -target-abi=n64 | FileCheck %s
 
 @gl = global i64 1250999896321, align 8
 
 ; CHECK: 8byte
 define i64 @foo1() nounwind readonly {
 entry:
-  %0 = load i64* @gl, align 8
+  %0 = load i64, i64* @gl, align 8
   ret i64 %0
 }
 
diff --git a/test/CodeGen/Mips/mips64ext.ll b/test/CodeGen/Mips/mips64ext.ll
index 22ea0eb7769c..9c1243b8f18c 100644
--- a/test/CodeGen/Mips/mips64ext.ll
+++ b/test/CodeGen/Mips/mips64ext.ll
@@ -1,5 +1,5 @@
-; RUN: llc  < %s -march=mips64el -mcpu=mips4 -mattr=n64 | FileCheck %s
-; RUN: llc  < %s -march=mips64el -mcpu=mips64 -mattr=n64 | FileCheck %s
+; RUN: llc  < %s -march=mips64el -mcpu=mips4 -target-abi=n64 | FileCheck %s
+; RUN: llc  < %s -march=mips64el -mcpu=mips64 -target-abi=n64 | FileCheck %s
 
 define i64 @zext64_32(i32 %a) nounwind readnone {
 entry:
diff --git a/test/CodeGen/Mips/mips64extins.ll b/test/CodeGen/Mips/mips64extins.ll
index 14f92ca86947..211cd5f8e7fd 100644
--- a/test/CodeGen/Mips/mips64extins.ll
+++ b/test/CodeGen/Mips/mips64extins.ll
@@ -1,4 +1,4 @@
-; RUN: llc  < %s -march=mips64el -mcpu=mips64r2 -mattr=n64 | FileCheck %s 
+; RUN: llc  < %s -march=mips64el -mcpu=mips64r2 -target-abi=n64 | FileCheck %s 
 
 define i64 @dext(i64 %i) nounwind readnone {
 entry:
diff --git a/test/CodeGen/Mips/mips64fpimm0.ll b/test/CodeGen/Mips/mips64fpimm0.ll
index 19e076d1ecda..0296cb523f98 100644
--- a/test/CodeGen/Mips/mips64fpimm0.ll
+++ b/test/CodeGen/Mips/mips64fpimm0.ll
@@ -1,5 +1,5 @@
-; RUN: llc  < %s -march=mips64el -mcpu=mips4 -mattr=n64 | FileCheck %s
-; RUN: llc  < %s -march=mips64el -mcpu=mips64 -mattr=n64 | FileCheck %s
+; RUN: llc  < %s -march=mips64el -mcpu=mips4 -target-abi=n64 | FileCheck %s
+; RUN: llc  < %s -march=mips64el -mcpu=mips64 -target-abi=n64 | FileCheck %s
 
 define double @foo1() nounwind readnone {
 entry:
diff --git a/test/CodeGen/Mips/mips64fpldst.ll b/test/CodeGen/Mips/mips64fpldst.ll
index 2f42270b645d..55d5c775cbba 100644
--- a/test/CodeGen/Mips/mips64fpldst.ll
+++ b/test/CodeGen/Mips/mips64fpldst.ll
@@ -1,7 +1,7 @@
-; RUN: llc  < %s -march=mips64el -mcpu=mips4 -mattr=-n64,n64 | FileCheck %s -check-prefix=CHECK-N64
-; RUN: llc  < %s -march=mips64el -mcpu=mips4 -mattr=-n64,n32 | FileCheck %s -check-prefix=CHECK-N32
-; RUN: llc  < %s -march=mips64el -mcpu=mips64 -mattr=-n64,n64 | FileCheck %s -check-prefix=CHECK-N64
-; RUN: llc  < %s -march=mips64el -mcpu=mips64 -mattr=-n64,n32 | FileCheck %s -check-prefix=CHECK-N32
+; RUN: llc  < %s -march=mips64el -mcpu=mips4 -target-abi n64 | FileCheck %s -check-prefix=CHECK-N64
+; RUN: llc  < %s -march=mips64el -mcpu=mips4 -target-abi n32 | FileCheck %s -check-prefix=CHECK-N32
+; RUN: llc  < %s -march=mips64el -mcpu=mips64 -target-abi n64 | FileCheck %s -check-prefix=CHECK-N64
+; RUN: llc  < %s -march=mips64el -mcpu=mips64 -target-abi n32 | FileCheck %s -check-prefix=CHECK-N32
 
 @f0 = common global float 0.000000e+00, align 4
 @d0 = common global double 0.000000e+00, align 8
@@ -16,7 +16,7 @@ entry:
 ; CHECK-N32: funcfl1
 ; CHECK-N32: lw $[[R0:[0-9]+]], %got_disp(f0)
 ; CHECK-N32: lwc1 $f{{[0-9]+}}, 0($[[R0]]) 
-  %0 = load float* @f0, align 4
+  %0 = load float, float* @f0, align 4
   ret float %0
 }
 
@@ -28,7 +28,7 @@ entry:
 ; CHECK-N32: funcfl2
 ; CHECK-N32: lw $[[R0:[0-9]+]], %got_disp(d0)
 ; CHECK-N32: ldc1 $f{{[0-9]+}}, 0($[[R0]]) 
-  %0 = load double* @d0, align 8 
+  %0 = load double, double* @d0, align 8 
   ret double %0
 }
 
@@ -40,7 +40,7 @@ entry:
 ; CHECK-N32: funcfs1
 ; CHECK-N32: lw $[[R0:[0-9]+]], %got_disp(f0)
 ; CHECK-N32: swc1 $f{{[0-9]+}}, 0($[[R0]]) 
-  %0 = load float* @f1, align 4 
+  %0 = load float, float* @f1, align 4 
   store float %0, float* @f0, align 4 
   ret void
 }
@@ -53,7 +53,7 @@ entry:
 ; CHECK-N32: funcfs2
 ; CHECK-N32: lw $[[R0:[0-9]+]], %got_disp(d0)
 ; CHECK-N32: sdc1 $f{{[0-9]+}}, 0($[[R0]]) 
-  %0 = load double* @d1, align 8 
+  %0 = load double, double* @d1, align 8 
   store double %0, double* @d0, align 8 
   ret void
 }
diff --git a/test/CodeGen/Mips/mips64instrs.ll b/test/CodeGen/Mips/mips64instrs.ll
index ed617be6532e..d64cdceb6b81 100644
--- a/test/CodeGen/Mips/mips64instrs.ll
+++ b/test/CodeGen/Mips/mips64instrs.ll
@@ -123,8 +123,8 @@ entry:
 ; GPRMULDIV:     ddiv $2, $[[T0]], $[[T1]]
 ; GPRMULDIV:     teq $[[T1]], $zero, 7
 
-  %0 = load i64* @gll0, align 8
-  %1 = load i64* @gll1, align 8
+  %0 = load i64, i64* @gll0, align 8
+  %1 = load i64, i64* @gll1, align 8
   %div = sdiv i64 %0, %1
   ret i64 %div
 }
@@ -144,8 +144,8 @@ entry:
 ; GPRMULDIV:     ddivu $2, $[[T0]], $[[T1]]
 ; GPRMULDIV:     teq $[[T1]], $zero, 7
 
-  %0 = load i64* @gll0, align 8
-  %1 = load i64* @gll1, align 8
+  %0 = load i64, i64* @gll0, align 8
+  %1 = load i64, i64* @gll1, align 8
   %div = udiv i64 %0, %1
   ret i64 %div
 }
diff --git a/test/CodeGen/Mips/mips64intldst.ll b/test/CodeGen/Mips/mips64intldst.ll
index c3607baeefeb..658ab88481c4 100644
--- a/test/CodeGen/Mips/mips64intldst.ll
+++ b/test/CodeGen/Mips/mips64intldst.ll
@@ -1,7 +1,7 @@
-; RUN: llc  < %s -march=mips64el -mcpu=mips4 -mattr=-n64,n64 | FileCheck %s -check-prefix=CHECK-N64
-; RUN: llc  < %s -march=mips64el -mcpu=mips4 -mattr=-n64,n32 | FileCheck %s -check-prefix=CHECK-N32
-; RUN: llc  < %s -march=mips64el -mcpu=mips64 -mattr=-n64,n64 | FileCheck %s -check-prefix=CHECK-N64
-; RUN: llc  < %s -march=mips64el -mcpu=mips64 -mattr=-n64,n32 | FileCheck %s -check-prefix=CHECK-N32
+; RUN: llc  < %s -march=mips64el -mcpu=mips4 -target-abi n64 | FileCheck %s -check-prefix=CHECK-N64
+; RUN: llc  < %s -march=mips64el -mcpu=mips4 -target-abi n32 | FileCheck %s -check-prefix=CHECK-N32
+; RUN: llc  < %s -march=mips64el -mcpu=mips64 -target-abi n64 | FileCheck %s -check-prefix=CHECK-N64
+; RUN: llc  < %s -march=mips64el -mcpu=mips64 -target-abi n32 | FileCheck %s -check-prefix=CHECK-N32
 
 @c = common global i8 0, align 4
 @s = common global i16 0, align 4
@@ -20,7 +20,7 @@ entry:
 ; CHECK-N32: func1
 ; CHECK-N32: lw $[[R0:[0-9]+]], %got_disp(c)
 ; CHECK-N32: lb ${{[0-9]+}}, 0($[[R0]])
-  %0 = load i8* @c, align 4
+  %0 = load i8, i8* @c, align 4
   %conv = sext i8 %0 to i64
   ret i64 %conv
 }
@@ -33,7 +33,7 @@ entry:
 ; CHECK-N32: func2
 ; CHECK-N32: lw $[[R0:[0-9]+]], %got_disp(s)
 ; CHECK-N32: lh ${{[0-9]+}}, 0($[[R0]])
-  %0 = load i16* @s, align 4
+  %0 = load i16, i16* @s, align 4
   %conv = sext i16 %0 to i64
   ret i64 %conv
 }
@@ -46,7 +46,7 @@ entry:
 ; CHECK-N32: func3
 ; CHECK-N32: lw $[[R0:[0-9]+]], %got_disp(i)
 ; CHECK-N32: lw ${{[0-9]+}}, 0($[[R0]])
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %conv = sext i32 %0 to i64
   ret i64 %conv
 }
@@ -59,7 +59,7 @@ entry:
 ; CHECK-N32: func4
 ; CHECK-N32: lw $[[R0:[0-9]+]], %got_disp(l)
 ; CHECK-N32: ld ${{[0-9]+}}, 0($[[R0]])
-  %0 = load i64* @l, align 8
+  %0 = load i64, i64* @l, align 8
   ret i64 %0
 }
 
@@ -71,7 +71,7 @@ entry:
 ; CHECK-N32: ufunc1
 ; CHECK-N32: lw $[[R0:[0-9]+]], %got_disp(uc)
 ; CHECK-N32: lbu ${{[0-9]+}}, 0($[[R0]])
-  %0 = load i8* @uc, align 4
+  %0 = load i8, i8* @uc, align 4
   %conv = zext i8 %0 to i64
   ret i64 %conv
 }
@@ -84,7 +84,7 @@ entry:
 ; CHECK-N32: ufunc2
 ; CHECK-N32: lw $[[R0:[0-9]+]], %got_disp(us)
 ; CHECK-N32: lhu ${{[0-9]+}}, 0($[[R0]])
-  %0 = load i16* @us, align 4
+  %0 = load i16, i16* @us, align 4
   %conv = zext i16 %0 to i64
   ret i64 %conv
 }
@@ -97,7 +97,7 @@ entry:
 ; CHECK-N32: ufunc3
 ; CHECK-N32: lw $[[R0:[0-9]+]], %got_disp(ui)
 ; CHECK-N32: lwu ${{[0-9]+}}, 0($[[R0]])
-  %0 = load i32* @ui, align 4
+  %0 = load i32, i32* @ui, align 4
   %conv = zext i32 %0 to i64
   ret i64 %conv
 }
@@ -110,7 +110,7 @@ entry:
 ; CHECK-N32: sfunc1
 ; CHECK-N32: lw $[[R0:[0-9]+]], %got_disp(c)
 ; CHECK-N32: sb ${{[0-9]+}}, 0($[[R0]])
-  %0 = load i64* @l1, align 8
+  %0 = load i64, i64* @l1, align 8
   %conv = trunc i64 %0 to i8
   store i8 %conv, i8* @c, align 4
   ret void
@@ -124,7 +124,7 @@ entry:
 ; CHECK-N32: sfunc2
 ; CHECK-N32: lw $[[R0:[0-9]+]], %got_disp(s)
 ; CHECK-N32: sh ${{[0-9]+}}, 0($[[R0]])
-  %0 = load i64* @l1, align 8
+  %0 = load i64, i64* @l1, align 8
   %conv = trunc i64 %0 to i16
   store i16 %conv, i16* @s, align 4
   ret void
@@ -138,7 +138,7 @@ entry:
 ; CHECK-N32: sfunc3
 ; CHECK-N32: lw $[[R0:[0-9]+]], %got_disp(i)
 ; CHECK-N32: sw ${{[0-9]+}}, 0($[[R0]])
-  %0 = load i64* @l1, align 8
+  %0 = load i64, i64* @l1, align 8
   %conv = trunc i64 %0 to i32
   store i32 %conv, i32* @i, align 4
   ret void
@@ -152,7 +152,7 @@ entry:
 ; CHECK-N32: sfunc4
 ; CHECK-N32: lw $[[R0:[0-9]+]], %got_disp(l)
 ; CHECK-N32: sd ${{[0-9]+}}, 0($[[R0]])
-  %0 = load i64* @l1, align 8
+  %0 = load i64, i64* @l1, align 8
   store i64 %0, i64* @l, align 8
   ret void
 }
diff --git a/test/CodeGen/Mips/mips64shift.ll b/test/CodeGen/Mips/mips64shift.ll
index 45d1c9532276..52c6f9066392 100644
--- a/test/CodeGen/Mips/mips64shift.ll
+++ b/test/CodeGen/Mips/mips64shift.ll
@@ -65,7 +65,8 @@ entry:
 
 define i64 @f9(i64 %a0, i64 %a1) nounwind readnone {
 entry:
-; CHECK: drotrv
+; CHECK-NOT:  sll
+; CHECK:      drotrv
   %shr = lshr i64 %a0, %a1
   %sub = sub i64 64, %a1
   %shl = shl i64 %a0, %sub
@@ -75,7 +76,8 @@ entry:
 
 define i64 @f10(i64 %a0, i64 %a1) nounwind readnone {
 entry:
-; CHECK: drotrv
+; CHECK-NOT:  sll
+; CHECK:      drotrv
   %shl = shl i64 %a0, %a1
   %sub = sub i64 64, %a1
   %shr = lshr i64 %a0, %sub
diff --git a/test/CodeGen/Mips/mips64signextendsesf.ll b/test/CodeGen/Mips/mips64signextendsesf.ll
index dec83b80afea..d0ce1b86bf56 100644
--- a/test/CodeGen/Mips/mips64signextendsesf.ll
+++ b/test/CodeGen/Mips/mips64signextendsesf.ll
@@ -1,11 +1,11 @@
-; RUN: llc -march=mips64 -mcpu=mips64r2 -soft-float -O2 < %s | FileCheck %s
+; RUN: llc -march=mips64 -mcpu=mips64r2 -mattr=+soft-float -O2 < %s | FileCheck %s
 
 define void @foosf() #0 {
 entry:
   %in = alloca float, align 4
   %out = alloca float, align 4
   store volatile float 0xBFD59E1380000000, float* %in, align 4
-  %in.0.in.0. = load volatile float* %in, align 4
+  %in.0.in.0. = load volatile float, float* %in, align 4
   %rintf = tail call float @rintf(float %in.0.in.0.) #1
   store volatile float %rintf, float* %out, align 4
   ret void
@@ -20,7 +20,7 @@ declare float @rintf(float)
 
 define float @foosf1(float* nocapture readonly %a) #0 {
 entry:
-  %0 = load float* %a, align 4
+  %0 = load float, float* %a, align 4
   %call = tail call float @roundf(float %0) #2
   ret float %call
 
@@ -34,7 +34,7 @@ declare float @roundf(float) #1
 
 define float @foosf2(float* nocapture readonly %a) #0 {
 entry:
-  %0 = load float* %a, align 4
+  %0 = load float, float* %a, align 4
   %call = tail call float @truncf(float %0) #2
   ret float %call
 
@@ -48,7 +48,7 @@ declare float @truncf(float) #1
 
 define float @foosf3(float* nocapture readonly %a) #0 {
 entry:
-  %0 = load float* %a, align 4
+  %0 = load float, float* %a, align 4
   %call = tail call float @floorf(float %0) #2
   ret float %call
 
@@ -62,7 +62,7 @@ declare float @floorf(float) #1
 
 define float @foosf4(float* nocapture readonly %a) #0 {
 entry:
-  %0 = load float* %a, align 4
+  %0 = load float, float* %a, align 4
   %call = tail call float @nearbyintf(float %0) #2
   ret float %call
 
@@ -76,7 +76,7 @@ declare float @nearbyintf(float) #1
 
 define float @foosf5(float* nocapture readonly %a) #0 {
 entry:
-  %0 = load float* %a, align 4
+  %0 = load float, float* %a, align 4
   %mul = fmul float %0, undef
   ret float %mul
 
@@ -88,7 +88,7 @@ entry:
 
 define float @foosf6(float* nocapture readonly %a) #0 {
 entry:
-  %0 = load float* %a, align 4
+  %0 = load float, float* %a, align 4
   %sub = fsub float %0, undef
   ret float %sub
 
@@ -100,7 +100,7 @@ entry:
 
 define float @foosf7(float* nocapture readonly %a) #0 {
 entry:
-  %0 = load float* %a, align 4
+  %0 = load float, float* %a, align 4
   %add = fadd float %0, undef
   ret float %add
 
@@ -113,8 +113,8 @@ entry:
 define float @foosf8(float* nocapture readonly %a) #0 {
 entry:
   %b = alloca float, align 4
-  %b.0.b.0. = load volatile float* %b, align 4
-  %0 = load float* %a, align 4
+  %b.0.b.0. = load volatile float, float* %b, align 4
+  %0 = load float, float* %a, align 4
   %div = fdiv float %b.0.b.0., %0
   ret float %div
 
@@ -127,9 +127,9 @@ entry:
 define float @foosf9() #0 {
 entry:
   %b = alloca float, align 4
-  %b.0.b.0. = load volatile float* %b, align 4
+  %b.0.b.0. = load volatile float, float* %b, align 4
   %conv = fpext float %b.0.b.0. to double
-  %b.0.b.0.3 = load volatile float* %b, align 4
+  %b.0.b.0.3 = load volatile float, float* %b, align 4
   %conv1 = fpext float %b.0.b.0.3 to double
   %call = tail call double @pow(double %conv, double %conv1) #1
   %conv2 = fptrunc double %call to float
@@ -146,7 +146,7 @@ declare double @pow(double, double) #0
 define float @foosf10() #0 {
 entry:
   %a = alloca float, align 4
-  %a.0.a.0. = load volatile float* %a, align 4
+  %a.0.a.0. = load volatile float, float* %a, align 4
   %conv = fpext float %a.0.a.0. to double
   %call = tail call double @sin(double %conv) #1
   %conv1 = fptrunc double %call to float
@@ -163,7 +163,7 @@ declare double @sin(double) #0
 define float @foosf11() #0 {
 entry:
   %b = alloca float, align 4
-  %b.0.b.0. = load volatile float* %b, align 4
+  %b.0.b.0. = load volatile float, float* %b, align 4
   %call = tail call float @ceilf(float %b.0.b.0.) #2
   ret float %call
 
@@ -179,8 +179,8 @@ define float @foosf12() #0 {
 entry:
   %b = alloca float, align 4
   %a = alloca float, align 4
-  %b.0.b.0. = load volatile float* %b, align 4
-  %a.0.a.0. = load volatile float* %a, align 4
+  %b.0.b.0. = load volatile float, float* %b, align 4
+  %a.0.a.0. = load volatile float, float* %a, align 4
   %call = tail call float @fmaxf(float %b.0.b.0., float %a.0.a.0.) #2
   ret float %call
 
@@ -196,8 +196,8 @@ define float @foosf13() #0 {
 entry:
   %b = alloca float, align 4
   %a = alloca float, align 4
-  %b.0.b.0. = load volatile float* %b, align 4
-  %a.0.a.0. = load volatile float* %a, align 4
+  %b.0.b.0. = load volatile float, float* %b, align 4
+  %a.0.a.0. = load volatile float, float* %a, align 4
   %call = tail call float @fminf(float %b.0.b.0., float %a.0.a.0.) #2
   ret float %call
 
@@ -211,4 +211,4 @@ declare float @fminf(float, float) #1
 
 
 attributes #0 = { nounwind "use-soft-float"="true" }
-attributes #1 = { nounwind readnone "use-soft-float"="true" }
+attributes #1 = { nounwind readnone "use-soft-float"="true" }
\ No newline at end of file
diff --git a/test/CodeGen/Mips/mips64sinttofpsf.ll b/test/CodeGen/Mips/mips64sinttofpsf.ll
index d3d46036f7da..87f803ecf552 100644
--- a/test/CodeGen/Mips/mips64sinttofpsf.ll
+++ b/test/CodeGen/Mips/mips64sinttofpsf.ll
@@ -1,11 +1,11 @@
-; RUN: llc -march=mips64 -mcpu=mips64r2 -soft-float -O0 < %s | FileCheck %s
+; RUN: llc -march=mips64 -mcpu=mips64r2 -mattr=+soft-float -O0 < %s | FileCheck %s
 
 
 define double @foo() #0 {
 entry:
   %x = alloca i32, align 4
   store volatile i32 -32, i32* %x, align 4
-  %0 = load volatile i32* %x, align 4
+  %0 = load volatile i32, i32* %x, align 4
   %conv = sitofp i32 %0 to double
   ret double %conv
 
diff --git a/test/CodeGen/Mips/mipslopat.ll b/test/CodeGen/Mips/mipslopat.ll
index 1f433b9870ce..63b68c1762b2 100644
--- a/test/CodeGen/Mips/mipslopat.ll
+++ b/test/CodeGen/Mips/mipslopat.ll
@@ -6,10 +6,10 @@
 
 define void @simple_vol_file() nounwind {
 entry:
-  %tmp = load volatile i32** @stat_vol_ptr_int, align 4
+  %tmp = load volatile i32*, i32** @stat_vol_ptr_int, align 4
   %0 = bitcast i32* %tmp to i8*
   call void @llvm.prefetch(i8* %0, i32 0, i32 0, i32 1)
-  %tmp1 = load i32** @stat_ptr_vol_int, align 4
+  %tmp1 = load i32*, i32** @stat_ptr_vol_int, align 4
   %1 = bitcast i32* %tmp1 to i8*
   call void @llvm.prefetch(i8* %1, i32 0, i32 0, i32 1)
   ret void
diff --git a/test/CodeGen/Mips/misha.ll b/test/CodeGen/Mips/misha.ll
index 65d3b7b5d874..23ad7f6057af 100644
--- a/test/CodeGen/Mips/misha.ll
+++ b/test/CodeGen/Mips/misha.ll
@@ -8,15 +8,15 @@ entry:
   br i1 %cmp8, label %for.end, label %for.body.lr.ph
 
 for.body.lr.ph:                                   ; preds = %entry
-  %.pre = load i8* %to, align 1
+  %.pre = load i8, i8* %to, align 1
   br label %for.body
 
 for.body:                                         ; preds = %for.body.lr.ph, %for.body
   %1 = phi i8 [ %.pre, %for.body.lr.ph ], [ %conv4, %for.body ]
   %i.010 = phi i32 [ 0, %for.body.lr.ph ], [ %inc, %for.body ]
   %from.addr.09 = phi i8* [ %from, %for.body.lr.ph ], [ %incdec.ptr, %for.body ]
-  %incdec.ptr = getelementptr inbounds i8* %from.addr.09, i32 1
-  %2 = load i8* %from.addr.09, align 1
+  %incdec.ptr = getelementptr inbounds i8, i8* %from.addr.09, i32 1
+  %2 = load i8, i8* %from.addr.09, align 1
   %conv27 = zext i8 %2 to i32
   %conv36 = zext i8 %1 to i32
   %add = add nsw i32 %conv36, %conv27
@@ -44,15 +44,15 @@ entry:
   br i1 %cmp8, label %for.end, label %for.body.lr.ph
 
 for.body.lr.ph:                                   ; preds = %entry
-  %.pre = load i16* %to, align 2
+  %.pre = load i16, i16* %to, align 2
   br label %for.body
 
 for.body:                                         ; preds = %for.body.lr.ph, %for.body
   %1 = phi i16 [ %.pre, %for.body.lr.ph ], [ %conv4, %for.body ]
   %i.010 = phi i32 [ 0, %for.body.lr.ph ], [ %inc, %for.body ]
   %from.addr.09 = phi i16* [ %from, %for.body.lr.ph ], [ %incdec.ptr, %for.body ]
-  %incdec.ptr = getelementptr inbounds i16* %from.addr.09, i32 1
-  %2 = load i16* %from.addr.09, align 2
+  %incdec.ptr = getelementptr inbounds i16, i16* %from.addr.09, i32 1
+  %2 = load i16, i16* %from.addr.09, align 2
   %conv27 = zext i16 %2 to i32
   %conv36 = zext i16 %1 to i32
   %add = add nsw i32 %conv36, %conv27
diff --git a/test/CodeGen/Mips/mno-ldc1-sdc1.ll b/test/CodeGen/Mips/mno-ldc1-sdc1.ll
index db653eadf2f7..c7eda3320bc6 100644
--- a/test/CodeGen/Mips/mno-ldc1-sdc1.ll
+++ b/test/CodeGen/Mips/mno-ldc1-sdc1.ll
@@ -111,7 +111,7 @@
 
 define double @test_ldc1() {
 entry:
-  %0 = load double* @g0, align 8
+  %0 = load double, double* @g0, align 8
   ret double %0
 }
 
@@ -212,8 +212,8 @@ entry:
 
 define double @test_ldxc1(double* nocapture readonly %a, i32 %i) {
 entry:
-  %arrayidx = getelementptr inbounds double* %a, i32 %i
-  %0 = load double* %arrayidx, align 8
+  %arrayidx = getelementptr inbounds double, double* %a, i32 %i
+  %0 = load double, double* %arrayidx, align 8
   ret double %0
 }
 
@@ -243,7 +243,7 @@ entry:
 
 define void @test_sdxc1(double %b, double* nocapture %a, i32 %i) {
 entry:
-  %arrayidx = getelementptr inbounds double* %a, i32 %i
+  %arrayidx = getelementptr inbounds double, double* %a, i32 %i
   store double %b, double* %arrayidx, align 8
   ret void
 }
diff --git a/test/CodeGen/Mips/msa/2r.ll b/test/CodeGen/Mips/msa/2r.ll
index da35ad82cad1..501936c76e73 100644
--- a/test/CodeGen/Mips/msa/2r.ll
+++ b/test/CodeGen/Mips/msa/2r.ll
@@ -8,7 +8,7 @@
 
 define void @llvm_mips_nloc_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_nloc_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_nloc_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.nloc.b(<16 x i8> %0)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_nloc_b_RES
   ret void
@@ -29,7 +29,7 @@ declare <16 x i8> @llvm.mips.nloc.b(<16 x i8>) nounwind
 
 define void @llvm_mips_nloc_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_nloc_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_nloc_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.nloc.h(<8 x i16> %0)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_nloc_h_RES
   ret void
@@ -50,7 +50,7 @@ declare <8 x i16> @llvm.mips.nloc.h(<8 x i16>) nounwind
 
 define void @llvm_mips_nloc_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_nloc_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_nloc_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.nloc.w(<4 x i32> %0)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_nloc_w_RES
   ret void
@@ -71,7 +71,7 @@ declare <4 x i32> @llvm.mips.nloc.w(<4 x i32>) nounwind
 
 define void @llvm_mips_nloc_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_nloc_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_nloc_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.nloc.d(<2 x i64> %0)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_nloc_d_RES
   ret void
@@ -92,7 +92,7 @@ declare <2 x i64> @llvm.mips.nloc.d(<2 x i64>) nounwind
 
 define void @llvm_mips_nlzc_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_nlzc_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_nlzc_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.nlzc.b(<16 x i8> %0)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_nlzc_b_RES
   ret void
@@ -113,7 +113,7 @@ declare <16 x i8> @llvm.mips.nlzc.b(<16 x i8>) nounwind
 
 define void @llvm_mips_nlzc_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_nlzc_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_nlzc_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.nlzc.h(<8 x i16> %0)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_nlzc_h_RES
   ret void
@@ -134,7 +134,7 @@ declare <8 x i16> @llvm.mips.nlzc.h(<8 x i16>) nounwind
 
 define void @llvm_mips_nlzc_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_nlzc_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_nlzc_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.nlzc.w(<4 x i32> %0)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_nlzc_w_RES
   ret void
@@ -155,7 +155,7 @@ declare <4 x i32> @llvm.mips.nlzc.w(<4 x i32>) nounwind
 
 define void @llvm_mips_nlzc_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_nlzc_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_nlzc_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.nlzc.d(<2 x i64> %0)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_nlzc_d_RES
   ret void
@@ -176,7 +176,7 @@ declare <2 x i64> @llvm.mips.nlzc.d(<2 x i64>) nounwind
 
 define void @llvm_mips_pcnt_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_pcnt_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_pcnt_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.pcnt.b(<16 x i8> %0)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_pcnt_b_RES
   ret void
@@ -197,7 +197,7 @@ declare <16 x i8> @llvm.mips.pcnt.b(<16 x i8>) nounwind
 
 define void @llvm_mips_pcnt_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_pcnt_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_pcnt_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.pcnt.h(<8 x i16> %0)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_pcnt_h_RES
   ret void
@@ -218,7 +218,7 @@ declare <8 x i16> @llvm.mips.pcnt.h(<8 x i16>) nounwind
 
 define void @llvm_mips_pcnt_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_pcnt_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_pcnt_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.pcnt.w(<4 x i32> %0)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_pcnt_w_RES
   ret void
@@ -239,7 +239,7 @@ declare <4 x i32> @llvm.mips.pcnt.w(<4 x i32>) nounwind
 
 define void @llvm_mips_pcnt_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_pcnt_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_pcnt_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.pcnt.d(<2 x i64> %0)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_pcnt_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/2r_vector_scalar.ll b/test/CodeGen/Mips/msa/2r_vector_scalar.ll
index 64e459e4d9a9..ddcd3cf757d9 100644
--- a/test/CodeGen/Mips/msa/2r_vector_scalar.ll
+++ b/test/CodeGen/Mips/msa/2r_vector_scalar.ll
@@ -15,7 +15,7 @@
 
 define void @llvm_mips_fill_b_test() nounwind {
 entry:
-  %0 = load i32* @llvm_mips_fill_b_ARG1
+  %0 = load i32, i32* @llvm_mips_fill_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.fill.b(i32 %0)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_fill_b_RES
   ret void
@@ -35,7 +35,7 @@ declare <16 x i8> @llvm.mips.fill.b(i32) nounwind
 
 define void @llvm_mips_fill_h_test() nounwind {
 entry:
-  %0 = load i32* @llvm_mips_fill_h_ARG1
+  %0 = load i32, i32* @llvm_mips_fill_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.fill.h(i32 %0)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_fill_h_RES
   ret void
@@ -55,7 +55,7 @@ declare <8 x i16> @llvm.mips.fill.h(i32) nounwind
 
 define void @llvm_mips_fill_w_test() nounwind {
 entry:
-  %0 = load i32* @llvm_mips_fill_w_ARG1
+  %0 = load i32, i32* @llvm_mips_fill_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.fill.w(i32 %0)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_fill_w_RES
   ret void
@@ -75,7 +75,7 @@ declare <4 x i32> @llvm.mips.fill.w(i32) nounwind
 
 define void @llvm_mips_fill_d_test() nounwind {
 entry:
-  %0 = load i64* @llvm_mips_fill_d_ARG1
+  %0 = load i64, i64* @llvm_mips_fill_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.fill.d(i64 %0)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_fill_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/2rf.ll b/test/CodeGen/Mips/msa/2rf.ll
index b361ef5eae21..1dbfbda1b612 100644
--- a/test/CodeGen/Mips/msa/2rf.ll
+++ b/test/CodeGen/Mips/msa/2rf.ll
@@ -8,7 +8,7 @@
 
 define void @llvm_mips_flog2_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_flog2_w_ARG1
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_flog2_w_ARG1
   %1 = tail call <4 x float> @llvm.mips.flog2.w(<4 x float> %0)
   store <4 x float> %1, <4 x float>* @llvm_mips_flog2_w_RES
   ret void
@@ -29,7 +29,7 @@ declare <4 x float> @llvm.mips.flog2.w(<4 x float>) nounwind
 
 define void @llvm_mips_flog2_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_flog2_d_ARG1
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_flog2_d_ARG1
   %1 = tail call <2 x double> @llvm.mips.flog2.d(<2 x double> %0)
   store <2 x double> %1, <2 x double>* @llvm_mips_flog2_d_RES
   ret void
@@ -47,7 +47,7 @@ declare <2 x double> @llvm.mips.flog2.d(<2 x double>) nounwind
 
 define void @flog2_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_flog2_w_ARG1
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_flog2_w_ARG1
   %1 = tail call <4 x float> @llvm.log2.v4f32(<4 x float> %0)
   store <4 x float> %1, <4 x float>* @llvm_mips_flog2_w_RES
   ret void
@@ -65,7 +65,7 @@ declare <4 x float> @llvm.log2.v4f32(<4 x float> %val)
 
 define void @flog2_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_flog2_d_ARG1
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_flog2_d_ARG1
   %1 = tail call <2 x double> @llvm.log2.v2f64(<2 x double> %0)
   store <2 x double> %1, <2 x double>* @llvm_mips_flog2_d_RES
   ret void
@@ -86,7 +86,7 @@ declare <2 x double> @llvm.log2.v2f64(<2 x double> %val)
 
 define void @llvm_mips_frint_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_frint_w_ARG1
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_frint_w_ARG1
   %1 = tail call <4 x float> @llvm.mips.frint.w(<4 x float> %0)
   store <4 x float> %1, <4 x float>* @llvm_mips_frint_w_RES
   ret void
@@ -107,7 +107,7 @@ declare <4 x float> @llvm.mips.frint.w(<4 x float>) nounwind
 
 define void @llvm_mips_frint_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_frint_d_ARG1
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_frint_d_ARG1
   %1 = tail call <2 x double> @llvm.mips.frint.d(<2 x double> %0)
   store <2 x double> %1, <2 x double>* @llvm_mips_frint_d_RES
   ret void
@@ -125,7 +125,7 @@ declare <2 x double> @llvm.mips.frint.d(<2 x double>) nounwind
 
 define void @frint_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_frint_w_ARG1
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_frint_w_ARG1
   %1 = tail call <4 x float> @llvm.rint.v4f32(<4 x float> %0)
   store <4 x float> %1, <4 x float>* @llvm_mips_frint_w_RES
   ret void
@@ -143,7 +143,7 @@ declare <4 x float> @llvm.rint.v4f32(<4 x float>) nounwind
 
 define void @frint_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_frint_d_ARG1
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_frint_d_ARG1
   %1 = tail call <2 x double> @llvm.rint.v2f64(<2 x double> %0)
   store <2 x double> %1, <2 x double>* @llvm_mips_frint_d_RES
   ret void
@@ -164,7 +164,7 @@ declare <2 x double> @llvm.rint.v2f64(<2 x double>) nounwind
 
 define void @llvm_mips_frcp_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_frcp_w_ARG1
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_frcp_w_ARG1
   %1 = tail call <4 x float> @llvm.mips.frcp.w(<4 x float> %0)
   store <4 x float> %1, <4 x float>* @llvm_mips_frcp_w_RES
   ret void
@@ -185,7 +185,7 @@ declare <4 x float> @llvm.mips.frcp.w(<4 x float>) nounwind
 
 define void @llvm_mips_frcp_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_frcp_d_ARG1
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_frcp_d_ARG1
   %1 = tail call <2 x double> @llvm.mips.frcp.d(<2 x double> %0)
   store <2 x double> %1, <2 x double>* @llvm_mips_frcp_d_RES
   ret void
@@ -206,7 +206,7 @@ declare <2 x double> @llvm.mips.frcp.d(<2 x double>) nounwind
 
 define void @llvm_mips_frsqrt_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_frsqrt_w_ARG1
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_frsqrt_w_ARG1
   %1 = tail call <4 x float> @llvm.mips.frsqrt.w(<4 x float> %0)
   store <4 x float> %1, <4 x float>* @llvm_mips_frsqrt_w_RES
   ret void
@@ -227,7 +227,7 @@ declare <4 x float> @llvm.mips.frsqrt.w(<4 x float>) nounwind
 
 define void @llvm_mips_frsqrt_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_frsqrt_d_ARG1
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_frsqrt_d_ARG1
   %1 = tail call <2 x double> @llvm.mips.frsqrt.d(<2 x double> %0)
   store <2 x double> %1, <2 x double>* @llvm_mips_frsqrt_d_RES
   ret void
@@ -248,7 +248,7 @@ declare <2 x double> @llvm.mips.frsqrt.d(<2 x double>) nounwind
 
 define void @llvm_mips_fsqrt_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fsqrt_w_ARG1
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fsqrt_w_ARG1
   %1 = tail call <4 x float> @llvm.mips.fsqrt.w(<4 x float> %0)
   store <4 x float> %1, <4 x float>* @llvm_mips_fsqrt_w_RES
   ret void
@@ -269,7 +269,7 @@ declare <4 x float> @llvm.mips.fsqrt.w(<4 x float>) nounwind
 
 define void @llvm_mips_fsqrt_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fsqrt_d_ARG1
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fsqrt_d_ARG1
   %1 = tail call <2 x double> @llvm.mips.fsqrt.d(<2 x double> %0)
   store <2 x double> %1, <2 x double>* @llvm_mips_fsqrt_d_RES
   ret void
@@ -287,7 +287,7 @@ declare <2 x double> @llvm.mips.fsqrt.d(<2 x double>) nounwind
 
 define void @fsqrt_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fsqrt_w_ARG1
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fsqrt_w_ARG1
   %1 = tail call <4 x float> @llvm.sqrt.v4f32(<4 x float> %0)
   store <4 x float> %1, <4 x float>* @llvm_mips_fsqrt_w_RES
   ret void
@@ -305,7 +305,7 @@ declare <4 x float> @llvm.sqrt.v4f32(<4 x float>) nounwind
 
 define void @fsqrt_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fsqrt_d_ARG1
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fsqrt_d_ARG1
   %1 = tail call <2 x double> @llvm.sqrt.v2f64(<2 x double> %0)
   store <2 x double> %1, <2 x double>* @llvm_mips_fsqrt_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/2rf_exup.ll b/test/CodeGen/Mips/msa/2rf_exup.ll
index 8d7cc367040a..fd81ff6d112b 100644
--- a/test/CodeGen/Mips/msa/2rf_exup.ll
+++ b/test/CodeGen/Mips/msa/2rf_exup.ll
@@ -9,7 +9,7 @@
 
 define void @llvm_mips_fexupl_w_test() nounwind {
 entry:
-  %0 = load <8 x half>* @llvm_mips_fexupl_w_ARG1
+  %0 = load <8 x half>, <8 x half>* @llvm_mips_fexupl_w_ARG1
   %1 = tail call <4 x float> @llvm.mips.fexupl.w(<8 x half> %0)
   store <4 x float> %1, <4 x float>* @llvm_mips_fexupl_w_RES
   ret void
@@ -28,7 +28,7 @@ declare <4 x float> @llvm.mips.fexupl.w(<8 x half>) nounwind
 
 define void @llvm_mips_fexupl_d_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fexupl_d_ARG1
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fexupl_d_ARG1
   %1 = tail call <2 x double> @llvm.mips.fexupl.d(<4 x float> %0)
   store <2 x double> %1, <2 x double>* @llvm_mips_fexupl_d_RES
   ret void
@@ -47,7 +47,7 @@ declare <2 x double> @llvm.mips.fexupl.d(<4 x float>) nounwind
 
 define void @llvm_mips_fexupr_w_test() nounwind {
 entry:
-  %0 = load <8 x half>* @llvm_mips_fexupr_w_ARG1
+  %0 = load <8 x half>, <8 x half>* @llvm_mips_fexupr_w_ARG1
   %1 = tail call <4 x float> @llvm.mips.fexupr.w(<8 x half> %0)
   store <4 x float> %1, <4 x float>* @llvm_mips_fexupr_w_RES
   ret void
@@ -66,7 +66,7 @@ declare <4 x float> @llvm.mips.fexupr.w(<8 x half>) nounwind
 
 define void @llvm_mips_fexupr_d_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fexupr_d_ARG1
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fexupr_d_ARG1
   %1 = tail call <2 x double> @llvm.mips.fexupr.d(<4 x float> %0)
   store <2 x double> %1, <2 x double>* @llvm_mips_fexupr_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/2rf_float_int.ll b/test/CodeGen/Mips/msa/2rf_float_int.ll
index 3b5dfda2d1e6..369015814b0e 100644
--- a/test/CodeGen/Mips/msa/2rf_float_int.ll
+++ b/test/CodeGen/Mips/msa/2rf_float_int.ll
@@ -9,7 +9,7 @@
 
 define void @llvm_mips_ffint_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_ffint_s_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ffint_s_w_ARG1
   %1 = tail call <4 x float> @llvm.mips.ffint.s.w(<4 x i32> %0)
   store <4 x float> %1, <4 x float>* @llvm_mips_ffint_s_w_RES
   ret void
@@ -30,7 +30,7 @@ declare <4 x float> @llvm.mips.ffint.s.w(<4 x i32>) nounwind
 
 define void @llvm_mips_ffint_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_ffint_s_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ffint_s_d_ARG1
   %1 = tail call <2 x double> @llvm.mips.ffint.s.d(<2 x i64> %0)
   store <2 x double> %1, <2 x double>* @llvm_mips_ffint_s_d_RES
   ret void
@@ -51,7 +51,7 @@ declare <2 x double> @llvm.mips.ffint.s.d(<2 x i64>) nounwind
 
 define void @llvm_mips_ffint_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_ffint_u_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ffint_u_w_ARG1
   %1 = tail call <4 x float> @llvm.mips.ffint.u.w(<4 x i32> %0)
   store <4 x float> %1, <4 x float>* @llvm_mips_ffint_u_w_RES
   ret void
@@ -72,7 +72,7 @@ declare <4 x float> @llvm.mips.ffint.u.w(<4 x i32>) nounwind
 
 define void @llvm_mips_ffint_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_ffint_u_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ffint_u_d_ARG1
   %1 = tail call <2 x double> @llvm.mips.ffint.u.d(<2 x i64> %0)
   store <2 x double> %1, <2 x double>* @llvm_mips_ffint_u_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/2rf_fq.ll b/test/CodeGen/Mips/msa/2rf_fq.ll
index 021dd937fad3..05c649ee918a 100644
--- a/test/CodeGen/Mips/msa/2rf_fq.ll
+++ b/test/CodeGen/Mips/msa/2rf_fq.ll
@@ -9,7 +9,7 @@
 
 define void @llvm_mips_ffql_w_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_ffql_w_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ffql_w_ARG1
   %1 = tail call <4 x float> @llvm.mips.ffql.w(<8 x i16> %0)
   store <4 x float> %1, <4 x float>* @llvm_mips_ffql_w_RES
   ret void
@@ -28,7 +28,7 @@ declare <4 x float> @llvm.mips.ffql.w(<8 x i16>) nounwind
 
 define void @llvm_mips_ffql_d_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_ffql_d_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ffql_d_ARG1
   %1 = tail call <2 x double> @llvm.mips.ffql.d(<4 x i32> %0)
   store <2 x double> %1, <2 x double>* @llvm_mips_ffql_d_RES
   ret void
@@ -47,7 +47,7 @@ declare <2 x double> @llvm.mips.ffql.d(<4 x i32>) nounwind
 
 define void @llvm_mips_ffqr_w_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_ffqr_w_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ffqr_w_ARG1
   %1 = tail call <4 x float> @llvm.mips.ffqr.w(<8 x i16> %0)
   store <4 x float> %1, <4 x float>* @llvm_mips_ffqr_w_RES
   ret void
@@ -66,7 +66,7 @@ declare <4 x float> @llvm.mips.ffqr.w(<8 x i16>) nounwind
 
 define void @llvm_mips_ffqr_d_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_ffqr_d_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ffqr_d_ARG1
   %1 = tail call <2 x double> @llvm.mips.ffqr.d(<4 x i32> %0)
   store <2 x double> %1, <2 x double>* @llvm_mips_ffqr_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/2rf_int_float.ll b/test/CodeGen/Mips/msa/2rf_int_float.ll
index 4665ae066a4f..77d1404f9cfa 100644
--- a/test/CodeGen/Mips/msa/2rf_int_float.ll
+++ b/test/CodeGen/Mips/msa/2rf_int_float.ll
@@ -10,7 +10,7 @@
 
 define void @llvm_mips_fclass_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fclass_w_ARG1
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fclass_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.fclass.w(<4 x float> %0)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_fclass_w_RES
   ret void
@@ -31,7 +31,7 @@ declare <4 x i32> @llvm.mips.fclass.w(<4 x float>) nounwind
 
 define void @llvm_mips_fclass_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fclass_d_ARG1
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fclass_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.fclass.d(<2 x double> %0)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_fclass_d_RES
   ret void
@@ -52,7 +52,7 @@ declare <2 x i64> @llvm.mips.fclass.d(<2 x double>) nounwind
 
 define void @llvm_mips_ftrunc_s_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_ftrunc_s_w_ARG1
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_ftrunc_s_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.ftrunc.s.w(<4 x float> %0)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_ftrunc_s_w_RES
   ret void
@@ -73,7 +73,7 @@ declare <4 x i32> @llvm.mips.ftrunc.s.w(<4 x float>) nounwind
 
 define void @llvm_mips_ftrunc_s_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_ftrunc_s_d_ARG1
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_ftrunc_s_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.ftrunc.s.d(<2 x double> %0)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_ftrunc_s_d_RES
   ret void
@@ -94,7 +94,7 @@ declare <2 x i64> @llvm.mips.ftrunc.s.d(<2 x double>) nounwind
 
 define void @llvm_mips_ftrunc_u_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_ftrunc_u_w_ARG1
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_ftrunc_u_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.ftrunc.u.w(<4 x float> %0)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_ftrunc_u_w_RES
   ret void
@@ -115,7 +115,7 @@ declare <4 x i32> @llvm.mips.ftrunc.u.w(<4 x float>) nounwind
 
 define void @llvm_mips_ftrunc_u_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_ftrunc_u_d_ARG1
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_ftrunc_u_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.ftrunc.u.d(<2 x double> %0)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_ftrunc_u_d_RES
   ret void
@@ -136,7 +136,7 @@ declare <2 x i64> @llvm.mips.ftrunc.u.d(<2 x double>) nounwind
 
 define void @llvm_mips_ftint_s_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_ftint_s_w_ARG1
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_ftint_s_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.ftint.s.w(<4 x float> %0)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_ftint_s_w_RES
   ret void
@@ -157,7 +157,7 @@ declare <4 x i32> @llvm.mips.ftint.s.w(<4 x float>) nounwind
 
 define void @llvm_mips_ftint_s_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_ftint_s_d_ARG1
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_ftint_s_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.ftint.s.d(<2 x double> %0)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_ftint_s_d_RES
   ret void
@@ -178,7 +178,7 @@ declare <2 x i64> @llvm.mips.ftint.s.d(<2 x double>) nounwind
 
 define void @llvm_mips_ftint_u_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_ftint_u_w_ARG1
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_ftint_u_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.ftint.u.w(<4 x float> %0)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_ftint_u_w_RES
   ret void
@@ -199,7 +199,7 @@ declare <4 x i32> @llvm.mips.ftint.u.w(<4 x float>) nounwind
 
 define void @llvm_mips_ftint_u_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_ftint_u_d_ARG1
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_ftint_u_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.ftint.u.d(<2 x double> %0)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_ftint_u_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/2rf_tq.ll b/test/CodeGen/Mips/msa/2rf_tq.ll
index 6f3c508f5b8c..9b7f02a5ef33 100644
--- a/test/CodeGen/Mips/msa/2rf_tq.ll
+++ b/test/CodeGen/Mips/msa/2rf_tq.ll
@@ -10,8 +10,8 @@
 
 define void @llvm_mips_ftq_h_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_ftq_h_ARG1
-  %1 = load <4 x float>* @llvm_mips_ftq_h_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_ftq_h_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_ftq_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.ftq.h(<4 x float> %0, <4 x float> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_ftq_h_RES
   ret void
@@ -32,8 +32,8 @@ declare <8 x i16> @llvm.mips.ftq.h(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_ftq_w_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_ftq_w_ARG1
-  %1 = load <2 x double>* @llvm_mips_ftq_w_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_ftq_w_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_ftq_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.ftq.w(<2 x double> %0, <2 x double> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_ftq_w_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3r-a.ll b/test/CodeGen/Mips/msa/3r-a.ll
index dab15b66b7ce..db772f918614 100644
--- a/test/CodeGen/Mips/msa/3r-a.ll
+++ b/test/CodeGen/Mips/msa/3r-a.ll
@@ -15,8 +15,8 @@
 
 define void @llvm_mips_add_a_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_add_a_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_add_a_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_add_a_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_add_a_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.add.a.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_add_a_b_RES
   ret void
@@ -40,8 +40,8 @@ declare <16 x i8> @llvm.mips.add.a.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_add_a_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_add_a_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_add_a_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_add_a_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_add_a_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.add.a.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_add_a_h_RES
   ret void
@@ -65,8 +65,8 @@ declare <8 x i16> @llvm.mips.add.a.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_add_a_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_add_a_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_add_a_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_add_a_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_add_a_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.add.a.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_add_a_w_RES
   ret void
@@ -90,8 +90,8 @@ declare <4 x i32> @llvm.mips.add.a.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_add_a_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_add_a_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_add_a_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_add_a_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_add_a_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.add.a.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_add_a_d_RES
   ret void
@@ -115,8 +115,8 @@ declare <2 x i64> @llvm.mips.add.a.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_adds_a_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_adds_a_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_adds_a_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_adds_a_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_adds_a_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.adds.a.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_adds_a_b_RES
   ret void
@@ -140,8 +140,8 @@ declare <16 x i8> @llvm.mips.adds.a.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_adds_a_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_adds_a_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_adds_a_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_adds_a_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_adds_a_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.adds.a.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_adds_a_h_RES
   ret void
@@ -165,8 +165,8 @@ declare <8 x i16> @llvm.mips.adds.a.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_adds_a_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_adds_a_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_adds_a_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_adds_a_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_adds_a_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.adds.a.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_adds_a_w_RES
   ret void
@@ -190,8 +190,8 @@ declare <4 x i32> @llvm.mips.adds.a.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_adds_a_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_adds_a_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_adds_a_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_adds_a_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_adds_a_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.adds.a.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_adds_a_d_RES
   ret void
@@ -215,8 +215,8 @@ declare <2 x i64> @llvm.mips.adds.a.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_adds_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_adds_s_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_adds_s_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_adds_s_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_adds_s_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.adds.s.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_adds_s_b_RES
   ret void
@@ -240,8 +240,8 @@ declare <16 x i8> @llvm.mips.adds.s.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_adds_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_adds_s_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_adds_s_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_adds_s_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_adds_s_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.adds.s.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_adds_s_h_RES
   ret void
@@ -265,8 +265,8 @@ declare <8 x i16> @llvm.mips.adds.s.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_adds_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_adds_s_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_adds_s_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_adds_s_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_adds_s_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.adds.s.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_adds_s_w_RES
   ret void
@@ -290,8 +290,8 @@ declare <4 x i32> @llvm.mips.adds.s.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_adds_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_adds_s_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_adds_s_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_adds_s_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_adds_s_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.adds.s.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_adds_s_d_RES
   ret void
@@ -315,8 +315,8 @@ declare <2 x i64> @llvm.mips.adds.s.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_adds_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_adds_u_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_adds_u_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_adds_u_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_adds_u_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.adds.u.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_adds_u_b_RES
   ret void
@@ -340,8 +340,8 @@ declare <16 x i8> @llvm.mips.adds.u.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_adds_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_adds_u_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_adds_u_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_adds_u_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_adds_u_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.adds.u.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_adds_u_h_RES
   ret void
@@ -365,8 +365,8 @@ declare <8 x i16> @llvm.mips.adds.u.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_adds_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_adds_u_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_adds_u_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_adds_u_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_adds_u_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.adds.u.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_adds_u_w_RES
   ret void
@@ -390,8 +390,8 @@ declare <4 x i32> @llvm.mips.adds.u.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_adds_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_adds_u_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_adds_u_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_adds_u_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_adds_u_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.adds.u.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_adds_u_d_RES
   ret void
@@ -415,8 +415,8 @@ declare <2 x i64> @llvm.mips.adds.u.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_addv_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_addv_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_addv_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_addv_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_addv_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_addv_b_RES
   ret void
@@ -440,8 +440,8 @@ declare <16 x i8> @llvm.mips.addv.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_addv_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_addv_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_addv_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_addv_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_addv_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_addv_h_RES
   ret void
@@ -465,8 +465,8 @@ declare <8 x i16> @llvm.mips.addv.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_addv_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_addv_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_addv_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_addv_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_addv_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_addv_w_RES
   ret void
@@ -490,8 +490,8 @@ declare <4 x i32> @llvm.mips.addv.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_addv_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_addv_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_addv_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_addv_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_addv_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_addv_d_RES
   ret void
@@ -512,8 +512,8 @@ declare <2 x i64> @llvm.mips.addv.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @addv_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_addv_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_addv_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_addv_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_addv_b_ARG2
   %2 = add <16 x i8> %0, %1
   store <16 x i8> %2, <16 x i8>* @llvm_mips_addv_b_RES
   ret void
@@ -532,8 +532,8 @@ entry:
 
 define void @addv_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_addv_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_addv_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_addv_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_addv_h_ARG2
   %2 = add <8 x i16> %0, %1
   store <8 x i16> %2, <8 x i16>* @llvm_mips_addv_h_RES
   ret void
@@ -552,8 +552,8 @@ entry:
 
 define void @addv_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_addv_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_addv_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_addv_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_addv_w_ARG2
   %2 = add <4 x i32> %0, %1
   store <4 x i32> %2, <4 x i32>* @llvm_mips_addv_w_RES
   ret void
@@ -572,8 +572,8 @@ entry:
 
 define void @addv_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_addv_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_addv_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_addv_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_addv_d_ARG2
   %2 = add <2 x i64> %0, %1
   store <2 x i64> %2, <2 x i64>* @llvm_mips_addv_d_RES
   ret void
@@ -595,8 +595,8 @@ entry:
 
 define void @llvm_mips_asub_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_asub_s_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_asub_s_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_asub_s_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_asub_s_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.asub.s.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_asub_s_b_RES
   ret void
@@ -620,8 +620,8 @@ declare <16 x i8> @llvm.mips.asub.s.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_asub_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_asub_s_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_asub_s_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_asub_s_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_asub_s_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.asub.s.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_asub_s_h_RES
   ret void
@@ -645,8 +645,8 @@ declare <8 x i16> @llvm.mips.asub.s.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_asub_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_asub_s_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_asub_s_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_asub_s_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_asub_s_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.asub.s.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_asub_s_w_RES
   ret void
@@ -670,8 +670,8 @@ declare <4 x i32> @llvm.mips.asub.s.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_asub_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_asub_s_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_asub_s_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_asub_s_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_asub_s_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.asub.s.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_asub_s_d_RES
   ret void
@@ -695,8 +695,8 @@ declare <2 x i64> @llvm.mips.asub.s.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_asub_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_asub_u_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_asub_u_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_asub_u_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_asub_u_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.asub.u.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_asub_u_b_RES
   ret void
@@ -720,8 +720,8 @@ declare <16 x i8> @llvm.mips.asub.u.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_asub_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_asub_u_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_asub_u_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_asub_u_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_asub_u_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.asub.u.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_asub_u_h_RES
   ret void
@@ -745,8 +745,8 @@ declare <8 x i16> @llvm.mips.asub.u.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_asub_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_asub_u_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_asub_u_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_asub_u_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_asub_u_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.asub.u.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_asub_u_w_RES
   ret void
@@ -770,8 +770,8 @@ declare <4 x i32> @llvm.mips.asub.u.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_asub_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_asub_u_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_asub_u_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_asub_u_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_asub_u_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.asub.u.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_asub_u_d_RES
   ret void
@@ -795,8 +795,8 @@ declare <2 x i64> @llvm.mips.asub.u.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_ave_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_ave_s_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_ave_s_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ave_s_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_ave_s_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.ave.s.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_ave_s_b_RES
   ret void
@@ -820,8 +820,8 @@ declare <16 x i8> @llvm.mips.ave.s.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_ave_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_ave_s_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_ave_s_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ave_s_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_ave_s_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.ave.s.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_ave_s_h_RES
   ret void
@@ -845,8 +845,8 @@ declare <8 x i16> @llvm.mips.ave.s.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_ave_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_ave_s_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_ave_s_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ave_s_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_ave_s_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.ave.s.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_ave_s_w_RES
   ret void
@@ -870,8 +870,8 @@ declare <4 x i32> @llvm.mips.ave.s.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_ave_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_ave_s_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_ave_s_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ave_s_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_ave_s_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.ave.s.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_ave_s_d_RES
   ret void
@@ -895,8 +895,8 @@ declare <2 x i64> @llvm.mips.ave.s.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_ave_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_ave_u_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_ave_u_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ave_u_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_ave_u_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.ave.u.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_ave_u_b_RES
   ret void
@@ -920,8 +920,8 @@ declare <16 x i8> @llvm.mips.ave.u.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_ave_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_ave_u_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_ave_u_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ave_u_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_ave_u_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.ave.u.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_ave_u_h_RES
   ret void
@@ -945,8 +945,8 @@ declare <8 x i16> @llvm.mips.ave.u.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_ave_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_ave_u_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_ave_u_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ave_u_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_ave_u_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.ave.u.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_ave_u_w_RES
   ret void
@@ -970,8 +970,8 @@ declare <4 x i32> @llvm.mips.ave.u.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_ave_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_ave_u_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_ave_u_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ave_u_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_ave_u_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.ave.u.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_ave_u_d_RES
   ret void
@@ -995,8 +995,8 @@ declare <2 x i64> @llvm.mips.ave.u.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_aver_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_aver_s_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_aver_s_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_aver_s_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_aver_s_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.aver.s.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_aver_s_b_RES
   ret void
@@ -1020,8 +1020,8 @@ declare <16 x i8> @llvm.mips.aver.s.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_aver_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_aver_s_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_aver_s_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_aver_s_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_aver_s_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.aver.s.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_aver_s_h_RES
   ret void
@@ -1045,8 +1045,8 @@ declare <8 x i16> @llvm.mips.aver.s.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_aver_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_aver_s_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_aver_s_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_aver_s_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_aver_s_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.aver.s.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_aver_s_w_RES
   ret void
@@ -1070,8 +1070,8 @@ declare <4 x i32> @llvm.mips.aver.s.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_aver_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_aver_s_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_aver_s_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_aver_s_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_aver_s_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.aver.s.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_aver_s_d_RES
   ret void
@@ -1095,8 +1095,8 @@ declare <2 x i64> @llvm.mips.aver.s.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_aver_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_aver_u_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_aver_u_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_aver_u_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_aver_u_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.aver.u.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_aver_u_b_RES
   ret void
@@ -1120,8 +1120,8 @@ declare <16 x i8> @llvm.mips.aver.u.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_aver_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_aver_u_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_aver_u_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_aver_u_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_aver_u_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.aver.u.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_aver_u_h_RES
   ret void
@@ -1145,8 +1145,8 @@ declare <8 x i16> @llvm.mips.aver.u.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_aver_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_aver_u_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_aver_u_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_aver_u_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_aver_u_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.aver.u.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_aver_u_w_RES
   ret void
@@ -1170,8 +1170,8 @@ declare <4 x i32> @llvm.mips.aver.u.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_aver_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_aver_u_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_aver_u_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_aver_u_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_aver_u_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.aver.u.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_aver_u_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3r-b.ll b/test/CodeGen/Mips/msa/3r-b.ll
index a05d19b4d490..2ecdc4290067 100644
--- a/test/CodeGen/Mips/msa/3r-b.ll
+++ b/test/CodeGen/Mips/msa/3r-b.ll
@@ -10,8 +10,8 @@
 
 define void @llvm_mips_bclr_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_bclr_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_bclr_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bclr_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bclr_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.bclr.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_bclr_b_RES
   ret void
@@ -32,8 +32,8 @@ declare <16 x i8> @llvm.mips.bclr.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_bclr_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_bclr_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_bclr_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bclr_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_bclr_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.bclr.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_bclr_h_RES
   ret void
@@ -54,8 +54,8 @@ declare <8 x i16> @llvm.mips.bclr.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_bclr_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_bclr_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_bclr_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bclr_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_bclr_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.bclr.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_bclr_w_RES
   ret void
@@ -76,8 +76,8 @@ declare <4 x i32> @llvm.mips.bclr.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_bclr_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_bclr_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_bclr_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bclr_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_bclr_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.bclr.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_bclr_d_RES
   ret void
@@ -99,9 +99,9 @@ declare <2 x i64> @llvm.mips.bclr.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_binsl_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_binsl_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_binsl_b_ARG2
-  %2 = load <16 x i8>* @llvm_mips_binsl_b_ARG3
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_binsl_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_binsl_b_ARG2
+  %2 = load <16 x i8>, <16 x i8>* @llvm_mips_binsl_b_ARG3
   %3 = tail call <16 x i8> @llvm.mips.binsl.b(<16 x i8> %0, <16 x i8> %1, <16 x i8> %2)
   store <16 x i8> %3, <16 x i8>* @llvm_mips_binsl_b_RES
   ret void
@@ -127,9 +127,9 @@ declare <16 x i8> @llvm.mips.binsl.b(<16 x i8>, <16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_binsl_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_binsl_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_binsl_h_ARG2
-  %2 = load <8 x i16>* @llvm_mips_binsl_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_binsl_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_binsl_h_ARG2
+  %2 = load <8 x i16>, <8 x i16>* @llvm_mips_binsl_h_ARG3
   %3 = tail call <8 x i16> @llvm.mips.binsl.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2)
   store <8 x i16> %3, <8 x i16>* @llvm_mips_binsl_h_RES
   ret void
@@ -155,9 +155,9 @@ declare <8 x i16> @llvm.mips.binsl.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_binsl_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_binsl_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_binsl_w_ARG2
-  %2 = load <4 x i32>* @llvm_mips_binsl_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_binsl_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_binsl_w_ARG2
+  %2 = load <4 x i32>, <4 x i32>* @llvm_mips_binsl_w_ARG3
   %3 = tail call <4 x i32> @llvm.mips.binsl.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2)
   store <4 x i32> %3, <4 x i32>* @llvm_mips_binsl_w_RES
   ret void
@@ -183,9 +183,9 @@ declare <4 x i32> @llvm.mips.binsl.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_binsl_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_binsl_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_binsl_d_ARG2
-  %2 = load <2 x i64>* @llvm_mips_binsl_d_ARG3
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_binsl_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_binsl_d_ARG2
+  %2 = load <2 x i64>, <2 x i64>* @llvm_mips_binsl_d_ARG3
   %3 = tail call <2 x i64> @llvm.mips.binsl.d(<2 x i64> %0, <2 x i64> %1, <2 x i64> %2)
   store <2 x i64> %3, <2 x i64>* @llvm_mips_binsl_d_RES
   ret void
@@ -211,9 +211,9 @@ declare <2 x i64> @llvm.mips.binsl.d(<2 x i64>, <2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_binsr_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_binsr_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_binsr_b_ARG2
-  %2 = load <16 x i8>* @llvm_mips_binsr_b_ARG3
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_binsr_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_binsr_b_ARG2
+  %2 = load <16 x i8>, <16 x i8>* @llvm_mips_binsr_b_ARG3
   %3 = tail call <16 x i8> @llvm.mips.binsr.b(<16 x i8> %0, <16 x i8> %1, <16 x i8> %2)
   store <16 x i8> %3, <16 x i8>* @llvm_mips_binsr_b_RES
   ret void
@@ -239,9 +239,9 @@ declare <16 x i8> @llvm.mips.binsr.b(<16 x i8>, <16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_binsr_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_binsr_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_binsr_h_ARG2
-  %2 = load <8 x i16>* @llvm_mips_binsr_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_binsr_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_binsr_h_ARG2
+  %2 = load <8 x i16>, <8 x i16>* @llvm_mips_binsr_h_ARG3
   %3 = tail call <8 x i16> @llvm.mips.binsr.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2)
   store <8 x i16> %3, <8 x i16>* @llvm_mips_binsr_h_RES
   ret void
@@ -267,9 +267,9 @@ declare <8 x i16> @llvm.mips.binsr.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_binsr_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_binsr_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_binsr_w_ARG2
-  %2 = load <4 x i32>* @llvm_mips_binsr_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_binsr_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_binsr_w_ARG2
+  %2 = load <4 x i32>, <4 x i32>* @llvm_mips_binsr_w_ARG3
   %3 = tail call <4 x i32> @llvm.mips.binsr.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2)
   store <4 x i32> %3, <4 x i32>* @llvm_mips_binsr_w_RES
   ret void
@@ -295,9 +295,9 @@ declare <4 x i32> @llvm.mips.binsr.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_binsr_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_binsr_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_binsr_d_ARG2
-  %2 = load <2 x i64>* @llvm_mips_binsr_d_ARG3
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_binsr_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_binsr_d_ARG2
+  %2 = load <2 x i64>, <2 x i64>* @llvm_mips_binsr_d_ARG3
   %3 = tail call <2 x i64> @llvm.mips.binsr.d(<2 x i64> %0, <2 x i64> %1, <2 x i64> %2)
   store <2 x i64> %3, <2 x i64>* @llvm_mips_binsr_d_RES
   ret void
@@ -322,8 +322,8 @@ declare <2 x i64> @llvm.mips.binsr.d(<2 x i64>, <2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_bneg_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_bneg_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_bneg_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bneg_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bneg_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.bneg.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_bneg_b_RES
   ret void
@@ -344,8 +344,8 @@ declare <16 x i8> @llvm.mips.bneg.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_bneg_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_bneg_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_bneg_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bneg_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_bneg_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.bneg.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_bneg_h_RES
   ret void
@@ -366,8 +366,8 @@ declare <8 x i16> @llvm.mips.bneg.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_bneg_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_bneg_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_bneg_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bneg_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_bneg_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.bneg.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_bneg_w_RES
   ret void
@@ -388,8 +388,8 @@ declare <4 x i32> @llvm.mips.bneg.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_bneg_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_bneg_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_bneg_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bneg_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_bneg_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.bneg.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_bneg_d_RES
   ret void
@@ -410,8 +410,8 @@ declare <2 x i64> @llvm.mips.bneg.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_bset_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_bset_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_bset_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bset_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bset_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.bset.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_bset_b_RES
   ret void
@@ -432,8 +432,8 @@ declare <16 x i8> @llvm.mips.bset.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_bset_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_bset_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_bset_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bset_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_bset_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.bset.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_bset_h_RES
   ret void
@@ -454,8 +454,8 @@ declare <8 x i16> @llvm.mips.bset.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_bset_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_bset_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_bset_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bset_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_bset_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.bset.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_bset_w_RES
   ret void
@@ -476,8 +476,8 @@ declare <4 x i32> @llvm.mips.bset.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_bset_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_bset_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_bset_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bset_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_bset_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.bset.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_bset_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3r-c.ll b/test/CodeGen/Mips/msa/3r-c.ll
index 6ec92c284fec..a3913e0a27fd 100644
--- a/test/CodeGen/Mips/msa/3r-c.ll
+++ b/test/CodeGen/Mips/msa/3r-c.ll
@@ -10,8 +10,8 @@
 
 define void @llvm_mips_ceq_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_ceq_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_ceq_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ceq_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_ceq_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.ceq.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_ceq_b_RES
   ret void
@@ -32,8 +32,8 @@ declare <16 x i8> @llvm.mips.ceq.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_ceq_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_ceq_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_ceq_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ceq_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_ceq_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.ceq.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_ceq_h_RES
   ret void
@@ -54,8 +54,8 @@ declare <8 x i16> @llvm.mips.ceq.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_ceq_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_ceq_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_ceq_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ceq_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_ceq_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.ceq.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_ceq_w_RES
   ret void
@@ -76,8 +76,8 @@ declare <4 x i32> @llvm.mips.ceq.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_ceq_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_ceq_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_ceq_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ceq_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_ceq_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.ceq.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_ceq_d_RES
   ret void
@@ -98,8 +98,8 @@ declare <2 x i64> @llvm.mips.ceq.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_cle_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_cle_s_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_cle_s_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_cle_s_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_cle_s_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.cle.s.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_cle_s_b_RES
   ret void
@@ -120,8 +120,8 @@ declare <16 x i8> @llvm.mips.cle.s.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_cle_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_cle_s_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_cle_s_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_cle_s_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_cle_s_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.cle.s.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_cle_s_h_RES
   ret void
@@ -142,8 +142,8 @@ declare <8 x i16> @llvm.mips.cle.s.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_cle_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_cle_s_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_cle_s_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_cle_s_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_cle_s_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.cle.s.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_cle_s_w_RES
   ret void
@@ -164,8 +164,8 @@ declare <4 x i32> @llvm.mips.cle.s.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_cle_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_cle_s_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_cle_s_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_cle_s_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_cle_s_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.cle.s.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_cle_s_d_RES
   ret void
@@ -186,8 +186,8 @@ declare <2 x i64> @llvm.mips.cle.s.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_cle_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_cle_u_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_cle_u_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_cle_u_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_cle_u_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.cle.u.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_cle_u_b_RES
   ret void
@@ -208,8 +208,8 @@ declare <16 x i8> @llvm.mips.cle.u.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_cle_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_cle_u_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_cle_u_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_cle_u_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_cle_u_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.cle.u.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_cle_u_h_RES
   ret void
@@ -230,8 +230,8 @@ declare <8 x i16> @llvm.mips.cle.u.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_cle_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_cle_u_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_cle_u_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_cle_u_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_cle_u_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.cle.u.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_cle_u_w_RES
   ret void
@@ -252,8 +252,8 @@ declare <4 x i32> @llvm.mips.cle.u.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_cle_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_cle_u_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_cle_u_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_cle_u_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_cle_u_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.cle.u.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_cle_u_d_RES
   ret void
@@ -274,8 +274,8 @@ declare <2 x i64> @llvm.mips.cle.u.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_clt_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_clt_s_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_clt_s_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_clt_s_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_clt_s_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.clt.s.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_clt_s_b_RES
   ret void
@@ -296,8 +296,8 @@ declare <16 x i8> @llvm.mips.clt.s.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_clt_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_clt_s_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_clt_s_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_clt_s_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_clt_s_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.clt.s.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_clt_s_h_RES
   ret void
@@ -318,8 +318,8 @@ declare <8 x i16> @llvm.mips.clt.s.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_clt_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_clt_s_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_clt_s_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_clt_s_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_clt_s_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.clt.s.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_clt_s_w_RES
   ret void
@@ -340,8 +340,8 @@ declare <4 x i32> @llvm.mips.clt.s.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_clt_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_clt_s_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_clt_s_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_clt_s_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_clt_s_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.clt.s.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_clt_s_d_RES
   ret void
@@ -362,8 +362,8 @@ declare <2 x i64> @llvm.mips.clt.s.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_clt_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_clt_u_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_clt_u_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_clt_u_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_clt_u_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.clt.u.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_clt_u_b_RES
   ret void
@@ -384,8 +384,8 @@ declare <16 x i8> @llvm.mips.clt.u.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_clt_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_clt_u_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_clt_u_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_clt_u_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_clt_u_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.clt.u.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_clt_u_h_RES
   ret void
@@ -406,8 +406,8 @@ declare <8 x i16> @llvm.mips.clt.u.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_clt_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_clt_u_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_clt_u_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_clt_u_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_clt_u_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.clt.u.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_clt_u_w_RES
   ret void
@@ -428,8 +428,8 @@ declare <4 x i32> @llvm.mips.clt.u.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_clt_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_clt_u_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_clt_u_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_clt_u_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_clt_u_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.clt.u.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_clt_u_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3r-d.ll b/test/CodeGen/Mips/msa/3r-d.ll
index 0099554a8eea..4fc32b76a7b2 100644
--- a/test/CodeGen/Mips/msa/3r-d.ll
+++ b/test/CodeGen/Mips/msa/3r-d.ll
@@ -10,8 +10,8 @@
 
 define void @llvm_mips_div_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_div_s_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_div_s_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_div_s_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_div_s_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.div.s.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_div_s_b_RES
   ret void
@@ -32,8 +32,8 @@ declare <16 x i8> @llvm.mips.div.s.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_div_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_div_s_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_div_s_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_div_s_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_div_s_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.div.s.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_div_s_h_RES
   ret void
@@ -54,8 +54,8 @@ declare <8 x i16> @llvm.mips.div.s.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_div_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_div_s_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_div_s_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_div_s_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_div_s_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.div.s.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_div_s_w_RES
   ret void
@@ -76,8 +76,8 @@ declare <4 x i32> @llvm.mips.div.s.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_div_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_div_s_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_div_s_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_div_s_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_div_s_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.div.s.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_div_s_d_RES
   ret void
@@ -95,8 +95,8 @@ declare <2 x i64> @llvm.mips.div.s.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @div_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_div_s_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_div_s_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_div_s_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_div_s_b_ARG2
   %2 = sdiv <16 x i8> %0, %1
   store <16 x i8> %2, <16 x i8>* @llvm_mips_div_s_b_RES
   ret void
@@ -111,8 +111,8 @@ entry:
 
 define void @div_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_div_s_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_div_s_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_div_s_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_div_s_h_ARG2
   %2 = sdiv <8 x i16> %0, %1
   store <8 x i16> %2, <8 x i16>* @llvm_mips_div_s_h_RES
   ret void
@@ -127,8 +127,8 @@ entry:
 
 define void @div_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_div_s_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_div_s_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_div_s_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_div_s_w_ARG2
   %2 = sdiv <4 x i32> %0, %1
   store <4 x i32> %2, <4 x i32>* @llvm_mips_div_s_w_RES
   ret void
@@ -143,8 +143,8 @@ entry:
 
 define void @div_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_div_s_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_div_s_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_div_s_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_div_s_d_ARG2
   %2 = sdiv <2 x i64> %0, %1
   store <2 x i64> %2, <2 x i64>* @llvm_mips_div_s_d_RES
   ret void
@@ -163,8 +163,8 @@ entry:
 
 define void @llvm_mips_div_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_div_u_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_div_u_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_div_u_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_div_u_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.div.u.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_div_u_b_RES
   ret void
@@ -185,8 +185,8 @@ declare <16 x i8> @llvm.mips.div.u.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_div_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_div_u_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_div_u_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_div_u_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_div_u_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.div.u.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_div_u_h_RES
   ret void
@@ -207,8 +207,8 @@ declare <8 x i16> @llvm.mips.div.u.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_div_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_div_u_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_div_u_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_div_u_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_div_u_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.div.u.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_div_u_w_RES
   ret void
@@ -229,8 +229,8 @@ declare <4 x i32> @llvm.mips.div.u.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_div_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_div_u_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_div_u_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_div_u_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_div_u_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.div.u.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_div_u_d_RES
   ret void
@@ -248,8 +248,8 @@ declare <2 x i64> @llvm.mips.div.u.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @div_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_div_u_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_div_u_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_div_u_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_div_u_b_ARG2
   %2 = udiv <16 x i8> %0, %1
   store <16 x i8> %2, <16 x i8>* @llvm_mips_div_u_b_RES
   ret void
@@ -264,8 +264,8 @@ entry:
 
 define void @div_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_div_u_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_div_u_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_div_u_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_div_u_h_ARG2
   %2 = udiv <8 x i16> %0, %1
   store <8 x i16> %2, <8 x i16>* @llvm_mips_div_u_h_RES
   ret void
@@ -280,8 +280,8 @@ entry:
 
 define void @div_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_div_u_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_div_u_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_div_u_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_div_u_w_ARG2
   %2 = udiv <4 x i32> %0, %1
   store <4 x i32> %2, <4 x i32>* @llvm_mips_div_u_w_RES
   ret void
@@ -296,8 +296,8 @@ entry:
 
 define void @div_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_div_u_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_div_u_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_div_u_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_div_u_d_ARG2
   %2 = udiv <2 x i64> %0, %1
   store <2 x i64> %2, <2 x i64>* @llvm_mips_div_u_d_RES
   ret void
@@ -326,8 +326,8 @@ entry:
 
 define void @llvm_mips_dotp_s_h_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_dotp_s_h_ARG1
-  %1 = load <16 x i8>* @llvm_mips_dotp_s_h_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_dotp_s_h_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_dotp_s_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.dotp.s.h(<16 x i8> %0, <16 x i8> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_dotp_s_h_RES
   ret void
@@ -353,8 +353,8 @@ declare <8 x i16> @llvm.mips.dotp.s.h(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_dotp_s_w_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_dotp_s_w_ARG1
-  %1 = load <8 x i16>* @llvm_mips_dotp_s_w_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_dotp_s_w_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_dotp_s_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.dotp.s.w(<8 x i16> %0, <8 x i16> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_dotp_s_w_RES
   ret void
@@ -377,8 +377,8 @@ declare <4 x i32> @llvm.mips.dotp.s.w(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_dotp_s_d_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_dotp_s_d_ARG1
-  %1 = load <4 x i32>* @llvm_mips_dotp_s_d_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_dotp_s_d_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_dotp_s_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.dotp.s.d(<4 x i32> %0, <4 x i32> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_dotp_s_d_RES
   ret void
@@ -409,8 +409,8 @@ declare <2 x i64> @llvm.mips.dotp.s.d(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_dotp_u_h_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_dotp_u_h_ARG1
-  %1 = load <16 x i8>* @llvm_mips_dotp_u_h_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_dotp_u_h_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_dotp_u_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.dotp.u.h(<16 x i8> %0, <16 x i8> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_dotp_u_h_RES
   ret void
@@ -436,8 +436,8 @@ declare <8 x i16> @llvm.mips.dotp.u.h(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_dotp_u_w_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_dotp_u_w_ARG1
-  %1 = load <8 x i16>* @llvm_mips_dotp_u_w_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_dotp_u_w_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_dotp_u_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.dotp.u.w(<8 x i16> %0, <8 x i16> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_dotp_u_w_RES
   ret void
@@ -460,8 +460,8 @@ declare <4 x i32> @llvm.mips.dotp.u.w(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_dotp_u_d_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_dotp_u_d_ARG1
-  %1 = load <4 x i32>* @llvm_mips_dotp_u_d_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_dotp_u_d_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_dotp_u_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.dotp.u.d(<4 x i32> %0, <4 x i32> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_dotp_u_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3r-i.ll b/test/CodeGen/Mips/msa/3r-i.ll
index 2ef30471b026..7147b756b15f 100644
--- a/test/CodeGen/Mips/msa/3r-i.ll
+++ b/test/CodeGen/Mips/msa/3r-i.ll
@@ -10,8 +10,8 @@
 
 define void @llvm_mips_ilvev_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_ilvev_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_ilvev_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ilvev_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_ilvev_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.ilvev.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_ilvev_b_RES
   ret void
@@ -32,8 +32,8 @@ declare <16 x i8> @llvm.mips.ilvev.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_ilvev_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_ilvev_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_ilvev_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ilvev_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_ilvev_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.ilvev.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_ilvev_h_RES
   ret void
@@ -54,8 +54,8 @@ declare <8 x i16> @llvm.mips.ilvev.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_ilvev_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_ilvev_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_ilvev_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ilvev_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_ilvev_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.ilvev.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_ilvev_w_RES
   ret void
@@ -76,8 +76,8 @@ declare <4 x i32> @llvm.mips.ilvev.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_ilvev_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_ilvev_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_ilvev_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ilvev_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_ilvev_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.ilvev.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_ilvev_d_RES
   ret void
@@ -98,8 +98,8 @@ declare <2 x i64> @llvm.mips.ilvev.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_ilvl_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_ilvl_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_ilvl_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ilvl_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_ilvl_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.ilvl.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_ilvl_b_RES
   ret void
@@ -120,8 +120,8 @@ declare <16 x i8> @llvm.mips.ilvl.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_ilvl_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_ilvl_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_ilvl_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ilvl_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_ilvl_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.ilvl.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_ilvl_h_RES
   ret void
@@ -142,8 +142,8 @@ declare <8 x i16> @llvm.mips.ilvl.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_ilvl_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_ilvl_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_ilvl_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ilvl_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_ilvl_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.ilvl.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_ilvl_w_RES
   ret void
@@ -164,8 +164,8 @@ declare <4 x i32> @llvm.mips.ilvl.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_ilvl_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_ilvl_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_ilvl_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ilvl_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_ilvl_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.ilvl.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_ilvl_d_RES
   ret void
@@ -186,8 +186,8 @@ declare <2 x i64> @llvm.mips.ilvl.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_ilvod_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_ilvod_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_ilvod_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ilvod_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_ilvod_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.ilvod.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_ilvod_b_RES
   ret void
@@ -208,8 +208,8 @@ declare <16 x i8> @llvm.mips.ilvod.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_ilvod_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_ilvod_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_ilvod_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ilvod_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_ilvod_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.ilvod.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_ilvod_h_RES
   ret void
@@ -230,8 +230,8 @@ declare <8 x i16> @llvm.mips.ilvod.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_ilvod_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_ilvod_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_ilvod_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ilvod_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_ilvod_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.ilvod.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_ilvod_w_RES
   ret void
@@ -252,8 +252,8 @@ declare <4 x i32> @llvm.mips.ilvod.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_ilvod_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_ilvod_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_ilvod_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ilvod_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_ilvod_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.ilvod.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_ilvod_d_RES
   ret void
@@ -274,8 +274,8 @@ declare <2 x i64> @llvm.mips.ilvod.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_ilvr_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_ilvr_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_ilvr_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ilvr_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_ilvr_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.ilvr.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_ilvr_b_RES
   ret void
@@ -296,8 +296,8 @@ declare <16 x i8> @llvm.mips.ilvr.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_ilvr_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_ilvr_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_ilvr_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ilvr_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_ilvr_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.ilvr.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_ilvr_h_RES
   ret void
@@ -318,8 +318,8 @@ declare <8 x i16> @llvm.mips.ilvr.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_ilvr_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_ilvr_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_ilvr_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ilvr_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_ilvr_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.ilvr.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_ilvr_w_RES
   ret void
@@ -340,8 +340,8 @@ declare <4 x i32> @llvm.mips.ilvr.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_ilvr_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_ilvr_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_ilvr_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ilvr_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_ilvr_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.ilvr.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_ilvr_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3r-m.ll b/test/CodeGen/Mips/msa/3r-m.ll
index ddfd720a2f84..39b4f7db3a48 100644
--- a/test/CodeGen/Mips/msa/3r-m.ll
+++ b/test/CodeGen/Mips/msa/3r-m.ll
@@ -10,8 +10,8 @@
 
 define void @llvm_mips_max_a_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_max_a_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_max_a_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_max_a_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_max_a_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.max.a.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_max_a_b_RES
   ret void
@@ -32,8 +32,8 @@ declare <16 x i8> @llvm.mips.max.a.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_max_a_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_max_a_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_max_a_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_max_a_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_max_a_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.max.a.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_max_a_h_RES
   ret void
@@ -54,8 +54,8 @@ declare <8 x i16> @llvm.mips.max.a.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_max_a_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_max_a_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_max_a_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_max_a_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_max_a_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.max.a.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_max_a_w_RES
   ret void
@@ -76,8 +76,8 @@ declare <4 x i32> @llvm.mips.max.a.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_max_a_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_max_a_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_max_a_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_max_a_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_max_a_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.max.a.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_max_a_d_RES
   ret void
@@ -98,8 +98,8 @@ declare <2 x i64> @llvm.mips.max.a.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_max_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_max_s_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_max_s_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_max_s_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_max_s_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.max.s.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_max_s_b_RES
   ret void
@@ -120,8 +120,8 @@ declare <16 x i8> @llvm.mips.max.s.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_max_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_max_s_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_max_s_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_max_s_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_max_s_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.max.s.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_max_s_h_RES
   ret void
@@ -142,8 +142,8 @@ declare <8 x i16> @llvm.mips.max.s.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_max_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_max_s_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_max_s_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_max_s_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_max_s_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.max.s.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_max_s_w_RES
   ret void
@@ -164,8 +164,8 @@ declare <4 x i32> @llvm.mips.max.s.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_max_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_max_s_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_max_s_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_max_s_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_max_s_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.max.s.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_max_s_d_RES
   ret void
@@ -186,8 +186,8 @@ declare <2 x i64> @llvm.mips.max.s.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_max_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_max_u_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_max_u_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_max_u_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_max_u_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.max.u.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_max_u_b_RES
   ret void
@@ -208,8 +208,8 @@ declare <16 x i8> @llvm.mips.max.u.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_max_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_max_u_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_max_u_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_max_u_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_max_u_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.max.u.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_max_u_h_RES
   ret void
@@ -230,8 +230,8 @@ declare <8 x i16> @llvm.mips.max.u.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_max_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_max_u_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_max_u_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_max_u_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_max_u_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.max.u.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_max_u_w_RES
   ret void
@@ -252,8 +252,8 @@ declare <4 x i32> @llvm.mips.max.u.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_max_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_max_u_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_max_u_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_max_u_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_max_u_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.max.u.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_max_u_d_RES
   ret void
@@ -274,8 +274,8 @@ declare <2 x i64> @llvm.mips.max.u.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_min_a_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_min_a_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_min_a_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_min_a_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_min_a_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.min.a.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_min_a_b_RES
   ret void
@@ -296,8 +296,8 @@ declare <16 x i8> @llvm.mips.min.a.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_min_a_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_min_a_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_min_a_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_min_a_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_min_a_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.min.a.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_min_a_h_RES
   ret void
@@ -318,8 +318,8 @@ declare <8 x i16> @llvm.mips.min.a.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_min_a_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_min_a_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_min_a_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_min_a_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_min_a_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.min.a.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_min_a_w_RES
   ret void
@@ -340,8 +340,8 @@ declare <4 x i32> @llvm.mips.min.a.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_min_a_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_min_a_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_min_a_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_min_a_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_min_a_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.min.a.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_min_a_d_RES
   ret void
@@ -362,8 +362,8 @@ declare <2 x i64> @llvm.mips.min.a.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_min_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_min_s_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_min_s_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_min_s_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_min_s_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.min.s.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_min_s_b_RES
   ret void
@@ -384,8 +384,8 @@ declare <16 x i8> @llvm.mips.min.s.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_min_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_min_s_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_min_s_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_min_s_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_min_s_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.min.s.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_min_s_h_RES
   ret void
@@ -406,8 +406,8 @@ declare <8 x i16> @llvm.mips.min.s.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_min_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_min_s_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_min_s_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_min_s_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_min_s_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.min.s.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_min_s_w_RES
   ret void
@@ -428,8 +428,8 @@ declare <4 x i32> @llvm.mips.min.s.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_min_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_min_s_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_min_s_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_min_s_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_min_s_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.min.s.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_min_s_d_RES
   ret void
@@ -450,8 +450,8 @@ declare <2 x i64> @llvm.mips.min.s.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_min_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_min_u_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_min_u_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_min_u_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_min_u_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.min.u.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_min_u_b_RES
   ret void
@@ -472,8 +472,8 @@ declare <16 x i8> @llvm.mips.min.u.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_min_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_min_u_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_min_u_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_min_u_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_min_u_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.min.u.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_min_u_h_RES
   ret void
@@ -494,8 +494,8 @@ declare <8 x i16> @llvm.mips.min.u.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_min_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_min_u_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_min_u_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_min_u_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_min_u_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.min.u.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_min_u_w_RES
   ret void
@@ -516,8 +516,8 @@ declare <4 x i32> @llvm.mips.min.u.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_min_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_min_u_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_min_u_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_min_u_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_min_u_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.min.u.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_min_u_d_RES
   ret void
@@ -538,8 +538,8 @@ declare <2 x i64> @llvm.mips.min.u.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_mod_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_mod_s_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_mod_s_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_mod_s_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_mod_s_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.mod.s.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_mod_s_b_RES
   ret void
@@ -560,8 +560,8 @@ declare <16 x i8> @llvm.mips.mod.s.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_mod_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_mod_s_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_mod_s_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_mod_s_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_mod_s_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.mod.s.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_mod_s_h_RES
   ret void
@@ -582,8 +582,8 @@ declare <8 x i16> @llvm.mips.mod.s.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_mod_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_mod_s_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_mod_s_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_mod_s_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_mod_s_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.mod.s.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_mod_s_w_RES
   ret void
@@ -604,8 +604,8 @@ declare <4 x i32> @llvm.mips.mod.s.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_mod_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_mod_s_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_mod_s_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_mod_s_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_mod_s_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.mod.s.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_mod_s_d_RES
   ret void
@@ -626,8 +626,8 @@ declare <2 x i64> @llvm.mips.mod.s.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_mod_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_mod_u_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_mod_u_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_mod_u_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_mod_u_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.mod.u.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_mod_u_b_RES
   ret void
@@ -648,8 +648,8 @@ declare <16 x i8> @llvm.mips.mod.u.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_mod_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_mod_u_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_mod_u_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_mod_u_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_mod_u_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.mod.u.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_mod_u_h_RES
   ret void
@@ -670,8 +670,8 @@ declare <8 x i16> @llvm.mips.mod.u.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_mod_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_mod_u_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_mod_u_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_mod_u_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_mod_u_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.mod.u.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_mod_u_w_RES
   ret void
@@ -692,8 +692,8 @@ declare <4 x i32> @llvm.mips.mod.u.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_mod_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_mod_u_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_mod_u_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_mod_u_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_mod_u_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.mod.u.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_mod_u_d_RES
   ret void
@@ -714,8 +714,8 @@ declare <2 x i64> @llvm.mips.mod.u.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_mulv_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_mulv_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_mulv_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_mulv_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_mulv_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.mulv.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_mulv_b_RES
   ret void
@@ -736,8 +736,8 @@ declare <16 x i8> @llvm.mips.mulv.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_mulv_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_mulv_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_mulv_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_mulv_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_mulv_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.mulv.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_mulv_h_RES
   ret void
@@ -758,8 +758,8 @@ declare <8 x i16> @llvm.mips.mulv.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_mulv_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_mulv_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_mulv_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_mulv_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_mulv_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.mulv.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_mulv_w_RES
   ret void
@@ -780,8 +780,8 @@ declare <4 x i32> @llvm.mips.mulv.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_mulv_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_mulv_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_mulv_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_mulv_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_mulv_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.mulv.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_mulv_d_RES
   ret void
@@ -798,8 +798,8 @@ declare <2 x i64> @llvm.mips.mulv.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @mulv_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_mulv_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_mulv_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_mulv_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_mulv_b_ARG2
   %2 = mul <16 x i8> %0, %1
   store <16 x i8> %2, <16 x i8>* @llvm_mips_mulv_b_RES
   ret void
@@ -814,8 +814,8 @@ entry:
 
 define void @mulv_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_mulv_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_mulv_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_mulv_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_mulv_h_ARG2
   %2 = mul <8 x i16> %0, %1
   store <8 x i16> %2, <8 x i16>* @llvm_mips_mulv_h_RES
   ret void
@@ -830,8 +830,8 @@ entry:
 
 define void @mulv_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_mulv_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_mulv_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_mulv_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_mulv_w_ARG2
   %2 = mul <4 x i32> %0, %1
   store <4 x i32> %2, <4 x i32>* @llvm_mips_mulv_w_RES
   ret void
@@ -846,8 +846,8 @@ entry:
 
 define void @mulv_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_mulv_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_mulv_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_mulv_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_mulv_d_ARG2
   %2 = mul <2 x i64> %0, %1
   store <2 x i64> %2, <2 x i64>* @llvm_mips_mulv_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3r-p.ll b/test/CodeGen/Mips/msa/3r-p.ll
index 852023b0824a..70b98aa8f9a1 100644
--- a/test/CodeGen/Mips/msa/3r-p.ll
+++ b/test/CodeGen/Mips/msa/3r-p.ll
@@ -10,8 +10,8 @@
 
 define void @llvm_mips_pckev_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_pckev_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_pckev_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_pckev_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_pckev_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.pckev.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_pckev_b_RES
   ret void
@@ -32,8 +32,8 @@ declare <16 x i8> @llvm.mips.pckev.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_pckev_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_pckev_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_pckev_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_pckev_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_pckev_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.pckev.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_pckev_h_RES
   ret void
@@ -54,8 +54,8 @@ declare <8 x i16> @llvm.mips.pckev.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_pckev_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_pckev_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_pckev_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_pckev_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_pckev_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.pckev.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_pckev_w_RES
   ret void
@@ -76,8 +76,8 @@ declare <4 x i32> @llvm.mips.pckev.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_pckev_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_pckev_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_pckev_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_pckev_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_pckev_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.pckev.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_pckev_d_RES
   ret void
@@ -98,8 +98,8 @@ declare <2 x i64> @llvm.mips.pckev.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_pckod_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_pckod_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_pckod_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_pckod_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_pckod_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.pckod.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_pckod_b_RES
   ret void
@@ -120,8 +120,8 @@ declare <16 x i8> @llvm.mips.pckod.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_pckod_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_pckod_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_pckod_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_pckod_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_pckod_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.pckod.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_pckod_h_RES
   ret void
@@ -142,8 +142,8 @@ declare <8 x i16> @llvm.mips.pckod.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_pckod_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_pckod_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_pckod_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_pckod_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_pckod_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.pckod.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_pckod_w_RES
   ret void
@@ -164,8 +164,8 @@ declare <4 x i32> @llvm.mips.pckod.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_pckod_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_pckod_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_pckod_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_pckod_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_pckod_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.pckod.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_pckod_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3r-s.ll b/test/CodeGen/Mips/msa/3r-s.ll
index 581c3bfd78af..d04c5ff165f2 100644
--- a/test/CodeGen/Mips/msa/3r-s.ll
+++ b/test/CodeGen/Mips/msa/3r-s.ll
@@ -11,9 +11,9 @@
 
 define void @llvm_mips_sld_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_sld_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_sld_b_ARG2
-  %2 = load i32* @llvm_mips_sld_b_ARG3
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_sld_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_sld_b_ARG2
+  %2 = load i32, i32* @llvm_mips_sld_b_ARG3
   %3 = tail call <16 x i8> @llvm.mips.sld.b(<16 x i8> %0, <16 x i8> %1, i32 %2)
   store <16 x i8> %3, <16 x i8>* @llvm_mips_sld_b_RES
   ret void
@@ -39,9 +39,9 @@ declare <16 x i8> @llvm.mips.sld.b(<16 x i8>, <16 x i8>, i32) nounwind
 
 define void @llvm_mips_sld_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_sld_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_sld_h_ARG2
-  %2 = load i32* @llvm_mips_sld_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_sld_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_sld_h_ARG2
+  %2 = load i32, i32* @llvm_mips_sld_h_ARG3
   %3 = tail call <8 x i16> @llvm.mips.sld.h(<8 x i16> %0, <8 x i16> %1, i32 %2)
   store <8 x i16> %3, <8 x i16>* @llvm_mips_sld_h_RES
   ret void
@@ -67,9 +67,9 @@ declare <8 x i16> @llvm.mips.sld.h(<8 x i16>, <8 x i16>, i32) nounwind
 
 define void @llvm_mips_sld_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_sld_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_sld_w_ARG2
-  %2 = load i32* @llvm_mips_sld_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_sld_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_sld_w_ARG2
+  %2 = load i32, i32* @llvm_mips_sld_w_ARG3
   %3 = tail call <4 x i32> @llvm.mips.sld.w(<4 x i32> %0, <4 x i32> %1, i32 %2)
   store <4 x i32> %3, <4 x i32>* @llvm_mips_sld_w_RES
   ret void
@@ -95,9 +95,9 @@ declare <4 x i32> @llvm.mips.sld.w(<4 x i32>, <4 x i32>, i32) nounwind
 
 define void @llvm_mips_sld_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_sld_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_sld_d_ARG2
-  %2 = load i32* @llvm_mips_sld_d_ARG3
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_sld_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_sld_d_ARG2
+  %2 = load i32, i32* @llvm_mips_sld_d_ARG3
   %3 = tail call <2 x i64> @llvm.mips.sld.d(<2 x i64> %0, <2 x i64> %1, i32 %2)
   store <2 x i64> %3, <2 x i64>* @llvm_mips_sld_d_RES
   ret void
@@ -122,8 +122,8 @@ declare <2 x i64> @llvm.mips.sld.d(<2 x i64>, <2 x i64>, i32) nounwind
 
 define void @llvm_mips_sll_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_sll_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_sll_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_sll_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_sll_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.sll.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_sll_b_RES
   ret void
@@ -146,8 +146,8 @@ declare <16 x i8> @llvm.mips.sll.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_sll_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_sll_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_sll_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_sll_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_sll_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.sll.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_sll_h_RES
   ret void
@@ -170,8 +170,8 @@ declare <8 x i16> @llvm.mips.sll.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_sll_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_sll_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_sll_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_sll_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_sll_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.sll.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_sll_w_RES
   ret void
@@ -194,8 +194,8 @@ declare <4 x i32> @llvm.mips.sll.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_sll_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_sll_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_sll_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_sll_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_sll_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.sll.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_sll_d_RES
   ret void
@@ -214,8 +214,8 @@ declare <2 x i64> @llvm.mips.sll.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @sll_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_sll_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_sll_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_sll_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_sll_b_ARG2
   %2 = shl <16 x i8> %0, %1
   store <16 x i8> %2, <16 x i8>* @llvm_mips_sll_b_RES
   ret void
@@ -232,8 +232,8 @@ entry:
 
 define void @sll_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_sll_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_sll_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_sll_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_sll_h_ARG2
   %2 = shl <8 x i16> %0, %1
   store <8 x i16> %2, <8 x i16>* @llvm_mips_sll_h_RES
   ret void
@@ -250,8 +250,8 @@ entry:
 
 define void @sll_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_sll_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_sll_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_sll_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_sll_w_ARG2
   %2 = shl <4 x i32> %0, %1
   store <4 x i32> %2, <4 x i32>* @llvm_mips_sll_w_RES
   ret void
@@ -268,8 +268,8 @@ entry:
 
 define void @sll_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_sll_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_sll_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_sll_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_sll_d_ARG2
   %2 = shl <2 x i64> %0, %1
   store <2 x i64> %2, <2 x i64>* @llvm_mips_sll_d_RES
   ret void
@@ -290,8 +290,8 @@ entry:
 
 define void @llvm_mips_sra_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_sra_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_sra_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_sra_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_sra_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.sra.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_sra_b_RES
   ret void
@@ -314,8 +314,8 @@ declare <16 x i8> @llvm.mips.sra.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_sra_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_sra_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_sra_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_sra_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_sra_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.sra.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_sra_h_RES
   ret void
@@ -338,8 +338,8 @@ declare <8 x i16> @llvm.mips.sra.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_sra_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_sra_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_sra_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_sra_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_sra_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.sra.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_sra_w_RES
   ret void
@@ -362,8 +362,8 @@ declare <4 x i32> @llvm.mips.sra.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_sra_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_sra_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_sra_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_sra_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_sra_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.sra.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_sra_d_RES
   ret void
@@ -383,8 +383,8 @@ declare <2 x i64> @llvm.mips.sra.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @sra_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_sra_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_sra_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_sra_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_sra_b_ARG2
   %2 = ashr <16 x i8> %0, %1
   store <16 x i8> %2, <16 x i8>* @llvm_mips_sra_b_RES
   ret void
@@ -401,8 +401,8 @@ entry:
 
 define void @sra_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_sra_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_sra_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_sra_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_sra_h_ARG2
   %2 = ashr <8 x i16> %0, %1
   store <8 x i16> %2, <8 x i16>* @llvm_mips_sra_h_RES
   ret void
@@ -419,8 +419,8 @@ entry:
 
 define void @sra_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_sra_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_sra_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_sra_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_sra_w_ARG2
   %2 = ashr <4 x i32> %0, %1
   store <4 x i32> %2, <4 x i32>* @llvm_mips_sra_w_RES
   ret void
@@ -437,8 +437,8 @@ entry:
 
 define void @sra_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_sra_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_sra_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_sra_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_sra_d_ARG2
   %2 = ashr <2 x i64> %0, %1
   store <2 x i64> %2, <2 x i64>* @llvm_mips_sra_d_RES
   ret void
@@ -459,8 +459,8 @@ entry:
 
 define void @llvm_mips_srar_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_srar_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_srar_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_srar_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_srar_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.srar.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_srar_b_RES
   ret void
@@ -483,8 +483,8 @@ declare <16 x i8> @llvm.mips.srar.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_srar_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_srar_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_srar_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_srar_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_srar_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.srar.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_srar_h_RES
   ret void
@@ -507,8 +507,8 @@ declare <8 x i16> @llvm.mips.srar.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_srar_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_srar_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_srar_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_srar_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_srar_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.srar.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_srar_w_RES
   ret void
@@ -531,8 +531,8 @@ declare <4 x i32> @llvm.mips.srar.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_srar_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_srar_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_srar_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_srar_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_srar_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.srar.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_srar_d_RES
   ret void
@@ -555,8 +555,8 @@ declare <2 x i64> @llvm.mips.srar.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_srl_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_srl_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_srl_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_srl_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_srl_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.srl.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_srl_b_RES
   ret void
@@ -579,8 +579,8 @@ declare <16 x i8> @llvm.mips.srl.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_srl_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_srl_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_srl_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_srl_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_srl_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.srl.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_srl_h_RES
   ret void
@@ -603,8 +603,8 @@ declare <8 x i16> @llvm.mips.srl.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_srl_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_srl_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_srl_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_srl_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_srl_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.srl.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_srl_w_RES
   ret void
@@ -627,8 +627,8 @@ declare <4 x i32> @llvm.mips.srl.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_srl_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_srl_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_srl_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_srl_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_srl_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.srl.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_srl_d_RES
   ret void
@@ -651,8 +651,8 @@ declare <2 x i64> @llvm.mips.srl.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_srlr_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_srlr_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_srlr_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_srlr_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_srlr_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.srlr.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_srlr_b_RES
   ret void
@@ -675,8 +675,8 @@ declare <16 x i8> @llvm.mips.srlr.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_srlr_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_srlr_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_srlr_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_srlr_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_srlr_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.srlr.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_srlr_h_RES
   ret void
@@ -699,8 +699,8 @@ declare <8 x i16> @llvm.mips.srlr.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_srlr_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_srlr_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_srlr_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_srlr_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_srlr_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.srlr.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_srlr_w_RES
   ret void
@@ -723,8 +723,8 @@ declare <4 x i32> @llvm.mips.srlr.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_srlr_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_srlr_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_srlr_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_srlr_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_srlr_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.srlr.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_srlr_d_RES
   ret void
@@ -744,8 +744,8 @@ declare <2 x i64> @llvm.mips.srlr.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @srl_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_srl_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_srl_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_srl_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_srl_b_ARG2
   %2 = lshr <16 x i8> %0, %1
   store <16 x i8> %2, <16 x i8>* @llvm_mips_srl_b_RES
   ret void
@@ -762,8 +762,8 @@ entry:
 
 define void @srl_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_srl_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_srl_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_srl_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_srl_h_ARG2
   %2 = lshr <8 x i16> %0, %1
   store <8 x i16> %2, <8 x i16>* @llvm_mips_srl_h_RES
   ret void
@@ -780,8 +780,8 @@ entry:
 
 define void @srl_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_srl_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_srl_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_srl_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_srl_w_ARG2
   %2 = lshr <4 x i32> %0, %1
   store <4 x i32> %2, <4 x i32>* @llvm_mips_srl_w_RES
   ret void
@@ -798,8 +798,8 @@ entry:
 
 define void @srl_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_srl_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_srl_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_srl_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_srl_d_ARG2
   %2 = lshr <2 x i64> %0, %1
   store <2 x i64> %2, <2 x i64>* @llvm_mips_srl_d_RES
   ret void
@@ -820,8 +820,8 @@ entry:
 
 define void @llvm_mips_subs_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_subs_s_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_subs_s_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_subs_s_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_subs_s_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.subs.s.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_subs_s_b_RES
   ret void
@@ -844,8 +844,8 @@ declare <16 x i8> @llvm.mips.subs.s.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_subs_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_subs_s_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_subs_s_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_subs_s_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_subs_s_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.subs.s.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_subs_s_h_RES
   ret void
@@ -868,8 +868,8 @@ declare <8 x i16> @llvm.mips.subs.s.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_subs_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_subs_s_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_subs_s_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_subs_s_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_subs_s_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.subs.s.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_subs_s_w_RES
   ret void
@@ -892,8 +892,8 @@ declare <4 x i32> @llvm.mips.subs.s.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_subs_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_subs_s_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_subs_s_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_subs_s_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_subs_s_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.subs.s.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_subs_s_d_RES
   ret void
@@ -916,8 +916,8 @@ declare <2 x i64> @llvm.mips.subs.s.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_subs_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_subs_u_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_subs_u_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_subs_u_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_subs_u_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.subs.u.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_subs_u_b_RES
   ret void
@@ -940,8 +940,8 @@ declare <16 x i8> @llvm.mips.subs.u.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_subs_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_subs_u_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_subs_u_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_subs_u_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_subs_u_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.subs.u.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_subs_u_h_RES
   ret void
@@ -964,8 +964,8 @@ declare <8 x i16> @llvm.mips.subs.u.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_subs_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_subs_u_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_subs_u_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_subs_u_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_subs_u_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.subs.u.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_subs_u_w_RES
   ret void
@@ -988,8 +988,8 @@ declare <4 x i32> @llvm.mips.subs.u.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_subs_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_subs_u_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_subs_u_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_subs_u_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_subs_u_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.subs.u.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_subs_u_d_RES
   ret void
@@ -1012,8 +1012,8 @@ declare <2 x i64> @llvm.mips.subs.u.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_subsus_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_subsus_u_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_subsus_u_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_subsus_u_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_subsus_u_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.subsus.u.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_subsus_u_b_RES
   ret void
@@ -1036,8 +1036,8 @@ declare <16 x i8> @llvm.mips.subsus.u.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_subsus_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_subsus_u_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_subsus_u_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_subsus_u_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_subsus_u_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.subsus.u.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_subsus_u_h_RES
   ret void
@@ -1060,8 +1060,8 @@ declare <8 x i16> @llvm.mips.subsus.u.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_subsus_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_subsus_u_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_subsus_u_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_subsus_u_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_subsus_u_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.subsus.u.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_subsus_u_w_RES
   ret void
@@ -1084,8 +1084,8 @@ declare <4 x i32> @llvm.mips.subsus.u.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_subsus_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_subsus_u_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_subsus_u_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_subsus_u_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_subsus_u_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.subsus.u.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_subsus_u_d_RES
   ret void
@@ -1108,8 +1108,8 @@ declare <2 x i64> @llvm.mips.subsus.u.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_subsuu_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_subsuu_s_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_subsuu_s_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_subsuu_s_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_subsuu_s_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.subsuu.s.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_subsuu_s_b_RES
   ret void
@@ -1132,8 +1132,8 @@ declare <16 x i8> @llvm.mips.subsuu.s.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_subsuu_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_subsuu_s_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_subsuu_s_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_subsuu_s_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_subsuu_s_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.subsuu.s.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_subsuu_s_h_RES
   ret void
@@ -1156,8 +1156,8 @@ declare <8 x i16> @llvm.mips.subsuu.s.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_subsuu_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_subsuu_s_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_subsuu_s_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_subsuu_s_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_subsuu_s_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.subsuu.s.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_subsuu_s_w_RES
   ret void
@@ -1180,8 +1180,8 @@ declare <4 x i32> @llvm.mips.subsuu.s.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_subsuu_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_subsuu_s_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_subsuu_s_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_subsuu_s_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_subsuu_s_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.subsuu.s.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_subsuu_s_d_RES
   ret void
@@ -1204,8 +1204,8 @@ declare <2 x i64> @llvm.mips.subsuu.s.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_subv_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_subv_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_subv_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_subv_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_subv_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.subv.b(<16 x i8> %0, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_subv_b_RES
   ret void
@@ -1228,8 +1228,8 @@ declare <16 x i8> @llvm.mips.subv.b(<16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_subv_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_subv_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_subv_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_subv_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_subv_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.subv.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_subv_h_RES
   ret void
@@ -1252,8 +1252,8 @@ declare <8 x i16> @llvm.mips.subv.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_subv_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_subv_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_subv_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_subv_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_subv_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.subv.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_subv_w_RES
   ret void
@@ -1276,8 +1276,8 @@ declare <4 x i32> @llvm.mips.subv.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_subv_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_subv_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_subv_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_subv_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_subv_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.subv.d(<2 x i64> %0, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_subv_d_RES
   ret void
@@ -1297,8 +1297,8 @@ declare <2 x i64> @llvm.mips.subv.d(<2 x i64>, <2 x i64>) nounwind
 
 define void @subv_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_subv_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_subv_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_subv_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_subv_b_ARG2
   %2 = sub <16 x i8> %0, %1
   store <16 x i8> %2, <16 x i8>* @llvm_mips_subv_b_RES
   ret void
@@ -1315,8 +1315,8 @@ entry:
 
 define void @subv_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_subv_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_subv_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_subv_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_subv_h_ARG2
   %2 = sub <8 x i16> %0, %1
   store <8 x i16> %2, <8 x i16>* @llvm_mips_subv_h_RES
   ret void
@@ -1333,8 +1333,8 @@ entry:
 
 define void @subv_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_subv_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_subv_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_subv_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_subv_w_ARG2
   %2 = sub <4 x i32> %0, %1
   store <4 x i32> %2, <4 x i32>* @llvm_mips_subv_w_RES
   ret void
@@ -1351,8 +1351,8 @@ entry:
 
 define void @subv_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_subv_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_subv_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_subv_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_subv_d_ARG2
   %2 = sub <2 x i64> %0, %1
   store <2 x i64> %2, <2 x i64>* @llvm_mips_subv_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3r-v.ll b/test/CodeGen/Mips/msa/3r-v.ll
index c9693f90d556..2d36da40d2b1 100644
--- a/test/CodeGen/Mips/msa/3r-v.ll
+++ b/test/CodeGen/Mips/msa/3r-v.ll
@@ -11,9 +11,9 @@
 
 define void @llvm_mips_vshf_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_vshf_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_vshf_b_ARG2
-  %2 = load <16 x i8>* @llvm_mips_vshf_b_ARG3
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_vshf_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_vshf_b_ARG2
+  %2 = load <16 x i8>, <16 x i8>* @llvm_mips_vshf_b_ARG3
   %3 = tail call <16 x i8> @llvm.mips.vshf.b(<16 x i8> %0, <16 x i8> %1, <16 x i8> %2)
   store <16 x i8> %3, <16 x i8>* @llvm_mips_vshf_b_RES
   ret void
@@ -36,9 +36,9 @@ declare <16 x i8> @llvm.mips.vshf.b(<16 x i8>, <16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_vshf_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_vshf_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_vshf_h_ARG2
-  %2 = load <8 x i16>* @llvm_mips_vshf_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_vshf_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_vshf_h_ARG2
+  %2 = load <8 x i16>, <8 x i16>* @llvm_mips_vshf_h_ARG3
   %3 = tail call <8 x i16> @llvm.mips.vshf.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2)
   store <8 x i16> %3, <8 x i16>* @llvm_mips_vshf_h_RES
   ret void
@@ -61,9 +61,9 @@ declare <8 x i16> @llvm.mips.vshf.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_vshf_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_vshf_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_vshf_w_ARG2
-  %2 = load <4 x i32>* @llvm_mips_vshf_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_vshf_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_vshf_w_ARG2
+  %2 = load <4 x i32>, <4 x i32>* @llvm_mips_vshf_w_ARG3
   %3 = tail call <4 x i32> @llvm.mips.vshf.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2)
   store <4 x i32> %3, <4 x i32>* @llvm_mips_vshf_w_RES
   ret void
@@ -86,9 +86,9 @@ declare <4 x i32> @llvm.mips.vshf.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_vshf_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_vshf_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_vshf_d_ARG2
-  %2 = load <2 x i64>* @llvm_mips_vshf_d_ARG3
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_vshf_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_vshf_d_ARG2
+  %2 = load <2 x i64>, <2 x i64>* @llvm_mips_vshf_d_ARG3
   %3 = tail call <2 x i64> @llvm.mips.vshf.d(<2 x i64> %0, <2 x i64> %1, <2 x i64> %2)
   store <2 x i64> %3, <2 x i64>* @llvm_mips_vshf_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3r_4r.ll b/test/CodeGen/Mips/msa/3r_4r.ll
index b7fd7283788c..73d104c68c53 100644
--- a/test/CodeGen/Mips/msa/3r_4r.ll
+++ b/test/CodeGen/Mips/msa/3r_4r.ll
@@ -11,9 +11,9 @@
 
 define void @llvm_mips_maddv_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_maddv_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_maddv_b_ARG2
-  %2 = load <16 x i8>* @llvm_mips_maddv_b_ARG3
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_maddv_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_maddv_b_ARG2
+  %2 = load <16 x i8>, <16 x i8>* @llvm_mips_maddv_b_ARG3
   %3 = tail call <16 x i8> @llvm.mips.maddv.b(<16 x i8> %0, <16 x i8> %1, <16 x i8> %2)
   store <16 x i8> %3, <16 x i8>* @llvm_mips_maddv_b_RES
   ret void
@@ -36,9 +36,9 @@ declare <16 x i8> @llvm.mips.maddv.b(<16 x i8>, <16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_maddv_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_maddv_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_maddv_h_ARG2
-  %2 = load <8 x i16>* @llvm_mips_maddv_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_maddv_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_maddv_h_ARG2
+  %2 = load <8 x i16>, <8 x i16>* @llvm_mips_maddv_h_ARG3
   %3 = tail call <8 x i16> @llvm.mips.maddv.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2)
   store <8 x i16> %3, <8 x i16>* @llvm_mips_maddv_h_RES
   ret void
@@ -61,9 +61,9 @@ declare <8 x i16> @llvm.mips.maddv.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_maddv_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_maddv_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_maddv_w_ARG2
-  %2 = load <4 x i32>* @llvm_mips_maddv_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_maddv_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_maddv_w_ARG2
+  %2 = load <4 x i32>, <4 x i32>* @llvm_mips_maddv_w_ARG3
   %3 = tail call <4 x i32> @llvm.mips.maddv.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2)
   store <4 x i32> %3, <4 x i32>* @llvm_mips_maddv_w_RES
   ret void
@@ -86,9 +86,9 @@ declare <4 x i32> @llvm.mips.maddv.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_maddv_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_maddv_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_maddv_d_ARG2
-  %2 = load <2 x i64>* @llvm_mips_maddv_d_ARG3
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_maddv_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_maddv_d_ARG2
+  %2 = load <2 x i64>, <2 x i64>* @llvm_mips_maddv_d_ARG3
   %3 = tail call <2 x i64> @llvm.mips.maddv.d(<2 x i64> %0, <2 x i64> %1, <2 x i64> %2)
   store <2 x i64> %3, <2 x i64>* @llvm_mips_maddv_d_RES
   ret void
@@ -111,9 +111,9 @@ declare <2 x i64> @llvm.mips.maddv.d(<2 x i64>, <2 x i64>, <2 x i64>) nounwind
 
 define void @llvm_mips_msubv_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_msubv_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_msubv_b_ARG2
-  %2 = load <16 x i8>* @llvm_mips_msubv_b_ARG3
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_msubv_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_msubv_b_ARG2
+  %2 = load <16 x i8>, <16 x i8>* @llvm_mips_msubv_b_ARG3
   %3 = tail call <16 x i8> @llvm.mips.msubv.b(<16 x i8> %0, <16 x i8> %1, <16 x i8> %2)
   store <16 x i8> %3, <16 x i8>* @llvm_mips_msubv_b_RES
   ret void
@@ -136,9 +136,9 @@ declare <16 x i8> @llvm.mips.msubv.b(<16 x i8>, <16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_msubv_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_msubv_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_msubv_h_ARG2
-  %2 = load <8 x i16>* @llvm_mips_msubv_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_msubv_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_msubv_h_ARG2
+  %2 = load <8 x i16>, <8 x i16>* @llvm_mips_msubv_h_ARG3
   %3 = tail call <8 x i16> @llvm.mips.msubv.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2)
   store <8 x i16> %3, <8 x i16>* @llvm_mips_msubv_h_RES
   ret void
@@ -161,9 +161,9 @@ declare <8 x i16> @llvm.mips.msubv.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_msubv_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_msubv_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_msubv_w_ARG2
-  %2 = load <4 x i32>* @llvm_mips_msubv_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_msubv_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_msubv_w_ARG2
+  %2 = load <4 x i32>, <4 x i32>* @llvm_mips_msubv_w_ARG3
   %3 = tail call <4 x i32> @llvm.mips.msubv.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2)
   store <4 x i32> %3, <4 x i32>* @llvm_mips_msubv_w_RES
   ret void
@@ -186,9 +186,9 @@ declare <4 x i32> @llvm.mips.msubv.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_msubv_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_msubv_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_msubv_d_ARG2
-  %2 = load <2 x i64>* @llvm_mips_msubv_d_ARG3
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_msubv_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_msubv_d_ARG2
+  %2 = load <2 x i64>, <2 x i64>* @llvm_mips_msubv_d_ARG3
   %3 = tail call <2 x i64> @llvm.mips.msubv.d(<2 x i64> %0, <2 x i64> %1, <2 x i64> %2)
   store <2 x i64> %3, <2 x i64>* @llvm_mips_msubv_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3r_4r_widen.ll b/test/CodeGen/Mips/msa/3r_4r_widen.ll
index 7063e4566a78..fe248eeb566b 100644
--- a/test/CodeGen/Mips/msa/3r_4r_widen.ll
+++ b/test/CodeGen/Mips/msa/3r_4r_widen.ll
@@ -12,9 +12,9 @@
 
 define void @llvm_mips_dpadd_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_dpadd_s_h_ARG1
-  %1 = load <16 x i8>* @llvm_mips_dpadd_s_h_ARG2
-  %2 = load <16 x i8>* @llvm_mips_dpadd_s_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_dpadd_s_h_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_dpadd_s_h_ARG2
+  %2 = load <16 x i8>, <16 x i8>* @llvm_mips_dpadd_s_h_ARG3
   %3 = tail call <8 x i16> @llvm.mips.dpadd.s.h(<8 x i16> %0, <16 x i8> %1, <16 x i8> %2)
   store <8 x i16> %3, <8 x i16>* @llvm_mips_dpadd_s_h_RES
   ret void
@@ -37,9 +37,9 @@ declare <8 x i16> @llvm.mips.dpadd.s.h(<8 x i16>, <16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_dpadd_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_dpadd_s_w_ARG1
-  %1 = load <8 x i16>* @llvm_mips_dpadd_s_w_ARG2
-  %2 = load <8 x i16>* @llvm_mips_dpadd_s_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_dpadd_s_w_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_dpadd_s_w_ARG2
+  %2 = load <8 x i16>, <8 x i16>* @llvm_mips_dpadd_s_w_ARG3
   %3 = tail call <4 x i32> @llvm.mips.dpadd.s.w(<4 x i32> %0, <8 x i16> %1, <8 x i16> %2)
   store <4 x i32> %3, <4 x i32>* @llvm_mips_dpadd_s_w_RES
   ret void
@@ -62,9 +62,9 @@ declare <4 x i32> @llvm.mips.dpadd.s.w(<4 x i32>, <8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_dpadd_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_dpadd_s_d_ARG1
-  %1 = load <4 x i32>* @llvm_mips_dpadd_s_d_ARG2
-  %2 = load <4 x i32>* @llvm_mips_dpadd_s_d_ARG3
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_dpadd_s_d_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_dpadd_s_d_ARG2
+  %2 = load <4 x i32>, <4 x i32>* @llvm_mips_dpadd_s_d_ARG3
   %3 = tail call <2 x i64> @llvm.mips.dpadd.s.d(<2 x i64> %0, <4 x i32> %1, <4 x i32> %2)
   store <2 x i64> %3, <2 x i64>* @llvm_mips_dpadd_s_d_RES
   ret void
@@ -87,9 +87,9 @@ declare <2 x i64> @llvm.mips.dpadd.s.d(<2 x i64>, <4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_dpadd_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_dpadd_u_h_ARG1
-  %1 = load <16 x i8>* @llvm_mips_dpadd_u_h_ARG2
-  %2 = load <16 x i8>* @llvm_mips_dpadd_u_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_dpadd_u_h_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_dpadd_u_h_ARG2
+  %2 = load <16 x i8>, <16 x i8>* @llvm_mips_dpadd_u_h_ARG3
   %3 = tail call <8 x i16> @llvm.mips.dpadd.u.h(<8 x i16> %0, <16 x i8> %1, <16 x i8> %2)
   store <8 x i16> %3, <8 x i16>* @llvm_mips_dpadd_u_h_RES
   ret void
@@ -112,9 +112,9 @@ declare <8 x i16> @llvm.mips.dpadd.u.h(<8 x i16>, <16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_dpadd_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_dpadd_u_w_ARG1
-  %1 = load <8 x i16>* @llvm_mips_dpadd_u_w_ARG2
-  %2 = load <8 x i16>* @llvm_mips_dpadd_u_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_dpadd_u_w_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_dpadd_u_w_ARG2
+  %2 = load <8 x i16>, <8 x i16>* @llvm_mips_dpadd_u_w_ARG3
   %3 = tail call <4 x i32> @llvm.mips.dpadd.u.w(<4 x i32> %0, <8 x i16> %1, <8 x i16> %2)
   store <4 x i32> %3, <4 x i32>* @llvm_mips_dpadd_u_w_RES
   ret void
@@ -137,9 +137,9 @@ declare <4 x i32> @llvm.mips.dpadd.u.w(<4 x i32>, <8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_dpadd_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_dpadd_u_d_ARG1
-  %1 = load <4 x i32>* @llvm_mips_dpadd_u_d_ARG2
-  %2 = load <4 x i32>* @llvm_mips_dpadd_u_d_ARG3
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_dpadd_u_d_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_dpadd_u_d_ARG2
+  %2 = load <4 x i32>, <4 x i32>* @llvm_mips_dpadd_u_d_ARG3
   %3 = tail call <2 x i64> @llvm.mips.dpadd.u.d(<2 x i64> %0, <4 x i32> %1, <4 x i32> %2)
   store <2 x i64> %3, <2 x i64>* @llvm_mips_dpadd_u_d_RES
   ret void
@@ -162,9 +162,9 @@ declare <2 x i64> @llvm.mips.dpadd.u.d(<2 x i64>, <4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_dpsub_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_dpsub_s_h_ARG1
-  %1 = load <16 x i8>* @llvm_mips_dpsub_s_h_ARG2
-  %2 = load <16 x i8>* @llvm_mips_dpsub_s_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_dpsub_s_h_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_dpsub_s_h_ARG2
+  %2 = load <16 x i8>, <16 x i8>* @llvm_mips_dpsub_s_h_ARG3
   %3 = tail call <8 x i16> @llvm.mips.dpsub.s.h(<8 x i16> %0, <16 x i8> %1, <16 x i8> %2)
   store <8 x i16> %3, <8 x i16>* @llvm_mips_dpsub_s_h_RES
   ret void
@@ -187,9 +187,9 @@ declare <8 x i16> @llvm.mips.dpsub.s.h(<8 x i16>, <16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_dpsub_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_dpsub_s_w_ARG1
-  %1 = load <8 x i16>* @llvm_mips_dpsub_s_w_ARG2
-  %2 = load <8 x i16>* @llvm_mips_dpsub_s_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_dpsub_s_w_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_dpsub_s_w_ARG2
+  %2 = load <8 x i16>, <8 x i16>* @llvm_mips_dpsub_s_w_ARG3
   %3 = tail call <4 x i32> @llvm.mips.dpsub.s.w(<4 x i32> %0, <8 x i16> %1, <8 x i16> %2)
   store <4 x i32> %3, <4 x i32>* @llvm_mips_dpsub_s_w_RES
   ret void
@@ -212,9 +212,9 @@ declare <4 x i32> @llvm.mips.dpsub.s.w(<4 x i32>, <8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_dpsub_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_dpsub_s_d_ARG1
-  %1 = load <4 x i32>* @llvm_mips_dpsub_s_d_ARG2
-  %2 = load <4 x i32>* @llvm_mips_dpsub_s_d_ARG3
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_dpsub_s_d_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_dpsub_s_d_ARG2
+  %2 = load <4 x i32>, <4 x i32>* @llvm_mips_dpsub_s_d_ARG3
   %3 = tail call <2 x i64> @llvm.mips.dpsub.s.d(<2 x i64> %0, <4 x i32> %1, <4 x i32> %2)
   store <2 x i64> %3, <2 x i64>* @llvm_mips_dpsub_s_d_RES
   ret void
@@ -237,9 +237,9 @@ declare <2 x i64> @llvm.mips.dpsub.s.d(<2 x i64>, <4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_dpsub_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_dpsub_u_h_ARG1
-  %1 = load <16 x i8>* @llvm_mips_dpsub_u_h_ARG2
-  %2 = load <16 x i8>* @llvm_mips_dpsub_u_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_dpsub_u_h_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_dpsub_u_h_ARG2
+  %2 = load <16 x i8>, <16 x i8>* @llvm_mips_dpsub_u_h_ARG3
   %3 = tail call <8 x i16> @llvm.mips.dpsub.u.h(<8 x i16> %0, <16 x i8> %1, <16 x i8> %2)
   store <8 x i16> %3, <8 x i16>* @llvm_mips_dpsub_u_h_RES
   ret void
@@ -262,9 +262,9 @@ declare <8 x i16> @llvm.mips.dpsub.u.h(<8 x i16>, <16 x i8>, <16 x i8>) nounwind
 
 define void @llvm_mips_dpsub_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_dpsub_u_w_ARG1
-  %1 = load <8 x i16>* @llvm_mips_dpsub_u_w_ARG2
-  %2 = load <8 x i16>* @llvm_mips_dpsub_u_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_dpsub_u_w_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_dpsub_u_w_ARG2
+  %2 = load <8 x i16>, <8 x i16>* @llvm_mips_dpsub_u_w_ARG3
   %3 = tail call <4 x i32> @llvm.mips.dpsub.u.w(<4 x i32> %0, <8 x i16> %1, <8 x i16> %2)
   store <4 x i32> %3, <4 x i32>* @llvm_mips_dpsub_u_w_RES
   ret void
@@ -287,9 +287,9 @@ declare <4 x i32> @llvm.mips.dpsub.u.w(<4 x i32>, <8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_dpsub_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_dpsub_u_d_ARG1
-  %1 = load <4 x i32>* @llvm_mips_dpsub_u_d_ARG2
-  %2 = load <4 x i32>* @llvm_mips_dpsub_u_d_ARG3
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_dpsub_u_d_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_dpsub_u_d_ARG2
+  %2 = load <4 x i32>, <4 x i32>* @llvm_mips_dpsub_u_d_ARG3
   %3 = tail call <2 x i64> @llvm.mips.dpsub.u.d(<2 x i64> %0, <4 x i32> %1, <4 x i32> %2)
   store <2 x i64> %3, <2 x i64>* @llvm_mips_dpsub_u_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3r_splat.ll b/test/CodeGen/Mips/msa/3r_splat.ll
index 6b0cb26f8c81..56d26b030de9 100644
--- a/test/CodeGen/Mips/msa/3r_splat.ll
+++ b/test/CodeGen/Mips/msa/3r_splat.ll
@@ -11,7 +11,7 @@
 
 define void @llvm_mips_splat_b_test(i32 %a) nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_splat_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_splat_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.splat.b(<16 x i8> %0, i32 %a)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_splat_b_RES
   ret void
@@ -32,7 +32,7 @@ declare <16 x i8> @llvm.mips.splat.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_splat_h_test(i32 %a) nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_splat_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_splat_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.splat.h(<8 x i16> %0, i32 %a)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_splat_h_RES
   ret void
@@ -53,7 +53,7 @@ declare <8 x i16> @llvm.mips.splat.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_splat_w_test(i32 %a) nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_splat_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_splat_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.splat.w(<4 x i32> %0, i32 %a)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_splat_w_RES
   ret void
@@ -74,7 +74,7 @@ declare <4 x i32> @llvm.mips.splat.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_splat_d_test(i32 %a) nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_splat_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_splat_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.splat.d(<2 x i64> %0, i32 %a)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_splat_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3rf.ll b/test/CodeGen/Mips/msa/3rf.ll
index ae665afcc950..dce0c275e8da 100644
--- a/test/CodeGen/Mips/msa/3rf.ll
+++ b/test/CodeGen/Mips/msa/3rf.ll
@@ -9,8 +9,8 @@
 
 define void @llvm_mips_fadd_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fadd_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fadd_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fadd_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fadd_w_ARG2
   %2 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %0, <4 x float> %1)
   store <4 x float> %2, <4 x float>* @llvm_mips_fadd_w_RES
   ret void
@@ -31,8 +31,8 @@ declare <4 x float> @llvm.mips.fadd.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fadd_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fadd_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fadd_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fadd_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fadd_d_ARG2
   %2 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %0, <2 x double> %1)
   store <2 x double> %2, <2 x double>* @llvm_mips_fadd_d_RES
   ret void
@@ -49,8 +49,8 @@ declare <2 x double> @llvm.mips.fadd.d(<2 x double>, <2 x double>) nounwind
 
 define void @fadd_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fadd_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fadd_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fadd_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fadd_w_ARG2
   %2 = fadd <4 x float> %0, %1
   store <4 x float> %2, <4 x float>* @llvm_mips_fadd_w_RES
   ret void
@@ -65,8 +65,8 @@ entry:
 
 define void @fadd_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fadd_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fadd_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fadd_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fadd_d_ARG2
   %2 = fadd <2 x double> %0, %1
   store <2 x double> %2, <2 x double>* @llvm_mips_fadd_d_RES
   ret void
@@ -85,8 +85,8 @@ entry:
 
 define void @llvm_mips_fdiv_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fdiv_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fdiv_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fdiv_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fdiv_w_ARG2
   %2 = tail call <4 x float> @llvm.mips.fdiv.w(<4 x float> %0, <4 x float> %1)
   store <4 x float> %2, <4 x float>* @llvm_mips_fdiv_w_RES
   ret void
@@ -107,8 +107,8 @@ declare <4 x float> @llvm.mips.fdiv.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fdiv_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fdiv_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fdiv_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fdiv_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fdiv_d_ARG2
   %2 = tail call <2 x double> @llvm.mips.fdiv.d(<2 x double> %0, <2 x double> %1)
   store <2 x double> %2, <2 x double>* @llvm_mips_fdiv_d_RES
   ret void
@@ -125,8 +125,8 @@ declare <2 x double> @llvm.mips.fdiv.d(<2 x double>, <2 x double>) nounwind
 
 define void @fdiv_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fdiv_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fdiv_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fdiv_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fdiv_w_ARG2
   %2 = fdiv <4 x float> %0, %1
   store <4 x float> %2, <4 x float>* @llvm_mips_fdiv_w_RES
   ret void
@@ -141,8 +141,8 @@ entry:
 
 define void @fdiv_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fdiv_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fdiv_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fdiv_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fdiv_d_ARG2
   %2 = fdiv <2 x double> %0, %1
   store <2 x double> %2, <2 x double>* @llvm_mips_fdiv_d_RES
   ret void
@@ -161,8 +161,8 @@ entry:
 
 define void @llvm_mips_fmin_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fmin_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fmin_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fmin_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fmin_w_ARG2
   %2 = tail call <4 x float> @llvm.mips.fmin.w(<4 x float> %0, <4 x float> %1)
   store <4 x float> %2, <4 x float>* @llvm_mips_fmin_w_RES
   ret void
@@ -183,8 +183,8 @@ declare <4 x float> @llvm.mips.fmin.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fmin_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fmin_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fmin_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fmin_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fmin_d_ARG2
   %2 = tail call <2 x double> @llvm.mips.fmin.d(<2 x double> %0, <2 x double> %1)
   store <2 x double> %2, <2 x double>* @llvm_mips_fmin_d_RES
   ret void
@@ -205,8 +205,8 @@ declare <2 x double> @llvm.mips.fmin.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fmin_a_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fmin_a_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fmin_a_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fmin_a_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fmin_a_w_ARG2
   %2 = tail call <4 x float> @llvm.mips.fmin.a.w(<4 x float> %0, <4 x float> %1)
   store <4 x float> %2, <4 x float>* @llvm_mips_fmin_a_w_RES
   ret void
@@ -227,8 +227,8 @@ declare <4 x float> @llvm.mips.fmin.a.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fmin_a_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fmin_a_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fmin_a_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fmin_a_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fmin_a_d_ARG2
   %2 = tail call <2 x double> @llvm.mips.fmin.a.d(<2 x double> %0, <2 x double> %1)
   store <2 x double> %2, <2 x double>* @llvm_mips_fmin_a_d_RES
   ret void
@@ -249,8 +249,8 @@ declare <2 x double> @llvm.mips.fmin.a.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fmax_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fmax_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fmax_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fmax_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fmax_w_ARG2
   %2 = tail call <4 x float> @llvm.mips.fmax.w(<4 x float> %0, <4 x float> %1)
   store <4 x float> %2, <4 x float>* @llvm_mips_fmax_w_RES
   ret void
@@ -271,8 +271,8 @@ declare <4 x float> @llvm.mips.fmax.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fmax_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fmax_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fmax_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fmax_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fmax_d_ARG2
   %2 = tail call <2 x double> @llvm.mips.fmax.d(<2 x double> %0, <2 x double> %1)
   store <2 x double> %2, <2 x double>* @llvm_mips_fmax_d_RES
   ret void
@@ -293,8 +293,8 @@ declare <2 x double> @llvm.mips.fmax.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fmax_a_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fmax_a_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fmax_a_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fmax_a_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fmax_a_w_ARG2
   %2 = tail call <4 x float> @llvm.mips.fmax.a.w(<4 x float> %0, <4 x float> %1)
   store <4 x float> %2, <4 x float>* @llvm_mips_fmax_a_w_RES
   ret void
@@ -315,8 +315,8 @@ declare <4 x float> @llvm.mips.fmax.a.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fmax_a_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fmax_a_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fmax_a_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fmax_a_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fmax_a_d_ARG2
   %2 = tail call <2 x double> @llvm.mips.fmax.a.d(<2 x double> %0, <2 x double> %1)
   store <2 x double> %2, <2 x double>* @llvm_mips_fmax_a_d_RES
   ret void
@@ -337,8 +337,8 @@ declare <2 x double> @llvm.mips.fmax.a.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fmul_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fmul_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fmul_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fmul_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fmul_w_ARG2
   %2 = tail call <4 x float> @llvm.mips.fmul.w(<4 x float> %0, <4 x float> %1)
   store <4 x float> %2, <4 x float>* @llvm_mips_fmul_w_RES
   ret void
@@ -359,8 +359,8 @@ declare <4 x float> @llvm.mips.fmul.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fmul_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fmul_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fmul_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fmul_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fmul_d_ARG2
   %2 = tail call <2 x double> @llvm.mips.fmul.d(<2 x double> %0, <2 x double> %1)
   store <2 x double> %2, <2 x double>* @llvm_mips_fmul_d_RES
   ret void
@@ -377,8 +377,8 @@ declare <2 x double> @llvm.mips.fmul.d(<2 x double>, <2 x double>) nounwind
 
 define void @fmul_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fmul_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fmul_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fmul_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fmul_w_ARG2
   %2 = fmul <4 x float> %0, %1
   store <4 x float> %2, <4 x float>* @llvm_mips_fmul_w_RES
   ret void
@@ -393,8 +393,8 @@ entry:
 
 define void @fmul_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fmul_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fmul_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fmul_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fmul_d_ARG2
   %2 = fmul <2 x double> %0, %1
   store <2 x double> %2, <2 x double>* @llvm_mips_fmul_d_RES
   ret void
@@ -413,8 +413,8 @@ entry:
 
 define void @llvm_mips_fsub_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fsub_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fsub_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fsub_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fsub_w_ARG2
   %2 = tail call <4 x float> @llvm.mips.fsub.w(<4 x float> %0, <4 x float> %1)
   store <4 x float> %2, <4 x float>* @llvm_mips_fsub_w_RES
   ret void
@@ -435,8 +435,8 @@ declare <4 x float> @llvm.mips.fsub.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fsub_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fsub_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fsub_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fsub_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fsub_d_ARG2
   %2 = tail call <2 x double> @llvm.mips.fsub.d(<2 x double> %0, <2 x double> %1)
   store <2 x double> %2, <2 x double>* @llvm_mips_fsub_d_RES
   ret void
@@ -454,8 +454,8 @@ declare <2 x double> @llvm.mips.fsub.d(<2 x double>, <2 x double>) nounwind
 
 define void @fsub_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fsub_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fsub_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fsub_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fsub_w_ARG2
   %2 = fsub <4 x float> %0, %1
   store <4 x float> %2, <4 x float>* @llvm_mips_fsub_w_RES
   ret void
@@ -470,8 +470,8 @@ entry:
 
 define void @fsub_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fsub_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fsub_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fsub_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fsub_d_ARG2
   %2 = fsub <2 x double> %0, %1
   store <2 x double> %2, <2 x double>* @llvm_mips_fsub_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3rf_4rf.ll b/test/CodeGen/Mips/msa/3rf_4rf.ll
index 67ef7fd2bae1..f1a3002e8179 100644
--- a/test/CodeGen/Mips/msa/3rf_4rf.ll
+++ b/test/CodeGen/Mips/msa/3rf_4rf.ll
@@ -11,9 +11,9 @@
 
 define void @llvm_mips_fmadd_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fmadd_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fmadd_w_ARG2
-  %2 = load <4 x float>* @llvm_mips_fmadd_w_ARG3
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fmadd_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fmadd_w_ARG2
+  %2 = load <4 x float>, <4 x float>* @llvm_mips_fmadd_w_ARG3
   %3 = tail call <4 x float> @llvm.mips.fmadd.w(<4 x float> %0, <4 x float> %1, <4 x float> %2)
   store <4 x float> %3, <4 x float>* @llvm_mips_fmadd_w_RES
   ret void
@@ -36,9 +36,9 @@ declare <4 x float> @llvm.mips.fmadd.w(<4 x float>, <4 x float>, <4 x float>) no
 
 define void @llvm_mips_fmadd_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fmadd_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fmadd_d_ARG2
-  %2 = load <2 x double>* @llvm_mips_fmadd_d_ARG3
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fmadd_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fmadd_d_ARG2
+  %2 = load <2 x double>, <2 x double>* @llvm_mips_fmadd_d_ARG3
   %3 = tail call <2 x double> @llvm.mips.fmadd.d(<2 x double> %0, <2 x double> %1, <2 x double> %2)
   store <2 x double> %3, <2 x double>* @llvm_mips_fmadd_d_RES
   ret void
@@ -61,9 +61,9 @@ declare <2 x double> @llvm.mips.fmadd.d(<2 x double>, <2 x double>, <2 x double>
 
 define void @llvm_mips_fmsub_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fmsub_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fmsub_w_ARG2
-  %2 = load <4 x float>* @llvm_mips_fmsub_w_ARG3
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fmsub_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fmsub_w_ARG2
+  %2 = load <4 x float>, <4 x float>* @llvm_mips_fmsub_w_ARG3
   %3 = tail call <4 x float> @llvm.mips.fmsub.w(<4 x float> %0, <4 x float> %1, <4 x float> %2)
   store <4 x float> %3, <4 x float>* @llvm_mips_fmsub_w_RES
   ret void
@@ -86,9 +86,9 @@ declare <4 x float> @llvm.mips.fmsub.w(<4 x float>, <4 x float>, <4 x float>) no
 
 define void @llvm_mips_fmsub_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fmsub_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fmsub_d_ARG2
-  %2 = load <2 x double>* @llvm_mips_fmsub_d_ARG3
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fmsub_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fmsub_d_ARG2
+  %2 = load <2 x double>, <2 x double>* @llvm_mips_fmsub_d_ARG3
   %3 = tail call <2 x double> @llvm.mips.fmsub.d(<2 x double> %0, <2 x double> %1, <2 x double> %2)
   store <2 x double> %3, <2 x double>* @llvm_mips_fmsub_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3rf_4rf_q.ll b/test/CodeGen/Mips/msa/3rf_4rf_q.ll
index de28be0b1c22..704c4b7e7cbc 100644
--- a/test/CodeGen/Mips/msa/3rf_4rf_q.ll
+++ b/test/CodeGen/Mips/msa/3rf_4rf_q.ll
@@ -11,9 +11,9 @@
 
 define void @llvm_mips_madd_q_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_madd_q_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_madd_q_h_ARG2
-  %2 = load <8 x i16>* @llvm_mips_madd_q_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_madd_q_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_madd_q_h_ARG2
+  %2 = load <8 x i16>, <8 x i16>* @llvm_mips_madd_q_h_ARG3
   %3 = tail call <8 x i16> @llvm.mips.madd.q.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2)
   store <8 x i16> %3, <8 x i16>* @llvm_mips_madd_q_h_RES
   ret void
@@ -36,9 +36,9 @@ declare <8 x i16> @llvm.mips.madd.q.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_madd_q_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_madd_q_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_madd_q_w_ARG2
-  %2 = load <4 x i32>* @llvm_mips_madd_q_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_madd_q_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_madd_q_w_ARG2
+  %2 = load <4 x i32>, <4 x i32>* @llvm_mips_madd_q_w_ARG3
   %3 = tail call <4 x i32> @llvm.mips.madd.q.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2)
   store <4 x i32> %3, <4 x i32>* @llvm_mips_madd_q_w_RES
   ret void
@@ -61,9 +61,9 @@ declare <4 x i32> @llvm.mips.madd.q.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_maddr_q_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_maddr_q_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_maddr_q_h_ARG2
-  %2 = load <8 x i16>* @llvm_mips_maddr_q_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_maddr_q_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_maddr_q_h_ARG2
+  %2 = load <8 x i16>, <8 x i16>* @llvm_mips_maddr_q_h_ARG3
   %3 = tail call <8 x i16> @llvm.mips.maddr.q.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2)
   store <8 x i16> %3, <8 x i16>* @llvm_mips_maddr_q_h_RES
   ret void
@@ -86,9 +86,9 @@ declare <8 x i16> @llvm.mips.maddr.q.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_maddr_q_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_maddr_q_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_maddr_q_w_ARG2
-  %2 = load <4 x i32>* @llvm_mips_maddr_q_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_maddr_q_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_maddr_q_w_ARG2
+  %2 = load <4 x i32>, <4 x i32>* @llvm_mips_maddr_q_w_ARG3
   %3 = tail call <4 x i32> @llvm.mips.maddr.q.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2)
   store <4 x i32> %3, <4 x i32>* @llvm_mips_maddr_q_w_RES
   ret void
@@ -111,9 +111,9 @@ declare <4 x i32> @llvm.mips.maddr.q.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_msub_q_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_msub_q_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_msub_q_h_ARG2
-  %2 = load <8 x i16>* @llvm_mips_msub_q_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_msub_q_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_msub_q_h_ARG2
+  %2 = load <8 x i16>, <8 x i16>* @llvm_mips_msub_q_h_ARG3
   %3 = tail call <8 x i16> @llvm.mips.msub.q.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2)
   store <8 x i16> %3, <8 x i16>* @llvm_mips_msub_q_h_RES
   ret void
@@ -136,9 +136,9 @@ declare <8 x i16> @llvm.mips.msub.q.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_msub_q_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_msub_q_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_msub_q_w_ARG2
-  %2 = load <4 x i32>* @llvm_mips_msub_q_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_msub_q_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_msub_q_w_ARG2
+  %2 = load <4 x i32>, <4 x i32>* @llvm_mips_msub_q_w_ARG3
   %3 = tail call <4 x i32> @llvm.mips.msub.q.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2)
   store <4 x i32> %3, <4 x i32>* @llvm_mips_msub_q_w_RES
   ret void
@@ -161,9 +161,9 @@ declare <4 x i32> @llvm.mips.msub.q.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_msubr_q_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_msubr_q_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_msubr_q_h_ARG2
-  %2 = load <8 x i16>* @llvm_mips_msubr_q_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_msubr_q_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_msubr_q_h_ARG2
+  %2 = load <8 x i16>, <8 x i16>* @llvm_mips_msubr_q_h_ARG3
   %3 = tail call <8 x i16> @llvm.mips.msubr.q.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2)
   store <8 x i16> %3, <8 x i16>* @llvm_mips_msubr_q_h_RES
   ret void
@@ -186,9 +186,9 @@ declare <8 x i16> @llvm.mips.msubr.q.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_msubr_q_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_msubr_q_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_msubr_q_w_ARG2
-  %2 = load <4 x i32>* @llvm_mips_msubr_q_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_msubr_q_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_msubr_q_w_ARG2
+  %2 = load <4 x i32>, <4 x i32>* @llvm_mips_msubr_q_w_ARG3
   %3 = tail call <4 x i32> @llvm.mips.msubr.q.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2)
   store <4 x i32> %3, <4 x i32>* @llvm_mips_msubr_q_w_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3rf_exdo.ll b/test/CodeGen/Mips/msa/3rf_exdo.ll
index 8a7f268a5069..1b1b2e9243ec 100644
--- a/test/CodeGen/Mips/msa/3rf_exdo.ll
+++ b/test/CodeGen/Mips/msa/3rf_exdo.ll
@@ -10,8 +10,8 @@
 
 define void @llvm_mips_fexdo_h_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fexdo_h_ARG1
-  %1 = load <4 x float>* @llvm_mips_fexdo_h_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fexdo_h_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fexdo_h_ARG2
   %2 = tail call <8 x half> @llvm.mips.fexdo.h(<4 x float> %0, <4 x float> %1)
   store <8 x half> %2, <8 x half>* @llvm_mips_fexdo_h_RES
   ret void
@@ -32,8 +32,8 @@ declare <8 x half> @llvm.mips.fexdo.h(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fexdo_w_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fexdo_w_ARG1
-  %1 = load <2 x double>* @llvm_mips_fexdo_w_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fexdo_w_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fexdo_w_ARG2
   %2 = tail call <4 x float> @llvm.mips.fexdo.w(<2 x double> %0, <2 x double> %1)
   store <4 x float> %2, <4 x float>* @llvm_mips_fexdo_w_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3rf_float_int.ll b/test/CodeGen/Mips/msa/3rf_float_int.ll
index 7b01e1721db9..2bd056d3cc8d 100644
--- a/test/CodeGen/Mips/msa/3rf_float_int.ll
+++ b/test/CodeGen/Mips/msa/3rf_float_int.ll
@@ -10,8 +10,8 @@
 
 define void @llvm_mips_fexp2_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fexp2_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_fexp2_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fexp2_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_fexp2_w_ARG2
   %2 = tail call <4 x float> @llvm.mips.fexp2.w(<4 x float> %0, <4 x i32> %1)
   store <4 x float> %2, <4 x float>* @llvm_mips_fexp2_w_RES
   ret void
@@ -32,8 +32,8 @@ declare <4 x float> @llvm.mips.fexp2.w(<4 x float>, <4 x i32>) nounwind
 
 define void @llvm_mips_fexp2_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fexp2_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_fexp2_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fexp2_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_fexp2_d_ARG2
   %2 = tail call <2 x double> @llvm.mips.fexp2.d(<2 x double> %0, <2 x i64> %1)
   store <2 x double> %2, <2 x double>* @llvm_mips_fexp2_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3rf_int_float.ll b/test/CodeGen/Mips/msa/3rf_int_float.ll
index 5624771b8357..545e5435d643 100644
--- a/test/CodeGen/Mips/msa/3rf_int_float.ll
+++ b/test/CodeGen/Mips/msa/3rf_int_float.ll
@@ -10,8 +10,8 @@
 
 define void @llvm_mips_fcaf_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fcaf_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fcaf_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fcaf_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fcaf_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fcaf.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fcaf_w_RES
   ret void
@@ -32,8 +32,8 @@ declare <4 x i32> @llvm.mips.fcaf.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fcaf_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fcaf_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fcaf_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fcaf_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fcaf_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fcaf.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fcaf_d_RES
   ret void
@@ -54,8 +54,8 @@ declare <2 x i64> @llvm.mips.fcaf.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fceq_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fceq_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fceq_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fceq_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fceq_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fceq.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fceq_w_RES
   ret void
@@ -76,8 +76,8 @@ declare <4 x i32> @llvm.mips.fceq.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fceq_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fceq_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fceq_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fceq_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fceq_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fceq.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fceq_d_RES
   ret void
@@ -98,8 +98,8 @@ declare <2 x i64> @llvm.mips.fceq.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fcle_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fcle_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fcle_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fcle_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fcle_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fcle.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fcle_w_RES
   ret void
@@ -120,8 +120,8 @@ declare <4 x i32> @llvm.mips.fcle.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fcle_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fcle_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fcle_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fcle_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fcle_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fcle.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fcle_d_RES
   ret void
@@ -142,8 +142,8 @@ declare <2 x i64> @llvm.mips.fcle.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fclt_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fclt_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fclt_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fclt_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fclt_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fclt.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fclt_w_RES
   ret void
@@ -164,8 +164,8 @@ declare <4 x i32> @llvm.mips.fclt.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fclt_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fclt_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fclt_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fclt_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fclt_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fclt.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fclt_d_RES
   ret void
@@ -186,8 +186,8 @@ declare <2 x i64> @llvm.mips.fclt.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fcor_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fcor_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fcor_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fcor_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fcor_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fcor.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fcor_w_RES
   ret void
@@ -208,8 +208,8 @@ declare <4 x i32> @llvm.mips.fcor.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fcor_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fcor_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fcor_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fcor_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fcor_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fcor.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fcor_d_RES
   ret void
@@ -230,8 +230,8 @@ declare <2 x i64> @llvm.mips.fcor.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fcne_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fcne_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fcne_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fcne_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fcne_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fcne.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fcne_w_RES
   ret void
@@ -252,8 +252,8 @@ declare <4 x i32> @llvm.mips.fcne.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fcne_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fcne_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fcne_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fcne_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fcne_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fcne.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fcne_d_RES
   ret void
@@ -274,8 +274,8 @@ declare <2 x i64> @llvm.mips.fcne.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fcueq_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fcueq_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fcueq_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fcueq_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fcueq_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fcueq.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fcueq_w_RES
   ret void
@@ -296,8 +296,8 @@ declare <4 x i32> @llvm.mips.fcueq.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fcueq_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fcueq_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fcueq_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fcueq_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fcueq_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fcueq.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fcueq_d_RES
   ret void
@@ -318,8 +318,8 @@ declare <2 x i64> @llvm.mips.fcueq.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fcult_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fcult_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fcult_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fcult_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fcult_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fcult.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fcult_w_RES
   ret void
@@ -340,8 +340,8 @@ declare <4 x i32> @llvm.mips.fcult.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fcult_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fcult_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fcult_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fcult_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fcult_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fcult.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fcult_d_RES
   ret void
@@ -362,8 +362,8 @@ declare <2 x i64> @llvm.mips.fcult.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fcule_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fcule_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fcule_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fcule_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fcule_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fcule.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fcule_w_RES
   ret void
@@ -384,8 +384,8 @@ declare <4 x i32> @llvm.mips.fcule.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fcule_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fcule_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fcule_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fcule_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fcule_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fcule.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fcule_d_RES
   ret void
@@ -406,8 +406,8 @@ declare <2 x i64> @llvm.mips.fcule.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fcun_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fcun_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fcun_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fcun_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fcun_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fcun.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fcun_w_RES
   ret void
@@ -428,8 +428,8 @@ declare <4 x i32> @llvm.mips.fcun.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fcun_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fcun_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fcun_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fcun_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fcun_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fcun.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fcun_d_RES
   ret void
@@ -450,8 +450,8 @@ declare <2 x i64> @llvm.mips.fcun.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fcune_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fcune_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fcune_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fcune_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fcune_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fcune.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fcune_w_RES
   ret void
@@ -472,8 +472,8 @@ declare <4 x i32> @llvm.mips.fcune.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fcune_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fcune_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fcune_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fcune_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fcune_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fcune.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fcune_d_RES
   ret void
@@ -494,8 +494,8 @@ declare <2 x i64> @llvm.mips.fcune.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fsaf_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fsaf_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fsaf_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fsaf_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fsaf_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fsaf.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fsaf_w_RES
   ret void
@@ -516,8 +516,8 @@ declare <4 x i32> @llvm.mips.fsaf.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fsaf_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fsaf_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fsaf_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fsaf_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fsaf_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fsaf.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fsaf_d_RES
   ret void
@@ -538,8 +538,8 @@ declare <2 x i64> @llvm.mips.fsaf.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fseq_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fseq_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fseq_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fseq_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fseq_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fseq.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fseq_w_RES
   ret void
@@ -560,8 +560,8 @@ declare <4 x i32> @llvm.mips.fseq.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fseq_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fseq_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fseq_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fseq_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fseq_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fseq.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fseq_d_RES
   ret void
@@ -582,8 +582,8 @@ declare <2 x i64> @llvm.mips.fseq.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fsle_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fsle_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fsle_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fsle_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fsle_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fsle.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fsle_w_RES
   ret void
@@ -604,8 +604,8 @@ declare <4 x i32> @llvm.mips.fsle.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fsle_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fsle_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fsle_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fsle_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fsle_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fsle.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fsle_d_RES
   ret void
@@ -626,8 +626,8 @@ declare <2 x i64> @llvm.mips.fsle.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fslt_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fslt_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fslt_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fslt_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fslt_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fslt.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fslt_w_RES
   ret void
@@ -648,8 +648,8 @@ declare <4 x i32> @llvm.mips.fslt.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fslt_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fslt_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fslt_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fslt_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fslt_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fslt.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fslt_d_RES
   ret void
@@ -670,8 +670,8 @@ declare <2 x i64> @llvm.mips.fslt.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fsor_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fsor_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fsor_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fsor_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fsor_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fsor.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fsor_w_RES
   ret void
@@ -692,8 +692,8 @@ declare <4 x i32> @llvm.mips.fsor.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fsor_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fsor_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fsor_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fsor_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fsor_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fsor.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fsor_d_RES
   ret void
@@ -714,8 +714,8 @@ declare <2 x i64> @llvm.mips.fsor.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fsne_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fsne_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fsne_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fsne_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fsne_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fsne.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fsne_w_RES
   ret void
@@ -736,8 +736,8 @@ declare <4 x i32> @llvm.mips.fsne.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fsne_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fsne_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fsne_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fsne_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fsne_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fsne.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fsne_d_RES
   ret void
@@ -758,8 +758,8 @@ declare <2 x i64> @llvm.mips.fsne.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fsueq_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fsueq_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fsueq_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fsueq_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fsueq_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fsueq.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fsueq_w_RES
   ret void
@@ -780,8 +780,8 @@ declare <4 x i32> @llvm.mips.fsueq.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fsueq_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fsueq_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fsueq_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fsueq_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fsueq_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fsueq.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fsueq_d_RES
   ret void
@@ -802,8 +802,8 @@ declare <2 x i64> @llvm.mips.fsueq.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fsult_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fsult_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fsult_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fsult_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fsult_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fsult.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fsult_w_RES
   ret void
@@ -824,8 +824,8 @@ declare <4 x i32> @llvm.mips.fsult.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fsult_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fsult_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fsult_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fsult_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fsult_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fsult.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fsult_d_RES
   ret void
@@ -846,8 +846,8 @@ declare <2 x i64> @llvm.mips.fsult.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fsule_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fsule_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fsule_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fsule_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fsule_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fsule.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fsule_w_RES
   ret void
@@ -868,8 +868,8 @@ declare <4 x i32> @llvm.mips.fsule.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fsule_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fsule_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fsule_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fsule_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fsule_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fsule.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fsule_d_RES
   ret void
@@ -890,8 +890,8 @@ declare <2 x i64> @llvm.mips.fsule.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fsun_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fsun_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fsun_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fsun_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fsun_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fsun.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fsun_w_RES
   ret void
@@ -912,8 +912,8 @@ declare <4 x i32> @llvm.mips.fsun.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fsun_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fsun_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fsun_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fsun_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fsun_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fsun.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fsun_d_RES
   ret void
@@ -934,8 +934,8 @@ declare <2 x i64> @llvm.mips.fsun.d(<2 x double>, <2 x double>) nounwind
 
 define void @llvm_mips_fsune_w_test() nounwind {
 entry:
-  %0 = load <4 x float>* @llvm_mips_fsune_w_ARG1
-  %1 = load <4 x float>* @llvm_mips_fsune_w_ARG2
+  %0 = load <4 x float>, <4 x float>* @llvm_mips_fsune_w_ARG1
+  %1 = load <4 x float>, <4 x float>* @llvm_mips_fsune_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.fsune.w(<4 x float> %0, <4 x float> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_fsune_w_RES
   ret void
@@ -956,8 +956,8 @@ declare <4 x i32> @llvm.mips.fsune.w(<4 x float>, <4 x float>) nounwind
 
 define void @llvm_mips_fsune_d_test() nounwind {
 entry:
-  %0 = load <2 x double>* @llvm_mips_fsune_d_ARG1
-  %1 = load <2 x double>* @llvm_mips_fsune_d_ARG2
+  %0 = load <2 x double>, <2 x double>* @llvm_mips_fsune_d_ARG1
+  %1 = load <2 x double>, <2 x double>* @llvm_mips_fsune_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.fsune.d(<2 x double> %0, <2 x double> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_fsune_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/3rf_q.ll b/test/CodeGen/Mips/msa/3rf_q.ll
index f7000ee913a6..c8b0a5000021 100644
--- a/test/CodeGen/Mips/msa/3rf_q.ll
+++ b/test/CodeGen/Mips/msa/3rf_q.ll
@@ -10,8 +10,8 @@
 
 define void @llvm_mips_mul_q_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_mul_q_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_mul_q_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_mul_q_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_mul_q_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.mul.q.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_mul_q_h_RES
   ret void
@@ -32,8 +32,8 @@ declare <8 x i16> @llvm.mips.mul.q.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_mul_q_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_mul_q_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_mul_q_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_mul_q_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_mul_q_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.mul.q.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_mul_q_w_RES
   ret void
@@ -54,8 +54,8 @@ declare <4 x i32> @llvm.mips.mul.q.w(<4 x i32>, <4 x i32>) nounwind
 
 define void @llvm_mips_mulr_q_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_mulr_q_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_mulr_q_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_mulr_q_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_mulr_q_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.mulr.q.h(<8 x i16> %0, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_mulr_q_h_RES
   ret void
@@ -76,8 +76,8 @@ declare <8 x i16> @llvm.mips.mulr.q.h(<8 x i16>, <8 x i16>) nounwind
 
 define void @llvm_mips_mulr_q_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_mulr_q_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_mulr_q_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_mulr_q_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_mulr_q_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.mulr.q.w(<4 x i32> %0, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_mulr_q_w_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/arithmetic.ll b/test/CodeGen/Mips/msa/arithmetic.ll
index 09ee5023c7b1..3ecd0e43589f 100644
--- a/test/CodeGen/Mips/msa/arithmetic.ll
+++ b/test/CodeGen/Mips/msa/arithmetic.ll
@@ -4,9 +4,9 @@
 define void @add_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: add_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = add <16 x i8> %1, %2
   ; CHECK-DAG: addv.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -20,9 +20,9 @@ define void @add_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @add_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: add_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = add <8 x i16> %1, %2
   ; CHECK-DAG: addv.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -36,9 +36,9 @@ define void @add_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @add_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: add_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = add <4 x i32> %1, %2
   ; CHECK-DAG: addv.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -52,9 +52,9 @@ define void @add_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @add_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: add_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = add <2 x i64> %1, %2
   ; CHECK-DAG: addv.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -68,7 +68,7 @@ define void @add_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @add_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: add_v16i8_i:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = add <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1,
                           i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
@@ -83,7 +83,7 @@ define void @add_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @add_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: add_v8i16_i:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = add <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1,
                           i16 1, i16 1, i16 1, i16 1>
@@ -98,7 +98,7 @@ define void @add_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @add_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: add_v4i32_i:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = add <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   ; CHECK-DAG: addvi.w [[R3:\$w[0-9]+]], [[R1]], 1
@@ -112,7 +112,7 @@ define void @add_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @add_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: add_v2i64_i:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = add <2 x i64> %1, <i64 1, i64 1>
   ; CHECK-DAG: addvi.d [[R3:\$w[0-9]+]], [[R1]], 1
@@ -126,9 +126,9 @@ define void @add_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @sub_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: sub_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = sub <16 x i8> %1, %2
   ; CHECK-DAG: subv.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -142,9 +142,9 @@ define void @sub_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @sub_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: sub_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = sub <8 x i16> %1, %2
   ; CHECK-DAG: subv.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -158,9 +158,9 @@ define void @sub_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @sub_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: sub_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = sub <4 x i32> %1, %2
   ; CHECK-DAG: subv.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -174,9 +174,9 @@ define void @sub_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @sub_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: sub_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = sub <2 x i64> %1, %2
   ; CHECK-DAG: subv.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -190,7 +190,7 @@ define void @sub_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @sub_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: sub_v16i8_i:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = sub <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1,
                           i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
@@ -205,7 +205,7 @@ define void @sub_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @sub_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: sub_v8i16_i:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = sub <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1,
                           i16 1, i16 1, i16 1, i16 1>
@@ -220,7 +220,7 @@ define void @sub_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @sub_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: sub_v4i32_i:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = sub <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   ; CHECK-DAG: subvi.w [[R3:\$w[0-9]+]], [[R1]], 1
@@ -234,7 +234,7 @@ define void @sub_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @sub_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: sub_v2i64_i:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = sub <2 x i64> %1, <i64 1, i64 1>
   ; CHECK-DAG: subvi.d [[R3:\$w[0-9]+]], [[R1]], 1
@@ -248,9 +248,9 @@ define void @sub_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @mul_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: mul_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = mul <16 x i8> %1, %2
   ; CHECK-DAG: mulv.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -264,9 +264,9 @@ define void @mul_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @mul_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: mul_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = mul <8 x i16> %1, %2
   ; CHECK-DAG: mulv.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -280,9 +280,9 @@ define void @mul_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @mul_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: mul_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = mul <4 x i32> %1, %2
   ; CHECK-DAG: mulv.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -296,9 +296,9 @@ define void @mul_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @mul_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: mul_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = mul <2 x i64> %1, %2
   ; CHECK-DAG: mulv.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -313,11 +313,11 @@ define void @maddv_v16i8(<16 x i8>* %d, <16 x i8>* %a, <16 x i8>* %b,
                          <16 x i8>* %c) nounwind {
   ; CHECK: maddv_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <16 x i8>* %c
+  %3 = load <16 x i8>, <16 x i8>* %c
   ; CHECK-DAG: ld.b [[R3:\$w[0-9]+]], 0($7)
   %4 = mul <16 x i8> %2, %3
   %5 = add <16 x i8> %4, %1
@@ -333,11 +333,11 @@ define void @maddv_v8i16(<8 x i16>* %d, <8 x i16>* %a, <8 x i16>* %b,
                          <8 x i16>* %c) nounwind {
   ; CHECK: maddv_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <8 x i16>* %c
+  %3 = load <8 x i16>, <8 x i16>* %c
   ; CHECK-DAG: ld.h [[R3:\$w[0-9]+]], 0($7)
   %4 = mul <8 x i16> %2, %3
   %5 = add <8 x i16> %4, %1
@@ -353,11 +353,11 @@ define void @maddv_v4i32(<4 x i32>* %d, <4 x i32>* %a, <4 x i32>* %b,
                          <4 x i32>* %c) nounwind {
   ; CHECK: maddv_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <4 x i32>* %c
+  %3 = load <4 x i32>, <4 x i32>* %c
   ; CHECK-DAG: ld.w [[R3:\$w[0-9]+]], 0($7)
   %4 = mul <4 x i32> %2, %3
   %5 = add <4 x i32> %4, %1
@@ -373,11 +373,11 @@ define void @maddv_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b,
                          <2 x i64>* %c) nounwind {
   ; CHECK: maddv_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <2 x i64>* %c
+  %3 = load <2 x i64>, <2 x i64>* %c
   ; CHECK-DAG: ld.d [[R3:\$w[0-9]+]], 0($7)
   %4 = mul <2 x i64> %2, %3
   %5 = add <2 x i64> %4, %1
@@ -393,11 +393,11 @@ define void @msubv_v16i8(<16 x i8>* %d, <16 x i8>* %a, <16 x i8>* %b,
                          <16 x i8>* %c) nounwind {
   ; CHECK: msubv_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <16 x i8>* %c
+  %3 = load <16 x i8>, <16 x i8>* %c
   ; CHECK-DAG: ld.b [[R3:\$w[0-9]+]], 0($7)
   %4 = mul <16 x i8> %2, %3
   %5 = sub <16 x i8> %1, %4
@@ -413,11 +413,11 @@ define void @msubv_v8i16(<8 x i16>* %d, <8 x i16>* %a, <8 x i16>* %b,
                          <8 x i16>* %c) nounwind {
   ; CHECK: msubv_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <8 x i16>* %c
+  %3 = load <8 x i16>, <8 x i16>* %c
   ; CHECK-DAG: ld.h [[R3:\$w[0-9]+]], 0($7)
   %4 = mul <8 x i16> %2, %3
   %5 = sub <8 x i16> %1, %4
@@ -433,11 +433,11 @@ define void @msubv_v4i32(<4 x i32>* %d, <4 x i32>* %a, <4 x i32>* %b,
                          <4 x i32>* %c) nounwind {
   ; CHECK: msubv_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <4 x i32>* %c
+  %3 = load <4 x i32>, <4 x i32>* %c
   ; CHECK-DAG: ld.w [[R3:\$w[0-9]+]], 0($7)
   %4 = mul <4 x i32> %2, %3
   %5 = sub <4 x i32> %1, %4
@@ -453,11 +453,11 @@ define void @msubv_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b,
                          <2 x i64>* %c) nounwind {
   ; CHECK: msubv_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <2 x i64>* %c
+  %3 = load <2 x i64>, <2 x i64>* %c
   ; CHECK-DAG: ld.d [[R3:\$w[0-9]+]], 0($7)
   %4 = mul <2 x i64> %2, %3
   %5 = sub <2 x i64> %1, %4
@@ -472,9 +472,9 @@ define void @msubv_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b,
 define void @div_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: div_s_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = sdiv <16 x i8> %1, %2
   ; CHECK-DAG: div_s.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -488,9 +488,9 @@ define void @div_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @div_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: div_s_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = sdiv <8 x i16> %1, %2
   ; CHECK-DAG: div_s.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -504,9 +504,9 @@ define void @div_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @div_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: div_s_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = sdiv <4 x i32> %1, %2
   ; CHECK-DAG: div_s.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -520,9 +520,9 @@ define void @div_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @div_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: div_s_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = sdiv <2 x i64> %1, %2
   ; CHECK-DAG: div_s.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -536,9 +536,9 @@ define void @div_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @div_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: div_u_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = udiv <16 x i8> %1, %2
   ; CHECK-DAG: div_u.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -552,9 +552,9 @@ define void @div_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @div_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: div_u_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = udiv <8 x i16> %1, %2
   ; CHECK-DAG: div_u.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -568,9 +568,9 @@ define void @div_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @div_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: div_u_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = udiv <4 x i32> %1, %2
   ; CHECK-DAG: div_u.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -584,9 +584,9 @@ define void @div_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @div_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: div_u_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = udiv <2 x i64> %1, %2
   ; CHECK-DAG: div_u.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -600,9 +600,9 @@ define void @div_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @mod_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: mod_s_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = srem <16 x i8> %1, %2
   ; CHECK-DAG: mod_s.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -616,9 +616,9 @@ define void @mod_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @mod_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: mod_s_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = srem <8 x i16> %1, %2
   ; CHECK-DAG: mod_s.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -632,9 +632,9 @@ define void @mod_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @mod_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: mod_s_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = srem <4 x i32> %1, %2
   ; CHECK-DAG: mod_s.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -648,9 +648,9 @@ define void @mod_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @mod_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: mod_s_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = srem <2 x i64> %1, %2
   ; CHECK-DAG: mod_s.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -664,9 +664,9 @@ define void @mod_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @mod_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: mod_u_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = urem <16 x i8> %1, %2
   ; CHECK-DAG: mod_u.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -680,9 +680,9 @@ define void @mod_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @mod_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: mod_u_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = urem <8 x i16> %1, %2
   ; CHECK-DAG: mod_u.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -696,9 +696,9 @@ define void @mod_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @mod_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: mod_u_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = urem <4 x i32> %1, %2
   ; CHECK-DAG: mod_u.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -712,9 +712,9 @@ define void @mod_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @mod_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: mod_u_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = urem <2 x i64> %1, %2
   ; CHECK-DAG: mod_u.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
diff --git a/test/CodeGen/Mips/msa/arithmetic_float.ll b/test/CodeGen/Mips/msa/arithmetic_float.ll
index 9aae284fe535..d2ead536804a 100644
--- a/test/CodeGen/Mips/msa/arithmetic_float.ll
+++ b/test/CodeGen/Mips/msa/arithmetic_float.ll
@@ -4,9 +4,9 @@
 define void @add_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: add_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fadd <4 x float> %1, %2
   ; CHECK-DAG: fadd.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -20,9 +20,9 @@ define void @add_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwi
 define void @add_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: add_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fadd <2 x double> %1, %2
   ; CHECK-DAG: fadd.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -36,9 +36,9 @@ define void @add_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nou
 define void @sub_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: sub_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fsub <4 x float> %1, %2
   ; CHECK-DAG: fsub.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -52,9 +52,9 @@ define void @sub_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwi
 define void @sub_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: sub_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fsub <2 x double> %1, %2
   ; CHECK-DAG: fsub.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -68,9 +68,9 @@ define void @sub_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nou
 define void @mul_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: mul_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fmul <4 x float> %1, %2
   ; CHECK-DAG: fmul.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -84,9 +84,9 @@ define void @mul_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwi
 define void @mul_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: mul_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fmul <2 x double> %1, %2
   ; CHECK-DAG: fmul.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -101,11 +101,11 @@ define void @fma_v4f32(<4 x float>* %d, <4 x float>* %a, <4 x float>* %b,
                        <4 x float>* %c) nounwind {
   ; CHECK: fma_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <4 x float>* %c
+  %3 = load <4 x float>, <4 x float>* %c
   ; CHECK-DAG: ld.w [[R3:\$w[0-9]+]], 0($7)
   %4 = tail call <4 x float> @llvm.fma.v4f32 (<4 x float> %1, <4 x float> %2,
                                               <4 x float> %3)
@@ -121,11 +121,11 @@ define void @fma_v2f64(<2 x double>* %d, <2 x double>* %a, <2 x double>* %b,
                        <2 x double>* %c) nounwind {
   ; CHECK: fma_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <2 x double>* %c
+  %3 = load <2 x double>, <2 x double>* %c
   ; CHECK-DAG: ld.d [[R3:\$w[0-9]+]], 0($7)
   %4 = tail call <2 x double> @llvm.fma.v2f64 (<2 x double> %1, <2 x double> %2,
                                                <2 x double> %3)
@@ -141,11 +141,11 @@ define void @fmsub_v4f32(<4 x float>* %d, <4 x float>* %a, <4 x float>* %b,
                        <4 x float>* %c) nounwind {
   ; CHECK: fmsub_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <4 x float>* %c
+  %3 = load <4 x float>, <4 x float>* %c
   ; CHECK-DAG: ld.w [[R3:\$w[0-9]+]], 0($7)
   %4 = fmul <4 x float> %2, %3
   %5 = fsub <4 x float> %1, %4
@@ -161,11 +161,11 @@ define void @fmsub_v2f64(<2 x double>* %d, <2 x double>* %a, <2 x double>* %b,
                        <2 x double>* %c) nounwind {
   ; CHECK: fmsub_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <2 x double>* %c
+  %3 = load <2 x double>, <2 x double>* %c
   ; CHECK-DAG: ld.d [[R3:\$w[0-9]+]], 0($7)
   %4 = fmul <2 x double> %2, %3
   %5 = fsub <2 x double> %1, %4
@@ -180,9 +180,9 @@ define void @fmsub_v2f64(<2 x double>* %d, <2 x double>* %a, <2 x double>* %b,
 define void @fdiv_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: fdiv_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fdiv <4 x float> %1, %2
   ; CHECK-DAG: fdiv.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -196,9 +196,9 @@ define void @fdiv_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounw
 define void @fdiv_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: fdiv_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fdiv <2 x double> %1, %2
   ; CHECK-DAG: fdiv.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -212,7 +212,7 @@ define void @fdiv_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) no
 define void @fabs_v4f32(<4 x float>* %c, <4 x float>* %a) nounwind {
   ; CHECK: fabs_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = tail call <4 x float> @llvm.fabs.v4f32 (<4 x float> %1)
   ; CHECK-DAG: fmax_a.w [[R3:\$w[0-9]+]], [[R1]], [[R1]]
@@ -226,7 +226,7 @@ define void @fabs_v4f32(<4 x float>* %c, <4 x float>* %a) nounwind {
 define void @fabs_v2f64(<2 x double>* %c, <2 x double>* %a) nounwind {
   ; CHECK: fabs_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = tail call <2 x double> @llvm.fabs.v2f64 (<2 x double> %1)
   ; CHECK-DAG: fmax_a.d [[R3:\$w[0-9]+]], [[R1]], [[R1]]
@@ -240,7 +240,7 @@ define void @fabs_v2f64(<2 x double>* %c, <2 x double>* %a) nounwind {
 define void @fexp2_v4f32(<4 x float>* %c, <4 x float>* %a) nounwind {
   ; CHECK: fexp2_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = tail call <4 x float> @llvm.exp2.v4f32 (<4 x float> %1)
   ; CHECK-DAG: ldi.w [[R3:\$w[0-9]+]], 1
@@ -256,7 +256,7 @@ define void @fexp2_v4f32(<4 x float>* %c, <4 x float>* %a) nounwind {
 define void @fexp2_v2f64(<2 x double>* %c, <2 x double>* %a) nounwind {
   ; CHECK: fexp2_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = tail call <2 x double> @llvm.exp2.v2f64 (<2 x double> %1)
   ; CHECK-DAG: ldi.d [[R3:\$w[0-9]+]], 1
@@ -272,7 +272,7 @@ define void @fexp2_v2f64(<2 x double>* %c, <2 x double>* %a) nounwind {
 define void @fexp2_v4f32_2(<4 x float>* %c, <4 x float>* %a) nounwind {
   ; CHECK: fexp2_v4f32_2:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = tail call <4 x float> @llvm.exp2.v4f32 (<4 x float> %1)
   %3 = fmul <4 x float> <float 2.0, float 2.0, float 2.0, float 2.0>, %2
@@ -289,7 +289,7 @@ define void @fexp2_v4f32_2(<4 x float>* %c, <4 x float>* %a) nounwind {
 define void @fexp2_v2f64_2(<2 x double>* %c, <2 x double>* %a) nounwind {
   ; CHECK: fexp2_v2f64_2:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = tail call <2 x double> @llvm.exp2.v2f64 (<2 x double> %1)
   %3 = fmul <2 x double> <double 2.0, double 2.0>, %2
@@ -306,7 +306,7 @@ define void @fexp2_v2f64_2(<2 x double>* %c, <2 x double>* %a) nounwind {
 define void @fsqrt_v4f32(<4 x float>* %c, <4 x float>* %a) nounwind {
   ; CHECK: fsqrt_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = tail call <4 x float> @llvm.sqrt.v4f32 (<4 x float> %1)
   ; CHECK-DAG: fsqrt.w [[R3:\$w[0-9]+]], [[R1]]
@@ -320,7 +320,7 @@ define void @fsqrt_v4f32(<4 x float>* %c, <4 x float>* %a) nounwind {
 define void @fsqrt_v2f64(<2 x double>* %c, <2 x double>* %a) nounwind {
   ; CHECK: fsqrt_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = tail call <2 x double> @llvm.sqrt.v2f64 (<2 x double> %1)
   ; CHECK-DAG: fsqrt.d [[R3:\$w[0-9]+]], [[R1]]
@@ -334,7 +334,7 @@ define void @fsqrt_v2f64(<2 x double>* %c, <2 x double>* %a) nounwind {
 define void @ffint_u_v4f32(<4 x float>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: ffint_u_v4f32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = uitofp <4 x i32> %1 to <4 x float>
   ; CHECK-DAG: ffint_u.w [[R3:\$w[0-9]+]], [[R1]]
@@ -348,7 +348,7 @@ define void @ffint_u_v4f32(<4 x float>* %c, <4 x i32>* %a) nounwind {
 define void @ffint_u_v2f64(<2 x double>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: ffint_u_v2f64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = uitofp <2 x i64> %1 to <2 x double>
   ; CHECK-DAG: ffint_u.d [[R3:\$w[0-9]+]], [[R1]]
@@ -362,7 +362,7 @@ define void @ffint_u_v2f64(<2 x double>* %c, <2 x i64>* %a) nounwind {
 define void @ffint_s_v4f32(<4 x float>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: ffint_s_v4f32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = sitofp <4 x i32> %1 to <4 x float>
   ; CHECK-DAG: ffint_s.w [[R3:\$w[0-9]+]], [[R1]]
@@ -376,7 +376,7 @@ define void @ffint_s_v4f32(<4 x float>* %c, <4 x i32>* %a) nounwind {
 define void @ffint_s_v2f64(<2 x double>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: ffint_s_v2f64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = sitofp <2 x i64> %1 to <2 x double>
   ; CHECK-DAG: ffint_s.d [[R3:\$w[0-9]+]], [[R1]]
@@ -390,7 +390,7 @@ define void @ffint_s_v2f64(<2 x double>* %c, <2 x i64>* %a) nounwind {
 define void @ftrunc_u_v4f32(<4 x i32>* %c, <4 x float>* %a) nounwind {
   ; CHECK: ftrunc_u_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = fptoui <4 x float> %1 to <4 x i32>
   ; CHECK-DAG: ftrunc_u.w [[R3:\$w[0-9]+]], [[R1]]
@@ -404,7 +404,7 @@ define void @ftrunc_u_v4f32(<4 x i32>* %c, <4 x float>* %a) nounwind {
 define void @ftrunc_u_v2f64(<2 x i64>* %c, <2 x double>* %a) nounwind {
   ; CHECK: ftrunc_u_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = fptoui <2 x double> %1 to <2 x i64>
   ; CHECK-DAG: ftrunc_u.d [[R3:\$w[0-9]+]], [[R1]]
@@ -418,7 +418,7 @@ define void @ftrunc_u_v2f64(<2 x i64>* %c, <2 x double>* %a) nounwind {
 define void @ftrunc_s_v4f32(<4 x i32>* %c, <4 x float>* %a) nounwind {
   ; CHECK: ftrunc_s_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = fptosi <4 x float> %1 to <4 x i32>
   ; CHECK-DAG: ftrunc_s.w [[R3:\$w[0-9]+]], [[R1]]
@@ -432,7 +432,7 @@ define void @ftrunc_s_v4f32(<4 x i32>* %c, <4 x float>* %a) nounwind {
 define void @ftrunc_s_v2f64(<2 x i64>* %c, <2 x double>* %a) nounwind {
   ; CHECK: ftrunc_s_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = fptosi <2 x double> %1 to <2 x i64>
   ; CHECK-DAG: ftrunc_s.d [[R3:\$w[0-9]+]], [[R1]]
diff --git a/test/CodeGen/Mips/msa/basic_operations.ll b/test/CodeGen/Mips/msa/basic_operations.ll
index dbdf42be49ca..2efec2911935 100644
--- a/test/CodeGen/Mips/msa/basic_operations.ll
+++ b/test/CodeGen/Mips/msa/basic_operations.ll
@@ -1,5 +1,9 @@
-; RUN: llc -march=mips -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=MIPS32-AE -check-prefix=MIPS32-BE %s
-; RUN: llc -march=mipsel -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=MIPS32-AE -check-prefix=MIPS32-LE %s
+; RUN: llc -march=mips -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=O32 -check-prefix=MIPS32 -check-prefix=ALL-BE %s
+; RUN: llc -march=mipsel -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=O32 -check-prefix=MIPS32 -check-prefix=ALL-LE %s
+; RUN: llc -march=mips64 -target-abi n32 -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=N32 -check-prefix=MIPS64 -check-prefix=ALL-BE %s
+; RUN: llc -march=mips64el -target-abi n32 -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=N32 -check-prefix=MIPS64 -check-prefix=ALL-LE %s
+; RUN: llc -march=mips64 -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=N64 -check-prefix=MIPS64 -check-prefix=ALL-BE %s
+; RUN: llc -march=mips64el -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=N64 -check-prefix=MIPS64 -check-prefix=ALL-LE %s
 
 @v4i8 = global <4 x i8> <i8 0, i8 0, i8 0, i8 0>
 @v16i8 = global <16 x i8> <i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0>
@@ -10,135 +14,153 @@
 @i64 = global i64 0
 
 define void @const_v16i8() nounwind {
-  ; MIPS32-AE-LABEL: const_v16i8:
+  ; ALL-LABEL: const_v16i8:
 
   store volatile <16 x i8> <i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0>, <16 x i8>*@v16i8
-  ; MIPS32-AE: ldi.b [[R1:\$w[0-9]+]], 0
+  ; ALL: ldi.b [[R1:\$w[0-9]+]], 0
 
   store volatile <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>, <16 x i8>*@v16i8
-  ; MIPS32-AE: ldi.b [[R1:\$w[0-9]+]], 1
+  ; ALL: ldi.b [[R1:\$w[0-9]+]], 1
 
   store volatile <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 31>, <16 x i8>*@v16i8
-  ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32-AE: ld.b  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; ALL: ld.b  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   store volatile <16 x i8> <i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 8, i8 9, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6>, <16 x i8>*@v16i8
-  ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32-AE: ld.b  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; ALL: ld.b  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   store volatile <16 x i8> <i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0>, <16 x i8>*@v16i8
-  ; MIPS32-BE: ldi.h [[R1:\$w[0-9]+]], 256
-  ; MIPS32-LE: ldi.h [[R1:\$w[0-9]+]], 1
+  ; ALL-BE: ldi.h [[R1:\$w[0-9]+]], 256
+  ; ALL-LE: ldi.h [[R1:\$w[0-9]+]], 1
 
   store volatile <16 x i8> <i8 1, i8 2, i8 3, i8 4, i8 1, i8 2, i8 3, i8 4, i8 1, i8 2, i8 3, i8 4, i8 1, i8 2, i8 3, i8 4>, <16 x i8>*@v16i8
-  ; MIPS32-BE-DAG: lui [[R2:\$[0-9]+]], 258
-  ; MIPS32-LE-DAG: lui [[R2:\$[0-9]+]], 1027
-  ; MIPS32-BE-DAG: ori [[R2]], [[R2]], 772
-  ; MIPS32-LE-DAG: ori [[R2]], [[R2]], 513
-  ; MIPS32-AE-DAG: fill.w [[R1:\$w[0-9]+]], [[R2]]
+  ; ALL-BE-DAG: lui [[R2:\$[0-9]+]], 258
+  ; ALL-LE-DAG: lui [[R2:\$[0-9]+]], 1027
+  ; ALL-BE-DAG: ori [[R2]], [[R2]], 772
+  ; ALL-LE-DAG: ori [[R2]], [[R2]], 513
+  ; ALL-DAG: fill.w [[R1:\$w[0-9]+]], [[R2]]
 
   store volatile <16 x i8> <i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 8, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 8>, <16 x i8>*@v16i8
-  ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32-AE: ld.b  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; ALL: ld.b  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   ret void
-  ; MIPS32-AE: .size const_v16i8
 }
 
 define void @const_v8i16() nounwind {
-  ; MIPS32-AE-LABEL: const_v8i16:
+  ; ALL-LABEL: const_v8i16:
 
   store volatile <8 x i16> <i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0>, <8 x i16>*@v8i16
-  ; MIPS32-AE: ldi.b [[R1:\$w[0-9]+]], 0
+  ; ALL: ldi.b [[R1:\$w[0-9]+]], 0
 
   store volatile <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>, <8 x i16>*@v8i16
-  ; MIPS32-AE: ldi.h [[R1:\$w[0-9]+]], 1
+  ; ALL: ldi.h [[R1:\$w[0-9]+]], 1
 
   store volatile <8 x i16> <i16 1, i16 1, i16 1, i16 2, i16 1, i16 1, i16 1, i16 31>, <8 x i16>*@v8i16
-  ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32-AE: ld.h  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; ALL: ld.h  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   store volatile <8 x i16> <i16 1028, i16 1028, i16 1028, i16 1028, i16 1028, i16 1028, i16 1028, i16 1028>, <8 x i16>*@v8i16
-  ; MIPS32-AE: ldi.b [[R1:\$w[0-9]+]], 4
+  ; ALL: ldi.b [[R1:\$w[0-9]+]], 4
 
   store volatile <8 x i16> <i16 1, i16 2, i16 1, i16 2, i16 1, i16 2, i16 1, i16 2>, <8 x i16>*@v8i16
-  ; MIPS32-BE-DAG: lui [[R2:\$[0-9]+]], 1
-  ; MIPS32-LE-DAG: lui [[R2:\$[0-9]+]], 2
-  ; MIPS32-BE-DAG: ori [[R2]], [[R2]], 2
-  ; MIPS32-LE-DAG: ori [[R2]], [[R2]], 1
-  ; MIPS32-AE-DAG: fill.w [[R1:\$w[0-9]+]], [[R2]]
+  ; ALL-BE-DAG: lui [[R2:\$[0-9]+]], 1
+  ; ALL-LE-DAG: lui [[R2:\$[0-9]+]], 2
+  ; ALL-BE-DAG: ori [[R2]], [[R2]], 2
+  ; ALL-LE-DAG: ori [[R2]], [[R2]], 1
+  ; ALL-DAG: fill.w [[R1:\$w[0-9]+]], [[R2]]
 
   store volatile <8 x i16> <i16 1, i16 2, i16 3, i16 4, i16 1, i16 2, i16 3, i16 4>, <8 x i16>*@v8i16
-  ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32-AE: ld.h  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; ALL: ld.h  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   ret void
-  ; MIPS32-AE: .size const_v8i16
 }
 
 define void @const_v4i32() nounwind {
-  ; MIPS32-AE-LABEL: const_v4i32:
+  ; ALL-LABEL: const_v4i32:
 
   store volatile <4 x i32> <i32 0, i32 0, i32 0, i32 0>, <4 x i32>*@v4i32
-  ; MIPS32-AE: ldi.b [[R1:\$w[0-9]+]], 0
+  ; ALL: ldi.b [[R1:\$w[0-9]+]], 0
 
   store volatile <4 x i32> <i32 1, i32 1, i32 1, i32 1>, <4 x i32>*@v4i32
-  ; MIPS32-AE: ldi.w [[R1:\$w[0-9]+]], 1
+  ; ALL: ldi.w [[R1:\$w[0-9]+]], 1
 
   store volatile <4 x i32> <i32 1, i32 1, i32 1, i32 31>, <4 x i32>*@v4i32
-  ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32-AE: ld.w  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; ALL: ld.w  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   store volatile <4 x i32> <i32 16843009, i32 16843009, i32 16843009, i32 16843009>, <4 x i32>*@v4i32
-  ; MIPS32-AE: ldi.b [[R1:\$w[0-9]+]], 1
+  ; ALL: ldi.b [[R1:\$w[0-9]+]], 1
 
   store volatile <4 x i32> <i32 65537, i32 65537, i32 65537, i32 65537>, <4 x i32>*@v4i32
-  ; MIPS32-AE: ldi.h [[R1:\$w[0-9]+]], 1
+  ; ALL: ldi.h [[R1:\$w[0-9]+]], 1
 
   store volatile <4 x i32> <i32 1, i32 2, i32 1, i32 2>, <4 x i32>*@v4i32
-  ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32-AE: ld.w  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; ALL: ld.w  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   store volatile <4 x i32> <i32 3, i32 4, i32 5, i32 6>, <4 x i32>*@v4i32
-  ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32-AE: ld.w  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; ALL: ld.w  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   ret void
-  ; MIPS32-AE: .size const_v4i32
 }
 
 define void @const_v2i64() nounwind {
-  ; MIPS32-AE-LABEL: const_v2i64:
+  ; ALL-LABEL: const_v2i64:
 
   store volatile <2 x i64> <i64 0, i64 0>, <2 x i64>*@v2i64
-  ; MIPS32-AE: ldi.b [[R1:\$w[0-9]+]], 0
+  ; ALL: ldi.b [[R1:\$w[0-9]+]], 0
 
   store volatile <2 x i64> <i64 72340172838076673, i64 72340172838076673>, <2 x i64>*@v2i64
-  ; MIPS32-AE: ldi.b [[R1:\$w[0-9]+]], 1
+  ; ALL: ldi.b [[R1:\$w[0-9]+]], 1
 
   store volatile <2 x i64> <i64 281479271743489, i64 281479271743489>, <2 x i64>*@v2i64
-  ; MIPS32-AE: ldi.h [[R1:\$w[0-9]+]], 1
+  ; ALL: ldi.h [[R1:\$w[0-9]+]], 1
 
   store volatile <2 x i64> <i64 4294967297, i64 4294967297>, <2 x i64>*@v2i64
-  ; MIPS32-AE: ldi.w [[R1:\$w[0-9]+]], 1
+  ; ALL: ldi.w [[R1:\$w[0-9]+]], 1
 
   store volatile <2 x i64> <i64 1, i64 1>, <2 x i64>*@v2i64
-  ; MIPS32-AE: ldi.d [[R1:\$w[0-9]+]], 1
+  ; ALL: ldi.d [[R1:\$w[0-9]+]], 1
 
   store volatile <2 x i64> <i64 1, i64 31>, <2 x i64>*@v2i64
-  ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32-AE: ld.w  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; MIPS32: ld.w [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; MIPS64: ld.d [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   store volatile <2 x i64> <i64 3, i64 4>, <2 x i64>*@v2i64
-  ; MIPS32-AE: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32-AE: ld.w  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; MIPS32: ld.w  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; MIPS64: ld.d  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   ret void
-  ; MIPS32-AE: .size const_v2i64
 }
 
-define void @nonconst_v16i8(i8 %a, i8 %b, i8 %c, i8 %d, i8 %e, i8 %f, i8 %g, i8 %h) nounwind {
-  ; MIPS32-AE-LABEL: nonconst_v16i8:
+define void @nonconst_v16i8(i8 signext %a, i8 signext %b, i8 signext %c, i8 signext %d, i8 signext %e, i8 signext %f, i8 signext %g, i8 signext %h) nounwind {
+  ; ALL-LABEL: nonconst_v16i8:
 
   %1 = insertelement <16 x i8> undef, i8 %a, i32 0
   %2 = insertelement <16 x i8> %1, i8 %b, i32 1
@@ -156,39 +178,38 @@ define void @nonconst_v16i8(i8 %a, i8 %b, i8 %c, i8 %d, i8 %e, i8 %f, i8 %g, i8
   %14 = insertelement <16 x i8> %13, i8 %h, i32 13
   %15 = insertelement <16 x i8> %14, i8 %h, i32 14
   %16 = insertelement <16 x i8> %15, i8 %h, i32 15
-  ; MIPS32-AE-DAG: insert.b [[R1:\$w[0-9]+]][0], $4
-  ; MIPS32-AE-DAG: insert.b [[R1]][1], $5
-  ; MIPS32-AE-DAG: insert.b [[R1]][2], $6
-  ; MIPS32-AE-DAG: insert.b [[R1]][3], $7
-  ; MIPS32-BE-DAG: lbu [[R2:\$[0-9]+]], 19($sp)
-  ; MIPS32-LE-DAG: lbu [[R2:\$[0-9]+]], 16($sp)
-  ; MIPS32-AE-DAG: insert.b [[R1]][4], [[R2]]
-  ; MIPS32-BE-DAG: lbu [[R3:\$[0-9]+]], 23($sp)
-  ; MIPS32-LE-DAG: lbu [[R3:\$[0-9]+]], 20($sp)
-  ; MIPS32-AE-DAG: insert.b [[R1]][5], [[R3]]
-  ; MIPS32-BE-DAG: lbu [[R4:\$[0-9]+]], 27($sp)
-  ; MIPS32-LE-DAG: lbu [[R4:\$[0-9]+]], 24($sp)
-  ; MIPS32-AE-DAG: insert.b [[R1]][6], [[R4]]
-  ; MIPS32-BE-DAG: lbu [[R5:\$[0-9]+]], 31($sp)
-  ; MIPS32-LE-DAG: lbu [[R5:\$[0-9]+]], 28($sp)
-  ; MIPS32-AE-DAG: insert.b [[R1]][7], [[R5]]
-  ; MIPS32-AE-DAG: insert.b [[R1]][8], [[R5]]
-  ; MIPS32-AE-DAG: insert.b [[R1]][9], [[R5]]
-  ; MIPS32-AE-DAG: insert.b [[R1]][10], [[R5]]
-  ; MIPS32-AE-DAG: insert.b [[R1]][11], [[R5]]
-  ; MIPS32-AE-DAG: insert.b [[R1]][12], [[R5]]
-  ; MIPS32-AE-DAG: insert.b [[R1]][13], [[R5]]
-  ; MIPS32-AE-DAG: insert.b [[R1]][14], [[R5]]
-  ; MIPS32-AE-DAG: insert.b [[R1]][15], [[R5]]
+  ; ALL-DAG: insert.b [[R1:\$w[0-9]+]][0], $4
+  ; ALL-DAG: insert.b [[R1]][1], $5
+  ; ALL-DAG: insert.b [[R1]][2], $6
+  ; ALL-DAG: insert.b [[R1]][3], $7
+  ; MIPS32-DAG: lw [[R2:\$[0-9]+]], 16($sp)
+  ; MIPS32-DAG: insert.b [[R1]][4], [[R2]]
+  ; MIPS64-DAG: insert.b [[R1]][4], $8
+  ; MIPS32-DAG: lw [[R3:\$[0-9]+]], 20($sp)
+  ; MIPS32-DAG: insert.b [[R1]][5], [[R3]]
+  ; MIPS64-DAG: insert.b [[R1]][5], $9
+  ; MIPS32-DAG: lw [[R4:\$[0-9]+]], 24($sp)
+  ; MIPS32-DAG: insert.b [[R1]][6], [[R4]]
+  ; MIPS64-DAG: insert.b [[R1]][6], $10
+  ; MIPS32-DAG: lw [[R5:\$[0-9]+]], 28($sp)
+  ; MIPS32-DAG: insert.b [[R1]][7], [[R5]]
+  ; MIPS64-DAG: insert.b [[R1]][7], [[R5:\$11]]
+  ; ALL-DAG: insert.b [[R1]][8], [[R5]]
+  ; ALL-DAG: insert.b [[R1]][9], [[R5]]
+  ; ALL-DAG: insert.b [[R1]][10], [[R5]]
+  ; ALL-DAG: insert.b [[R1]][11], [[R5]]
+  ; ALL-DAG: insert.b [[R1]][12], [[R5]]
+  ; ALL-DAG: insert.b [[R1]][13], [[R5]]
+  ; ALL-DAG: insert.b [[R1]][14], [[R5]]
+  ; ALL-DAG: insert.b [[R1]][15], [[R5]]
 
   store volatile <16 x i8> %16, <16 x i8>*@v16i8
 
   ret void
-  ; MIPS32-AE: .size nonconst_v16i8
 }
 
-define void @nonconst_v8i16(i16 %a, i16 %b, i16 %c, i16 %d, i16 %e, i16 %f, i16 %g, i16 %h) nounwind {
-  ; MIPS32-AE-LABEL: nonconst_v8i16:
+define void @nonconst_v8i16(i16 signext %a, i16 signext %b, i16 signext %c, i16 signext %d, i16 signext %e, i16 signext %f, i16 signext %g, i16 signext %h) nounwind {
+  ; ALL-LABEL: nonconst_v8i16:
 
   %1 = insertelement <8 x i16> undef, i16 %a, i32 0
   %2 = insertelement <8 x i16> %1, i16 %b, i32 1
@@ -198,610 +219,648 @@ define void @nonconst_v8i16(i16 %a, i16 %b, i16 %c, i16 %d, i16 %e, i16 %f, i16
   %6 = insertelement <8 x i16> %5, i16 %f, i32 5
   %7 = insertelement <8 x i16> %6, i16 %g, i32 6
   %8 = insertelement <8 x i16> %7, i16 %h, i32 7
-  ; MIPS32-AE-DAG: insert.h [[R1:\$w[0-9]+]][0], $4
-  ; MIPS32-AE-DAG: insert.h [[R1]][1], $5
-  ; MIPS32-AE-DAG: insert.h [[R1]][2], $6
-  ; MIPS32-AE-DAG: insert.h [[R1]][3], $7
-  ; MIPS32-BE-DAG: lhu [[R2:\$[0-9]+]], 18($sp)
-  ; MIPS32-LE-DAG: lhu [[R2:\$[0-9]+]], 16($sp)
-  ; MIPS32-AE-DAG: insert.h [[R1]][4], [[R2]]
-  ; MIPS32-BE-DAG: lhu [[R2:\$[0-9]+]], 22($sp)
-  ; MIPS32-LE-DAG: lhu [[R2:\$[0-9]+]], 20($sp)
-  ; MIPS32-AE-DAG: insert.h [[R1]][5], [[R2]]
-  ; MIPS32-BE-DAG: lhu [[R2:\$[0-9]+]], 26($sp)
-  ; MIPS32-LE-DAG: lhu [[R2:\$[0-9]+]], 24($sp)
-  ; MIPS32-AE-DAG: insert.h [[R1]][6], [[R2]]
-  ; MIPS32-BE-DAG: lhu [[R2:\$[0-9]+]], 30($sp)
-  ; MIPS32-LE-DAG: lhu [[R2:\$[0-9]+]], 28($sp)
-  ; MIPS32-AE-DAG: insert.h [[R1]][7], [[R2]]
+  ; ALL-DAG: insert.h [[R1:\$w[0-9]+]][0], $4
+  ; ALL-DAG: insert.h [[R1]][1], $5
+  ; ALL-DAG: insert.h [[R1]][2], $6
+  ; ALL-DAG: insert.h [[R1]][3], $7
+  ; MIPS32-DAG: lw [[R2:\$[0-9]+]], 16($sp)
+  ; MIPS32-DAG: insert.h [[R1]][4], [[R2]]
+  ; MIPS64-DAG: insert.h [[R1]][4], $8
+  ; MIPS32-DAG: lw [[R2:\$[0-9]+]], 20($sp)
+  ; MIPS32-DAG: insert.h [[R1]][5], [[R2]]
+  ; MIPS64-DAG: insert.h [[R1]][5], $9
+  ; MIPS32-DAG: lw [[R2:\$[0-9]+]], 24($sp)
+  ; MIPS32-DAG: insert.h [[R1]][6], [[R2]]
+  ; MIPS64-DAG: insert.h [[R1]][6], $10
+  ; MIPS32-DAG: lw [[R2:\$[0-9]+]], 28($sp)
+  ; MIPS32-DAG: insert.h [[R1]][7], [[R2]]
+  ; MIPS64-DAG: insert.h [[R1]][7], $11
 
   store volatile <8 x i16> %8, <8 x i16>*@v8i16
 
   ret void
-  ; MIPS32-AE: .size nonconst_v8i16
 }
 
-define void @nonconst_v4i32(i32 %a, i32 %b, i32 %c, i32 %d) nounwind {
-  ; MIPS32-AE-LABEL: nonconst_v4i32:
+define void @nonconst_v4i32(i32 signext %a, i32 signext %b, i32 signext %c, i32 signext %d) nounwind {
+  ; ALL-LABEL: nonconst_v4i32:
 
   %1 = insertelement <4 x i32> undef, i32 %a, i32 0
   %2 = insertelement <4 x i32> %1, i32 %b, i32 1
   %3 = insertelement <4 x i32> %2, i32 %c, i32 2
   %4 = insertelement <4 x i32> %3, i32 %d, i32 3
-  ; MIPS32-AE: insert.w [[R1:\$w[0-9]+]][0], $4
-  ; MIPS32-AE: insert.w [[R1]][1], $5
-  ; MIPS32-AE: insert.w [[R1]][2], $6
-  ; MIPS32-AE: insert.w [[R1]][3], $7
+  ; ALL: insert.w [[R1:\$w[0-9]+]][0], $4
+  ; ALL: insert.w [[R1]][1], $5
+  ; ALL: insert.w [[R1]][2], $6
+  ; ALL: insert.w [[R1]][3], $7
 
   store volatile <4 x i32> %4, <4 x i32>*@v4i32
 
   ret void
-  ; MIPS32-AE: .size nonconst_v4i32
 }
 
-define void @nonconst_v2i64(i64 %a, i64 %b) nounwind {
-  ; MIPS32-AE-LABEL: nonconst_v2i64:
+define void @nonconst_v2i64(i64 signext %a, i64 signext %b) nounwind {
+  ; ALL-LABEL: nonconst_v2i64:
 
   %1 = insertelement <2 x i64> undef, i64 %a, i32 0
   %2 = insertelement <2 x i64> %1, i64 %b, i32 1
-  ; MIPS32-AE: insert.w [[R1:\$w[0-9]+]][0], $4
-  ; MIPS32-AE: insert.w [[R1]][1], $5
-  ; MIPS32-AE: insert.w [[R1]][2], $6
-  ; MIPS32-AE: insert.w [[R1]][3], $7
+  ; MIPS32: insert.w [[R1:\$w[0-9]+]][0], $4
+  ; MIPS32: insert.w [[R1]][1], $5
+  ; MIPS32: insert.w [[R1]][2], $6
+  ; MIPS32: insert.w [[R1]][3], $7
+  ; MIPS64: insert.d [[R1:\$w[0-9]+]][0], $4
+  ; MIPS64: insert.d [[R1]][1], $5
 
   store volatile <2 x i64> %2, <2 x i64>*@v2i64
 
   ret void
-  ; MIPS32-AE: .size nonconst_v2i64
 }
 
 define i32 @extract_sext_v16i8() nounwind {
-  ; MIPS32-AE-LABEL: extract_sext_v16i8:
+  ; ALL-LABEL: extract_sext_v16i8:
 
-  %1 = load <16 x i8>* @v16i8
-  ; MIPS32-AE-DAG: ld.b [[R1:\$w[0-9]+]],
+  %1 = load <16 x i8>, <16 x i8>* @v16i8
+  ; ALL-DAG: ld.b [[R1:\$w[0-9]+]],
 
   %2 = add <16 x i8> %1, %1
-  ; MIPS32-AE-DAG: addv.b [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: addv.b [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
   %3 = extractelement <16 x i8> %2, i32 1
   %4 = sext i8 %3 to i32
-  ; MIPS32-AE-DAG: copy_s.b [[R3:\$[0-9]+]], [[R1]][1]
-  ; MIPS32-AE-NOT: sll
-  ; MIPS32-AE-NOT: sra
+  ; ALL-DAG: copy_s.b [[R3:\$[0-9]+]], [[R1]][1]
+  ; ALL-NOT: sll
+  ; ALL-NOT: sra
 
   ret i32 %4
-  ; MIPS32-AE: .size extract_sext_v16i8
 }
 
 define i32 @extract_sext_v8i16() nounwind {
-  ; MIPS32-AE-LABEL: extract_sext_v8i16:
+  ; ALL-LABEL: extract_sext_v8i16:
 
-  %1 = load <8 x i16>* @v8i16
-  ; MIPS32-AE-DAG: ld.h [[R1:\$w[0-9]+]],
+  %1 = load <8 x i16>, <8 x i16>* @v8i16
+  ; ALL-DAG: ld.h [[R1:\$w[0-9]+]],
 
   %2 = add <8 x i16> %1, %1
-  ; MIPS32-AE-DAG: addv.h [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: addv.h [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
   %3 = extractelement <8 x i16> %2, i32 1
   %4 = sext i16 %3 to i32
-  ; MIPS32-AE-DAG: copy_s.h [[R3:\$[0-9]+]], [[R1]][1]
-  ; MIPS32-AE-NOT: sll
-  ; MIPS32-AE-NOT: sra
+  ; ALL-DAG: copy_s.h [[R3:\$[0-9]+]], [[R1]][1]
+  ; ALL-NOT: sll
+  ; ALL-NOT: sra
 
   ret i32 %4
-  ; MIPS32-AE: .size extract_sext_v8i16
 }
 
 define i32 @extract_sext_v4i32() nounwind {
-  ; MIPS32-AE-LABEL: extract_sext_v4i32:
+  ; ALL-LABEL: extract_sext_v4i32:
 
-  %1 = load <4 x i32>* @v4i32
-  ; MIPS32-AE-DAG: ld.w [[R1:\$w[0-9]+]],
+  %1 = load <4 x i32>, <4 x i32>* @v4i32
+  ; ALL-DAG: ld.w [[R1:\$w[0-9]+]],
 
   %2 = add <4 x i32> %1, %1
-  ; MIPS32-AE-DAG: addv.w [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: addv.w [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
   %3 = extractelement <4 x i32> %2, i32 1
-  ; MIPS32-AE-DAG: copy_s.w [[R3:\$[0-9]+]], [[R1]][1]
+  ; ALL-DAG: copy_s.w [[R3:\$[0-9]+]], [[R1]][1]
 
   ret i32 %3
-  ; MIPS32-AE: .size extract_sext_v4i32
 }
 
 define i64 @extract_sext_v2i64() nounwind {
-  ; MIPS32-AE-LABEL: extract_sext_v2i64:
+  ; ALL-LABEL: extract_sext_v2i64:
 
-  %1 = load <2 x i64>* @v2i64
-  ; MIPS32-AE-DAG: ld.d [[R1:\$w[0-9]+]],
+  %1 = load <2 x i64>, <2 x i64>* @v2i64
+  ; ALL-DAG: ld.d [[R1:\$w[0-9]+]],
 
   %2 = add <2 x i64> %1, %1
-  ; MIPS32-AE-DAG: addv.d [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: addv.d [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
   %3 = extractelement <2 x i64> %2, i32 1
-  ; MIPS32-AE-DAG: copy_s.w [[R3:\$[0-9]+]], [[R1]][2]
-  ; MIPS32-AE-DAG: copy_s.w [[R4:\$[0-9]+]], [[R1]][3]
-  ; MIPS32-AE-NOT: sll
-  ; MIPS32-AE-NOT: sra
+  ; MIPS32-DAG: copy_s.w [[R3:\$[0-9]+]], [[R1]][2]
+  ; MIPS32-DAG: copy_s.w [[R4:\$[0-9]+]], [[R1]][3]
+  ; MIPS64-DAG: copy_s.d [[R3:\$[0-9]+]], [[R1]][1]
+  ; ALL-NOT: sll
+  ; ALL-NOT: sra
 
   ret i64 %3
-  ; MIPS32-AE: .size extract_sext_v2i64
 }
 
 define i32 @extract_zext_v16i8() nounwind {
-  ; MIPS32-AE-LABEL: extract_zext_v16i8:
+  ; ALL-LABEL: extract_zext_v16i8:
 
-  %1 = load <16 x i8>* @v16i8
-  ; MIPS32-AE-DAG: ld.b [[R1:\$w[0-9]+]],
+  %1 = load <16 x i8>, <16 x i8>* @v16i8
+  ; ALL-DAG: ld.b [[R1:\$w[0-9]+]],
 
   %2 = add <16 x i8> %1, %1
-  ; MIPS32-AE-DAG: addv.b [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: addv.b [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
   %3 = extractelement <16 x i8> %2, i32 1
   %4 = zext i8 %3 to i32
-  ; MIPS32-AE-DAG: copy_u.b [[R3:\$[0-9]+]], [[R1]][1]
-  ; MIPS32-AE-NOT: andi
+  ; ALL-DAG: copy_u.b [[R3:\$[0-9]+]], [[R1]][1]
+  ; ALL-NOT: andi
 
   ret i32 %4
-  ; MIPS32-AE: .size extract_zext_v16i8
 }
 
 define i32 @extract_zext_v8i16() nounwind {
-  ; MIPS32-AE-LABEL: extract_zext_v8i16:
+  ; ALL-LABEL: extract_zext_v8i16:
 
-  %1 = load <8 x i16>* @v8i16
-  ; MIPS32-AE-DAG: ld.h [[R1:\$w[0-9]+]],
+  %1 = load <8 x i16>, <8 x i16>* @v8i16
+  ; ALL-DAG: ld.h [[R1:\$w[0-9]+]],
 
   %2 = add <8 x i16> %1, %1
-  ; MIPS32-AE-DAG: addv.h [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: addv.h [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
   %3 = extractelement <8 x i16> %2, i32 1
   %4 = zext i16 %3 to i32
-  ; MIPS32-AE-DAG: copy_u.h [[R3:\$[0-9]+]], [[R1]][1]
-  ; MIPS32-AE-NOT: andi
+  ; ALL-DAG: copy_u.h [[R3:\$[0-9]+]], [[R1]][1]
+  ; ALL-NOT: andi
 
   ret i32 %4
-  ; MIPS32-AE: .size extract_zext_v8i16
 }
 
 define i32 @extract_zext_v4i32() nounwind {
-  ; MIPS32-AE-LABEL: extract_zext_v4i32:
+  ; ALL-LABEL: extract_zext_v4i32:
 
-  %1 = load <4 x i32>* @v4i32
-  ; MIPS32-AE-DAG: ld.w [[R1:\$w[0-9]+]],
+  %1 = load <4 x i32>, <4 x i32>* @v4i32
+  ; ALL-DAG: ld.w [[R1:\$w[0-9]+]],
 
   %2 = add <4 x i32> %1, %1
-  ; MIPS32-AE-DAG: addv.w [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: addv.w [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
   %3 = extractelement <4 x i32> %2, i32 1
-  ; MIPS32-AE-DAG: copy_{{[su]}}.w [[R3:\$[0-9]+]], [[R1]][1]
+  ; ALL-DAG: copy_{{[su]}}.w [[R3:\$[0-9]+]], [[R1]][1]
 
   ret i32 %3
-  ; MIPS32-AE: .size extract_zext_v4i32
 }
 
 define i64 @extract_zext_v2i64() nounwind {
-  ; MIPS32-AE-LABEL: extract_zext_v2i64:
+  ; ALL-LABEL: extract_zext_v2i64:
 
-  %1 = load <2 x i64>* @v2i64
-  ; MIPS32-AE-DAG: ld.d [[R1:\$w[0-9]+]],
+  %1 = load <2 x i64>, <2 x i64>* @v2i64
+  ; ALL-DAG: ld.d [[R1:\$w[0-9]+]],
 
   %2 = add <2 x i64> %1, %1
-  ; MIPS32-AE-DAG: addv.d [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: addv.d [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
   %3 = extractelement <2 x i64> %2, i32 1
-  ; MIPS32-AE-DAG: copy_{{[su]}}.w [[R3:\$[0-9]+]], [[R1]][2]
-  ; MIPS32-AE-DAG: copy_{{[su]}}.w [[R4:\$[0-9]+]], [[R1]][3]
-  ; MIPS32-AE-NOT: andi
+  ; MIPS32-DAG: copy_{{[su]}}.w [[R3:\$[0-9]+]], [[R1]][2]
+  ; MIPS32-DAG: copy_{{[su]}}.w [[R4:\$[0-9]+]], [[R1]][3]
+  ; MIPS64-DAG: copy_{{[su]}}.d [[R3:\$[0-9]+]], [[R1]][1]
+  ; ALL-NOT: andi
 
   ret i64 %3
-  ; MIPS32-AE: .size extract_zext_v2i64
 }
 
 define i32 @extract_sext_v16i8_vidx() nounwind {
-  ; MIPS32-AE-LABEL: extract_sext_v16i8_vidx:
+  ; ALL-LABEL: extract_sext_v16i8_vidx:
 
-  %1 = load <16 x i8>* @v16i8
-  ; MIPS32-AE-DAG: lw [[PTR_V:\$[0-9]+]], %got(v16i8)(
-  ; MIPS32-AE-DAG: ld.b [[R1:\$w[0-9]+]], 0([[PTR_V]])
+  %1 = load <16 x i8>, <16 x i8>* @v16i8
+  ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v16i8)(
+  ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v16i8)(
+  ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v16i8)(
+  ; ALL-DAG: ld.b [[R1:\$w[0-9]+]], 0([[PTR_V]])
 
   %2 = add <16 x i8> %1, %1
-  ; MIPS32-AE-DAG: addv.b [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: addv.b [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
-  %3 = load i32* @i32
-  ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
-  ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
+  %3 = load i32, i32* @i32
+  ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
+  ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
 
   %4 = extractelement <16 x i8> %2, i32 %3
   %5 = sext i8 %4 to i32
-  ; MIPS32-AE-DAG: splat.b $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
-  ; MIPS32-AE-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]]
-  ; MIPS32-AE-DAG: sra [[R6:\$[0-9]+]], [[R5]], 24
+  ; ALL-DAG: splat.b $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
+  ; ALL-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]]
+  ; ALL-DAG: sra [[R6:\$[0-9]+]], [[R5]], 24
 
   ret i32 %5
-  ; MIPS32-AE: .size extract_sext_v16i8_vidx
 }
 
 define i32 @extract_sext_v8i16_vidx() nounwind {
-  ; MIPS32-AE-LABEL: extract_sext_v8i16_vidx:
+  ; ALL-LABEL: extract_sext_v8i16_vidx:
 
-  %1 = load <8 x i16>* @v8i16
-  ; MIPS32-AE-DAG: lw [[PTR_V:\$[0-9]+]], %got(v8i16)(
-  ; MIPS32-AE-DAG: ld.h [[R1:\$w[0-9]+]], 0([[PTR_V]])
+  %1 = load <8 x i16>, <8 x i16>* @v8i16
+  ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v8i16)(
+  ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v8i16)(
+  ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v8i16)(
+  ; ALL-DAG: ld.h [[R1:\$w[0-9]+]], 0([[PTR_V]])
 
   %2 = add <8 x i16> %1, %1
-  ; MIPS32-AE-DAG: addv.h [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: addv.h [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
-  %3 = load i32* @i32
-  ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
-  ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
+  %3 = load i32, i32* @i32
+  ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
+  ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
 
   %4 = extractelement <8 x i16> %2, i32 %3
   %5 = sext i16 %4 to i32
-  ; MIPS32-AE-DAG: splat.h $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
-  ; MIPS32-AE-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]]
-  ; MIPS32-AE-DAG: sra [[R6:\$[0-9]+]], [[R5]], 16
+  ; ALL-DAG: splat.h $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
+  ; ALL-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]]
+  ; ALL-DAG: sra [[R6:\$[0-9]+]], [[R5]], 16
 
   ret i32 %5
-  ; MIPS32-AE: .size extract_sext_v8i16_vidx
 }
 
 define i32 @extract_sext_v4i32_vidx() nounwind {
-  ; MIPS32-AE-LABEL: extract_sext_v4i32_vidx:
+  ; ALL-LABEL: extract_sext_v4i32_vidx:
 
-  %1 = load <4 x i32>* @v4i32
-  ; MIPS32-AE-DAG: lw [[PTR_V:\$[0-9]+]], %got(v4i32)(
-  ; MIPS32-AE-DAG: ld.w [[R1:\$w[0-9]+]], 0([[PTR_V]])
+  %1 = load <4 x i32>, <4 x i32>* @v4i32
+  ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v4i32)(
+  ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v4i32)(
+  ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v4i32)(
+  ; ALL-DAG: ld.w [[R1:\$w[0-9]+]], 0([[PTR_V]])
 
   %2 = add <4 x i32> %1, %1
-  ; MIPS32-AE-DAG: addv.w [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: addv.w [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
-  %3 = load i32* @i32
-  ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
-  ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
+  %3 = load i32, i32* @i32
+  ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
+  ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
 
   %4 = extractelement <4 x i32> %2, i32 %3
-  ; MIPS32-AE-DAG: splat.w $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
-  ; MIPS32-AE-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]]
-  ; MIPS32-AE-NOT: sra
+  ; ALL-DAG: splat.w $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
+  ; ALL-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]]
+  ; ALL-NOT: sra
 
   ret i32 %4
-  ; MIPS32-AE: .size extract_sext_v4i32_vidx
 }
 
 define i64 @extract_sext_v2i64_vidx() nounwind {
-  ; MIPS32-AE-LABEL: extract_sext_v2i64_vidx:
+  ; ALL-LABEL: extract_sext_v2i64_vidx:
 
-  %1 = load <2 x i64>* @v2i64
-  ; MIPS32-AE-DAG: lw [[PTR_V:\$[0-9]+]], %got(v2i64)(
-  ; MIPS32-AE-DAG: ld.d [[R1:\$w[0-9]+]], 0([[PTR_V]])
+  %1 = load <2 x i64>, <2 x i64>* @v2i64
+  ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v2i64)(
+  ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v2i64)(
+  ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v2i64)(
+  ; ALL-DAG: ld.d [[R1:\$w[0-9]+]], 0([[PTR_V]])
 
   %2 = add <2 x i64> %1, %1
-  ; MIPS32-AE-DAG: addv.d [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: addv.d [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
-  %3 = load i32* @i32
-  ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
-  ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
+  %3 = load i32, i32* @i32
+  ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
+  ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
 
   %4 = extractelement <2 x i64> %2, i32 %3
-  ; MIPS32-AE-DAG: splat.w $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
-  ; MIPS32-AE-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]]
-  ; MIPS32-AE-DAG: splat.w $w[[R4:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
-  ; MIPS32-AE-DAG: mfc1 [[R6:\$[0-9]+]], $f[[R4]]
-  ; MIPS32-AE-NOT: sra
+  ; MIPS32-DAG: splat.w $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
+  ; MIPS32-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]]
+  ; MIPS32-DAG: splat.w $w[[R4:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
+  ; MIPS32-DAG: mfc1 [[R6:\$[0-9]+]], $f[[R4]]
+  ; MIPS64-DAG: splat.d $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
+  ; MIPS64-DAG: dmfc1 [[R5:\$[0-9]+]], $f[[R3]]
+  ; ALL-NOT: sra
 
   ret i64 %4
-  ; MIPS32-AE: .size extract_sext_v2i64_vidx
 }
 
 define i32 @extract_zext_v16i8_vidx() nounwind {
-  ; MIPS32-AE-LABEL: extract_zext_v16i8_vidx:
+  ; ALL-LABEL: extract_zext_v16i8_vidx:
 
-  %1 = load <16 x i8>* @v16i8
-  ; MIPS32-AE-DAG: lw [[PTR_V:\$[0-9]+]], %got(v16i8)(
-  ; MIPS32-AE-DAG: ld.b [[R1:\$w[0-9]+]], 0([[PTR_V]])
+  %1 = load <16 x i8>, <16 x i8>* @v16i8
+  ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v16i8)(
+  ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v16i8)(
+  ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v16i8)(
+  ; ALL-DAG: ld.b [[R1:\$w[0-9]+]], 0([[PTR_V]])
 
   %2 = add <16 x i8> %1, %1
-  ; MIPS32-AE-DAG: addv.b [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: addv.b [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
-  %3 = load i32* @i32
-  ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
-  ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
+  %3 = load i32, i32* @i32
+  ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
+  ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
 
   %4 = extractelement <16 x i8> %2, i32 %3
   %5 = zext i8 %4 to i32
-  ; MIPS32-AE-DAG: splat.b $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
-  ; MIPS32-AE-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]]
-  ; MIPS32-AE-DAG: srl [[R6:\$[0-9]+]], [[R5]], 24
+  ; ALL-DAG: splat.b $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
+  ; ALL-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]]
+  ; ALL-DAG: srl [[R6:\$[0-9]+]], [[R5]], 24
 
   ret i32 %5
-  ; MIPS32-AE: .size extract_zext_v16i8_vidx
 }
 
 define i32 @extract_zext_v8i16_vidx() nounwind {
-  ; MIPS32-AE-LABEL: extract_zext_v8i16_vidx:
+  ; ALL-LABEL: extract_zext_v8i16_vidx:
 
-  %1 = load <8 x i16>* @v8i16
-  ; MIPS32-AE-DAG: lw [[PTR_V:\$[0-9]+]], %got(v8i16)(
-  ; MIPS32-AE-DAG: ld.h [[R1:\$w[0-9]+]], 0([[PTR_V]])
+  %1 = load <8 x i16>, <8 x i16>* @v8i16
+  ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v8i16)(
+  ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v8i16)(
+  ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v8i16)(
+  ; ALL-DAG: ld.h [[R1:\$w[0-9]+]], 0([[PTR_V]])
 
   %2 = add <8 x i16> %1, %1
-  ; MIPS32-AE-DAG: addv.h [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: addv.h [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
-  %3 = load i32* @i32
-  ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
-  ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
+  %3 = load i32, i32* @i32
+  ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
+  ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
 
   %4 = extractelement <8 x i16> %2, i32 %3
   %5 = zext i16 %4 to i32
-  ; MIPS32-AE-DAG: splat.h $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
-  ; MIPS32-AE-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]]
-  ; MIPS32-AE-DAG: srl [[R6:\$[0-9]+]], [[R5]], 16
+  ; ALL-DAG: splat.h $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
+  ; ALL-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]]
+  ; ALL-DAG: srl [[R6:\$[0-9]+]], [[R5]], 16
 
   ret i32 %5
-  ; MIPS32-AE: .size extract_zext_v8i16_vidx
 }
 
 define i32 @extract_zext_v4i32_vidx() nounwind {
-  ; MIPS32-AE-LABEL: extract_zext_v4i32_vidx:
+  ; ALL-LABEL: extract_zext_v4i32_vidx:
 
-  %1 = load <4 x i32>* @v4i32
-  ; MIPS32-AE-DAG: lw [[PTR_V:\$[0-9]+]], %got(v4i32)(
-  ; MIPS32-AE-DAG: ld.w [[R1:\$w[0-9]+]], 0([[PTR_V]])
+  %1 = load <4 x i32>, <4 x i32>* @v4i32
+  ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v4i32)(
+  ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v4i32)(
+  ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v4i32)(
+  ; ALL-DAG: ld.w [[R1:\$w[0-9]+]], 0([[PTR_V]])
 
   %2 = add <4 x i32> %1, %1
-  ; MIPS32-AE-DAG: addv.w [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: addv.w [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
-  %3 = load i32* @i32
-  ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
-  ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
+  %3 = load i32, i32* @i32
+  ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
+  ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
 
   %4 = extractelement <4 x i32> %2, i32 %3
-  ; MIPS32-AE-DAG: splat.w $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
-  ; MIPS32-AE-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]]
-  ; MIPS32-AE-NOT: srl
+  ; ALL-DAG: splat.w $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
+  ; ALL-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]]
+  ; ALL-NOT: srl
 
   ret i32 %4
-  ; MIPS32-AE: .size extract_zext_v4i32_vidx
 }
 
 define i64 @extract_zext_v2i64_vidx() nounwind {
-  ; MIPS32-AE-LABEL: extract_zext_v2i64_vidx:
+  ; ALL-LABEL: extract_zext_v2i64_vidx:
 
-  %1 = load <2 x i64>* @v2i64
-  ; MIPS32-AE-DAG: lw [[PTR_V:\$[0-9]+]], %got(v2i64)(
-  ; MIPS32-AE-DAG: ld.d [[R1:\$w[0-9]+]], 0([[PTR_V]])
+  %1 = load <2 x i64>, <2 x i64>* @v2i64
+  ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v2i64)(
+  ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v2i64)(
+  ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v2i64)(
+  ; ALL-DAG: ld.d [[R1:\$w[0-9]+]], 0([[PTR_V]])
 
   %2 = add <2 x i64> %1, %1
-  ; MIPS32-AE-DAG: addv.d [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: addv.d [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
-  %3 = load i32* @i32
-  ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
-  ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
+  %3 = load i32, i32* @i32
+  ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
+  ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
 
   %4 = extractelement <2 x i64> %2, i32 %3
-  ; MIPS32-AE-DAG: splat.w $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
-  ; MIPS32-AE-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]]
-  ; MIPS32-AE-DAG: splat.w $w[[R4:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
-  ; MIPS32-AE-DAG: mfc1 [[R6:\$[0-9]+]], $f[[R4]]
-  ; MIPS32-AE-NOT: srl
+  ; MIPS32-DAG: splat.w $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
+  ; MIPS32-DAG: mfc1 [[R5:\$[0-9]+]], $f[[R3]]
+  ; MIPS32-DAG: splat.w $w[[R4:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
+  ; MIPS32-DAG: mfc1 [[R6:\$[0-9]+]], $f[[R4]]
+  ; MIPS64-DAG: splat.d $w[[R3:[0-9]+]], [[R1]]{{\[}}[[IDX]]]
+  ; MIPS64-DAG: dmfc1 [[R5:\$[0-9]+]], $f[[R3]]
+  ; ALL-NOT: srl
 
   ret i64 %4
-  ; MIPS32-AE: .size extract_zext_v2i64_vidx
 }
 
-define void @insert_v16i8(i32 %a) nounwind {
-  ; MIPS32-AE-LABEL: insert_v16i8:
+define void @insert_v16i8(i32 signext %a) nounwind {
+  ; ALL-LABEL: insert_v16i8:
 
-  %1 = load <16 x i8>* @v16i8
-  ; MIPS32-AE-DAG: ld.b [[R1:\$w[0-9]+]],
+  %1 = load <16 x i8>, <16 x i8>* @v16i8
+  ; ALL-DAG: ld.b [[R1:\$w[0-9]+]],
 
   %a2 = trunc i32 %a to i8
   %a3 = sext i8 %a2 to i32
   %a4 = trunc i32 %a3 to i8
-  ; MIPS32-AE-NOT: andi
-  ; MIPS32-AE-NOT: sra
+  ; ALL-NOT: andi
+  ; ALL-NOT: sra
 
   %2 = insertelement <16 x i8> %1, i8 %a4, i32 1
-  ; MIPS32-AE-DAG: insert.b [[R1]][1], $4
+  ; ALL-DAG: insert.b [[R1]][1], $4
 
   store <16 x i8> %2, <16 x i8>* @v16i8
-  ; MIPS32-AE-DAG: st.b [[R1]]
+  ; ALL-DAG: st.b [[R1]]
 
   ret void
-  ; MIPS32-AE: .size insert_v16i8
 }
 
-define void @insert_v8i16(i32 %a) nounwind {
-  ; MIPS32-AE-LABEL: insert_v8i16:
+define void @insert_v8i16(i32 signext %a) nounwind {
+  ; ALL-LABEL: insert_v8i16:
 
-  %1 = load <8 x i16>* @v8i16
-  ; MIPS32-AE-DAG: ld.h [[R1:\$w[0-9]+]],
+  %1 = load <8 x i16>, <8 x i16>* @v8i16
+  ; ALL-DAG: ld.h [[R1:\$w[0-9]+]],
 
   %a2 = trunc i32 %a to i16
   %a3 = sext i16 %a2 to i32
   %a4 = trunc i32 %a3 to i16
-  ; MIPS32-AE-NOT: andi
-  ; MIPS32-AE-NOT: sra
+  ; ALL-NOT: andi
+  ; ALL-NOT: sra
 
   %2 = insertelement <8 x i16> %1, i16 %a4, i32 1
-  ; MIPS32-AE-DAG: insert.h [[R1]][1], $4
+  ; ALL-DAG: insert.h [[R1]][1], $4
 
   store <8 x i16> %2, <8 x i16>* @v8i16
-  ; MIPS32-AE-DAG: st.h [[R1]]
+  ; ALL-DAG: st.h [[R1]]
 
   ret void
-  ; MIPS32-AE: .size insert_v8i16
 }
 
-define void @insert_v4i32(i32 %a) nounwind {
-  ; MIPS32-AE-LABEL: insert_v4i32:
+define void @insert_v4i32(i32 signext %a) nounwind {
+  ; ALL-LABEL: insert_v4i32:
 
-  %1 = load <4 x i32>* @v4i32
-  ; MIPS32-AE-DAG: ld.w [[R1:\$w[0-9]+]],
+  %1 = load <4 x i32>, <4 x i32>* @v4i32
+  ; ALL-DAG: ld.w [[R1:\$w[0-9]+]],
 
-  ; MIPS32-AE-NOT: andi
-  ; MIPS32-AE-NOT: sra
+  ; ALL-NOT: andi
+  ; ALL-NOT: sra
 
   %2 = insertelement <4 x i32> %1, i32 %a, i32 1
-  ; MIPS32-AE-DAG: insert.w [[R1]][1], $4
+  ; ALL-DAG: insert.w [[R1]][1], $4
 
   store <4 x i32> %2, <4 x i32>* @v4i32
-  ; MIPS32-AE-DAG: st.w [[R1]]
+  ; ALL-DAG: st.w [[R1]]
 
   ret void
-  ; MIPS32-AE: .size insert_v4i32
 }
 
-define void @insert_v2i64(i64 %a) nounwind {
-  ; MIPS32-AE-LABEL: insert_v2i64:
+define void @insert_v2i64(i64 signext %a) nounwind {
+  ; ALL-LABEL: insert_v2i64:
 
-  %1 = load <2 x i64>* @v2i64
-  ; MIPS32-AE-DAG: ld.w [[R1:\$w[0-9]+]],
+  %1 = load <2 x i64>, <2 x i64>* @v2i64
+  ; MIPS32-DAG: ld.w [[R1:\$w[0-9]+]],
+  ; MIPS64-DAG: ld.d [[R1:\$w[0-9]+]],
 
-  ; MIPS32-AE-NOT: andi
-  ; MIPS32-AE-NOT: sra
+  ; ALL-NOT: andi
+  ; ALL-NOT: sra
 
   %2 = insertelement <2 x i64> %1, i64 %a, i32 1
-  ; MIPS32-AE-DAG: insert.w [[R1]][2], $4
-  ; MIPS32-AE-DAG: insert.w [[R1]][3], $5
+  ; MIPS32-DAG: insert.w [[R1]][2], $4
+  ; MIPS32-DAG: insert.w [[R1]][3], $5
+  ; MIPS64-DAG: insert.d [[R1]][1], $4
 
   store <2 x i64> %2, <2 x i64>* @v2i64
-  ; MIPS32-AE-DAG: st.w [[R1]]
+  ; MIPS32-DAG: st.w [[R1]]
+  ; MIPS64-DAG: st.d [[R1]]
 
   ret void
-  ; MIPS32-AE: .size insert_v2i64
 }
 
-define void @insert_v16i8_vidx(i32 %a) nounwind {
-  ; MIPS32-AE: insert_v16i8_vidx:
+define void @insert_v16i8_vidx(i32 signext %a) nounwind {
+  ; ALL-LABEL: insert_v16i8_vidx:
 
-  %1 = load <16 x i8>* @v16i8
-  ; MIPS32-AE-DAG: ld.b [[R1:\$w[0-9]+]],
+  %1 = load <16 x i8>, <16 x i8>* @v16i8
+  ; ALL-DAG: ld.b [[R1:\$w[0-9]+]],
 
-  %2 = load i32* @i32
-  ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
-  ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
+  %2 = load i32, i32* @i32
+  ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
+  ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
 
   %a2 = trunc i32 %a to i8
   %a3 = sext i8 %a2 to i32
   %a4 = trunc i32 %a3 to i8
-  ; MIPS32-AE-NOT: andi
-  ; MIPS32-AE-NOT: sra
+  ; ALL-NOT: andi
+  ; ALL-NOT: sra
 
   %3 = insertelement <16 x i8> %1, i8 %a4, i32 %2
-  ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[IDX]]]
-  ; MIPS32-AE-DAG: insert.b [[R1]][0], $4
-  ; MIPS32-AE-DAG: neg [[NIDX:\$[0-9]+]], [[IDX]]
-  ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]]
+  ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[IDX]]]
+  ; ALL-DAG: insert.b [[R1]][0], $4
+  ; O32-DAG: neg [[NIDX:\$[0-9]+]], [[IDX]]
+  ; N32-DAG: neg [[NIDX:\$[0-9]+]], [[IDX]]
+  ; N64-DAG: dneg [[NIDX:\$[0-9]+]], [[IDX]]
+  ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]]
 
   store <16 x i8> %3, <16 x i8>* @v16i8
-  ; MIPS32-AE-DAG: st.b [[R1]]
+  ; ALL-DAG: st.b [[R1]]
 
   ret void
-  ; MIPS32-AE: .size insert_v16i8_vidx
 }
 
-define void @insert_v8i16_vidx(i32 %a) nounwind {
-  ; MIPS32-AE: insert_v8i16_vidx:
+define void @insert_v8i16_vidx(i32 signext %a) nounwind {
+  ; ALL-LABEL: insert_v8i16_vidx:
 
-  %1 = load <8 x i16>* @v8i16
-  ; MIPS32-AE-DAG: ld.h [[R1:\$w[0-9]+]],
+  %1 = load <8 x i16>, <8 x i16>* @v8i16
+  ; ALL-DAG: ld.h [[R1:\$w[0-9]+]],
 
-  %2 = load i32* @i32
-  ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
-  ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
+  %2 = load i32, i32* @i32
+  ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
+  ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
 
   %a2 = trunc i32 %a to i16
   %a3 = sext i16 %a2 to i32
   %a4 = trunc i32 %a3 to i16
-  ; MIPS32-AE-NOT: andi
-  ; MIPS32-AE-NOT: sra
+  ; ALL-NOT: andi
+  ; ALL-NOT: sra
 
   %3 = insertelement <8 x i16> %1, i16 %a4, i32 %2
-  ; MIPS32-AE-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 1
-  ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]]
-  ; MIPS32-AE-DAG: insert.h [[R1]][0], $4
-  ; MIPS32-AE-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]]
-  ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]]
+  ; ALL-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 1
+  ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]]
+  ; ALL-DAG: insert.h [[R1]][0], $4
+  ; O32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]]
+  ; N32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]]
+  ; N64-DAG: dneg [[NIDX:\$[0-9]+]], [[BIDX]]
+  ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]]
 
   store <8 x i16> %3, <8 x i16>* @v8i16
-  ; MIPS32-AE-DAG: st.h [[R1]]
+  ; ALL-DAG: st.h [[R1]]
 
   ret void
-  ; MIPS32-AE: .size insert_v8i16_vidx
 }
 
-define void @insert_v4i32_vidx(i32 %a) nounwind {
-  ; MIPS32-AE: insert_v4i32_vidx:
+define void @insert_v4i32_vidx(i32 signext %a) nounwind {
+  ; ALL-LABEL: insert_v4i32_vidx:
 
-  %1 = load <4 x i32>* @v4i32
-  ; MIPS32-AE-DAG: ld.w [[R1:\$w[0-9]+]],
+  %1 = load <4 x i32>, <4 x i32>* @v4i32
+  ; ALL-DAG: ld.w [[R1:\$w[0-9]+]],
 
-  %2 = load i32* @i32
-  ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
-  ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
+  %2 = load i32, i32* @i32
+  ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
+  ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
 
-  ; MIPS32-AE-NOT: andi
-  ; MIPS32-AE-NOT: sra
+  ; ALL-NOT: andi
+  ; ALL-NOT: sra
 
   %3 = insertelement <4 x i32> %1, i32 %a, i32 %2
-  ; MIPS32-AE-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 2
-  ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]]
-  ; MIPS32-AE-DAG: insert.w [[R1]][0], $4
-  ; MIPS32-AE-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]]
-  ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]]
+  ; ALL-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 2
+  ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]]
+  ; ALL-DAG: insert.w [[R1]][0], $4
+  ; O32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]]
+  ; N32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]]
+  ; N64-DAG: dneg [[NIDX:\$[0-9]+]], [[BIDX]]
+  ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]]
 
   store <4 x i32> %3, <4 x i32>* @v4i32
-  ; MIPS32-AE-DAG: st.w [[R1]]
+  ; ALL-DAG: st.w [[R1]]
 
   ret void
-  ; MIPS32-AE: .size insert_v4i32_vidx
 }
 
-define void @insert_v2i64_vidx(i64 %a) nounwind {
-  ; MIPS32-AE: insert_v2i64_vidx:
+define void @insert_v2i64_vidx(i64 signext %a) nounwind {
+  ; ALL-LABEL: insert_v2i64_vidx:
 
-  %1 = load <2 x i64>* @v2i64
-  ; MIPS32-AE-DAG: ld.w [[R1:\$w[0-9]+]],
+  %1 = load <2 x i64>, <2 x i64>* @v2i64
+  ; MIPS32-DAG: ld.w [[R1:\$w[0-9]+]],
+  ; MIPS64-DAG: ld.d [[R1:\$w[0-9]+]],
 
-  %2 = load i32* @i32
-  ; MIPS32-AE-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
-  ; MIPS32-AE-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
+  %2 = load i32, i32* @i32
+  ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
+  ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
 
-  ; MIPS32-AE-NOT: andi
-  ; MIPS32-AE-NOT: sra
+  ; ALL-NOT: andi
+  ; ALL-NOT: sra
 
   %3 = insertelement <2 x i64> %1, i64 %a, i32 %2
   ; TODO: This code could be a lot better but it works. The legalizer splits
   ; 64-bit inserts into two 32-bit inserts because there is no i64 type on
   ; MIPS32. The obvious optimisation is to perform both insert.w's at once while
   ; the vector is rotated.
-  ; MIPS32-AE-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 2
-  ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]]
-  ; MIPS32-AE-DAG: insert.w [[R1]][0], $4
-  ; MIPS32-AE-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]]
-  ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]]
-  ; MIPS32-AE-DAG: addiu [[IDX2:\$[0-9]+]], [[IDX]], 1
-  ; MIPS32-AE-DAG: sll [[BIDX:\$[0-9]+]], [[IDX2]], 2
-  ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]]
-  ; MIPS32-AE-DAG: insert.w [[R1]][0], $5
-  ; MIPS32-AE-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]]
-  ; MIPS32-AE-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]]
+  ; MIPS32-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 2
+  ; MIPS32-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]]
+  ; MIPS32-DAG: insert.w [[R1]][0], $4
+  ; MIPS32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]]
+  ; MIPS32-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]]
+  ; MIPS32-DAG: addiu [[IDX2:\$[0-9]+]], [[IDX]], 1
+  ; MIPS32-DAG: sll [[BIDX:\$[0-9]+]], [[IDX2]], 2
+  ; MIPS32-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]]
+  ; MIPS32-DAG: insert.w [[R1]][0], $5
+  ; MIPS32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]]
+  ; MIPS32-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]]
+
+  ; MIPS64-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 3
+  ; MIPS64-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]]
+  ; MIPS64-DAG: insert.d [[R1]][0], $4
+  ; N32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]]
+  ; N64-DAG: dneg [[NIDX:\$[0-9]+]], [[BIDX]]
+  ; MIPS64-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]]
 
   store <2 x i64> %3, <2 x i64>* @v2i64
-  ; MIPS32-AE-DAG: st.w [[R1]]
+  ; MIPS32-DAG: st.w [[R1]]
+  ; MIPS64-DAG: st.d [[R1]]
 
   ret void
-  ; MIPS32-AE: .size insert_v2i64_vidx
 }
 
 define void @truncstore() nounwind {
-  ; MIPS32-AE-LABEL: truncstore:
+  ; ALL-LABEL: truncstore:
 
   store volatile <4 x i8> <i8 -1, i8 -1, i8 -1, i8 -1>, <4 x i8>*@v4i8
   ; TODO: What code should be emitted?
 
   ret void
-  ; MIPS32-AE: .size truncstore
 }
diff --git a/test/CodeGen/Mips/msa/basic_operations_float.ll b/test/CodeGen/Mips/msa/basic_operations_float.ll
index a0c9d29e231a..f19cb9b7c2e5 100644
--- a/test/CodeGen/Mips/msa/basic_operations_float.ll
+++ b/test/CodeGen/Mips/msa/basic_operations_float.ll
@@ -1,5 +1,9 @@
-; RUN: llc -march=mips -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=MIPS32 %s
-; RUN: llc -march=mipsel -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=MIPS32 %s
+; RUN: llc -march=mips -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=O32 %s
+; RUN: llc -march=mipsel -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=O32 %s
+; RUN: llc -march=mips64 -target-abi=n32 -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=N32 %s
+; RUN: llc -march=mips64el -target-abi=n32 -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=N32 %s
+; RUN: llc -march=mips64 -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=N64 %s
+; RUN: llc -march=mips64el -mattr=+msa,+fp64 < %s | FileCheck -check-prefix=ALL -check-prefix=N64 %s
 
 @v4f32 = global <4 x float> <float 0.0, float 0.0, float 0.0, float 0.0>
 @v2f64 = global <2 x double> <double 0.0, double 0.0>
@@ -8,322 +12,341 @@
 @f64 = global double 0.0
 
 define void @const_v4f32() nounwind {
-  ; MIPS32-LABEL: const_v4f32:
+  ; ALL-LABEL: const_v4f32:
 
   store volatile <4 x float> <float 0.0, float 0.0, float 0.0, float 0.0>, <4 x float>*@v4f32
-  ; MIPS32: ldi.b  [[R1:\$w[0-9]+]], 0
+  ; ALL: ldi.b  [[R1:\$w[0-9]+]], 0
 
   store volatile <4 x float> <float 1.0, float 1.0, float 1.0, float 1.0>, <4 x float>*@v4f32
-  ; MIPS32: lui     [[R1:\$[0-9]+]], 16256
-  ; MIPS32: fill.w  [[R2:\$w[0-9]+]], [[R1]]
+  ; ALL: lui     [[R1:\$[0-9]+]], 16256
+  ; ALL: fill.w  [[R2:\$w[0-9]+]], [[R1]]
 
   store volatile <4 x float> <float 1.0, float 1.0, float 1.0, float 31.0>, <4 x float>*@v4f32
-  ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32: ld.w  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; ALL: ld.w  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   store volatile <4 x float> <float 65537.0, float 65537.0, float 65537.0, float 65537.0>, <4 x float>*@v4f32
-  ; MIPS32: lui     [[R1:\$[0-9]+]], 18304
-  ; MIPS32: ori     [[R2:\$[0-9]+]], [[R1]], 128
-  ; MIPS32: fill.w  [[R3:\$w[0-9]+]], [[R2]]
+  ; ALL: lui     [[R1:\$[0-9]+]], 18304
+  ; ALL: ori     [[R2:\$[0-9]+]], [[R1]], 128
+  ; ALL: fill.w  [[R3:\$w[0-9]+]], [[R2]]
 
   store volatile <4 x float> <float 1.0, float 2.0, float 1.0, float 2.0>, <4 x float>*@v4f32
-  ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32: ld.w  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; ALL: ld.w  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   store volatile <4 x float> <float 3.0, float 4.0, float 5.0, float 6.0>, <4 x float>*@v4f32
-  ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32: ld.w  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; ALL: ld.w  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   ret void
-  ; MIPS32: .size const_v4f32
 }
 
 define void @const_v2f64() nounwind {
-  ; MIPS32-LABEL: const_v2f64:
+  ; ALL-LABEL: const_v2f64:
 
   store volatile <2 x double> <double 0.0, double 0.0>, <2 x double>*@v2f64
-  ; MIPS32: ldi.b  [[R1:\$w[0-9]+]], 0
+  ; ALL: ldi.b  [[R1:\$w[0-9]+]], 0
 
   store volatile <2 x double> <double 72340172838076673.0, double 72340172838076673.0>, <2 x double>*@v2f64
-  ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32: ld.d  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; ALL: ld.d  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   store volatile <2 x double> <double 281479271743489.0, double 281479271743489.0>, <2 x double>*@v2f64
-  ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32: ld.d  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; ALL: ld.d  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   store volatile <2 x double> <double 4294967297.0, double 4294967297.0>, <2 x double>*@v2f64
-  ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32: ld.d  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; ALL: ld.d  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   store volatile <2 x double> <double 1.0, double 1.0>, <2 x double>*@v2f64
-  ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32: ld.d  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; ALL: ld.d  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   store volatile <2 x double> <double 1.0, double 31.0>, <2 x double>*@v2f64
-  ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32: ld.d  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; ALL: ld.d  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   store volatile <2 x double> <double 3.0, double 4.0>, <2 x double>*@v2f64
-  ; MIPS32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
-  ; MIPS32: ld.d  [[R1:\$w[0-9]+]], 0([[G_PTR]])
+  ; O32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %lo($
+  ; N32: addiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; N64: daddiu [[G_PTR:\$[0-9]+]], {{.*}}, %got_ofst($
+  ; ALL: ld.d  [[R1:\$w[0-9]+]], 0([[G_PTR]])
 
   ret void
-  ; MIPS32: .size const_v2f64
 }
 
 define void @nonconst_v4f32() nounwind {
-  ; MIPS32-LABEL: nonconst_v4f32:
+  ; ALL-LABEL: nonconst_v4f32:
 
-  %1 = load float *@f32
+  %1 = load float , float *@f32
   %2 = insertelement <4 x float> undef, float %1, i32 0
   %3 = insertelement <4 x float> %2, float %1, i32 1
   %4 = insertelement <4 x float> %3, float %1, i32 2
   %5 = insertelement <4 x float> %4, float %1, i32 3
   store volatile <4 x float> %5, <4 x float>*@v4f32
-  ; MIPS32: lwc1 $f[[R1:[0-9]+]], 0(
-  ; MIPS32: splati.w [[R2:\$w[0-9]+]], $w[[R1]]
+  ; ALL: lwc1 $f[[R1:[0-9]+]], 0(
+  ; ALL: splati.w [[R2:\$w[0-9]+]], $w[[R1]]
 
   ret void
-  ; MIPS32: .size nonconst_v4f32
 }
 
 define void @nonconst_v2f64() nounwind {
-  ; MIPS32-LABEL: nonconst_v2f64:
+  ; ALL-LABEL: nonconst_v2f64:
 
-  %1 = load double *@f64
+  %1 = load double , double *@f64
   %2 = insertelement <2 x double> undef, double %1, i32 0
   %3 = insertelement <2 x double> %2, double %1, i32 1
   store volatile <2 x double> %3, <2 x double>*@v2f64
-  ; MIPS32: ldc1 $f[[R1:[0-9]+]], 0(
-  ; MIPS32: splati.d [[R2:\$w[0-9]+]], $w[[R1]]
+  ; ALL: ldc1 $f[[R1:[0-9]+]], 0(
+  ; ALL: splati.d [[R2:\$w[0-9]+]], $w[[R1]]
 
   ret void
-  ; MIPS32: .size nonconst_v2f64
 }
 
 define float @extract_v4f32() nounwind {
-  ; MIPS32-LABEL: extract_v4f32:
+  ; ALL-LABEL: extract_v4f32:
 
-  %1 = load <4 x float>* @v4f32
-  ; MIPS32-DAG: ld.w [[R1:\$w[0-9]+]],
+  %1 = load <4 x float>, <4 x float>* @v4f32
+  ; ALL-DAG: ld.w [[R1:\$w[0-9]+]],
 
   %2 = fadd <4 x float> %1, %1
-  ; MIPS32-DAG: fadd.w [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: fadd.w [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
   %3 = extractelement <4 x float> %2, i32 1
   ; Element 1 can be obtained by splatting it across the vector and extracting
   ; $w0:sub_lo
-  ; MIPS32-DAG: splati.w $w0, [[R1]][1]
+  ; ALL-DAG: splati.w $w0, [[R1]][1]
 
   ret float %3
-  ; MIPS32: .size extract_v4f32
 }
 
 define float @extract_v4f32_elt0() nounwind {
-  ; MIPS32-LABEL: extract_v4f32_elt0:
+  ; ALL-LABEL: extract_v4f32_elt0:
 
-  %1 = load <4 x float>* @v4f32
-  ; MIPS32-DAG: ld.w [[R1:\$w[0-9]+]],
+  %1 = load <4 x float>, <4 x float>* @v4f32
+  ; ALL-DAG: ld.w [[R1:\$w[0-9]+]],
 
   %2 = fadd <4 x float> %1, %1
-  ; MIPS32-DAG: fadd.w $w0, [[R1]], [[R1]]
+  ; ALL-DAG: fadd.w $w0, [[R1]], [[R1]]
 
   %3 = extractelement <4 x float> %2, i32 0
   ; Element 0 can be obtained by extracting $w0:sub_lo ($f0)
-  ; MIPS32-NOT: copy_u.w
-  ; MIPS32-NOT: mtc1
+  ; ALL-NOT: copy_u.w
+  ; ALL-NOT: mtc1
 
   ret float %3
-  ; MIPS32: .size extract_v4f32_elt0
 }
 
 define float @extract_v4f32_elt2() nounwind {
-  ; MIPS32-LABEL: extract_v4f32_elt2:
+  ; ALL-LABEL: extract_v4f32_elt2:
 
-  %1 = load <4 x float>* @v4f32
-  ; MIPS32-DAG: ld.w [[R1:\$w[0-9]+]],
+  %1 = load <4 x float>, <4 x float>* @v4f32
+  ; ALL-DAG: ld.w [[R1:\$w[0-9]+]],
 
   %2 = fadd <4 x float> %1, %1
-  ; MIPS32-DAG: fadd.w [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: fadd.w [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
   %3 = extractelement <4 x float> %2, i32 2
   ; Element 2 can be obtained by splatting it across the vector and extracting
   ; $w0:sub_lo
-  ; MIPS32-DAG: splati.w $w0, [[R1]][2]
+  ; ALL-DAG: splati.w $w0, [[R1]][2]
 
   ret float %3
-  ; MIPS32: .size extract_v4f32_elt2
 }
 
 define float @extract_v4f32_vidx() nounwind {
-  ; MIPS32-LABEL: extract_v4f32_vidx:
+  ; ALL-LABEL: extract_v4f32_vidx:
 
-  %1 = load <4 x float>* @v4f32
-  ; MIPS32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v4f32)(
-  ; MIPS32-DAG: ld.w [[R1:\$w[0-9]+]], 0([[PTR_V]])
+  %1 = load <4 x float>, <4 x float>* @v4f32
+  ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v4f32)(
+  ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v4f32)(
+  ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v4f32)(
+  ; ALL-DAG: ld.w [[R1:\$w[0-9]+]], 0([[PTR_V]])
 
   %2 = fadd <4 x float> %1, %1
-  ; MIPS32-DAG: fadd.w [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: fadd.w [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
-  %3 = load i32* @i32
-  ; MIPS32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
-  ; MIPS32-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
+  %3 = load i32, i32* @i32
+  ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
+  ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
 
   %4 = extractelement <4 x float> %2, i32 %3
-  ; MIPS32-DAG: splat.w $w0, [[R1]]{{\[}}[[IDX]]]
+  ; ALL-DAG: splat.w $w0, [[R1]]{{\[}}[[IDX]]]
 
   ret float %4
-  ; MIPS32: .size extract_v4f32_vidx
 }
 
 define double @extract_v2f64() nounwind {
-  ; MIPS32-LABEL: extract_v2f64:
+  ; ALL-LABEL: extract_v2f64:
 
-  %1 = load <2 x double>* @v2f64
-  ; MIPS32-DAG: ld.d [[R1:\$w[0-9]+]],
+  %1 = load <2 x double>, <2 x double>* @v2f64
+  ; ALL-DAG: ld.d [[R1:\$w[0-9]+]],
 
   %2 = fadd <2 x double> %1, %1
-  ; MIPS32-DAG: fadd.d [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: fadd.d [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
   %3 = extractelement <2 x double> %2, i32 1
   ; Element 1 can be obtained by splatting it across the vector and extracting
   ; $w0:sub_64
-  ; MIPS32-DAG: splati.d $w0, [[R1]][1]
-  ; MIPS32-NOT: copy_u.w
-  ; MIPS32-NOT: mtc1
-  ; MIPS32-NOT: mthc1
-  ; MIPS32-NOT: sll
-  ; MIPS32-NOT: sra
+  ; ALL-DAG: splati.d $w0, [[R1]][1]
+  ; ALL-NOT: copy_u.w
+  ; ALL-NOT: mtc1
+  ; ALL-NOT: mthc1
+  ; ALL-NOT: sll
+  ; ALL-NOT: sra
 
   ret double %3
-  ; MIPS32: .size extract_v2f64
 }
 
 define double @extract_v2f64_elt0() nounwind {
-  ; MIPS32-LABEL: extract_v2f64_elt0:
+  ; ALL-LABEL: extract_v2f64_elt0:
 
-  %1 = load <2 x double>* @v2f64
-  ; MIPS32-DAG: ld.d [[R1:\$w[0-9]+]],
+  %1 = load <2 x double>, <2 x double>* @v2f64
+  ; ALL-DAG: ld.d [[R1:\$w[0-9]+]],
 
   %2 = fadd <2 x double> %1, %1
-  ; MIPS32-DAG: fadd.d $w0, [[R1]], [[R1]]
+  ; ALL-DAG: fadd.d $w0, [[R1]], [[R1]]
 
   %3 = extractelement <2 x double> %2, i32 0
   ; Element 0 can be obtained by extracting $w0:sub_64 ($f0)
-  ; MIPS32-NOT: copy_u.w
-  ; MIPS32-NOT: mtc1
-  ; MIPS32-NOT: mthc1
-  ; MIPS32-NOT: sll
-  ; MIPS32-NOT: sra
+  ; ALL-NOT: copy_u.w
+  ; ALL-NOT: mtc1
+  ; ALL-NOT: mthc1
+  ; ALL-NOT: sll
+  ; ALL-NOT: sra
 
   ret double %3
-  ; MIPS32: .size extract_v2f64_elt0
 }
 
 define double @extract_v2f64_vidx() nounwind {
-  ; MIPS32-LABEL: extract_v2f64_vidx:
+  ; ALL-LABEL: extract_v2f64_vidx:
 
-  %1 = load <2 x double>* @v2f64
-  ; MIPS32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v2f64)(
-  ; MIPS32-DAG: ld.d [[R1:\$w[0-9]+]], 0([[PTR_V]])
+  %1 = load <2 x double>, <2 x double>* @v2f64
+  ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v2f64)(
+  ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v2f64)(
+  ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v2f64)(
+  ; ALL-DAG: ld.d [[R1:\$w[0-9]+]], 0([[PTR_V]])
 
   %2 = fadd <2 x double> %1, %1
-  ; MIPS32-DAG: fadd.d [[R2:\$w[0-9]+]], [[R1]], [[R1]]
+  ; ALL-DAG: fadd.d [[R2:\$w[0-9]+]], [[R1]], [[R1]]
 
-  %3 = load i32* @i32
-  ; MIPS32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
-  ; MIPS32-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
+  %3 = load i32, i32* @i32
+  ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
+  ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
 
   %4 = extractelement <2 x double> %2, i32 %3
-  ; MIPS32-DAG: splat.d $w0, [[R1]]{{\[}}[[IDX]]]
+  ; ALL-DAG: splat.d $w0, [[R1]]{{\[}}[[IDX]]]
 
   ret double %4
-  ; MIPS32: .size extract_v2f64_vidx
 }
 
 define void @insert_v4f32(float %a) nounwind {
-  ; MIPS32-LABEL: insert_v4f32:
+  ; ALL-LABEL: insert_v4f32:
 
-  %1 = load <4 x float>* @v4f32
-  ; MIPS32-DAG: ld.w [[R1:\$w[0-9]+]],
+  %1 = load <4 x float>, <4 x float>* @v4f32
+  ; ALL-DAG: ld.w [[R1:\$w[0-9]+]],
 
   %2 = insertelement <4 x float> %1, float %a, i32 1
   ; float argument passed in $f12
-  ; MIPS32-DAG: insve.w [[R1]][1], $w12[0]
+  ; ALL-DAG: insve.w [[R1]][1], $w12[0]
 
   store <4 x float> %2, <4 x float>* @v4f32
-  ; MIPS32-DAG: st.w [[R1]]
+  ; ALL-DAG: st.w [[R1]]
 
   ret void
-  ; MIPS32: .size insert_v4f32
 }
 
 define void @insert_v2f64(double %a) nounwind {
-  ; MIPS32-LABEL: insert_v2f64:
+  ; ALL-LABEL: insert_v2f64:
 
-  %1 = load <2 x double>* @v2f64
-  ; MIPS32-DAG: ld.d [[R1:\$w[0-9]+]],
+  %1 = load <2 x double>, <2 x double>* @v2f64
+  ; ALL-DAG: ld.d [[R1:\$w[0-9]+]],
 
   %2 = insertelement <2 x double> %1, double %a, i32 1
   ; double argument passed in $f12
-  ; MIPS32-DAG: insve.d [[R1]][1], $w12[0]
+  ; ALL-DAG: insve.d [[R1]][1], $w12[0]
 
   store <2 x double> %2, <2 x double>* @v2f64
-  ; MIPS32-DAG: st.d [[R1]]
+  ; ALL-DAG: st.d [[R1]]
 
   ret void
-  ; MIPS32: .size insert_v2f64
 }
 
 define void @insert_v4f32_vidx(float %a) nounwind {
-  ; MIPS32-LABEL: insert_v4f32_vidx:
+  ; ALL-LABEL: insert_v4f32_vidx:
 
-  %1 = load <4 x float>* @v4f32
-  ; MIPS32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v4f32)(
-  ; MIPS32-DAG: ld.w [[R1:\$w[0-9]+]], 0([[PTR_V]])
+  %1 = load <4 x float>, <4 x float>* @v4f32
+  ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v4f32)(
+  ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v4f32)(
+  ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v4f32)(
+  ; ALL-DAG: ld.w [[R1:\$w[0-9]+]], 0([[PTR_V]])
 
-  %2 = load i32* @i32
-  ; MIPS32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
-  ; MIPS32-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
+  %2 = load i32, i32* @i32
+  ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
+  ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
 
   %3 = insertelement <4 x float> %1, float %a, i32 %2
   ; float argument passed in $f12
-  ; MIPS32-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 2
-  ; MIPS32-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]]
-  ; MIPS32-DAG: insve.w [[R1]][0], $w12[0]
-  ; MIPS32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]]
-  ; MIPS32-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]]
+  ; ALL-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 2
+  ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]]
+  ; ALL-DAG: insve.w [[R1]][0], $w12[0]
+  ; ALL-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]]
+  ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]]
 
   store <4 x float> %3, <4 x float>* @v4f32
-  ; MIPS32-DAG: st.w [[R1]]
+  ; ALL-DAG: st.w [[R1]]
 
   ret void
-  ; MIPS32: .size insert_v4f32_vidx
 }
 
 define void @insert_v2f64_vidx(double %a) nounwind {
-  ; MIPS32-LABEL: insert_v2f64_vidx:
+  ; ALL-LABEL: insert_v2f64_vidx:
 
-  %1 = load <2 x double>* @v2f64
-  ; MIPS32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v2f64)(
-  ; MIPS32-DAG: ld.d [[R1:\$w[0-9]+]], 0([[PTR_V]])
+  %1 = load <2 x double>, <2 x double>* @v2f64
+  ; O32-DAG: lw [[PTR_V:\$[0-9]+]], %got(v2f64)(
+  ; N32-DAG: lw [[PTR_V:\$[0-9]+]], %got_disp(v2f64)(
+  ; N64-DAG: ld [[PTR_V:\$[0-9]+]], %got_disp(v2f64)(
+  ; ALL-DAG: ld.d [[R1:\$w[0-9]+]], 0([[PTR_V]])
 
-  %2 = load i32* @i32
-  ; MIPS32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
-  ; MIPS32-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
+  %2 = load i32, i32* @i32
+  ; O32-DAG: lw [[PTR_I:\$[0-9]+]], %got(i32)(
+  ; N32-DAG: lw [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; N64-DAG: ld [[PTR_I:\$[0-9]+]], %got_disp(i32)(
+  ; ALL-DAG: lw [[IDX:\$[0-9]+]], 0([[PTR_I]])
 
   %3 = insertelement <2 x double> %1, double %a, i32 %2
   ; double argument passed in $f12
-  ; MIPS32-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 3
-  ; MIPS32-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]]
-  ; MIPS32-DAG: insve.d [[R1]][0], $w12[0]
-  ; MIPS32-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]]
-  ; MIPS32-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]]
+  ; ALL-DAG: sll [[BIDX:\$[0-9]+]], [[IDX]], 3
+  ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[BIDX]]]
+  ; ALL-DAG: insve.d [[R1]][0], $w12[0]
+  ; ALL-DAG: neg [[NIDX:\$[0-9]+]], [[BIDX]]
+  ; ALL-DAG: sld.b [[R1]], [[R1]]{{\[}}[[NIDX]]]
 
   store <2 x double> %3, <2 x double>* @v2f64
-  ; MIPS32-DAG: st.d [[R1]]
+  ; ALL-DAG: st.d [[R1]]
 
   ret void
-  ; MIPS32: .size insert_v2f64_vidx
 }
diff --git a/test/CodeGen/Mips/msa/bit.ll b/test/CodeGen/Mips/msa/bit.ll
index 59ddbe17a33f..f0057307bbf7 100644
--- a/test/CodeGen/Mips/msa/bit.ll
+++ b/test/CodeGen/Mips/msa/bit.ll
@@ -8,7 +8,7 @@
 
 define void @llvm_mips_sat_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_sat_s_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_sat_s_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.sat.s.b(<16 x i8> %0, i32 7)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_sat_s_b_RES
   ret void
@@ -27,7 +27,7 @@ declare <16 x i8> @llvm.mips.sat.s.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_sat_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_sat_s_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_sat_s_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.sat.s.h(<8 x i16> %0, i32 7)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_sat_s_h_RES
   ret void
@@ -46,7 +46,7 @@ declare <8 x i16> @llvm.mips.sat.s.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_sat_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_sat_s_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_sat_s_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.sat.s.w(<4 x i32> %0, i32 7)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_sat_s_w_RES
   ret void
@@ -65,7 +65,7 @@ declare <4 x i32> @llvm.mips.sat.s.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_sat_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_sat_s_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_sat_s_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.sat.s.d(<2 x i64> %0, i32 7)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_sat_s_d_RES
   ret void
@@ -84,7 +84,7 @@ declare <2 x i64> @llvm.mips.sat.s.d(<2 x i64>, i32) nounwind
 
 define void @llvm_mips_sat_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_sat_u_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_sat_u_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.sat.u.b(<16 x i8> %0, i32 7)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_sat_u_b_RES
   ret void
@@ -103,7 +103,7 @@ declare <16 x i8> @llvm.mips.sat.u.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_sat_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_sat_u_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_sat_u_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.sat.u.h(<8 x i16> %0, i32 7)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_sat_u_h_RES
   ret void
@@ -122,7 +122,7 @@ declare <8 x i16> @llvm.mips.sat.u.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_sat_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_sat_u_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_sat_u_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.sat.u.w(<4 x i32> %0, i32 7)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_sat_u_w_RES
   ret void
@@ -141,7 +141,7 @@ declare <4 x i32> @llvm.mips.sat.u.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_sat_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_sat_u_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_sat_u_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.sat.u.d(<2 x i64> %0, i32 7)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_sat_u_d_RES
   ret void
@@ -160,7 +160,7 @@ declare <2 x i64> @llvm.mips.sat.u.d(<2 x i64>, i32) nounwind
 
 define void @llvm_mips_slli_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_slli_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_slli_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.slli.b(<16 x i8> %0, i32 7)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_slli_b_RES
   ret void
@@ -179,7 +179,7 @@ declare <16 x i8> @llvm.mips.slli.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_slli_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_slli_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_slli_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.slli.h(<8 x i16> %0, i32 7)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_slli_h_RES
   ret void
@@ -198,7 +198,7 @@ declare <8 x i16> @llvm.mips.slli.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_slli_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_slli_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_slli_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.slli.w(<4 x i32> %0, i32 7)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_slli_w_RES
   ret void
@@ -217,7 +217,7 @@ declare <4 x i32> @llvm.mips.slli.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_slli_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_slli_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_slli_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.slli.d(<2 x i64> %0, i32 7)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_slli_d_RES
   ret void
@@ -236,7 +236,7 @@ declare <2 x i64> @llvm.mips.slli.d(<2 x i64>, i32) nounwind
 
 define void @llvm_mips_srai_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_srai_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_srai_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.srai.b(<16 x i8> %0, i32 7)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_srai_b_RES
   ret void
@@ -255,7 +255,7 @@ declare <16 x i8> @llvm.mips.srai.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_srai_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_srai_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_srai_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.srai.h(<8 x i16> %0, i32 7)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_srai_h_RES
   ret void
@@ -274,7 +274,7 @@ declare <8 x i16> @llvm.mips.srai.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_srai_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_srai_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_srai_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.srai.w(<4 x i32> %0, i32 7)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_srai_w_RES
   ret void
@@ -293,7 +293,7 @@ declare <4 x i32> @llvm.mips.srai.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_srai_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_srai_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_srai_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.srai.d(<2 x i64> %0, i32 7)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_srai_d_RES
   ret void
@@ -312,7 +312,7 @@ declare <2 x i64> @llvm.mips.srai.d(<2 x i64>, i32) nounwind
 
 define void @llvm_mips_srari_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_srari_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_srari_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.srari.b(<16 x i8> %0, i32 7)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_srari_b_RES
   ret void
@@ -331,7 +331,7 @@ declare <16 x i8> @llvm.mips.srari.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_srari_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_srari_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_srari_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.srari.h(<8 x i16> %0, i32 7)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_srari_h_RES
   ret void
@@ -350,7 +350,7 @@ declare <8 x i16> @llvm.mips.srari.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_srari_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_srari_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_srari_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.srari.w(<4 x i32> %0, i32 7)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_srari_w_RES
   ret void
@@ -369,7 +369,7 @@ declare <4 x i32> @llvm.mips.srari.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_srari_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_srari_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_srari_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.srari.d(<2 x i64> %0, i32 7)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_srari_d_RES
   ret void
@@ -388,7 +388,7 @@ declare <2 x i64> @llvm.mips.srari.d(<2 x i64>, i32) nounwind
 
 define void @llvm_mips_srli_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_srli_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_srli_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.srli.b(<16 x i8> %0, i32 7)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_srli_b_RES
   ret void
@@ -407,7 +407,7 @@ declare <16 x i8> @llvm.mips.srli.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_srli_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_srli_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_srli_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.srli.h(<8 x i16> %0, i32 7)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_srli_h_RES
   ret void
@@ -426,7 +426,7 @@ declare <8 x i16> @llvm.mips.srli.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_srli_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_srli_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_srli_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.srli.w(<4 x i32> %0, i32 7)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_srli_w_RES
   ret void
@@ -445,7 +445,7 @@ declare <4 x i32> @llvm.mips.srli.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_srli_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_srli_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_srli_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.srli.d(<2 x i64> %0, i32 7)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_srli_d_RES
   ret void
@@ -464,7 +464,7 @@ declare <2 x i64> @llvm.mips.srli.d(<2 x i64>, i32) nounwind
 
 define void @llvm_mips_srlri_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_srlri_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_srlri_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.srlri.b(<16 x i8> %0, i32 7)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_srlri_b_RES
   ret void
@@ -483,7 +483,7 @@ declare <16 x i8> @llvm.mips.srlri.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_srlri_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_srlri_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_srlri_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.srlri.h(<8 x i16> %0, i32 7)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_srlri_h_RES
   ret void
@@ -502,7 +502,7 @@ declare <8 x i16> @llvm.mips.srlri.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_srlri_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_srlri_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_srlri_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.srlri.w(<4 x i32> %0, i32 7)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_srlri_w_RES
   ret void
@@ -521,7 +521,7 @@ declare <4 x i32> @llvm.mips.srlri.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_srlri_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_srlri_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_srlri_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.srlri.d(<2 x i64> %0, i32 7)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_srlri_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/bitcast.ll b/test/CodeGen/Mips/msa/bitcast.ll
index 8e880ecd9afb..837cc28aa82d 100644
--- a/test/CodeGen/Mips/msa/bitcast.ll
+++ b/test/CodeGen/Mips/msa/bitcast.ll
@@ -5,7 +5,7 @@
 
 define void @v16i8_to_v16i8(<16 x i8>* %src, <16 x i8>* %dst) nounwind {
 entry:
-  %0 = load volatile <16 x i8>* %src
+  %0 = load volatile <16 x i8>, <16 x i8>* %src
   %1 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0, <16 x i8> %0)
   %2 = bitcast <16 x i8> %1 to <16 x i8>
   %3 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %2, <16 x i8> %2)
@@ -29,7 +29,7 @@ entry:
 
 define void @v16i8_to_v8i16(<16 x i8>* %src, <8 x i16>* %dst) nounwind {
 entry:
-  %0 = load volatile <16 x i8>* %src
+  %0 = load volatile <16 x i8>, <16 x i8>* %src
   %1 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0, <16 x i8> %0)
   %2 = bitcast <16 x i8> %1 to <8 x i16>
   %3 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %2, <8 x i16> %2)
@@ -56,7 +56,7 @@ entry:
 ; are no operations for v8f16 to put in the way.
 define void @v16i8_to_v8f16(<16 x i8>* %src, <8 x half>* %dst) nounwind {
 entry:
-  %0 = load volatile <16 x i8>* %src
+  %0 = load volatile <16 x i8>, <16 x i8>* %src
   %1 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0, <16 x i8> %0)
   %2 = bitcast <16 x i8> %1 to <8 x half>
   store <8 x half> %2, <8 x half>* %dst
@@ -77,7 +77,7 @@ entry:
 
 define void @v16i8_to_v4i32(<16 x i8>* %src, <4 x i32>* %dst) nounwind {
 entry:
-  %0 = load volatile <16 x i8>* %src
+  %0 = load volatile <16 x i8>, <16 x i8>* %src
   %1 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0, <16 x i8> %0)
   %2 = bitcast <16 x i8> %1 to <4 x i32>
   %3 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %2, <4 x i32> %2)
@@ -102,7 +102,7 @@ entry:
 
 define void @v16i8_to_v4f32(<16 x i8>* %src, <4 x float>* %dst) nounwind {
 entry:
-  %0 = load volatile <16 x i8>* %src
+  %0 = load volatile <16 x i8>, <16 x i8>* %src
   %1 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0, <16 x i8> %0)
   %2 = bitcast <16 x i8> %1 to <4 x float>
   %3 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %2, <4 x float> %2)
@@ -127,7 +127,7 @@ entry:
 
 define void @v16i8_to_v2i64(<16 x i8>* %src, <2 x i64>* %dst) nounwind {
 entry:
-  %0 = load volatile <16 x i8>* %src
+  %0 = load volatile <16 x i8>, <16 x i8>* %src
   %1 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0, <16 x i8> %0)
   %2 = bitcast <16 x i8> %1 to <2 x i64>
   %3 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %2, <2 x i64> %2)
@@ -153,7 +153,7 @@ entry:
 
 define void @v16i8_to_v2f64(<16 x i8>* %src, <2 x double>* %dst) nounwind {
 entry:
-  %0 = load volatile <16 x i8>* %src
+  %0 = load volatile <16 x i8>, <16 x i8>* %src
   %1 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0, <16 x i8> %0)
   %2 = bitcast <16 x i8> %1 to <2 x double>
   %3 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %2, <2 x double> %2)
@@ -179,7 +179,7 @@ entry:
 
 define void @v8i16_to_v16i8(<8 x i16>* %src, <16 x i8>* %dst) nounwind {
 entry:
-  %0 = load volatile <8 x i16>* %src
+  %0 = load volatile <8 x i16>, <8 x i16>* %src
   %1 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0, <8 x i16> %0)
   %2 = bitcast <8 x i16> %1 to <16 x i8>
   %3 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %2, <16 x i8> %2)
@@ -204,7 +204,7 @@ entry:
 
 define void @v8i16_to_v8i16(<8 x i16>* %src, <8 x i16>* %dst) nounwind {
 entry:
-  %0 = load volatile <8 x i16>* %src
+  %0 = load volatile <8 x i16>, <8 x i16>* %src
   %1 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0, <8 x i16> %0)
   %2 = bitcast <8 x i16> %1 to <8 x i16>
   %3 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %2, <8 x i16> %2)
@@ -230,7 +230,7 @@ entry:
 ; are no operations for v8f16 to put in the way.
 define void @v8i16_to_v8f16(<8 x i16>* %src, <8 x half>* %dst) nounwind {
 entry:
-  %0 = load volatile <8 x i16>* %src
+  %0 = load volatile <8 x i16>, <8 x i16>* %src
   %1 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0, <8 x i16> %0)
   %2 = bitcast <8 x i16> %1 to <8 x half>
   store <8 x half> %2, <8 x half>* %dst
@@ -251,7 +251,7 @@ entry:
 
 define void @v8i16_to_v4i32(<8 x i16>* %src, <4 x i32>* %dst) nounwind {
 entry:
-  %0 = load volatile <8 x i16>* %src
+  %0 = load volatile <8 x i16>, <8 x i16>* %src
   %1 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0, <8 x i16> %0)
   %2 = bitcast <8 x i16> %1 to <4 x i32>
   %3 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %2, <4 x i32> %2)
@@ -276,7 +276,7 @@ entry:
 
 define void @v8i16_to_v4f32(<8 x i16>* %src, <4 x float>* %dst) nounwind {
 entry:
-  %0 = load volatile <8 x i16>* %src
+  %0 = load volatile <8 x i16>, <8 x i16>* %src
   %1 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0, <8 x i16> %0)
   %2 = bitcast <8 x i16> %1 to <4 x float>
   %3 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %2, <4 x float> %2)
@@ -301,7 +301,7 @@ entry:
 
 define void @v8i16_to_v2i64(<8 x i16>* %src, <2 x i64>* %dst) nounwind {
 entry:
-  %0 = load volatile <8 x i16>* %src
+  %0 = load volatile <8 x i16>, <8 x i16>* %src
   %1 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0, <8 x i16> %0)
   %2 = bitcast <8 x i16> %1 to <2 x i64>
   %3 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %2, <2 x i64> %2)
@@ -326,7 +326,7 @@ entry:
 
 define void @v8i16_to_v2f64(<8 x i16>* %src, <2 x double>* %dst) nounwind {
 entry:
-  %0 = load volatile <8 x i16>* %src
+  %0 = load volatile <8 x i16>, <8 x i16>* %src
   %1 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0, <8 x i16> %0)
   %2 = bitcast <8 x i16> %1 to <2 x double>
   %3 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %2, <2 x double> %2)
@@ -354,7 +354,7 @@ entry:
 ; are no operations for v8f16 to put in the way.
 define void @v8f16_to_v16i8(<8 x half>* %src, <16 x i8>* %dst) nounwind {
 entry:
-  %0 = load volatile <8 x half>* %src
+  %0 = load volatile <8 x half>, <8 x half>* %src
   %1 = bitcast <8 x half> %0 to <16 x i8>
   %2 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %1, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* %dst
@@ -378,7 +378,7 @@ entry:
 ; are no operations for v8f16 to put in the way.
 define void @v8f16_to_v8i16(<8 x half>* %src, <8 x i16>* %dst) nounwind {
 entry:
-  %0 = load volatile <8 x half>* %src
+  %0 = load volatile <8 x half>, <8 x half>* %src
   %1 = bitcast <8 x half> %0 to <8 x i16>
   %2 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %1, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* %dst
@@ -403,7 +403,7 @@ entry:
 ; are no operations for v8f16 to put in the way.
 define void @v8f16_to_v8f16(<8 x half>* %src, <8 x half>* %dst) nounwind {
 entry:
-  %0 = load volatile <8 x half>* %src
+  %0 = load volatile <8 x half>, <8 x half>* %src
   %1 = bitcast <8 x half> %0 to <8 x half>
   store <8 x half> %1, <8 x half>* %dst
   ret void
@@ -423,7 +423,7 @@ entry:
 ; are no operations for v8f16 to put in the way.
 define void @v8f16_to_v4i32(<8 x half>* %src, <4 x i32>* %dst) nounwind {
 entry:
-  %0 = load volatile <8 x half>* %src
+  %0 = load volatile <8 x half>, <8 x half>* %src
   %1 = bitcast <8 x half> %0 to <4 x i32>
   %2 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %1, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* %dst
@@ -447,7 +447,7 @@ entry:
 ; are no operations for v8f16 to put in the way.
 define void @v8f16_to_v4f32(<8 x half>* %src, <4 x float>* %dst) nounwind {
 entry:
-  %0 = load volatile <8 x half>* %src
+  %0 = load volatile <8 x half>, <8 x half>* %src
   %1 = bitcast <8 x half> %0 to <4 x float>
   %2 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %1, <4 x float> %1)
   store <4 x float> %2, <4 x float>* %dst
@@ -471,7 +471,7 @@ entry:
 ; are no operations for v8f16 to put in the way.
 define void @v8f16_to_v2i64(<8 x half>* %src, <2 x i64>* %dst) nounwind {
 entry:
-  %0 = load volatile <8 x half>* %src
+  %0 = load volatile <8 x half>, <8 x half>* %src
   %1 = bitcast <8 x half> %0 to <2 x i64>
   %2 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %1, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* %dst
@@ -495,7 +495,7 @@ entry:
 ; are no operations for v8f16 to put in the way.
 define void @v8f16_to_v2f64(<8 x half>* %src, <2 x double>* %dst) nounwind {
 entry:
-  %0 = load volatile <8 x half>* %src
+  %0 = load volatile <8 x half>, <8 x half>* %src
   %1 = bitcast <8 x half> %0 to <2 x double>
   %2 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %1, <2 x double> %1)
   store <2 x double> %2, <2 x double>* %dst
@@ -518,7 +518,7 @@ entry:
 
 define void @v4i32_to_v16i8(<4 x i32>* %src, <16 x i8>* %dst) nounwind {
 entry:
-  %0 = load volatile <4 x i32>* %src
+  %0 = load volatile <4 x i32>, <4 x i32>* %src
   %1 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %0)
   %2 = bitcast <4 x i32> %1 to <16 x i8>
   %3 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %2, <16 x i8> %2)
@@ -543,7 +543,7 @@ entry:
 
 define void @v4i32_to_v8i16(<4 x i32>* %src, <8 x i16>* %dst) nounwind {
 entry:
-  %0 = load volatile <4 x i32>* %src
+  %0 = load volatile <4 x i32>, <4 x i32>* %src
   %1 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %0)
   %2 = bitcast <4 x i32> %1 to <8 x i16>
   %3 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %2, <8 x i16> %2)
@@ -570,7 +570,7 @@ entry:
 ; are no operations for v8f16 to put in the way.
 define void @v4i32_to_v8f16(<4 x i32>* %src, <8 x half>* %dst) nounwind {
 entry:
-  %0 = load volatile <4 x i32>* %src
+  %0 = load volatile <4 x i32>, <4 x i32>* %src
   %1 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %0)
   %2 = bitcast <4 x i32> %1 to <8 x half>
   store <8 x half> %2, <8 x half>* %dst
@@ -591,7 +591,7 @@ entry:
 
 define void @v4i32_to_v4i32(<4 x i32>* %src, <4 x i32>* %dst) nounwind {
 entry:
-  %0 = load volatile <4 x i32>* %src
+  %0 = load volatile <4 x i32>, <4 x i32>* %src
   %1 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %0)
   %2 = bitcast <4 x i32> %1 to <4 x i32>
   %3 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %2, <4 x i32> %2)
@@ -615,7 +615,7 @@ entry:
 
 define void @v4i32_to_v4f32(<4 x i32>* %src, <4 x float>* %dst) nounwind {
 entry:
-  %0 = load volatile <4 x i32>* %src
+  %0 = load volatile <4 x i32>, <4 x i32>* %src
   %1 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %0)
   %2 = bitcast <4 x i32> %1 to <4 x float>
   %3 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %2, <4 x float> %2)
@@ -639,7 +639,7 @@ entry:
 
 define void @v4i32_to_v2i64(<4 x i32>* %src, <2 x i64>* %dst) nounwind {
 entry:
-  %0 = load volatile <4 x i32>* %src
+  %0 = load volatile <4 x i32>, <4 x i32>* %src
   %1 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %0)
   %2 = bitcast <4 x i32> %1 to <2 x i64>
   %3 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %2, <2 x i64> %2)
@@ -664,7 +664,7 @@ entry:
 
 define void @v4i32_to_v2f64(<4 x i32>* %src, <2 x double>* %dst) nounwind {
 entry:
-  %0 = load volatile <4 x i32>* %src
+  %0 = load volatile <4 x i32>, <4 x i32>* %src
   %1 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %0)
   %2 = bitcast <4 x i32> %1 to <2 x double>
   %3 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %2, <2 x double> %2)
@@ -689,7 +689,7 @@ entry:
 
 define void @v4f32_to_v16i8(<4 x float>* %src, <16 x i8>* %dst) nounwind {
 entry:
-  %0 = load volatile <4 x float>* %src
+  %0 = load volatile <4 x float>, <4 x float>* %src
   %1 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %0, <4 x float> %0)
   %2 = bitcast <4 x float> %1 to <16 x i8>
   %3 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %2, <16 x i8> %2)
@@ -714,7 +714,7 @@ entry:
 
 define void @v4f32_to_v8i16(<4 x float>* %src, <8 x i16>* %dst) nounwind {
 entry:
-  %0 = load volatile <4 x float>* %src
+  %0 = load volatile <4 x float>, <4 x float>* %src
   %1 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %0, <4 x float> %0)
   %2 = bitcast <4 x float> %1 to <8 x i16>
   %3 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %2, <8 x i16> %2)
@@ -741,7 +741,7 @@ entry:
 ; are no operations for v8f16 to put in the way.
 define void @v4f32_to_v8f16(<4 x float>* %src, <8 x half>* %dst) nounwind {
 entry:
-  %0 = load volatile <4 x float>* %src
+  %0 = load volatile <4 x float>, <4 x float>* %src
   %1 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %0, <4 x float> %0)
   %2 = bitcast <4 x float> %1 to <8 x half>
   store <8 x half> %2, <8 x half>* %dst
@@ -762,7 +762,7 @@ entry:
 
 define void @v4f32_to_v4i32(<4 x float>* %src, <4 x i32>* %dst) nounwind {
 entry:
-  %0 = load volatile <4 x float>* %src
+  %0 = load volatile <4 x float>, <4 x float>* %src
   %1 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %0, <4 x float> %0)
   %2 = bitcast <4 x float> %1 to <4 x i32>
   %3 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %2, <4 x i32> %2)
@@ -786,7 +786,7 @@ entry:
 
 define void @v4f32_to_v4f32(<4 x float>* %src, <4 x float>* %dst) nounwind {
 entry:
-  %0 = load volatile <4 x float>* %src
+  %0 = load volatile <4 x float>, <4 x float>* %src
   %1 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %0, <4 x float> %0)
   %2 = bitcast <4 x float> %1 to <4 x float>
   %3 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %2, <4 x float> %2)
@@ -810,7 +810,7 @@ entry:
 
 define void @v4f32_to_v2i64(<4 x float>* %src, <2 x i64>* %dst) nounwind {
 entry:
-  %0 = load volatile <4 x float>* %src
+  %0 = load volatile <4 x float>, <4 x float>* %src
   %1 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %0, <4 x float> %0)
   %2 = bitcast <4 x float> %1 to <2 x i64>
   %3 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %2, <2 x i64> %2)
@@ -835,7 +835,7 @@ entry:
 
 define void @v4f32_to_v2f64(<4 x float>* %src, <2 x double>* %dst) nounwind {
 entry:
-  %0 = load volatile <4 x float>* %src
+  %0 = load volatile <4 x float>, <4 x float>* %src
   %1 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %0, <4 x float> %0)
   %2 = bitcast <4 x float> %1 to <2 x double>
   %3 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %2, <2 x double> %2)
@@ -860,7 +860,7 @@ entry:
 
 define void @v2i64_to_v16i8(<2 x i64>* %src, <16 x i8>* %dst) nounwind {
 entry:
-  %0 = load volatile <2 x i64>* %src
+  %0 = load volatile <2 x i64>, <2 x i64>* %src
   %1 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0, <2 x i64> %0)
   %2 = bitcast <2 x i64> %1 to <16 x i8>
   %3 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %2, <16 x i8> %2)
@@ -886,7 +886,7 @@ entry:
 
 define void @v2i64_to_v8i16(<2 x i64>* %src, <8 x i16>* %dst) nounwind {
 entry:
-  %0 = load volatile <2 x i64>* %src
+  %0 = load volatile <2 x i64>, <2 x i64>* %src
   %1 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0, <2 x i64> %0)
   %2 = bitcast <2 x i64> %1 to <8 x i16>
   %3 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %2, <8 x i16> %2)
@@ -913,7 +913,7 @@ entry:
 ; are no operations for v8f16 to put in the way.
 define void @v2i64_to_v8f16(<2 x i64>* %src, <8 x half>* %dst) nounwind {
 entry:
-  %0 = load volatile <2 x i64>* %src
+  %0 = load volatile <2 x i64>, <2 x i64>* %src
   %1 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0, <2 x i64> %0)
   %2 = bitcast <2 x i64> %1 to <8 x half>
   store <8 x half> %2, <8 x half>* %dst
@@ -934,7 +934,7 @@ entry:
 
 define void @v2i64_to_v4i32(<2 x i64>* %src, <4 x i32>* %dst) nounwind {
 entry:
-  %0 = load volatile <2 x i64>* %src
+  %0 = load volatile <2 x i64>, <2 x i64>* %src
   %1 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0, <2 x i64> %0)
   %2 = bitcast <2 x i64> %1 to <4 x i32>
   %3 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %2, <4 x i32> %2)
@@ -959,7 +959,7 @@ entry:
 
 define void @v2i64_to_v4f32(<2 x i64>* %src, <4 x float>* %dst) nounwind {
 entry:
-  %0 = load volatile <2 x i64>* %src
+  %0 = load volatile <2 x i64>, <2 x i64>* %src
   %1 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0, <2 x i64> %0)
   %2 = bitcast <2 x i64> %1 to <4 x float>
   %3 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %2, <4 x float> %2)
@@ -984,7 +984,7 @@ entry:
 
 define void @v2i64_to_v2i64(<2 x i64>* %src, <2 x i64>* %dst) nounwind {
 entry:
-  %0 = load volatile <2 x i64>* %src
+  %0 = load volatile <2 x i64>, <2 x i64>* %src
   %1 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0, <2 x i64> %0)
   %2 = bitcast <2 x i64> %1 to <2 x i64>
   %3 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %2, <2 x i64> %2)
@@ -1008,7 +1008,7 @@ entry:
 
 define void @v2i64_to_v2f64(<2 x i64>* %src, <2 x double>* %dst) nounwind {
 entry:
-  %0 = load volatile <2 x i64>* %src
+  %0 = load volatile <2 x i64>, <2 x i64>* %src
   %1 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0, <2 x i64> %0)
   %2 = bitcast <2 x i64> %1 to <2 x double>
   %3 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %2, <2 x double> %2)
@@ -1032,7 +1032,7 @@ entry:
 
 define void @v2f64_to_v16i8(<2 x double>* %src, <16 x i8>* %dst) nounwind {
 entry:
-  %0 = load volatile <2 x double>* %src
+  %0 = load volatile <2 x double>, <2 x double>* %src
   %1 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %0, <2 x double> %0)
   %2 = bitcast <2 x double> %1 to <16 x i8>
   %3 = tail call <16 x i8> @llvm.mips.addv.b(<16 x i8> %2, <16 x i8> %2)
@@ -1058,7 +1058,7 @@ entry:
 
 define void @v2f64_to_v8i16(<2 x double>* %src, <8 x i16>* %dst) nounwind {
 entry:
-  %0 = load volatile <2 x double>* %src
+  %0 = load volatile <2 x double>, <2 x double>* %src
   %1 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %0, <2 x double> %0)
   %2 = bitcast <2 x double> %1 to <8 x i16>
   %3 = tail call <8 x i16> @llvm.mips.addv.h(<8 x i16> %2, <8 x i16> %2)
@@ -1085,7 +1085,7 @@ entry:
 ; are no operations for v8f16 to put in the way.
 define void @v2f64_to_v8f16(<2 x double>* %src, <8 x half>* %dst) nounwind {
 entry:
-  %0 = load volatile <2 x double>* %src
+  %0 = load volatile <2 x double>, <2 x double>* %src
   %1 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %0, <2 x double> %0)
   %2 = bitcast <2 x double> %1 to <8 x half>
   store <8 x half> %2, <8 x half>* %dst
@@ -1106,7 +1106,7 @@ entry:
 
 define void @v2f64_to_v4i32(<2 x double>* %src, <4 x i32>* %dst) nounwind {
 entry:
-  %0 = load volatile <2 x double>* %src
+  %0 = load volatile <2 x double>, <2 x double>* %src
   %1 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %0, <2 x double> %0)
   %2 = bitcast <2 x double> %1 to <4 x i32>
   %3 = tail call <4 x i32> @llvm.mips.addv.w(<4 x i32> %2, <4 x i32> %2)
@@ -1131,7 +1131,7 @@ entry:
 
 define void @v2f64_to_v4f32(<2 x double>* %src, <4 x float>* %dst) nounwind {
 entry:
-  %0 = load volatile <2 x double>* %src
+  %0 = load volatile <2 x double>, <2 x double>* %src
   %1 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %0, <2 x double> %0)
   %2 = bitcast <2 x double> %1 to <4 x float>
   %3 = tail call <4 x float> @llvm.mips.fadd.w(<4 x float> %2, <4 x float> %2)
@@ -1156,7 +1156,7 @@ entry:
 
 define void @v2f64_to_v2i64(<2 x double>* %src, <2 x i64>* %dst) nounwind {
 entry:
-  %0 = load volatile <2 x double>* %src
+  %0 = load volatile <2 x double>, <2 x double>* %src
   %1 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %0, <2 x double> %0)
   %2 = bitcast <2 x double> %1 to <2 x i64>
   %3 = tail call <2 x i64> @llvm.mips.addv.d(<2 x i64> %2, <2 x i64> %2)
@@ -1180,7 +1180,7 @@ entry:
 
 define void @v2f64_to_v2f64(<2 x double>* %src, <2 x double>* %dst) nounwind {
 entry:
-  %0 = load volatile <2 x double>* %src
+  %0 = load volatile <2 x double>, <2 x double>* %src
   %1 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %0, <2 x double> %0)
   %2 = bitcast <2 x double> %1 to <2 x double>
   %3 = tail call <2 x double> @llvm.mips.fadd.d(<2 x double> %2, <2 x double> %2)
diff --git a/test/CodeGen/Mips/msa/bitwise.ll b/test/CodeGen/Mips/msa/bitwise.ll
index 5d57198a9355..2a260b2c5733 100644
--- a/test/CodeGen/Mips/msa/bitwise.ll
+++ b/test/CodeGen/Mips/msa/bitwise.ll
@@ -4,9 +4,9 @@
 define void @and_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: and_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = and <16 x i8> %1, %2
   ; CHECK-DAG: and.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -20,9 +20,9 @@ define void @and_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @and_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: and_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = and <8 x i16> %1, %2
   ; CHECK-DAG: and.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -36,9 +36,9 @@ define void @and_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @and_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: and_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = and <4 x i32> %1, %2
   ; CHECK-DAG: and.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -52,9 +52,9 @@ define void @and_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @and_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: and_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = and <2 x i64> %1, %2
   ; CHECK-DAG: and.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -68,7 +68,7 @@ define void @and_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @and_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: and_v16i8_i:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = and <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   ; CHECK-DAG: andi.b [[R4:\$w[0-9]+]], [[R1]], 1
@@ -82,7 +82,7 @@ define void @and_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @and_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: and_v8i16_i:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = and <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   ; CHECK-DAG: ldi.h [[R3:\$w[0-9]+]], 1
@@ -97,7 +97,7 @@ define void @and_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @and_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: and_v4i32_i:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = and <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   ; CHECK-DAG: ldi.w [[R3:\$w[0-9]+]], 1
@@ -112,7 +112,7 @@ define void @and_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @and_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: and_v2i64_i:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = and <2 x i64> %1, <i64 1, i64 1>
   ; CHECK-DAG: ldi.d [[R3:\$w[0-9]+]], 1
@@ -127,9 +127,9 @@ define void @and_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @or_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: or_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = or <16 x i8> %1, %2
   ; CHECK-DAG: or.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -143,9 +143,9 @@ define void @or_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @or_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: or_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = or <8 x i16> %1, %2
   ; CHECK-DAG: or.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -159,9 +159,9 @@ define void @or_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @or_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: or_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = or <4 x i32> %1, %2
   ; CHECK-DAG: or.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -175,9 +175,9 @@ define void @or_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @or_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: or_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = or <2 x i64> %1, %2
   ; CHECK-DAG: or.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -191,7 +191,7 @@ define void @or_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @or_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: or_v16i8_i:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = or <16 x i8> %1, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>
   ; CHECK-DAG: ori.b [[R4:\$w[0-9]+]], [[R1]], 3
@@ -205,7 +205,7 @@ define void @or_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @or_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: or_v8i16_i:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = or <8 x i16> %1, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
   ; CHECK-DAG: ldi.h [[R3:\$w[0-9]+]], 3
@@ -220,7 +220,7 @@ define void @or_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @or_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: or_v4i32_i:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = or <4 x i32> %1, <i32 3, i32 3, i32 3, i32 3>
   ; CHECK-DAG: ldi.w [[R3:\$w[0-9]+]], 3
@@ -235,7 +235,7 @@ define void @or_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @or_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: or_v2i64_i:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = or <2 x i64> %1, <i64 3, i64 3>
   ; CHECK-DAG: ldi.d [[R3:\$w[0-9]+]], 3
@@ -250,9 +250,9 @@ define void @or_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @nor_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: nor_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = or <16 x i8> %1, %2
   %4 = xor <16 x i8> %3, <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
@@ -267,9 +267,9 @@ define void @nor_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @nor_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: nor_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = or <8 x i16> %1, %2
   %4 = xor <8 x i16> %3, <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1>
@@ -284,9 +284,9 @@ define void @nor_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @nor_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: nor_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = or <4 x i32> %1, %2
   %4 = xor <4 x i32> %3, <i32 -1, i32 -1, i32 -1, i32 -1>
@@ -301,9 +301,9 @@ define void @nor_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @nor_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: nor_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = or <2 x i64> %1, %2
   %4 = xor <2 x i64> %3, <i64 -1, i64 -1>
@@ -318,7 +318,7 @@ define void @nor_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @nor_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: nor_v16i8_i:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = or <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   %3 = xor <16 x i8> %2, <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
@@ -333,7 +333,7 @@ define void @nor_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @nor_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: nor_v8i16_i:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = or <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   %3 = xor <8 x i16> %2, <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1>
@@ -349,7 +349,7 @@ define void @nor_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @nor_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: nor_v4i32_i:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = or <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   %3 = xor <4 x i32> %2, <i32 -1, i32 -1, i32 -1, i32 -1>
@@ -365,7 +365,7 @@ define void @nor_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @nor_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: nor_v2i64_i:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = or <2 x i64> %1, <i64 1, i64 1>
   %3 = xor <2 x i64> %2, <i64 -1, i64 -1>
@@ -381,9 +381,9 @@ define void @nor_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @xor_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: xor_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = xor <16 x i8> %1, %2
   ; CHECK-DAG: xor.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -397,9 +397,9 @@ define void @xor_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @xor_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: xor_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = xor <8 x i16> %1, %2
   ; CHECK-DAG: xor.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -413,9 +413,9 @@ define void @xor_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @xor_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: xor_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = xor <4 x i32> %1, %2
   ; CHECK-DAG: xor.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -429,9 +429,9 @@ define void @xor_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @xor_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: xor_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = xor <2 x i64> %1, %2
   ; CHECK-DAG: xor.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -445,7 +445,7 @@ define void @xor_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @xor_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: xor_v16i8_i:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = xor <16 x i8> %1, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>
   ; CHECK-DAG: xori.b [[R4:\$w[0-9]+]], [[R1]], 3
@@ -459,7 +459,7 @@ define void @xor_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @xor_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: xor_v8i16_i:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = xor <8 x i16> %1, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
   ; CHECK-DAG: ldi.h [[R3:\$w[0-9]+]], 3
@@ -474,7 +474,7 @@ define void @xor_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @xor_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: xor_v4i32_i:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = xor <4 x i32> %1, <i32 3, i32 3, i32 3, i32 3>
   ; CHECK-DAG: ldi.w [[R3:\$w[0-9]+]], 3
@@ -489,7 +489,7 @@ define void @xor_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @xor_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: xor_v2i64_i:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = xor <2 x i64> %1, <i64 3, i64 3>
   ; CHECK-DAG: ldi.d [[R3:\$w[0-9]+]], 3
@@ -504,9 +504,9 @@ define void @xor_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @sll_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: sll_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = shl <16 x i8> %1, %2
   ; CHECK-DAG: sll.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -520,9 +520,9 @@ define void @sll_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @sll_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: sll_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = shl <8 x i16> %1, %2
   ; CHECK-DAG: sll.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -536,9 +536,9 @@ define void @sll_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @sll_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: sll_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = shl <4 x i32> %1, %2
   ; CHECK-DAG: sll.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -552,9 +552,9 @@ define void @sll_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @sll_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: sll_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = shl <2 x i64> %1, %2
   ; CHECK-DAG: sll.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -568,7 +568,7 @@ define void @sll_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @sll_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: sll_v16i8_i:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = shl <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   ; CHECK-DAG: slli.b [[R4:\$w[0-9]+]], [[R1]], 1
@@ -582,7 +582,7 @@ define void @sll_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @sll_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: sll_v8i16_i:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = shl <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   ; CHECK-DAG: slli.h [[R4:\$w[0-9]+]], [[R1]], 1
@@ -596,7 +596,7 @@ define void @sll_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @sll_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: sll_v4i32_i:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = shl <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   ; CHECK-DAG: slli.w [[R4:\$w[0-9]+]], [[R1]], 1
@@ -610,7 +610,7 @@ define void @sll_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @sll_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: sll_v2i64_i:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = shl <2 x i64> %1, <i64 1, i64 1>
   ; CHECK-DAG: slli.d [[R4:\$w[0-9]+]], [[R1]], 1
@@ -624,9 +624,9 @@ define void @sll_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @sra_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: sra_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = ashr <16 x i8> %1, %2
   ; CHECK-DAG: sra.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -640,9 +640,9 @@ define void @sra_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @sra_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: sra_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = ashr <8 x i16> %1, %2
   ; CHECK-DAG: sra.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -656,9 +656,9 @@ define void @sra_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @sra_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: sra_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = ashr <4 x i32> %1, %2
   ; CHECK-DAG: sra.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -672,9 +672,9 @@ define void @sra_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @sra_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: sra_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = ashr <2 x i64> %1, %2
   ; CHECK-DAG: sra.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -688,7 +688,7 @@ define void @sra_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @sra_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: sra_v16i8_i:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = ashr <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   ; CHECK-DAG: srai.b [[R4:\$w[0-9]+]], [[R1]], 1
@@ -702,7 +702,7 @@ define void @sra_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @sra_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: sra_v8i16_i:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = ashr <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   ; CHECK-DAG: srai.h [[R4:\$w[0-9]+]], [[R1]], 1
@@ -716,7 +716,7 @@ define void @sra_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @sra_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: sra_v4i32_i:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = ashr <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   ; CHECK-DAG: srai.w [[R4:\$w[0-9]+]], [[R1]], 1
@@ -730,7 +730,7 @@ define void @sra_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @sra_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: sra_v2i64_i:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = ashr <2 x i64> %1, <i64 1, i64 1>
   ; CHECK-DAG: srai.d [[R4:\$w[0-9]+]], [[R1]], 1
@@ -744,9 +744,9 @@ define void @sra_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @srl_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: srl_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = lshr <16 x i8> %1, %2
   ; CHECK-DAG: srl.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -760,9 +760,9 @@ define void @srl_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @srl_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: srl_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = lshr <8 x i16> %1, %2
   ; CHECK-DAG: srl.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -776,9 +776,9 @@ define void @srl_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @srl_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: srl_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = lshr <4 x i32> %1, %2
   ; CHECK-DAG: srl.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -792,9 +792,9 @@ define void @srl_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @srl_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: srl_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = lshr <2 x i64> %1, %2
   ; CHECK-DAG: srl.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -808,7 +808,7 @@ define void @srl_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @srl_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: srl_v16i8_i:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = lshr <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   ; CHECK-DAG: srli.b [[R4:\$w[0-9]+]], [[R1]], 1
@@ -822,7 +822,7 @@ define void @srl_v16i8_i(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @srl_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: srl_v8i16_i:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = lshr <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   ; CHECK-DAG: srli.h [[R4:\$w[0-9]+]], [[R1]], 1
@@ -836,7 +836,7 @@ define void @srl_v8i16_i(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @srl_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: srl_v4i32_i:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = lshr <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   ; CHECK-DAG: srli.w [[R4:\$w[0-9]+]], [[R1]], 1
@@ -850,7 +850,7 @@ define void @srl_v4i32_i(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @srl_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: srl_v2i64_i:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = lshr <2 x i64> %1, <i64 1, i64 1>
   ; CHECK-DAG: srli.d [[R4:\$w[0-9]+]], [[R1]], 1
@@ -864,7 +864,7 @@ define void @srl_v2i64_i(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @ctpop_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: ctpop_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = tail call <16 x i8> @llvm.ctpop.v16i8 (<16 x i8> %1)
   ; CHECK-DAG: pcnt.b [[R3:\$w[0-9]+]], [[R1]]
@@ -878,7 +878,7 @@ define void @ctpop_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @ctpop_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: ctpop_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = tail call <8 x i16> @llvm.ctpop.v8i16 (<8 x i16> %1)
   ; CHECK-DAG: pcnt.h [[R3:\$w[0-9]+]], [[R1]]
@@ -892,7 +892,7 @@ define void @ctpop_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @ctpop_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: ctpop_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = tail call <4 x i32> @llvm.ctpop.v4i32 (<4 x i32> %1)
   ; CHECK-DAG: pcnt.w [[R3:\$w[0-9]+]], [[R1]]
@@ -906,7 +906,7 @@ define void @ctpop_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @ctpop_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: ctpop_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = tail call <2 x i64> @llvm.ctpop.v2i64 (<2 x i64> %1)
   ; CHECK-DAG: pcnt.d [[R3:\$w[0-9]+]], [[R1]]
@@ -920,7 +920,7 @@ define void @ctpop_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @ctlz_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: ctlz_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = tail call <16 x i8> @llvm.ctlz.v16i8 (<16 x i8> %1)
   ; CHECK-DAG: nlzc.b [[R3:\$w[0-9]+]], [[R1]]
@@ -934,7 +934,7 @@ define void @ctlz_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @ctlz_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: ctlz_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = tail call <8 x i16> @llvm.ctlz.v8i16 (<8 x i16> %1)
   ; CHECK-DAG: nlzc.h [[R3:\$w[0-9]+]], [[R1]]
@@ -948,7 +948,7 @@ define void @ctlz_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @ctlz_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: ctlz_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = tail call <4 x i32> @llvm.ctlz.v4i32 (<4 x i32> %1)
   ; CHECK-DAG: nlzc.w [[R3:\$w[0-9]+]], [[R1]]
@@ -962,7 +962,7 @@ define void @ctlz_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @ctlz_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: ctlz_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = tail call <2 x i64> @llvm.ctlz.v2i64 (<2 x i64> %1)
   ; CHECK-DAG: nlzc.d [[R3:\$w[0-9]+]], [[R1]]
@@ -976,11 +976,11 @@ define void @ctlz_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @bsel_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b, <16 x i8>* %m) nounwind {
   ; CHECK: bsel_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <16 x i8>* %m
+  %3 = load <16 x i8>, <16 x i8>* %m
   ; CHECK-DAG: ld.b [[R3:\$w[0-9]+]], 0($7)
   %4 = xor <16 x i8> %3, <i8 -1, i8 -1, i8 -1, i8 -1,
                           i8 -1, i8 -1, i8 -1, i8 -1,
@@ -1002,9 +1002,9 @@ define void @bsel_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b, <16 x i8>*
 define void @bsel_v16i8_i(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %m) nounwind {
   ; CHECK: bsel_v16i8_i:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %m
+  %2 = load <16 x i8>, <16 x i8>* %m
   ; CHECK-DAG: ld.b [[R3:\$w[0-9]+]], 0($6)
   %3 = xor <16 x i8> %2, <i8 -1, i8 -1, i8 -1, i8 -1,
                           i8 -1, i8 -1, i8 -1, i8 -1,
@@ -1027,9 +1027,9 @@ define void @bsel_v16i8_i(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %m) nounwind
 define void @bsel_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: bsel_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = and <8 x i16> %1, <i16 6, i16 6, i16 6, i16 6,
                           i16 6, i16 6, i16 6, i16 6>
@@ -1048,9 +1048,9 @@ define void @bsel_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @bsel_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: bsel_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = and <4 x i32> %1, <i32 6, i32 6, i32 6, i32 6>
   %4 = and <4 x i32> %2, <i32 4294967289, i32 4294967289, i32 4294967289, i32 4294967289>
@@ -1067,9 +1067,9 @@ define void @bsel_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @bsel_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: bsel_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = and <2 x i64> %1, <i64 6, i64 6>
   %4 = and <2 x i64> %2, <i64 18446744073709551609, i64 18446744073709551609>
@@ -1086,9 +1086,9 @@ define void @bsel_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @binsl_v16i8_i(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: binsl_v16i8_i:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = and <16 x i8> %1, <i8 192, i8 192, i8 192, i8 192,
                           i8 192, i8 192, i8 192, i8 192,
@@ -1110,9 +1110,9 @@ define void @binsl_v16i8_i(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind
 define void @binsl_v8i16_i(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: binsl_v8i16_i:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = and <8 x i16> %1, <i16 49152, i16 49152, i16 49152, i16 49152,
                           i16 49152, i16 49152, i16 49152, i16 49152>
@@ -1130,9 +1130,9 @@ define void @binsl_v8i16_i(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind
 define void @binsl_v4i32_i(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: binsl_v4i32_i:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = and <4 x i32> %1, <i32 3221225472, i32 3221225472, i32 3221225472, i32 3221225472>
   %4 = and <4 x i32> %2, <i32 1073741823, i32 1073741823, i32 1073741823, i32 1073741823>
@@ -1148,9 +1148,9 @@ define void @binsl_v4i32_i(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind
 define void @binsl_v2i64_i(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: binsl_v2i64_i:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = and <2 x i64> %1, <i64 18446744073709551608, i64 18446744073709551608>
   %4 = and <2 x i64> %2, <i64 7, i64 7>
@@ -1170,9 +1170,9 @@ define void @binsl_v2i64_i(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind
 define void @binsr_v16i8_i(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: binsr_v16i8_i:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = and <16 x i8> %1, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3,
                           i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>
@@ -1192,9 +1192,9 @@ define void @binsr_v16i8_i(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind
 define void @binsr_v8i16_i(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: binsr_v8i16_i:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = and <8 x i16> %1, <i16 3, i16 3, i16 3, i16 3,
                           i16 3, i16 3, i16 3, i16 3>
@@ -1212,9 +1212,9 @@ define void @binsr_v8i16_i(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind
 define void @binsr_v4i32_i(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: binsr_v4i32_i:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = and <4 x i32> %1, <i32 3, i32 3, i32 3, i32 3>
   %4 = and <4 x i32> %2, <i32 4294967292, i32 4294967292, i32 4294967292, i32 4294967292>
@@ -1230,9 +1230,9 @@ define void @binsr_v4i32_i(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind
 define void @binsr_v2i64_i(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: binsr_v2i64_i:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = and <2 x i64> %1, <i64 3, i64 3>
   %4 = and <2 x i64> %2, <i64 18446744073709551612, i64 18446744073709551612>
@@ -1248,9 +1248,9 @@ define void @binsr_v2i64_i(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind
 define void @bclr_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: bclr_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = shl <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>, %2
   %4 = xor <16 x i8> %3, <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
@@ -1266,9 +1266,9 @@ define void @bclr_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @bclr_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: bclr_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = shl <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>, %2
   %4 = xor <8 x i16> %3, <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1>
@@ -1284,9 +1284,9 @@ define void @bclr_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @bclr_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: bclr_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = shl <4 x i32> <i32 1, i32 1, i32 1, i32 1>, %2
   %4 = xor <4 x i32> %3, <i32 -1, i32 -1, i32 -1, i32 -1>
@@ -1302,9 +1302,9 @@ define void @bclr_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @bclr_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: bclr_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = shl <2 x i64> <i64 1, i64 1>, %2
   %4 = xor <2 x i64> %3, <i64 -1, i64 -1>
@@ -1320,9 +1320,9 @@ define void @bclr_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @bset_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: bset_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = shl <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>, %2
   %4 = or <16 x i8> %1, %3
@@ -1337,9 +1337,9 @@ define void @bset_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @bset_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: bset_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = shl <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>, %2
   %4 = or <8 x i16> %1, %3
@@ -1354,9 +1354,9 @@ define void @bset_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @bset_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: bset_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = shl <4 x i32> <i32 1, i32 1, i32 1, i32 1>, %2
   %4 = or <4 x i32> %1, %3
@@ -1371,9 +1371,9 @@ define void @bset_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @bset_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: bset_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = shl <2 x i64> <i64 1, i64 1>, %2
   %4 = or <2 x i64> %1, %3
@@ -1388,9 +1388,9 @@ define void @bset_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @bneg_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: bneg_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = shl <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>, %2
   %4 = xor <16 x i8> %1, %3
@@ -1405,9 +1405,9 @@ define void @bneg_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @bneg_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: bneg_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = shl <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>, %2
   %4 = xor <8 x i16> %1, %3
@@ -1422,9 +1422,9 @@ define void @bneg_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @bneg_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: bneg_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = shl <4 x i32> <i32 1, i32 1, i32 1, i32 1>, %2
   %4 = xor <4 x i32> %1, %3
@@ -1439,9 +1439,9 @@ define void @bneg_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @bneg_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: bneg_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = shl <2 x i64> <i64 1, i64 1>, %2
   %4 = xor <2 x i64> %1, %3
@@ -1456,7 +1456,7 @@ define void @bneg_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @bclri_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: bclri_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = xor <16 x i8> <i8  8, i8  8, i8  8, i8  8, i8  8, i8  8, i8  8, i8  8, i8  8, i8  8, i8  8, i8  8, i8  8, i8  8, i8  8, i8  8>,
                      <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
@@ -1473,7 +1473,7 @@ define void @bclri_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @bclri_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: bclri_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = xor <8 x i16> <i16  8, i16  8, i16  8, i16  8, i16  8, i16  8, i16  8, i16  8>,
                      <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1>
@@ -1489,7 +1489,7 @@ define void @bclri_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @bclri_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: bclri_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = xor <4 x i32> <i32  8, i32  8, i32  8, i32  8>,
                      <i32 -1, i32 -1, i32 -1, i32 -1>
@@ -1505,7 +1505,7 @@ define void @bclri_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @bclri_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: bclri_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = xor <2 x i64> <i64  8, i64  8>,
                      <i64 -1, i64 -1>
@@ -1521,7 +1521,7 @@ define void @bclri_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @bseti_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: bseti_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = or <16 x i8> %1, <i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8>
   ; CHECK-DAG: bseti.b [[R3:\$w[0-9]+]], [[R1]], 3
@@ -1535,7 +1535,7 @@ define void @bseti_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @bseti_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: bseti_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = or <8 x i16> %1, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
   ; CHECK-DAG: bseti.h [[R3:\$w[0-9]+]], [[R1]], 3
@@ -1549,7 +1549,7 @@ define void @bseti_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @bseti_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: bseti_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = or <4 x i32> %1, <i32 8, i32 8, i32 8, i32 8>
   ; CHECK-DAG: bseti.w [[R3:\$w[0-9]+]], [[R1]], 3
@@ -1563,7 +1563,7 @@ define void @bseti_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @bseti_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: bseti_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = or <2 x i64> %1, <i64 8, i64 8>
   ; CHECK-DAG: bseti.d [[R3:\$w[0-9]+]], [[R1]], 3
@@ -1577,7 +1577,7 @@ define void @bseti_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @bnegi_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: bnegi_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = xor <16 x i8> %1, <i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8>
   ; CHECK-DAG: bnegi.b [[R3:\$w[0-9]+]], [[R1]], 3
@@ -1591,7 +1591,7 @@ define void @bnegi_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @bnegi_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: bnegi_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = xor <8 x i16> %1, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
   ; CHECK-DAG: bnegi.h [[R3:\$w[0-9]+]], [[R1]], 3
@@ -1605,7 +1605,7 @@ define void @bnegi_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @bnegi_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: bnegi_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = xor <4 x i32> %1, <i32 8, i32 8, i32 8, i32 8>
   ; CHECK-DAG: bnegi.w [[R3:\$w[0-9]+]], [[R1]], 3
@@ -1619,7 +1619,7 @@ define void @bnegi_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @bnegi_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: bnegi_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = xor <2 x i64> %1, <i64 8, i64 8>
   ; CHECK-DAG: bnegi.d [[R3:\$w[0-9]+]], [[R1]], 3
diff --git a/test/CodeGen/Mips/msa/compare.ll b/test/CodeGen/Mips/msa/compare.ll
index 87ca1482da81..bc4f6e7e3946 100644
--- a/test/CodeGen/Mips/msa/compare.ll
+++ b/test/CodeGen/Mips/msa/compare.ll
@@ -4,9 +4,9 @@
 define void @ceq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: ceq_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp eq <16 x i8> %1, %2
   %4 = sext <16 x i1> %3 to <16 x i8>
@@ -21,9 +21,9 @@ define void @ceq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @ceq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: ceq_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp eq <8 x i16> %1, %2
   %4 = sext <8 x i1> %3 to <8 x i16>
@@ -38,9 +38,9 @@ define void @ceq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @ceq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: ceq_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp eq <4 x i32> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -55,9 +55,9 @@ define void @ceq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @ceq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: ceq_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp eq <2 x i64> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -72,9 +72,9 @@ define void @ceq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @cle_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: cle_s_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sle <16 x i8> %1, %2
   %4 = sext <16 x i1> %3 to <16 x i8>
@@ -89,9 +89,9 @@ define void @cle_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @cle_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: cle_s_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sle <8 x i16> %1, %2
   %4 = sext <8 x i1> %3 to <8 x i16>
@@ -106,9 +106,9 @@ define void @cle_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @cle_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: cle_s_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sle <4 x i32> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -123,9 +123,9 @@ define void @cle_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @cle_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: cle_s_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sle <2 x i64> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -140,9 +140,9 @@ define void @cle_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @cle_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: cle_u_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ule <16 x i8> %1, %2
   %4 = sext <16 x i1> %3 to <16 x i8>
@@ -157,9 +157,9 @@ define void @cle_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @cle_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: cle_u_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ule <8 x i16> %1, %2
   %4 = sext <8 x i1> %3 to <8 x i16>
@@ -174,9 +174,9 @@ define void @cle_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @cle_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: cle_u_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ule <4 x i32> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -191,9 +191,9 @@ define void @cle_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @cle_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: cle_u_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ule <2 x i64> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -208,9 +208,9 @@ define void @cle_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @clt_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: clt_s_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp slt <16 x i8> %1, %2
   %4 = sext <16 x i1> %3 to <16 x i8>
@@ -225,9 +225,9 @@ define void @clt_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @clt_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: clt_s_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp slt <8 x i16> %1, %2
   %4 = sext <8 x i1> %3 to <8 x i16>
@@ -242,9 +242,9 @@ define void @clt_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @clt_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: clt_s_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp slt <4 x i32> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -259,9 +259,9 @@ define void @clt_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @clt_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: clt_s_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp slt <2 x i64> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -276,9 +276,9 @@ define void @clt_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @clt_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: clt_u_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ult <16 x i8> %1, %2
   %4 = sext <16 x i1> %3 to <16 x i8>
@@ -293,9 +293,9 @@ define void @clt_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @clt_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: clt_u_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ult <8 x i16> %1, %2
   %4 = sext <8 x i1> %3 to <8 x i16>
@@ -310,9 +310,9 @@ define void @clt_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @clt_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: clt_u_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ult <4 x i32> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -327,9 +327,9 @@ define void @clt_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @clt_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: clt_u_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ult <2 x i64> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -345,9 +345,9 @@ define void @clt_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 ; issues in this area.
 define void @cne_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: cne_v16i8:
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ne <16 x i8> %1, %2
   %4 = sext <16 x i1> %3 to <16 x i8>
@@ -365,9 +365,9 @@ define void @cne_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @cne_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: cne_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ne <8 x i16> %1, %2
   %4 = sext <8 x i1> %3 to <8 x i16>
@@ -387,9 +387,9 @@ define void @cne_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @cne_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: cne_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ne <4 x i32> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -409,9 +409,9 @@ define void @cne_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @cne_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: cne_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ne <2 x i64> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -429,7 +429,7 @@ define void @cne_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @ceqi_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: ceqi_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp eq <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   %3 = sext <16 x i1> %2 to <16 x i8>
@@ -444,7 +444,7 @@ define void @ceqi_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @ceqi_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: ceqi_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp eq <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   %3 = sext <8 x i1> %2 to <8 x i16>
@@ -459,7 +459,7 @@ define void @ceqi_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @ceqi_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: ceqi_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp eq <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   %3 = sext <4 x i1> %2 to <4 x i32>
@@ -474,7 +474,7 @@ define void @ceqi_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @ceqi_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: ceqi_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp eq <2 x i64> %1, <i64 1, i64 1>
   %3 = sext <2 x i1> %2 to <2 x i64>
@@ -489,7 +489,7 @@ define void @ceqi_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @clei_s_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: clei_s_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp sle <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   %3 = sext <16 x i1> %2 to <16 x i8>
@@ -504,7 +504,7 @@ define void @clei_s_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @clei_s_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: clei_s_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp sle <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   %3 = sext <8 x i1> %2 to <8 x i16>
@@ -519,7 +519,7 @@ define void @clei_s_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @clei_s_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: clei_s_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp sle <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   %3 = sext <4 x i1> %2 to <4 x i32>
@@ -534,7 +534,7 @@ define void @clei_s_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @clei_s_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: clei_s_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp sle <2 x i64> %1, <i64 1, i64 1>
   %3 = sext <2 x i1> %2 to <2 x i64>
@@ -549,7 +549,7 @@ define void @clei_s_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @clei_u_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: clei_u_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ule <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   %3 = sext <16 x i1> %2 to <16 x i8>
@@ -564,7 +564,7 @@ define void @clei_u_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @clei_u_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: clei_u_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ule <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   %3 = sext <8 x i1> %2 to <8 x i16>
@@ -579,7 +579,7 @@ define void @clei_u_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @clei_u_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: clei_u_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ule <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   %3 = sext <4 x i1> %2 to <4 x i32>
@@ -594,7 +594,7 @@ define void @clei_u_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @clei_u_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: clei_u_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ule <2 x i64> %1, <i64 1, i64 1>
   %3 = sext <2 x i1> %2 to <2 x i64>
@@ -609,7 +609,7 @@ define void @clei_u_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @clti_s_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: clti_s_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp slt <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   %3 = sext <16 x i1> %2 to <16 x i8>
@@ -624,7 +624,7 @@ define void @clti_s_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @clti_s_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: clti_s_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp slt <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   %3 = sext <8 x i1> %2 to <8 x i16>
@@ -639,7 +639,7 @@ define void @clti_s_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @clti_s_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: clti_s_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp slt <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   %3 = sext <4 x i1> %2 to <4 x i32>
@@ -654,7 +654,7 @@ define void @clti_s_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @clti_s_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: clti_s_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp slt <2 x i64> %1, <i64 1, i64 1>
   %3 = sext <2 x i1> %2 to <2 x i64>
@@ -669,7 +669,7 @@ define void @clti_s_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @clti_u_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: clti_u_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ult <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   %3 = sext <16 x i1> %2 to <16 x i8>
@@ -684,7 +684,7 @@ define void @clti_u_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @clti_u_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: clti_u_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ult <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   %3 = sext <8 x i1> %2 to <8 x i16>
@@ -699,7 +699,7 @@ define void @clti_u_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @clti_u_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: clti_u_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ult <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   %3 = sext <4 x i1> %2 to <4 x i32>
@@ -714,7 +714,7 @@ define void @clti_u_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @clti_u_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: clti_u_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ult <2 x i64> %1, <i64 1, i64 1>
   %3 = sext <2 x i1> %2 to <2 x i64>
@@ -730,11 +730,11 @@ define void @bsel_s_v16i8(<16 x i8>* %d, <16 x i8>* %a, <16 x i8>* %b,
                         <16 x i8>* %c) nounwind {
   ; CHECK: bsel_s_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <16 x i8>* %c
+  %3 = load <16 x i8>, <16 x i8>* %c
   ; CHECK-DAG: ld.b [[R3:\$w[0-9]+]], 0($7)
   %4 = icmp sgt <16 x i8> %1, %2
   ; CHECK-DAG: clt_s.b [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -752,11 +752,11 @@ define void @bsel_s_v8i16(<8 x i16>* %d, <8 x i16>* %a, <8 x i16>* %b,
                         <8 x i16>* %c) nounwind {
   ; CHECK: bsel_s_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <8 x i16>* %c
+  %3 = load <8 x i16>, <8 x i16>* %c
   ; CHECK-DAG: ld.h [[R3:\$w[0-9]+]], 0($7)
   %4 = icmp sgt <8 x i16> %1, %2
   ; CHECK-DAG: clt_s.h [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -774,11 +774,11 @@ define void @bsel_s_v4i32(<4 x i32>* %d, <4 x i32>* %a, <4 x i32>* %b,
                         <4 x i32>* %c) nounwind {
   ; CHECK: bsel_s_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <4 x i32>* %c
+  %3 = load <4 x i32>, <4 x i32>* %c
   ; CHECK-DAG: ld.w [[R3:\$w[0-9]+]], 0($7)
   %4 = icmp sgt <4 x i32> %1, %2
   ; CHECK-DAG: clt_s.w [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -796,11 +796,11 @@ define void @bsel_s_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b,
                         <2 x i64>* %c) nounwind {
   ; CHECK: bsel_s_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <2 x i64>* %c
+  %3 = load <2 x i64>, <2 x i64>* %c
   ; CHECK-DAG: ld.d [[R3:\$w[0-9]+]], 0($7)
   %4 = icmp sgt <2 x i64> %1, %2
   ; CHECK-DAG: clt_s.d [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -818,11 +818,11 @@ define void @bsel_u_v16i8(<16 x i8>* %d, <16 x i8>* %a, <16 x i8>* %b,
                         <16 x i8>* %c) nounwind {
   ; CHECK: bsel_u_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <16 x i8>* %c
+  %3 = load <16 x i8>, <16 x i8>* %c
   ; CHECK-DAG: ld.b [[R3:\$w[0-9]+]], 0($7)
   %4 = icmp ugt <16 x i8> %1, %2
   ; CHECK-DAG: clt_u.b [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -840,11 +840,11 @@ define void @bsel_u_v8i16(<8 x i16>* %d, <8 x i16>* %a, <8 x i16>* %b,
                         <8 x i16>* %c) nounwind {
   ; CHECK: bsel_u_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <8 x i16>* %c
+  %3 = load <8 x i16>, <8 x i16>* %c
   ; CHECK-DAG: ld.h [[R3:\$w[0-9]+]], 0($7)
   %4 = icmp ugt <8 x i16> %1, %2
   ; CHECK-DAG: clt_u.h [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -862,11 +862,11 @@ define void @bsel_u_v4i32(<4 x i32>* %d, <4 x i32>* %a, <4 x i32>* %b,
                         <4 x i32>* %c) nounwind {
   ; CHECK: bsel_u_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <4 x i32>* %c
+  %3 = load <4 x i32>, <4 x i32>* %c
   ; CHECK-DAG: ld.w [[R3:\$w[0-9]+]], 0($7)
   %4 = icmp ugt <4 x i32> %1, %2
   ; CHECK-DAG: clt_u.w [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -884,11 +884,11 @@ define void @bsel_u_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b,
                         <2 x i64>* %c) nounwind {
   ; CHECK: bsel_u_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <2 x i64>* %c
+  %3 = load <2 x i64>, <2 x i64>* %c
   ; CHECK-DAG: ld.d [[R3:\$w[0-9]+]], 0($7)
   %4 = icmp ugt <2 x i64> %1, %2
   ; CHECK-DAG: clt_u.d [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -906,9 +906,9 @@ define void @bseli_s_v16i8(<16 x i8>* %d, <16 x i8>* %a, <16 x i8>* %b,
                         <16 x i8>* %c) nounwind {
   ; CHECK: bseli_s_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sgt <16 x i8> %1, %2
   ; CHECK-DAG: clt_s.b [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -925,9 +925,9 @@ define void @bseli_s_v8i16(<8 x i16>* %d, <8 x i16>* %a, <8 x i16>* %b,
                         <8 x i16>* %c) nounwind {
   ; CHECK: bseli_s_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sgt <8 x i16> %1, %2
   ; CHECK-DAG: clt_s.h [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -945,9 +945,9 @@ define void @bseli_s_v4i32(<4 x i32>* %d, <4 x i32>* %a, <4 x i32>* %b,
                         <4 x i32>* %c) nounwind {
   ; CHECK: bseli_s_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sgt <4 x i32> %1, %2
   ; CHECK-DAG: clt_s.w [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -965,9 +965,9 @@ define void @bseli_s_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b,
                         <2 x i64>* %c) nounwind {
   ; CHECK: bseli_s_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sgt <2 x i64> %1, %2
   ; CHECK-DAG: clt_s.d [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -985,9 +985,9 @@ define void @bseli_u_v16i8(<16 x i8>* %d, <16 x i8>* %a, <16 x i8>* %b,
                         <16 x i8>* %c) nounwind {
   ; CHECK: bseli_u_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ugt <16 x i8> %1, %2
   ; CHECK-DAG: clt_u.b [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -1004,9 +1004,9 @@ define void @bseli_u_v8i16(<8 x i16>* %d, <8 x i16>* %a, <8 x i16>* %b,
                         <8 x i16>* %c) nounwind {
   ; CHECK: bseli_u_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ugt <8 x i16> %1, %2
   ; CHECK-DAG: clt_u.h [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -1024,9 +1024,9 @@ define void @bseli_u_v4i32(<4 x i32>* %d, <4 x i32>* %a, <4 x i32>* %b,
                         <4 x i32>* %c) nounwind {
   ; CHECK: bseli_u_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ugt <4 x i32> %1, %2
   ; CHECK-DAG: clt_u.w [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -1044,9 +1044,9 @@ define void @bseli_u_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b,
                         <2 x i64>* %c) nounwind {
   ; CHECK: bseli_u_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ugt <2 x i64> %1, %2
   ; CHECK-DAG: clt_u.d [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -1063,9 +1063,9 @@ define void @bseli_u_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b,
 define void @max_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: max_s_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sgt <16 x i8> %1, %2
   %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> %2
@@ -1080,9 +1080,9 @@ define void @max_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @max_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: max_s_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sgt <8 x i16> %1, %2
   %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> %2
@@ -1097,9 +1097,9 @@ define void @max_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @max_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: max_s_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sgt <4 x i32> %1, %2
   %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> %2
@@ -1114,9 +1114,9 @@ define void @max_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @max_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: max_s_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sgt <2 x i64> %1, %2
   %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> %2
@@ -1131,9 +1131,9 @@ define void @max_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @max_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: max_u_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ugt <16 x i8> %1, %2
   %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> %2
@@ -1148,9 +1148,9 @@ define void @max_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @max_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: max_u_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ugt <8 x i16> %1, %2
   %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> %2
@@ -1165,9 +1165,9 @@ define void @max_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @max_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: max_u_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ugt <4 x i32> %1, %2
   %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> %2
@@ -1182,9 +1182,9 @@ define void @max_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @max_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: max_u_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ugt <2 x i64> %1, %2
   %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> %2
@@ -1199,9 +1199,9 @@ define void @max_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @max_s_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: max_s_eq_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sge <16 x i8> %1, %2
   %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> %2
@@ -1216,9 +1216,9 @@ define void @max_s_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwin
 define void @max_s_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: max_s_eq_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sge <8 x i16> %1, %2
   %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> %2
@@ -1233,9 +1233,9 @@ define void @max_s_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwin
 define void @max_s_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: max_s_eq_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sge <4 x i32> %1, %2
   %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> %2
@@ -1250,9 +1250,9 @@ define void @max_s_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwin
 define void @max_s_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: max_s_eq_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sge <2 x i64> %1, %2
   %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> %2
@@ -1267,9 +1267,9 @@ define void @max_s_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwin
 define void @max_u_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: max_u_eq_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp uge <16 x i8> %1, %2
   %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> %2
@@ -1284,9 +1284,9 @@ define void @max_u_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwin
 define void @max_u_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: max_u_eq_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp uge <8 x i16> %1, %2
   %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> %2
@@ -1301,9 +1301,9 @@ define void @max_u_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwin
 define void @max_u_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: max_u_eq_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp uge <4 x i32> %1, %2
   %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> %2
@@ -1318,9 +1318,9 @@ define void @max_u_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwin
 define void @max_u_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: max_u_eq_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp uge <2 x i64> %1, %2
   %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> %2
@@ -1335,7 +1335,7 @@ define void @max_u_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwin
 define void @maxi_s_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: maxi_s_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp sgt <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   %3 = select <16 x i1> %2, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
@@ -1350,7 +1350,7 @@ define void @maxi_s_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @maxi_s_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: maxi_s_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp sgt <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   %3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
@@ -1365,7 +1365,7 @@ define void @maxi_s_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @maxi_s_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: maxi_s_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp sgt <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   %3 = select <4 x i1> %2, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
@@ -1380,7 +1380,7 @@ define void @maxi_s_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @maxi_s_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: maxi_s_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp sgt <2 x i64> %1, <i64 1, i64 1>
   %3 = select <2 x i1> %2, <2 x i64> %1, <2 x i64> <i64 1, i64 1>
@@ -1395,7 +1395,7 @@ define void @maxi_s_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @maxi_u_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: maxi_u_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ugt <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   %3 = select <16 x i1> %2, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
@@ -1410,7 +1410,7 @@ define void @maxi_u_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @maxi_u_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: maxi_u_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ugt <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   %3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
@@ -1425,7 +1425,7 @@ define void @maxi_u_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @maxi_u_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: maxi_u_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ugt <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   %3 = select <4 x i1> %2, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
@@ -1440,7 +1440,7 @@ define void @maxi_u_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @maxi_u_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: maxi_u_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ugt <2 x i64> %1, <i64 1, i64 1>
   %3 = select <2 x i1> %2, <2 x i64> %1, <2 x i64> <i64 1, i64 1>
@@ -1455,7 +1455,7 @@ define void @maxi_u_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @maxi_s_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: maxi_s_eq_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp sge <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   %3 = select <16 x i1> %2, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
@@ -1470,7 +1470,7 @@ define void @maxi_s_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @maxi_s_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: maxi_s_eq_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp sge <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   %3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
@@ -1485,7 +1485,7 @@ define void @maxi_s_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @maxi_s_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: maxi_s_eq_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp sge <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   %3 = select <4 x i1> %2, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
@@ -1500,7 +1500,7 @@ define void @maxi_s_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @maxi_s_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: maxi_s_eq_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp sge <2 x i64> %1, <i64 1, i64 1>
   %3 = select <2 x i1> %2, <2 x i64> %1, <2 x i64> <i64 1, i64 1>
@@ -1515,7 +1515,7 @@ define void @maxi_s_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @maxi_u_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: maxi_u_eq_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp uge <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   %3 = select <16 x i1> %2, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
@@ -1530,7 +1530,7 @@ define void @maxi_u_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @maxi_u_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: maxi_u_eq_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp uge <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   %3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
@@ -1545,7 +1545,7 @@ define void @maxi_u_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @maxi_u_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: maxi_u_eq_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp uge <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   %3 = select <4 x i1> %2, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
@@ -1560,7 +1560,7 @@ define void @maxi_u_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @maxi_u_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: maxi_u_eq_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp uge <2 x i64> %1, <i64 1, i64 1>
   %3 = select <2 x i1> %2, <2 x i64> %1, <2 x i64> <i64 1, i64 1>
@@ -1575,9 +1575,9 @@ define void @maxi_u_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @min_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: min_s_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sle <16 x i8> %1, %2
   %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> %2
@@ -1592,9 +1592,9 @@ define void @min_s_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @min_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: min_s_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp slt <8 x i16> %1, %2
   %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> %2
@@ -1609,9 +1609,9 @@ define void @min_s_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @min_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: min_s_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp slt <4 x i32> %1, %2
   %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> %2
@@ -1626,9 +1626,9 @@ define void @min_s_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @min_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: min_s_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp slt <2 x i64> %1, %2
   %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> %2
@@ -1643,9 +1643,9 @@ define void @min_s_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @min_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: min_u_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ult <16 x i8> %1, %2
   %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> %2
@@ -1660,9 +1660,9 @@ define void @min_u_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
 define void @min_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: min_u_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ult <8 x i16> %1, %2
   %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> %2
@@ -1677,9 +1677,9 @@ define void @min_u_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
 define void @min_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: min_u_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ult <4 x i32> %1, %2
   %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> %2
@@ -1694,9 +1694,9 @@ define void @min_u_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
 define void @min_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: min_u_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ult <2 x i64> %1, %2
   %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> %2
@@ -1711,9 +1711,9 @@ define void @min_u_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 define void @min_s_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: min_s_eq_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sle <16 x i8> %1, %2
   %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> %2
@@ -1728,9 +1728,9 @@ define void @min_s_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwin
 define void @min_s_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: min_s_eq_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sle <8 x i16> %1, %2
   %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> %2
@@ -1745,9 +1745,9 @@ define void @min_s_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwin
 define void @min_s_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: min_s_eq_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sle <4 x i32> %1, %2
   %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> %2
@@ -1762,9 +1762,9 @@ define void @min_s_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwin
 define void @min_s_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: min_s_eq_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp sle <2 x i64> %1, %2
   %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> %2
@@ -1779,9 +1779,9 @@ define void @min_s_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwin
 define void @min_u_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK: min_u_eq_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ule <16 x i8> %1, %2
   %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> %2
@@ -1796,9 +1796,9 @@ define void @min_u_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwin
 define void @min_u_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK: min_u_eq_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ule <8 x i16> %1, %2
   %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> %2
@@ -1813,9 +1813,9 @@ define void @min_u_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwin
 define void @min_u_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK: min_u_eq_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ule <4 x i32> %1, %2
   %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> %2
@@ -1830,9 +1830,9 @@ define void @min_u_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwin
 define void @min_u_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
   ; CHECK: min_u_eq_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = icmp ule <2 x i64> %1, %2
   %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> %2
@@ -1847,7 +1847,7 @@ define void @min_u_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwin
 define void @mini_s_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: mini_s_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp slt <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   %3 = select <16 x i1> %2, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
@@ -1862,7 +1862,7 @@ define void @mini_s_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @mini_s_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: mini_s_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp slt <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   %3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
@@ -1877,7 +1877,7 @@ define void @mini_s_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @mini_s_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: mini_s_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp slt <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   %3 = select <4 x i1> %2, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
@@ -1892,7 +1892,7 @@ define void @mini_s_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @mini_s_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: mini_s_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp slt <2 x i64> %1, <i64 1, i64 1>
   %3 = select <2 x i1> %2, <2 x i64> %1, <2 x i64> <i64 1, i64 1>
@@ -1907,7 +1907,7 @@ define void @mini_s_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @mini_u_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: mini_u_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ult <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   %3 = select <16 x i1> %2, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
@@ -1922,7 +1922,7 @@ define void @mini_u_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @mini_u_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: mini_u_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ult <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   %3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
@@ -1937,7 +1937,7 @@ define void @mini_u_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @mini_u_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: mini_u_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ult <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   %3 = select <4 x i1> %2, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
@@ -1952,7 +1952,7 @@ define void @mini_u_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @mini_u_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: mini_u_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ult <2 x i64> %1, <i64 1, i64 1>
   %3 = select <2 x i1> %2, <2 x i64> %1, <2 x i64> <i64 1, i64 1>
@@ -1967,7 +1967,7 @@ define void @mini_u_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @mini_s_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: mini_s_eq_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp sle <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   %3 = select <16 x i1> %2, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
@@ -1982,7 +1982,7 @@ define void @mini_s_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @mini_s_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: mini_s_eq_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp sle <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   %3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
@@ -1997,7 +1997,7 @@ define void @mini_s_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @mini_s_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: mini_s_eq_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp sle <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   %3 = select <4 x i1> %2, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
@@ -2012,7 +2012,7 @@ define void @mini_s_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @mini_s_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: mini_s_eq_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp sle <2 x i64> %1, <i64 1, i64 1>
   %3 = select <2 x i1> %2, <2 x i64> %1, <2 x i64> <i64 1, i64 1>
@@ -2027,7 +2027,7 @@ define void @mini_s_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
 define void @mini_u_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK: mini_u_eq_v16i8:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ule <16 x i8> %1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
   %3 = select <16 x i1> %2, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
@@ -2042,7 +2042,7 @@ define void @mini_u_eq_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
 define void @mini_u_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK: mini_u_eq_v8i16:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ule <8 x i16> %1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
   %3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
@@ -2057,7 +2057,7 @@ define void @mini_u_eq_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
 define void @mini_u_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
   ; CHECK: mini_u_eq_v4i32:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ule <4 x i32> %1, <i32 1, i32 1, i32 1, i32 1>
   %3 = select <4 x i1> %2, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
@@ -2072,7 +2072,7 @@ define void @mini_u_eq_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
 define void @mini_u_eq_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK: mini_u_eq_v2i64:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = icmp ule <2 x i64> %1, <i64 1, i64 1>
   %3 = select <2 x i1> %2, <2 x i64> %1, <2 x i64> <i64 1, i64 1>
diff --git a/test/CodeGen/Mips/msa/compare_float.ll b/test/CodeGen/Mips/msa/compare_float.ll
index e93221b93612..3229d027d95d 100644
--- a/test/CodeGen/Mips/msa/compare_float.ll
+++ b/test/CodeGen/Mips/msa/compare_float.ll
@@ -9,8 +9,8 @@ declare <2 x double> @llvm.mips.fmin.d(<2 x double>, <2 x double>) nounwind
 define void @false_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: false_v4f32:
 
-  %1 = load <4 x float>* %a
-  %2 = load <4 x float>* %b
+  %1 = load <4 x float>, <4 x float>* %a
+  %2 = load <4 x float>, <4 x float>* %b
   %3 = fcmp false <4 x float> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
   store <4 x i32> %4, <4 x i32>* %c
@@ -25,8 +25,8 @@ define void @false_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwi
 define void @false_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: false_v2f64:
 
-  %1 = load <2 x double>* %a
-  %2 = load <2 x double>* %b
+  %1 = load <2 x double>, <2 x double>* %a
+  %2 = load <2 x double>, <2 x double>* %b
   %3 = fcmp false <2 x double> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
   store <2 x i64> %4, <2 x i64>* %c
@@ -41,9 +41,9 @@ define void @false_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) noun
 define void @oeq_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: oeq_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp oeq <4 x float> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -58,9 +58,9 @@ define void @oeq_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind
 define void @oeq_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: oeq_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp oeq <2 x double> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -75,9 +75,9 @@ define void @oeq_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi
 define void @oge_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: oge_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp oge <4 x float> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -92,9 +92,9 @@ define void @oge_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind
 define void @oge_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: oge_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp oge <2 x double> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -109,9 +109,9 @@ define void @oge_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi
 define void @ogt_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: ogt_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp ogt <4 x float> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -126,9 +126,9 @@ define void @ogt_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind
 define void @ogt_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: ogt_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp ogt <2 x double> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -143,9 +143,9 @@ define void @ogt_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi
 define void @ole_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: ole_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp ole <4 x float> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -160,9 +160,9 @@ define void @ole_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind
 define void @ole_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: ole_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp ole <2 x double> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -177,9 +177,9 @@ define void @ole_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi
 define void @olt_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: olt_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp olt <4 x float> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -194,9 +194,9 @@ define void @olt_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind
 define void @olt_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: olt_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp olt <2 x double> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -211,9 +211,9 @@ define void @olt_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi
 define void @one_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: one_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp one <4 x float> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -228,9 +228,9 @@ define void @one_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind
 define void @one_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: one_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp one <2 x double> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -245,9 +245,9 @@ define void @one_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi
 define void @ord_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: ord_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp ord <4 x float> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -262,9 +262,9 @@ define void @ord_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind
 define void @ord_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: ord_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp ord <2 x double> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -279,9 +279,9 @@ define void @ord_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi
 define void @ueq_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: ueq_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp ueq <4 x float> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -296,9 +296,9 @@ define void @ueq_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind
 define void @ueq_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: ueq_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp ueq <2 x double> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -313,9 +313,9 @@ define void @ueq_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi
 define void @uge_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: uge_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp uge <4 x float> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -330,9 +330,9 @@ define void @uge_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind
 define void @uge_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: uge_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp uge <2 x double> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -347,9 +347,9 @@ define void @uge_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi
 define void @ugt_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: ugt_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp ugt <4 x float> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -364,9 +364,9 @@ define void @ugt_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind
 define void @ugt_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: ugt_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp ugt <2 x double> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -381,9 +381,9 @@ define void @ugt_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi
 define void @ule_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: ule_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp ule <4 x float> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -398,9 +398,9 @@ define void @ule_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind
 define void @ule_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: ule_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp ule <2 x double> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -415,9 +415,9 @@ define void @ule_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi
 define void @ult_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: ult_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp ult <4 x float> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -432,9 +432,9 @@ define void @ult_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind
 define void @ult_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: ult_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp ult <2 x double> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -449,9 +449,9 @@ define void @ult_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi
 define void @uno_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: uno_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp uno <4 x float> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
@@ -466,9 +466,9 @@ define void @uno_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind
 define void @uno_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: uno_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp uno <2 x double> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
@@ -483,8 +483,8 @@ define void @uno_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwi
 define void @true_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: true_v4f32:
 
-  %1 = load <4 x float>* %a
-  %2 = load <4 x float>* %b
+  %1 = load <4 x float>, <4 x float>* %a
+  %2 = load <4 x float>, <4 x float>* %b
   %3 = fcmp true <4 x float> %1, %2
   %4 = sext <4 x i1> %3 to <4 x i32>
   store <4 x i32> %4, <4 x i32>* %c
@@ -499,8 +499,8 @@ define void @true_v4f32(<4 x i32>* %c, <4 x float>* %a, <4 x float>* %b) nounwin
 define void @true_v2f64(<2 x i64>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: true_v2f64:
 
-  %1 = load <2 x double>* %a
-  %2 = load <2 x double>* %b
+  %1 = load <2 x double>, <2 x double>* %a
+  %2 = load <2 x double>, <2 x double>* %b
   %3 = fcmp true <2 x double> %1, %2
   %4 = sext <2 x i1> %3 to <2 x i64>
   store <2 x i64> %4, <2 x i64>* %c
@@ -516,11 +516,11 @@ define void @bsel_v4f32(<4 x float>* %d, <4 x float>* %a, <4 x float>* %b,
                           <4 x float>* %c) nounwind {
   ; CHECK: bsel_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <4 x float>* %c
+  %3 = load <4 x float>, <4 x float>* %c
   ; CHECK-DAG: ld.w [[R3:\$w[0-9]+]], 0($7)
   %4 = fcmp ogt <4 x float> %1, %2
   ; CHECK-DAG: fclt.w [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -538,11 +538,11 @@ define void @bsel_v2f64(<2 x double>* %d, <2 x double>* %a, <2 x double>* %b,
                           <2 x double>* %c) nounwind {
   ; CHECK: bsel_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
-  %3 = load <2 x double>* %c
+  %3 = load <2 x double>, <2 x double>* %c
   ; CHECK-DAG: ld.d [[R3:\$w[0-9]+]], 0($7)
   %4 = fcmp ogt <2 x double> %1, %2
   ; CHECK-DAG: fclt.d [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -560,9 +560,9 @@ define void @bseli_v4f32(<4 x float>* %d, <4 x float>* %a, <4 x float>* %b,
                           <4 x float>* %c) nounwind {
   ; CHECK: bseli_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp ogt <4 x float> %1, %2
   ; CHECK-DAG: fclt.w [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -580,9 +580,9 @@ define void @bseli_v2f64(<2 x double>* %d, <2 x double>* %a, <2 x double>* %b,
                           <2 x double>* %c) nounwind {
   ; CHECK: bseli_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = fcmp ogt <2 x double> %1, %2
   ; CHECK-DAG: fclt.d [[R4:\$w[0-9]+]], [[R2]], [[R1]]
@@ -599,9 +599,9 @@ define void @bseli_v2f64(<2 x double>* %d, <2 x double>* %a, <2 x double>* %b,
 define void @max_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: max_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = tail call <4 x float> @llvm.mips.fmax.w(<4 x float> %1, <4 x float> %2)
   ; CHECK-DAG: fmax.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -615,9 +615,9 @@ define void @max_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwi
 define void @max_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: max_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = tail call <2 x double> @llvm.mips.fmax.d(<2 x double> %1, <2 x double> %2)
   ; CHECK-DAG: fmax.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -631,9 +631,9 @@ define void @max_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nou
 define void @min_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwind {
   ; CHECK: min_v4f32:
 
-  %1 = load <4 x float>* %a
+  %1 = load <4 x float>, <4 x float>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x float>* %b
+  %2 = load <4 x float>, <4 x float>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = tail call <4 x float> @llvm.mips.fmin.w(<4 x float> %1, <4 x float> %2)
   ; CHECK-DAG: fmin.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -647,9 +647,9 @@ define void @min_v4f32(<4 x float>* %c, <4 x float>* %a, <4 x float>* %b) nounwi
 define void @min_v2f64(<2 x double>* %c, <2 x double>* %a, <2 x double>* %b) nounwind {
   ; CHECK: min_v2f64:
 
-  %1 = load <2 x double>* %a
+  %1 = load <2 x double>, <2 x double>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x double>* %b
+  %2 = load <2 x double>, <2 x double>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = tail call <2 x double> @llvm.mips.fmin.d(<2 x double> %1, <2 x double> %2)
   ; CHECK-DAG: fmin.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
diff --git a/test/CodeGen/Mips/msa/elm_copy.ll b/test/CodeGen/Mips/msa/elm_copy.ll
index 0dd75fa3db12..2a0d74f44524 100644
--- a/test/CodeGen/Mips/msa/elm_copy.ll
+++ b/test/CodeGen/Mips/msa/elm_copy.ll
@@ -15,7 +15,7 @@
 
 define void @llvm_mips_copy_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_copy_s_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_copy_s_b_ARG1
   %1 = tail call i32 @llvm.mips.copy.s.b(<16 x i8> %0, i32 1)
   store i32 %1, i32* @llvm_mips_copy_s_b_RES
   ret void
@@ -38,7 +38,7 @@ declare i32 @llvm.mips.copy.s.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_copy_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_copy_s_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_copy_s_h_ARG1
   %1 = tail call i32 @llvm.mips.copy.s.h(<8 x i16> %0, i32 1)
   store i32 %1, i32* @llvm_mips_copy_s_h_RES
   ret void
@@ -61,7 +61,7 @@ declare i32 @llvm.mips.copy.s.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_copy_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_copy_s_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_copy_s_w_ARG1
   %1 = tail call i32 @llvm.mips.copy.s.w(<4 x i32> %0, i32 1)
   store i32 %1, i32* @llvm_mips_copy_s_w_RES
   ret void
@@ -84,7 +84,7 @@ declare i32 @llvm.mips.copy.s.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_copy_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_copy_s_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_copy_s_d_ARG1
   %1 = tail call i64 @llvm.mips.copy.s.d(<2 x i64> %0, i32 1)
   store i64 %1, i64* @llvm_mips_copy_s_d_RES
   ret void
@@ -112,7 +112,7 @@ declare i64 @llvm.mips.copy.s.d(<2 x i64>, i32) nounwind
 
 define void @llvm_mips_copy_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_copy_u_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_copy_u_b_ARG1
   %1 = tail call i32 @llvm.mips.copy.u.b(<16 x i8> %0, i32 1)
   store i32 %1, i32* @llvm_mips_copy_u_b_RES
   ret void
@@ -135,7 +135,7 @@ declare i32 @llvm.mips.copy.u.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_copy_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_copy_u_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_copy_u_h_ARG1
   %1 = tail call i32 @llvm.mips.copy.u.h(<8 x i16> %0, i32 1)
   store i32 %1, i32* @llvm_mips_copy_u_h_RES
   ret void
@@ -158,7 +158,7 @@ declare i32 @llvm.mips.copy.u.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_copy_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_copy_u_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_copy_u_w_ARG1
   %1 = tail call i32 @llvm.mips.copy.u.w(<4 x i32> %0, i32 1)
   store i32 %1, i32* @llvm_mips_copy_u_w_RES
   ret void
@@ -181,7 +181,7 @@ declare i32 @llvm.mips.copy.u.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_copy_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_copy_u_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_copy_u_d_ARG1
   %1 = tail call i64 @llvm.mips.copy.u.d(<2 x i64> %0, i32 1)
   store i64 %1, i64* @llvm_mips_copy_u_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/elm_insv.ll b/test/CodeGen/Mips/msa/elm_insv.ll
index c746e523def6..46e6289189df 100644
--- a/test/CodeGen/Mips/msa/elm_insv.ll
+++ b/test/CodeGen/Mips/msa/elm_insv.ll
@@ -16,8 +16,8 @@
 
 define void @llvm_mips_insert_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_insert_b_ARG1
-  %1 = load i32* @llvm_mips_insert_b_ARG3
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_insert_b_ARG1
+  %1 = load i32, i32* @llvm_mips_insert_b_ARG3
   %2 = tail call <16 x i8> @llvm.mips.insert.b(<16 x i8> %0, i32 1, i32 %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_insert_b_RES
   ret void
@@ -38,8 +38,8 @@ declare <16 x i8> @llvm.mips.insert.b(<16 x i8>, i32, i32) nounwind
 
 define void @llvm_mips_insert_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_insert_h_ARG1
-  %1 = load i32* @llvm_mips_insert_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_insert_h_ARG1
+  %1 = load i32, i32* @llvm_mips_insert_h_ARG3
   %2 = tail call <8 x i16> @llvm.mips.insert.h(<8 x i16> %0, i32 1, i32 %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_insert_h_RES
   ret void
@@ -60,8 +60,8 @@ declare <8 x i16> @llvm.mips.insert.h(<8 x i16>, i32, i32) nounwind
 
 define void @llvm_mips_insert_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_insert_w_ARG1
-  %1 = load i32* @llvm_mips_insert_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_insert_w_ARG1
+  %1 = load i32, i32* @llvm_mips_insert_w_ARG3
   %2 = tail call <4 x i32> @llvm.mips.insert.w(<4 x i32> %0, i32 1, i32 %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_insert_w_RES
   ret void
@@ -82,8 +82,8 @@ declare <4 x i32> @llvm.mips.insert.w(<4 x i32>, i32, i32) nounwind
 
 define void @llvm_mips_insert_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_insert_d_ARG1
-  %1 = load i64* @llvm_mips_insert_d_ARG3
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_insert_d_ARG1
+  %1 = load i64, i64* @llvm_mips_insert_d_ARG3
   %2 = tail call <2 x i64> @llvm.mips.insert.d(<2 x i64> %0, i32 1, i64 %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_insert_d_RES
   ret void
@@ -110,8 +110,8 @@ declare <2 x i64> @llvm.mips.insert.d(<2 x i64>, i32, i64) nounwind
 
 define void @llvm_mips_insve_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_insve_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_insve_b_ARG3
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_insve_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_insve_b_ARG3
   %2 = tail call <16 x i8> @llvm.mips.insve.b(<16 x i8> %0, i32 1, <16 x i8> %1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_insve_b_RES
   ret void
@@ -136,8 +136,8 @@ declare <16 x i8> @llvm.mips.insve.b(<16 x i8>, i32, <16 x i8>) nounwind
 
 define void @llvm_mips_insve_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_insve_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_insve_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_insve_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_insve_h_ARG3
   %2 = tail call <8 x i16> @llvm.mips.insve.h(<8 x i16> %0, i32 1, <8 x i16> %1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_insve_h_RES
   ret void
@@ -162,8 +162,8 @@ declare <8 x i16> @llvm.mips.insve.h(<8 x i16>, i32, <8 x i16>) nounwind
 
 define void @llvm_mips_insve_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_insve_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_insve_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_insve_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_insve_w_ARG3
   %2 = tail call <4 x i32> @llvm.mips.insve.w(<4 x i32> %0, i32 1, <4 x i32> %1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_insve_w_RES
   ret void
@@ -188,8 +188,8 @@ declare <4 x i32> @llvm.mips.insve.w(<4 x i32>, i32, <4 x i32>) nounwind
 
 define void @llvm_mips_insve_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_insve_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_insve_d_ARG3
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_insve_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_insve_d_ARG3
   %2 = tail call <2 x i64> @llvm.mips.insve.d(<2 x i64> %0, i32 1, <2 x i64> %1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_insve_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/elm_move.ll b/test/CodeGen/Mips/msa/elm_move.ll
index 98c06c732c36..9665b6d688f3 100644
--- a/test/CodeGen/Mips/msa/elm_move.ll
+++ b/test/CodeGen/Mips/msa/elm_move.ll
@@ -9,7 +9,7 @@
 
 define void @llvm_mips_move_vb_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_move_vb_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_move_vb_ARG1
   %1 = tail call <16 x i8> @llvm.mips.move.v(<16 x i8> %0)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_move_vb_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/elm_shift_slide.ll b/test/CodeGen/Mips/msa/elm_shift_slide.ll
index 00a6544b1207..87f15f1a8c92 100644
--- a/test/CodeGen/Mips/msa/elm_shift_slide.ll
+++ b/test/CodeGen/Mips/msa/elm_shift_slide.ll
@@ -10,8 +10,8 @@
 
 define void @llvm_mips_sldi_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_sldi_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_sldi_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_sldi_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_sldi_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.sldi.b(<16 x i8> %0, <16 x i8> %1, i32 1)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_sldi_b_RES
   ret void
@@ -31,8 +31,8 @@ declare <16 x i8> @llvm.mips.sldi.b(<16 x i8>, <16 x i8>, i32) nounwind
 
 define void @llvm_mips_sldi_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_sldi_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_sldi_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_sldi_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_sldi_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.sldi.h(<8 x i16> %0, <8 x i16> %1, i32 1)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_sldi_h_RES
   ret void
@@ -52,8 +52,8 @@ declare <8 x i16> @llvm.mips.sldi.h(<8 x i16>, <8 x i16>, i32) nounwind
 
 define void @llvm_mips_sldi_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_sldi_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_sldi_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_sldi_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_sldi_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.sldi.w(<4 x i32> %0, <4 x i32> %1, i32 1)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_sldi_w_RES
   ret void
@@ -73,8 +73,8 @@ declare <4 x i32> @llvm.mips.sldi.w(<4 x i32>, <4 x i32>, i32) nounwind
 
 define void @llvm_mips_sldi_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_sldi_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_sldi_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_sldi_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_sldi_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.sldi.d(<2 x i64> %0, <2 x i64> %1, i32 1)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_sldi_d_RES
   ret void
@@ -93,7 +93,7 @@ declare <2 x i64> @llvm.mips.sldi.d(<2 x i64>, <2 x i64>, i32) nounwind
 
 define void @llvm_mips_splati_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_splati_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_splati_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.splati.b(<16 x i8> %0, i32 1)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_splati_b_RES
   ret void
@@ -112,7 +112,7 @@ declare <16 x i8> @llvm.mips.splati.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_splati_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_splati_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_splati_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.splati.h(<8 x i16> %0, i32 1)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_splati_h_RES
   ret void
@@ -131,7 +131,7 @@ declare <8 x i16> @llvm.mips.splati.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_splati_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_splati_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_splati_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.splati.w(<4 x i32> %0, i32 1)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_splati_w_RES
   ret void
@@ -150,7 +150,7 @@ declare <4 x i32> @llvm.mips.splati.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_splati_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_splati_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_splati_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.splati.d(<2 x i64> %0, i32 1)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_splati_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/frameindex.ll b/test/CodeGen/Mips/msa/frameindex.ll
index ebec465a3e33..afd28ae184dd 100644
--- a/test/CodeGen/Mips/msa/frameindex.ll
+++ b/test/CodeGen/Mips/msa/frameindex.ll
@@ -5,7 +5,7 @@ define void @loadstore_v16i8_near() nounwind {
   ; MIPS32-AE: loadstore_v16i8_near:
 
   %1 = alloca <16 x i8>
-  %2 = load volatile <16 x i8>* %1
+  %2 = load volatile <16 x i8>, <16 x i8>* %1
   ; MIPS32-AE: ld.b [[R1:\$w[0-9]+]], 0($sp)
   store volatile <16 x i8> %2, <16 x i8>* %1
   ; MIPS32-AE: st.b [[R1]], 0($sp)
@@ -20,7 +20,7 @@ define void @loadstore_v16i8_just_under_simm10() nounwind {
   %1 = alloca <16 x i8>
   %2 = alloca [496 x i8] ; Push the frame right up to 512 bytes
 
-  %3 = load volatile <16 x i8>* %1
+  %3 = load volatile <16 x i8>, <16 x i8>* %1
   ; MIPS32-AE: ld.b [[R1:\$w[0-9]+]], 496($sp)
   store volatile <16 x i8> %3, <16 x i8>* %1
   ; MIPS32-AE: st.b [[R1]], 496($sp)
@@ -35,7 +35,7 @@ define void @loadstore_v16i8_just_over_simm10() nounwind {
   %1 = alloca <16 x i8>
   %2 = alloca [497 x i8] ; Push the frame just over 512 bytes
 
-  %3 = load volatile <16 x i8>* %1
+  %3 = load volatile <16 x i8>, <16 x i8>* %1
   ; MIPS32-AE: addiu [[BASE:\$([0-9]+|gp)]], $sp, 512
   ; MIPS32-AE: ld.b [[R1:\$w[0-9]+]], 0([[BASE]])
   store volatile <16 x i8> %3, <16 x i8>* %1
@@ -52,7 +52,7 @@ define void @loadstore_v16i8_just_under_simm16() nounwind {
   %1 = alloca <16 x i8>
   %2 = alloca [32752 x i8] ; Push the frame right up to 32768 bytes
 
-  %3 = load volatile <16 x i8>* %1
+  %3 = load volatile <16 x i8>, <16 x i8>* %1
   ; MIPS32-AE: ori [[R2:\$([0-9]+|gp)]], $zero, 32768
   ; MIPS32-AE: addu [[BASE:\$([0-9]+|gp)]], $sp, [[R2]]
   ; MIPS32-AE: ld.b [[R1:\$w[0-9]+]], 0([[BASE]])
@@ -71,7 +71,7 @@ define void @loadstore_v16i8_just_over_simm16() nounwind {
   %1 = alloca <16 x i8>
   %2 = alloca [32753 x i8] ; Push the frame just over 32768 bytes
 
-  %3 = load volatile <16 x i8>* %1
+  %3 = load volatile <16 x i8>, <16 x i8>* %1
   ; MIPS32-AE: ori [[R2:\$([0-9]+|gp)]], $zero, 32768
   ; MIPS32-AE: addu [[BASE:\$([0-9]+|gp)]], $sp, [[R2]]
   ; MIPS32-AE: ld.b [[R1:\$w[0-9]+]], 0([[BASE]])
@@ -88,7 +88,7 @@ define void @loadstore_v8i16_near() nounwind {
   ; MIPS32-AE: loadstore_v8i16_near:
 
   %1 = alloca <8 x i16>
-  %2 = load volatile <8 x i16>* %1
+  %2 = load volatile <8 x i16>, <8 x i16>* %1
   ; MIPS32-AE: ld.h [[R1:\$w[0-9]+]], 0($sp)
   store volatile <8 x i16> %2, <8 x i16>* %1
   ; MIPS32-AE: st.h [[R1]], 0($sp)
@@ -102,11 +102,11 @@ define void @loadstore_v8i16_unaligned() nounwind {
 
   %1 = alloca [2 x <8 x i16>]
   %2 = bitcast [2 x <8 x i16>]* %1 to i8*
-  %3 = getelementptr i8* %2, i32 1
+  %3 = getelementptr i8, i8* %2, i32 1
   %4 = bitcast i8* %3 to [2 x <8 x i16>]*
-  %5 = getelementptr [2 x <8 x i16>]* %4, i32 0, i32 0
+  %5 = getelementptr [2 x <8 x i16>], [2 x <8 x i16>]* %4, i32 0, i32 0
 
-  %6 = load volatile <8 x i16>* %5
+  %6 = load volatile <8 x i16>, <8 x i16>* %5
   ; MIPS32-AE: addiu [[BASE:\$([0-9]+|gp)]], $sp, 1
   ; MIPS32-AE: ld.h [[R1:\$w[0-9]+]], 0([[BASE]])
   store volatile <8 x i16> %6, <8 x i16>* %5
@@ -123,7 +123,7 @@ define void @loadstore_v8i16_just_under_simm10() nounwind {
   %1 = alloca <8 x i16>
   %2 = alloca [1008 x i8] ; Push the frame right up to 1024 bytes
 
-  %3 = load volatile <8 x i16>* %1
+  %3 = load volatile <8 x i16>, <8 x i16>* %1
   ; MIPS32-AE: ld.h [[R1:\$w[0-9]+]], 1008($sp)
   store volatile <8 x i16> %3, <8 x i16>* %1
   ; MIPS32-AE: st.h [[R1]], 1008($sp)
@@ -138,7 +138,7 @@ define void @loadstore_v8i16_just_over_simm10() nounwind {
   %1 = alloca <8 x i16>
   %2 = alloca [1009 x i8] ; Push the frame just over 1024 bytes
 
-  %3 = load volatile <8 x i16>* %1
+  %3 = load volatile <8 x i16>, <8 x i16>* %1
   ; MIPS32-AE: addiu [[BASE:\$([0-9]+|gp)]], $sp, 1024
   ; MIPS32-AE: ld.h [[R1:\$w[0-9]+]], 0([[BASE]])
   store volatile <8 x i16> %3, <8 x i16>* %1
@@ -155,7 +155,7 @@ define void @loadstore_v8i16_just_under_simm16() nounwind {
   %1 = alloca <8 x i16>
   %2 = alloca [32752 x i8] ; Push the frame right up to 32768 bytes
 
-  %3 = load volatile <8 x i16>* %1
+  %3 = load volatile <8 x i16>, <8 x i16>* %1
   ; MIPS32-AE: ori [[R2:\$([0-9]+|gp)]], $zero, 32768
   ; MIPS32-AE: addu [[BASE:\$([0-9]+|gp)]], $sp, [[R2]]
   ; MIPS32-AE: ld.h [[R1:\$w[0-9]+]], 0([[BASE]])
@@ -174,7 +174,7 @@ define void @loadstore_v8i16_just_over_simm16() nounwind {
   %1 = alloca <8 x i16>
   %2 = alloca [32753 x i8] ; Push the frame just over 32768 bytes
 
-  %3 = load volatile <8 x i16>* %1
+  %3 = load volatile <8 x i16>, <8 x i16>* %1
   ; MIPS32-AE: ori [[R2:\$([0-9]+|gp)]], $zero, 32768
   ; MIPS32-AE: addu [[BASE:\$([0-9]+|gp)]], $sp, [[R2]]
   ; MIPS32-AE: ld.h [[R1:\$w[0-9]+]], 0([[BASE]])
@@ -191,7 +191,7 @@ define void @loadstore_v4i32_near() nounwind {
   ; MIPS32-AE: loadstore_v4i32_near:
 
   %1 = alloca <4 x i32>
-  %2 = load volatile <4 x i32>* %1
+  %2 = load volatile <4 x i32>, <4 x i32>* %1
   ; MIPS32-AE: ld.w [[R1:\$w[0-9]+]], 0($sp)
   store volatile <4 x i32> %2, <4 x i32>* %1
   ; MIPS32-AE: st.w [[R1]], 0($sp)
@@ -205,11 +205,11 @@ define void @loadstore_v4i32_unaligned() nounwind {
 
   %1 = alloca [2 x <4 x i32>]
   %2 = bitcast [2 x <4 x i32>]* %1 to i8*
-  %3 = getelementptr i8* %2, i32 1
+  %3 = getelementptr i8, i8* %2, i32 1
   %4 = bitcast i8* %3 to [2 x <4 x i32>]*
-  %5 = getelementptr [2 x <4 x i32>]* %4, i32 0, i32 0
+  %5 = getelementptr [2 x <4 x i32>], [2 x <4 x i32>]* %4, i32 0, i32 0
 
-  %6 = load volatile <4 x i32>* %5
+  %6 = load volatile <4 x i32>, <4 x i32>* %5
   ; MIPS32-AE: addiu [[BASE:\$([0-9]+|gp)]], $sp, 1
   ; MIPS32-AE: ld.w [[R1:\$w[0-9]+]], 0([[BASE]])
   store volatile <4 x i32> %6, <4 x i32>* %5
@@ -226,7 +226,7 @@ define void @loadstore_v4i32_just_under_simm10() nounwind {
   %1 = alloca <4 x i32>
   %2 = alloca [2032 x i8] ; Push the frame right up to 2048 bytes
 
-  %3 = load volatile <4 x i32>* %1
+  %3 = load volatile <4 x i32>, <4 x i32>* %1
   ; MIPS32-AE: ld.w [[R1:\$w[0-9]+]], 2032($sp)
   store volatile <4 x i32> %3, <4 x i32>* %1
   ; MIPS32-AE: st.w [[R1]], 2032($sp)
@@ -241,7 +241,7 @@ define void @loadstore_v4i32_just_over_simm10() nounwind {
   %1 = alloca <4 x i32>
   %2 = alloca [2033 x i8] ; Push the frame just over 2048 bytes
 
-  %3 = load volatile <4 x i32>* %1
+  %3 = load volatile <4 x i32>, <4 x i32>* %1
   ; MIPS32-AE: addiu [[BASE:\$([0-9]+|gp)]], $sp, 2048
   ; MIPS32-AE: ld.w [[R1:\$w[0-9]+]], 0([[BASE]])
   store volatile <4 x i32> %3, <4 x i32>* %1
@@ -258,7 +258,7 @@ define void @loadstore_v4i32_just_under_simm16() nounwind {
   %1 = alloca <4 x i32>
   %2 = alloca [32752 x i8] ; Push the frame right up to 32768 bytes
 
-  %3 = load volatile <4 x i32>* %1
+  %3 = load volatile <4 x i32>, <4 x i32>* %1
   ; MIPS32-AE: ori [[R2:\$([0-9]+|gp)]], $zero, 32768
   ; MIPS32-AE: addu [[BASE:\$([0-9]+|gp)]], $sp, [[R2]]
   ; MIPS32-AE: ld.w [[R1:\$w[0-9]+]], 0([[BASE]])
@@ -277,7 +277,7 @@ define void @loadstore_v4i32_just_over_simm16() nounwind {
   %1 = alloca <4 x i32>
   %2 = alloca [32753 x i8] ; Push the frame just over 32768 bytes
 
-  %3 = load volatile <4 x i32>* %1
+  %3 = load volatile <4 x i32>, <4 x i32>* %1
   ; MIPS32-AE: ori [[R2:\$([0-9]+|gp)]], $zero, 32768
   ; MIPS32-AE: addu [[BASE:\$([0-9]+|gp)]], $sp, [[R2]]
   ; MIPS32-AE: ld.w [[R1:\$w[0-9]+]], 0([[BASE]])
@@ -294,7 +294,7 @@ define void @loadstore_v2i64_near() nounwind {
   ; MIPS32-AE: loadstore_v2i64_near:
 
   %1 = alloca <2 x i64>
-  %2 = load volatile <2 x i64>* %1
+  %2 = load volatile <2 x i64>, <2 x i64>* %1
   ; MIPS32-AE: ld.d [[R1:\$w[0-9]+]], 0($sp)
   store volatile <2 x i64> %2, <2 x i64>* %1
   ; MIPS32-AE: st.d [[R1]], 0($sp)
@@ -308,11 +308,11 @@ define void @loadstore_v2i64_unaligned() nounwind {
 
   %1 = alloca [2 x <2 x i64>]
   %2 = bitcast [2 x <2 x i64>]* %1 to i8*
-  %3 = getelementptr i8* %2, i32 1
+  %3 = getelementptr i8, i8* %2, i32 1
   %4 = bitcast i8* %3 to [2 x <2 x i64>]*
-  %5 = getelementptr [2 x <2 x i64>]* %4, i32 0, i32 0
+  %5 = getelementptr [2 x <2 x i64>], [2 x <2 x i64>]* %4, i32 0, i32 0
 
-  %6 = load volatile <2 x i64>* %5
+  %6 = load volatile <2 x i64>, <2 x i64>* %5
   ; MIPS32-AE: addiu [[BASE:\$([0-9]+|gp)]], $sp, 1
   ; MIPS32-AE: ld.d [[R1:\$w[0-9]+]], 0([[BASE]])
   store volatile <2 x i64> %6, <2 x i64>* %5
@@ -329,7 +329,7 @@ define void @loadstore_v2i64_just_under_simm10() nounwind {
   %1 = alloca <2 x i64>
   %2 = alloca [4080 x i8] ; Push the frame right up to 4096 bytes
 
-  %3 = load volatile <2 x i64>* %1
+  %3 = load volatile <2 x i64>, <2 x i64>* %1
   ; MIPS32-AE: ld.d [[R1:\$w[0-9]+]], 4080($sp)
   store volatile <2 x i64> %3, <2 x i64>* %1
   ; MIPS32-AE: st.d [[R1]], 4080($sp)
@@ -344,7 +344,7 @@ define void @loadstore_v2i64_just_over_simm10() nounwind {
   %1 = alloca <2 x i64>
   %2 = alloca [4081 x i8] ; Push the frame just over 4096 bytes
 
-  %3 = load volatile <2 x i64>* %1
+  %3 = load volatile <2 x i64>, <2 x i64>* %1
   ; MIPS32-AE: addiu [[BASE:\$([0-9]+|gp)]], $sp, 4096
   ; MIPS32-AE: ld.d [[R1:\$w[0-9]+]], 0([[BASE]])
   store volatile <2 x i64> %3, <2 x i64>* %1
@@ -361,7 +361,7 @@ define void @loadstore_v2i64_just_under_simm16() nounwind {
   %1 = alloca <2 x i64>
   %2 = alloca [32752 x i8] ; Push the frame right up to 32768 bytes
 
-  %3 = load volatile <2 x i64>* %1
+  %3 = load volatile <2 x i64>, <2 x i64>* %1
   ; MIPS32-AE: ori [[R2:\$([0-9]+|gp)]], $zero, 32768
   ; MIPS32-AE: addu [[BASE:\$([0-9]+|gp)]], $sp, [[R2]]
   ; MIPS32-AE: ld.d [[R1:\$w[0-9]+]], 0([[BASE]])
@@ -380,7 +380,7 @@ define void @loadstore_v2i64_just_over_simm16() nounwind {
   %1 = alloca <2 x i64>
   %2 = alloca [32753 x i8] ; Push the frame just over 32768 bytes
 
-  %3 = load volatile <2 x i64>* %1
+  %3 = load volatile <2 x i64>, <2 x i64>* %1
   ; MIPS32-AE: ori [[R2:\$([0-9]+|gp)]], $zero, 32768
   ; MIPS32-AE: addu [[BASE:\$([0-9]+|gp)]], $sp, [[R2]]
   ; MIPS32-AE: ld.d [[R1:\$w[0-9]+]], 0([[BASE]])
diff --git a/test/CodeGen/Mips/msa/i10.ll b/test/CodeGen/Mips/msa/i10.ll
index c5a96174a734..204884bbf025 100644
--- a/test/CodeGen/Mips/msa/i10.ll
+++ b/test/CodeGen/Mips/msa/i10.ll
@@ -7,7 +7,7 @@
 
 define i32 @llvm_mips_bnz_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_bnz_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bnz_b_ARG1
   %1 = tail call i32 @llvm.mips.bnz.b(<16 x i8> %0)
   %2 = icmp eq i32 %1, 0
   br i1 %2, label %true, label %false
@@ -28,7 +28,7 @@ declare i32 @llvm.mips.bnz.b(<16 x i8>) nounwind
 
 define i32 @llvm_mips_bnz_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_bnz_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bnz_h_ARG1
   %1 = tail call i32 @llvm.mips.bnz.h(<8 x i16> %0)
   %2 = icmp eq i32 %1, 0
   br i1 %2, label %true, label %false
@@ -49,7 +49,7 @@ declare i32 @llvm.mips.bnz.h(<8 x i16>) nounwind
 
 define i32 @llvm_mips_bnz_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_bnz_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bnz_w_ARG1
   %1 = tail call i32 @llvm.mips.bnz.w(<4 x i32> %0)
   %2 = icmp eq i32 %1, 0
   br i1 %2, label %true, label %false
@@ -70,7 +70,7 @@ declare i32 @llvm.mips.bnz.w(<4 x i32>) nounwind
 
 define i32 @llvm_mips_bnz_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_bnz_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bnz_d_ARG1
   %1 = tail call i32 @llvm.mips.bnz.d(<2 x i64> %0)
   %2 = icmp eq i32 %1, 0
   br i1 %2, label %true, label %false
diff --git a/test/CodeGen/Mips/msa/i5-a.ll b/test/CodeGen/Mips/msa/i5-a.ll
index 0b507208f429..f9486b17e0ad 100644
--- a/test/CodeGen/Mips/msa/i5-a.ll
+++ b/test/CodeGen/Mips/msa/i5-a.ll
@@ -9,7 +9,7 @@
 
 define void @llvm_mips_addvi_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_addvi_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_addvi_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.addvi.b(<16 x i8> %0, i32 14)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_addvi_b_RES
   ret void
@@ -28,7 +28,7 @@ declare <16 x i8> @llvm.mips.addvi.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_addvi_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_addvi_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_addvi_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.addvi.h(<8 x i16> %0, i32 14)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_addvi_h_RES
   ret void
@@ -47,7 +47,7 @@ declare <8 x i16> @llvm.mips.addvi.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_addvi_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_addvi_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_addvi_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.addvi.w(<4 x i32> %0, i32 14)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_addvi_w_RES
   ret void
@@ -66,7 +66,7 @@ declare <4 x i32> @llvm.mips.addvi.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_addvi_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_addvi_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_addvi_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.addvi.d(<2 x i64> %0, i32 14)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_addvi_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/i5-b.ll b/test/CodeGen/Mips/msa/i5-b.ll
index da6be669f0dd..40ab095f6809 100644
--- a/test/CodeGen/Mips/msa/i5-b.ll
+++ b/test/CodeGen/Mips/msa/i5-b.ll
@@ -9,7 +9,7 @@
 
 define void @llvm_mips_bclri_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_bclri_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bclri_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.bclri.b(<16 x i8> %0, i32 7)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_bclri_b_RES
   ret void
@@ -29,7 +29,7 @@ declare <16 x i8> @llvm.mips.bclri.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_bclri_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_bclri_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bclri_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.bclri.h(<8 x i16> %0, i32 7)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_bclri_h_RES
   ret void
@@ -48,7 +48,7 @@ declare <8 x i16> @llvm.mips.bclri.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_bclri_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_bclri_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bclri_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.bclri.w(<4 x i32> %0, i32 7)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_bclri_w_RES
   ret void
@@ -67,7 +67,7 @@ declare <4 x i32> @llvm.mips.bclri.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_bclri_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_bclri_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bclri_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.bclri.d(<2 x i64> %0, i32 7)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_bclri_d_RES
   ret void
@@ -87,8 +87,8 @@ declare <2 x i64> @llvm.mips.bclri.d(<2 x i64>, i32) nounwind
 
 define void @llvm_mips_binsli_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_binsli_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_binsli_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_binsli_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_binsli_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.binsli.b(<16 x i8> %0, <16 x i8> %1, i32 7)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_binsli_b_RES
   ret void
@@ -112,8 +112,8 @@ declare <16 x i8> @llvm.mips.binsli.b(<16 x i8>, <16 x i8>, i32) nounwind
 
 define void @llvm_mips_binsli_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_binsli_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_binsli_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_binsli_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_binsli_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.binsli.h(<8 x i16> %0, <8 x i16> %1, i32 7)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_binsli_h_RES
   ret void
@@ -137,8 +137,8 @@ declare <8 x i16> @llvm.mips.binsli.h(<8 x i16>, <8 x i16>, i32) nounwind
 
 define void @llvm_mips_binsli_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_binsli_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_binsli_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_binsli_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_binsli_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.binsli.w(<4 x i32> %0, <4 x i32> %1, i32 7)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_binsli_w_RES
   ret void
@@ -162,8 +162,8 @@ declare <4 x i32> @llvm.mips.binsli.w(<4 x i32>, <4 x i32>, i32) nounwind
 
 define void @llvm_mips_binsli_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_binsli_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_binsli_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_binsli_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_binsli_d_ARG2
   ; TODO: We use a particularly wide mask here to work around a legalization
   ;       issue. If the mask doesn't fit within a 10-bit immediate, it gets
   ;       legalized into a constant pool. We should add a test to cover the
@@ -191,8 +191,8 @@ declare <2 x i64> @llvm.mips.binsli.d(<2 x i64>, <2 x i64>, i32) nounwind
 
 define void @llvm_mips_binsri_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_binsri_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_binsri_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_binsri_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_binsri_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.binsri.b(<16 x i8> %0, <16 x i8> %1, i32 7)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_binsri_b_RES
   ret void
@@ -216,8 +216,8 @@ declare <16 x i8> @llvm.mips.binsri.b(<16 x i8>, <16 x i8>, i32) nounwind
 
 define void @llvm_mips_binsri_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_binsri_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_binsri_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_binsri_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_binsri_h_ARG2
   %2 = tail call <8 x i16> @llvm.mips.binsri.h(<8 x i16> %0, <8 x i16> %1, i32 7)
   store <8 x i16> %2, <8 x i16>* @llvm_mips_binsri_h_RES
   ret void
@@ -241,8 +241,8 @@ declare <8 x i16> @llvm.mips.binsri.h(<8 x i16>, <8 x i16>, i32) nounwind
 
 define void @llvm_mips_binsri_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_binsri_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_binsri_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_binsri_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_binsri_w_ARG2
   %2 = tail call <4 x i32> @llvm.mips.binsri.w(<4 x i32> %0, <4 x i32> %1, i32 7)
   store <4 x i32> %2, <4 x i32>* @llvm_mips_binsri_w_RES
   ret void
@@ -266,8 +266,8 @@ declare <4 x i32> @llvm.mips.binsri.w(<4 x i32>, <4 x i32>, i32) nounwind
 
 define void @llvm_mips_binsri_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_binsri_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_binsri_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_binsri_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_binsri_d_ARG2
   %2 = tail call <2 x i64> @llvm.mips.binsri.d(<2 x i64> %0, <2 x i64> %1, i32 7)
   store <2 x i64> %2, <2 x i64>* @llvm_mips_binsri_d_RES
   ret void
@@ -290,7 +290,7 @@ declare <2 x i64> @llvm.mips.binsri.d(<2 x i64>, <2 x i64>, i32) nounwind
 
 define void @llvm_mips_bnegi_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_bnegi_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bnegi_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.bnegi.b(<16 x i8> %0, i32 7)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_bnegi_b_RES
   ret void
@@ -309,7 +309,7 @@ declare <16 x i8> @llvm.mips.bnegi.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_bnegi_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_bnegi_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bnegi_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.bnegi.h(<8 x i16> %0, i32 7)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_bnegi_h_RES
   ret void
@@ -328,7 +328,7 @@ declare <8 x i16> @llvm.mips.bnegi.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_bnegi_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_bnegi_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bnegi_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.bnegi.w(<4 x i32> %0, i32 7)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_bnegi_w_RES
   ret void
@@ -347,7 +347,7 @@ declare <4 x i32> @llvm.mips.bnegi.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_bnegi_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_bnegi_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bnegi_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.bnegi.d(<2 x i64> %0, i32 7)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_bnegi_d_RES
   ret void
@@ -366,7 +366,7 @@ declare <2 x i64> @llvm.mips.bnegi.d(<2 x i64>, i32) nounwind
 
 define void @llvm_mips_bseti_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_bseti_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bseti_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.bseti.b(<16 x i8> %0, i32 7)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_bseti_b_RES
   ret void
@@ -385,7 +385,7 @@ declare <16 x i8> @llvm.mips.bseti.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_bseti_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_bseti_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bseti_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.bseti.h(<8 x i16> %0, i32 7)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_bseti_h_RES
   ret void
@@ -404,7 +404,7 @@ declare <8 x i16> @llvm.mips.bseti.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_bseti_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_bseti_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bseti_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.bseti.w(<4 x i32> %0, i32 7)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_bseti_w_RES
   ret void
@@ -423,7 +423,7 @@ declare <4 x i32> @llvm.mips.bseti.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_bseti_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_bseti_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bseti_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.bseti.d(<2 x i64> %0, i32 7)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_bseti_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/i5-c.ll b/test/CodeGen/Mips/msa/i5-c.ll
index bf1578f30f32..815825013ea1 100644
--- a/test/CodeGen/Mips/msa/i5-c.ll
+++ b/test/CodeGen/Mips/msa/i5-c.ll
@@ -9,7 +9,7 @@
 
 define void @llvm_mips_ceqi_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_ceqi_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ceqi_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.ceqi.b(<16 x i8> %0, i32 14)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_ceqi_b_RES
   ret void
@@ -28,7 +28,7 @@ declare <16 x i8> @llvm.mips.ceqi.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_ceqi_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_ceqi_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_ceqi_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.ceqi.h(<8 x i16> %0, i32 14)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_ceqi_h_RES
   ret void
@@ -47,7 +47,7 @@ declare <8 x i16> @llvm.mips.ceqi.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_ceqi_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_ceqi_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_ceqi_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.ceqi.w(<4 x i32> %0, i32 14)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_ceqi_w_RES
   ret void
@@ -66,7 +66,7 @@ declare <4 x i32> @llvm.mips.ceqi.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_ceqi_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_ceqi_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_ceqi_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.ceqi.d(<2 x i64> %0, i32 14)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_ceqi_d_RES
   ret void
@@ -85,7 +85,7 @@ declare <2 x i64> @llvm.mips.ceqi.d(<2 x i64>, i32) nounwind
 
 define void @llvm_mips_clei_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_clei_s_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_clei_s_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.clei.s.b(<16 x i8> %0, i32 14)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_clei_s_b_RES
   ret void
@@ -104,7 +104,7 @@ declare <16 x i8> @llvm.mips.clei.s.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_clei_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_clei_s_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_clei_s_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.clei.s.h(<8 x i16> %0, i32 14)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_clei_s_h_RES
   ret void
@@ -123,7 +123,7 @@ declare <8 x i16> @llvm.mips.clei.s.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_clei_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_clei_s_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_clei_s_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.clei.s.w(<4 x i32> %0, i32 14)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_clei_s_w_RES
   ret void
@@ -142,7 +142,7 @@ declare <4 x i32> @llvm.mips.clei.s.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_clei_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_clei_s_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_clei_s_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.clei.s.d(<2 x i64> %0, i32 14)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_clei_s_d_RES
   ret void
@@ -161,7 +161,7 @@ declare <2 x i64> @llvm.mips.clei.s.d(<2 x i64>, i32) nounwind
 
 define void @llvm_mips_clei_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_clei_u_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_clei_u_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.clei.u.b(<16 x i8> %0, i32 14)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_clei_u_b_RES
   ret void
@@ -180,7 +180,7 @@ declare <16 x i8> @llvm.mips.clei.u.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_clei_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_clei_u_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_clei_u_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.clei.u.h(<8 x i16> %0, i32 14)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_clei_u_h_RES
   ret void
@@ -199,7 +199,7 @@ declare <8 x i16> @llvm.mips.clei.u.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_clei_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_clei_u_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_clei_u_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.clei.u.w(<4 x i32> %0, i32 14)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_clei_u_w_RES
   ret void
@@ -218,7 +218,7 @@ declare <4 x i32> @llvm.mips.clei.u.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_clei_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_clei_u_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_clei_u_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.clei.u.d(<2 x i64> %0, i32 14)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_clei_u_d_RES
   ret void
@@ -237,7 +237,7 @@ declare <2 x i64> @llvm.mips.clei.u.d(<2 x i64>, i32) nounwind
 
 define void @llvm_mips_clti_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_clti_s_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_clti_s_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.clti.s.b(<16 x i8> %0, i32 14)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_clti_s_b_RES
   ret void
@@ -256,7 +256,7 @@ declare <16 x i8> @llvm.mips.clti.s.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_clti_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_clti_s_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_clti_s_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.clti.s.h(<8 x i16> %0, i32 14)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_clti_s_h_RES
   ret void
@@ -275,7 +275,7 @@ declare <8 x i16> @llvm.mips.clti.s.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_clti_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_clti_s_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_clti_s_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.clti.s.w(<4 x i32> %0, i32 14)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_clti_s_w_RES
   ret void
@@ -294,7 +294,7 @@ declare <4 x i32> @llvm.mips.clti.s.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_clti_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_clti_s_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_clti_s_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.clti.s.d(<2 x i64> %0, i32 14)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_clti_s_d_RES
   ret void
@@ -313,7 +313,7 @@ declare <2 x i64> @llvm.mips.clti.s.d(<2 x i64>, i32) nounwind
 
 define void @llvm_mips_clti_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_clti_u_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_clti_u_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.clti.u.b(<16 x i8> %0, i32 14)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_clti_u_b_RES
   ret void
@@ -332,7 +332,7 @@ declare <16 x i8> @llvm.mips.clti.u.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_clti_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_clti_u_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_clti_u_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.clti.u.h(<8 x i16> %0, i32 14)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_clti_u_h_RES
   ret void
@@ -351,7 +351,7 @@ declare <8 x i16> @llvm.mips.clti.u.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_clti_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_clti_u_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_clti_u_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.clti.u.w(<4 x i32> %0, i32 14)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_clti_u_w_RES
   ret void
@@ -370,7 +370,7 @@ declare <4 x i32> @llvm.mips.clti.u.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_clti_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_clti_u_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_clti_u_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.clti.u.d(<2 x i64> %0, i32 14)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_clti_u_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/i5-m.ll b/test/CodeGen/Mips/msa/i5-m.ll
index 27663494324d..ba6e9d2384a7 100644
--- a/test/CodeGen/Mips/msa/i5-m.ll
+++ b/test/CodeGen/Mips/msa/i5-m.ll
@@ -9,7 +9,7 @@
 
 define void @llvm_mips_maxi_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_maxi_s_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_maxi_s_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.maxi.s.b(<16 x i8> %0, i32 14)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_maxi_s_b_RES
   ret void
@@ -28,7 +28,7 @@ declare <16 x i8> @llvm.mips.maxi.s.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_maxi_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_maxi_s_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_maxi_s_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.maxi.s.h(<8 x i16> %0, i32 14)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_maxi_s_h_RES
   ret void
@@ -47,7 +47,7 @@ declare <8 x i16> @llvm.mips.maxi.s.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_maxi_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_maxi_s_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_maxi_s_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.maxi.s.w(<4 x i32> %0, i32 14)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_maxi_s_w_RES
   ret void
@@ -66,7 +66,7 @@ declare <4 x i32> @llvm.mips.maxi.s.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_maxi_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_maxi_s_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_maxi_s_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.maxi.s.d(<2 x i64> %0, i32 14)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_maxi_s_d_RES
   ret void
@@ -85,7 +85,7 @@ declare <2 x i64> @llvm.mips.maxi.s.d(<2 x i64>, i32) nounwind
 
 define void @llvm_mips_maxi_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_maxi_u_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_maxi_u_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.maxi.u.b(<16 x i8> %0, i32 14)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_maxi_u_b_RES
   ret void
@@ -104,7 +104,7 @@ declare <16 x i8> @llvm.mips.maxi.u.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_maxi_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_maxi_u_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_maxi_u_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.maxi.u.h(<8 x i16> %0, i32 14)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_maxi_u_h_RES
   ret void
@@ -123,7 +123,7 @@ declare <8 x i16> @llvm.mips.maxi.u.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_maxi_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_maxi_u_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_maxi_u_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.maxi.u.w(<4 x i32> %0, i32 14)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_maxi_u_w_RES
   ret void
@@ -142,7 +142,7 @@ declare <4 x i32> @llvm.mips.maxi.u.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_maxi_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_maxi_u_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_maxi_u_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.maxi.u.d(<2 x i64> %0, i32 14)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_maxi_u_d_RES
   ret void
@@ -161,7 +161,7 @@ declare <2 x i64> @llvm.mips.maxi.u.d(<2 x i64>, i32) nounwind
 
 define void @llvm_mips_mini_s_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_mini_s_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_mini_s_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.mini.s.b(<16 x i8> %0, i32 14)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_mini_s_b_RES
   ret void
@@ -180,7 +180,7 @@ declare <16 x i8> @llvm.mips.mini.s.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_mini_s_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_mini_s_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_mini_s_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.mini.s.h(<8 x i16> %0, i32 14)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_mini_s_h_RES
   ret void
@@ -199,7 +199,7 @@ declare <8 x i16> @llvm.mips.mini.s.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_mini_s_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_mini_s_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_mini_s_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.mini.s.w(<4 x i32> %0, i32 14)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_mini_s_w_RES
   ret void
@@ -218,7 +218,7 @@ declare <4 x i32> @llvm.mips.mini.s.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_mini_s_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_mini_s_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_mini_s_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.mini.s.d(<2 x i64> %0, i32 14)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_mini_s_d_RES
   ret void
@@ -237,7 +237,7 @@ declare <2 x i64> @llvm.mips.mini.s.d(<2 x i64>, i32) nounwind
 
 define void @llvm_mips_mini_u_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_mini_u_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_mini_u_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.mini.u.b(<16 x i8> %0, i32 14)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_mini_u_b_RES
   ret void
@@ -256,7 +256,7 @@ declare <16 x i8> @llvm.mips.mini.u.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_mini_u_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_mini_u_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_mini_u_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.mini.u.h(<8 x i16> %0, i32 14)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_mini_u_h_RES
   ret void
@@ -275,7 +275,7 @@ declare <8 x i16> @llvm.mips.mini.u.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_mini_u_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_mini_u_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_mini_u_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.mini.u.w(<4 x i32> %0, i32 14)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_mini_u_w_RES
   ret void
@@ -294,7 +294,7 @@ declare <4 x i32> @llvm.mips.mini.u.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_mini_u_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_mini_u_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_mini_u_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.mini.u.d(<2 x i64> %0, i32 14)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_mini_u_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/i5-s.ll b/test/CodeGen/Mips/msa/i5-s.ll
index 184172f63b85..db331b1476ce 100644
--- a/test/CodeGen/Mips/msa/i5-s.ll
+++ b/test/CodeGen/Mips/msa/i5-s.ll
@@ -9,7 +9,7 @@
 
 define void @llvm_mips_subvi_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_subvi_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_subvi_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.subvi.b(<16 x i8> %0, i32 14)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_subvi_b_RES
   ret void
@@ -28,7 +28,7 @@ declare <16 x i8> @llvm.mips.subvi.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_subvi_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_subvi_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_subvi_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.subvi.h(<8 x i16> %0, i32 14)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_subvi_h_RES
   ret void
@@ -47,7 +47,7 @@ declare <8 x i16> @llvm.mips.subvi.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_subvi_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_subvi_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_subvi_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.subvi.w(<4 x i32> %0, i32 14)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_subvi_w_RES
   ret void
@@ -66,7 +66,7 @@ declare <4 x i32> @llvm.mips.subvi.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_subvi_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_subvi_d_ARG1
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_subvi_d_ARG1
   %1 = tail call <2 x i64> @llvm.mips.subvi.d(<2 x i64> %0, i32 14)
   store <2 x i64> %1, <2 x i64>* @llvm_mips_subvi_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/i5_ld_st.ll b/test/CodeGen/Mips/msa/i5_ld_st.ll
index 7cc55f2904be..991bb8436b33 100644
--- a/test/CodeGen/Mips/msa/i5_ld_st.ll
+++ b/test/CodeGen/Mips/msa/i5_ld_st.ll
@@ -81,7 +81,7 @@ declare <2 x i64> @llvm.mips.ld.d(i8*, i32) nounwind
 
 define void @llvm_mips_st_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_st_b_ARG
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_st_b_ARG
   %1 = bitcast <16 x i8>* @llvm_mips_st_b_RES to i8*
   tail call void @llvm.mips.st.b(<16 x i8> %0, i8* %1, i32 16)
   ret void
@@ -99,7 +99,7 @@ declare void @llvm.mips.st.b(<16 x i8>, i8*, i32) nounwind
 
 define void @llvm_mips_st_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_st_h_ARG
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_st_h_ARG
   %1 = bitcast <8 x i16>* @llvm_mips_st_h_RES to i8*
   tail call void @llvm.mips.st.h(<8 x i16> %0, i8* %1, i32 16)
   ret void
@@ -117,7 +117,7 @@ declare void @llvm.mips.st.h(<8 x i16>, i8*, i32) nounwind
 
 define void @llvm_mips_st_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_st_w_ARG
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_st_w_ARG
   %1 = bitcast <4 x i32>* @llvm_mips_st_w_RES to i8*
   tail call void @llvm.mips.st.w(<4 x i32> %0, i8* %1, i32 16)
   ret void
@@ -135,7 +135,7 @@ declare void @llvm.mips.st.w(<4 x i32>, i8*, i32) nounwind
 
 define void @llvm_mips_st_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_st_d_ARG
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_st_d_ARG
   %1 = bitcast <2 x i64>* @llvm_mips_st_d_RES to i8*
   tail call void @llvm.mips.st.d(<2 x i64> %0, i8* %1, i32 16)
   ret void
diff --git a/test/CodeGen/Mips/msa/i8.ll b/test/CodeGen/Mips/msa/i8.ll
index d2931a72feaa..4af9c588fdef 100644
--- a/test/CodeGen/Mips/msa/i8.ll
+++ b/test/CodeGen/Mips/msa/i8.ll
@@ -8,7 +8,7 @@
 
 define void @llvm_mips_andi_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_andi_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_andi_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.andi.b(<16 x i8> %0, i32 25)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_andi_b_RES
   ret void
@@ -28,8 +28,8 @@ declare <16 x i8> @llvm.mips.andi.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_bmnzi_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_bmnzi_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_bmnzi_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bmnzi_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bmnzi_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.bmnzi.b(<16 x i8> %0, <16 x i8> %1, i32 25)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_bmnzi_b_RES
   ret void
@@ -52,8 +52,8 @@ declare <16 x i8> @llvm.mips.bmnzi.b(<16 x i8>, <16 x i8>, i32) nounwind
 
 define void @llvm_mips_bmzi_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_bmzi_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_bmzi_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bmzi_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bmzi_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.bmzi.b(<16 x i8> %0, <16 x i8> %1, i32 25)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_bmzi_b_RES
   ret void
@@ -77,8 +77,8 @@ declare <16 x i8> @llvm.mips.bmzi.b(<16 x i8>, <16 x i8>, i32) nounwind
 
 define void @llvm_mips_bseli_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_bseli_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_bseli_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bseli_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bseli_b_ARG2
   %2 = tail call <16 x i8> @llvm.mips.bseli.b(<16 x i8> %0, <16 x i8> %1, i32 25)
   store <16 x i8> %2, <16 x i8>* @llvm_mips_bseli_b_RES
   ret void
@@ -100,7 +100,7 @@ declare <16 x i8> @llvm.mips.bseli.b(<16 x i8>, <16 x i8>, i32) nounwind
 
 define void @llvm_mips_nori_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_nori_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_nori_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.nori.b(<16 x i8> %0, i32 25)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_nori_b_RES
   ret void
@@ -119,7 +119,7 @@ declare <16 x i8> @llvm.mips.nori.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_ori_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_ori_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_ori_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.ori.b(<16 x i8> %0, i32 25)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_ori_b_RES
   ret void
@@ -138,7 +138,7 @@ declare <16 x i8> @llvm.mips.ori.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_shf_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_shf_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_shf_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.shf.b(<16 x i8> %0, i32 25)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_shf_b_RES
   ret void
@@ -157,7 +157,7 @@ declare <16 x i8> @llvm.mips.shf.b(<16 x i8>, i32) nounwind
 
 define void @llvm_mips_shf_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_shf_h_ARG1
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_shf_h_ARG1
   %1 = tail call <8 x i16> @llvm.mips.shf.h(<8 x i16> %0, i32 25)
   store <8 x i16> %1, <8 x i16>* @llvm_mips_shf_h_RES
   ret void
@@ -176,7 +176,7 @@ declare <8 x i16> @llvm.mips.shf.h(<8 x i16>, i32) nounwind
 
 define void @llvm_mips_shf_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_shf_w_ARG1
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_shf_w_ARG1
   %1 = tail call <4 x i32> @llvm.mips.shf.w(<4 x i32> %0, i32 25)
   store <4 x i32> %1, <4 x i32>* @llvm_mips_shf_w_RES
   ret void
@@ -195,7 +195,7 @@ declare <4 x i32> @llvm.mips.shf.w(<4 x i32>, i32) nounwind
 
 define void @llvm_mips_xori_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_xori_b_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_xori_b_ARG1
   %1 = tail call <16 x i8> @llvm.mips.xori.b(<16 x i8> %0, i32 25)
   store <16 x i8> %1, <16 x i8>* @llvm_mips_xori_b_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/inline-asm.ll b/test/CodeGen/Mips/msa/inline-asm.ll
index 4a34273f3c00..85da87b6f8a7 100644
--- a/test/CodeGen/Mips/msa/inline-asm.ll
+++ b/test/CodeGen/Mips/msa/inline-asm.ll
@@ -16,7 +16,7 @@ entry:
 define void @test2() nounwind {
 entry:
   ; CHECK-LABEL: test2:
-  %0 = load <4 x i32>* @v4i32_r
+  %0 = load <4 x i32>, <4 x i32>* @v4i32_r
   %1 = call <4 x i32> asm "addvi.w ${0:w}, ${1:w}, 1", "=f,f"(<4 x i32> %0)
   ; CHECK: addvi.w $w{{[1-3]?[0-9]}}, $w{{[1-3]?[0-9]}}, 1
   store <4 x i32> %1, <4 x i32>* @v4i32_r
@@ -26,7 +26,7 @@ entry:
 define void @test3() nounwind {
 entry:
   ; CHECK-LABEL: test3:
-  %0 = load <4 x i32>* @v4i32_r
+  %0 = load <4 x i32>, <4 x i32>* @v4i32_r
   %1 = call <4 x i32> asm sideeffect "addvi.w ${0:w}, ${1:w}, 1", "=f,f,~{$w0}"(<4 x i32> %0)
   ; CHECK: addvi.w $w{{([1-9]|[1-3][0-9])}}, $w{{([1-9]|[1-3][0-9])}}, 1
   store <4 x i32> %1, <4 x i32>* @v4i32_r
diff --git a/test/CodeGen/Mips/msa/llvm-stress-s1704963983.ll b/test/CodeGen/Mips/msa/llvm-stress-s1704963983.ll
index 4beaaa9c1841..beb361bc9f3d 100644
--- a/test/CodeGen/Mips/msa/llvm-stress-s1704963983.ll
+++ b/test/CodeGen/Mips/msa/llvm-stress-s1704963983.ll
@@ -14,7 +14,7 @@ BB:
   %A2 = alloca <1 x double>
   %A1 = alloca double
   %A = alloca i32
-  %L = load i8* %0
+  %L = load i8, i8* %0
   store i8 77, i8* %0
   %E = extractelement <8 x i64> zeroinitializer, i32 2
   %Shuff = shufflevector <8 x i64> zeroinitializer, <8 x i64> zeroinitializer, <8 x i32> <i32 5, i32 7, i32 undef, i32 undef, i32 13, i32 15, i32 1, i32 3>
@@ -24,7 +24,7 @@ BB:
   br label %CF
 
 CF:                                               ; preds = %CF, %CF78, %BB
-  %L5 = load i8* %Sl
+  %L5 = load i8, i8* %Sl
   store i8 %L, i8* %Sl
   %E6 = extractelement <8 x i32> zeroinitializer, i32 2
   %Shuff7 = shufflevector <8 x i64> zeroinitializer, <8 x i64> %Shuff, <8 x i32> <i32 13, i32 15, i32 1, i32 3, i32 5, i32 7, i32 9, i32 undef>
@@ -33,7 +33,7 @@ CF:                                               ; preds = %CF, %CF78, %BB
   %FC = sitofp <8 x i64> zeroinitializer to <8 x float>
   %Sl9 = select i1 %Cmp, i8 77, i8 77
   %Cmp10 = icmp uge <8 x i64> %Shuff, zeroinitializer
-  %L11 = load i8* %0
+  %L11 = load i8, i8* %0
   store i8 %Sl9, i8* %0
   %E12 = extractelement <1 x i16> zeroinitializer, i32 0
   %Shuff13 = shufflevector <8 x i64> zeroinitializer, <8 x i64> %Shuff, <8 x i32> <i32 9, i32 11, i32 13, i32 15, i32 undef, i32 3, i32 5, i32 7>
@@ -42,7 +42,7 @@ CF:                                               ; preds = %CF, %CF78, %BB
   %Tr = trunc <8 x i64> %Shuff to <8 x i32>
   %Sl16 = select i1 %Cmp, i8 77, i8 %5
   %Cmp17 = icmp ult <8 x i1> %Cmp10, %Cmp10
-  %L18 = load i8* %Sl
+  %L18 = load i8, i8* %Sl
   store i8 -1, i8* %Sl
   %E19 = extractelement <8 x i32> zeroinitializer, i32 3
   %Shuff20 = shufflevector <8 x float> %FC, <8 x float> %FC, <8 x i32> <i32 6, i32 8, i32 undef, i32 12, i32 14, i32 0, i32 2, i32 undef>
@@ -54,7 +54,7 @@ CF:                                               ; preds = %CF, %CF78, %BB
   br i1 %Cmp25, label %CF, label %CF78
 
 CF78:                                             ; preds = %CF
-  %L26 = load i8* %Sl
+  %L26 = load i8, i8* %Sl
   store i32 50347, i32* %A
   %E27 = extractelement <8 x i1> %Cmp10, i32 2
   br i1 %E27, label %CF, label %CF77
@@ -65,7 +65,7 @@ CF77:                                             ; preds = %CF77, %CF81, %CF78
   %B30 = urem <8 x i32> %Tr, zeroinitializer
   %Tr31 = trunc i32 0 to i16
   %Sl32 = select i1 %Cmp, <2 x i1> zeroinitializer, <2 x i1> zeroinitializer
-  %L33 = load i8* %Sl
+  %L33 = load i8, i8* %Sl
   store i8 %L26, i8* %Sl
   %E34 = extractelement <4 x i32> zeroinitializer, i32 0
   %Shuff35 = shufflevector <1 x i16> zeroinitializer, <1 x i16> %B, <1 x i32> undef
@@ -73,7 +73,7 @@ CF77:                                             ; preds = %CF77, %CF81, %CF78
   %B37 = srem <1 x i16> %I29, zeroinitializer
   %FC38 = sitofp <8 x i32> %B30 to <8 x double>
   %Sl39 = select i1 %Cmp, double 0.000000e+00, double %Sl24
-  %L40 = load i8* %Sl
+  %L40 = load i8, i8* %Sl
   store i8 %Sl16, i8* %Sl
   %E41 = extractelement <1 x i16> zeroinitializer, i32 0
   %Shuff42 = shufflevector <8 x i1> %Cmp17, <8 x i1> %Cmp10, <8 x i32> <i32 14, i32 undef, i32 2, i32 4, i32 undef, i32 8, i32 10, i32 12>
@@ -85,7 +85,7 @@ CF77:                                             ; preds = %CF77, %CF81, %CF78
   br i1 %Cmp46, label %CF77, label %CF80
 
 CF80:                                             ; preds = %CF80, %CF77
-  %L47 = load i64* %PC
+  %L47 = load i64, i64* %PC
   store i8 77, i8* %Sl
   %E48 = extractelement <8 x i64> zeroinitializer, i32 2
   %Shuff49 = shufflevector <8 x i64> zeroinitializer, <8 x i64> %Shuff7, <8 x i32> <i32 5, i32 7, i32 9, i32 undef, i32 undef, i32 undef, i32 undef, i32 3>
@@ -97,7 +97,7 @@ CF80:                                             ; preds = %CF80, %CF77
   br i1 %Cmp54, label %CF80, label %CF81
 
 CF81:                                             ; preds = %CF80
-  %L55 = load i8* %Sl
+  %L55 = load i8, i8* %Sl
   store i8 %Sl16, i8* %Sl
   %E56 = extractelement <1 x i16> %B, i32 0
   %Shuff57 = shufflevector <1 x i16> zeroinitializer, <1 x i16> zeroinitializer, <1 x i32> <i32 1>
@@ -105,7 +105,7 @@ CF81:                                             ; preds = %CF80
   %B59 = srem i32 %E19, %E19
   %Sl60 = select i1 %Cmp, i8 77, i8 77
   %Cmp61 = icmp ult <1 x i16> zeroinitializer, %B
-  %L62 = load i8* %Sl
+  %L62 = load i8, i8* %Sl
   store i64 %L47, i64* %PC52
   %E63 = extractelement <4 x i32> %I43, i32 2
   %Shuff64 = shufflevector <4 x i1> zeroinitializer, <4 x i1> zeroinitializer, <4 x i32> <i32 undef, i32 undef, i32 1, i32 3>
@@ -117,7 +117,7 @@ CF81:                                             ; preds = %CF80
   br i1 %Cmp69, label %CF77, label %CF79
 
 CF79:                                             ; preds = %CF81
-  %L70 = load i32* %A
+  %L70 = load i32, i32* %A
   store i64 %4, i64* %PC
   %E71 = extractelement <4 x i32> zeroinitializer, i32 0
   %Shuff72 = shufflevector <8 x i32> zeroinitializer, <8 x i32> %B44, <8 x i32> <i32 11, i32 undef, i32 15, i32 1, i32 3, i32 undef, i32 7, i32 9>
diff --git a/test/CodeGen/Mips/msa/llvm-stress-s1935737938.ll b/test/CodeGen/Mips/msa/llvm-stress-s1935737938.ll
index f9cab037e7cc..bdf6eafdf4ef 100644
--- a/test/CodeGen/Mips/msa/llvm-stress-s1935737938.ll
+++ b/test/CodeGen/Mips/msa/llvm-stress-s1935737938.ll
@@ -14,7 +14,7 @@ BB:
   %A2 = alloca i64
   %A1 = alloca i32
   %A = alloca <2 x i64>
-  %L = load i8* %0
+  %L = load i8, i8* %0
   store i8 -1, i8* %0
   %E = extractelement <2 x i32> zeroinitializer, i32 0
   %Shuff = shufflevector <2 x i32> zeroinitializer, <2 x i32> zeroinitializer, <2 x i32> <i32 1, i32 3>
@@ -22,7 +22,7 @@ BB:
   %B = lshr i8 %L, -69
   %ZE = fpext float 0xBF2AA5FE80000000 to double
   %Sl = select i1 true, <1 x i64> <i64 -1>, <1 x i64> <i64 -1>
-  %L5 = load i8* %0
+  %L5 = load i8, i8* %0
   store i8 -69, i8* %0
   %E6 = extractelement <16 x i64> <i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1>, i32 14
   %Shuff7 = shufflevector <2 x i32> zeroinitializer, <2 x i32> zeroinitializer, <2 x i32> <i32 1, i32 3>
@@ -31,7 +31,7 @@ BB:
   %FC = uitofp i32 %3 to double
   %Sl10 = select i1 true, <1 x i1> zeroinitializer, <1 x i1> zeroinitializer
   %Cmp = icmp ne <1 x i64> %I, <i64 -1>
-  %L11 = load i8* %0
+  %L11 = load i8, i8* %0
   store i8 %L11, i8* %0
   %E12 = extractelement <1 x i64> <i64 -1>, i32 0
   %Shuff13 = shufflevector <1 x i64> %Sl, <1 x i64> <i64 -1>, <1 x i32> <i32 1>
@@ -42,7 +42,7 @@ BB:
   br label %CF74
 
 CF74:                                             ; preds = %CF74, %CF80, %CF76, %BB
-  %L18 = load i8* %0
+  %L18 = load i8, i8* %0
   store i8 -69, i8* %0
   %E19 = extractelement <1 x i64> %Sl, i32 0
   %Shuff20 = shufflevector <8 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>, <8 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>, <8 x i32> <i32 12, i32 14, i32 0, i32 2, i32 4, i32 6, i32 8, i32 10>
@@ -50,7 +50,7 @@ CF74:                                             ; preds = %CF74, %CF80, %CF76,
   %B22 = urem i32 135673, %3
   %FC23 = sitofp i8 %L to float
   %Sl24 = select i1 true, i8 %B, i8 %L18
-  %L25 = load i8* %0
+  %L25 = load i8, i8* %0
   store i8 %L, i8* %0
   %E26 = extractelement <2 x i32> %Shuff, i32 1
   %Shuff27 = shufflevector <2 x i32> zeroinitializer, <2 x i32> zeroinitializer, <2 x i32> <i32 2, i32 0>
@@ -62,7 +62,7 @@ CF74:                                             ; preds = %CF74, %CF80, %CF76,
   br i1 %Cmp31, label %CF74, label %CF80
 
 CF80:                                             ; preds = %CF74
-  %L32 = load i8* %0
+  %L32 = load i8, i8* %0
   store i8 -1, i8* %0
   %E33 = extractelement <2 x i32> zeroinitializer, i32 1
   %Shuff34 = shufflevector <1 x i64> %Shuff13, <1 x i64> <i64 -1>, <1 x i32> zeroinitializer
@@ -70,7 +70,7 @@ CF80:                                             ; preds = %CF74
   %FC36 = sitofp <1 x i1> %Cmp to <1 x float>
   %Sl37 = select i1 true, <8 x i8> %Shuff20, <8 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
   %Cmp38 = icmp sgt <2 x i32> %I21, %Shuff27
-  %L39 = load i8* %0
+  %L39 = load i8, i8* %0
   store i8 %Sl24, i8* %0
   %E40 = extractelement <8 x i64> zeroinitializer, i32 1
   %Shuff41 = shufflevector <2 x i1> zeroinitializer, <2 x i1> %Cmp38, <2 x i32> <i32 0, i32 2>
@@ -81,7 +81,7 @@ CF80:                                             ; preds = %CF74
   br i1 %Cmp45, label %CF74, label %CF76
 
 CF76:                                             ; preds = %CF80
-  %L46 = load i8* %0
+  %L46 = load i8, i8* %0
   store i8 %L39, i8* %0
   %E47 = extractelement <2 x i32> %Shuff27, i32 0
   %Shuff48 = shufflevector <1 x i1> %Sl10, <1 x i1> %Sl10, <1 x i32> <i32 1>
@@ -92,7 +92,7 @@ CF76:                                             ; preds = %CF80
   br i1 %Cmp52, label %CF74, label %CF75
 
 CF75:                                             ; preds = %CF75, %CF76
-  %L53 = load i8* %0
+  %L53 = load i8, i8* %0
   store i8 %L18, i8* %0
   %E54 = extractelement <8 x i8> %Shuff20, i32 5
   %Shuff55 = shufflevector <2 x i32> %Shuff, <2 x i32> zeroinitializer, <2 x i32> <i32 0, i32 2>
@@ -103,7 +103,7 @@ CF75:                                             ; preds = %CF75, %CF76
   br i1 %Cmp59, label %CF75, label %CF78
 
 CF78:                                             ; preds = %CF75
-  %L60 = load i8* %0
+  %L60 = load i8, i8* %0
   store i8 -69, i8* %0
   %E61 = extractelement <2 x i32> zeroinitializer, i32 0
   %Shuff62 = shufflevector <2 x i32> %Shuff7, <2 x i32> %I21, <2 x i32> <i32 1, i32 3>
@@ -115,7 +115,7 @@ CF78:                                             ; preds = %CF75
   br label %CF
 
 CF:                                               ; preds = %CF, %CF78
-  %L68 = load i8* %0
+  %L68 = load i8, i8* %0
   store i64 %B57, i64* %2
   %E69 = extractelement <2 x i1> %Shuff41, i32 1
   br i1 %E69, label %CF, label %CF77
diff --git a/test/CodeGen/Mips/msa/llvm-stress-s2704903805.ll b/test/CodeGen/Mips/msa/llvm-stress-s2704903805.ll
index e14f405320cb..8f23a8ca5177 100644
--- a/test/CodeGen/Mips/msa/llvm-stress-s2704903805.ll
+++ b/test/CodeGen/Mips/msa/llvm-stress-s2704903805.ll
@@ -13,7 +13,7 @@ BB:
   %A2 = alloca i8
   %A1 = alloca i32
   %A = alloca i8
-  %L = load i8* %0
+  %L = load i8, i8* %0
   store i8 %5, i8* %0
   %E = extractelement <2 x i16> zeroinitializer, i32 0
   %Shuff = shufflevector <1 x i8> <i8 -1>, <1 x i8> <i8 -1>, <1 x i32> undef
@@ -25,7 +25,7 @@ BB:
   br label %CF83
 
 CF83:                                             ; preds = %BB
-  %L5 = load i8* %0
+  %L5 = load i8, i8* %0
   store i8 85, i8* %0
   %E6 = extractelement <1 x i8> <i8 -1>, i32 0
   %Shuff7 = shufflevector <2 x i16> zeroinitializer, <2 x i16> zeroinitializer, <2 x i32> <i32 1, i32 3>
@@ -37,7 +37,7 @@ CF83:                                             ; preds = %BB
   br label %CF
 
 CF:                                               ; preds = %CF, %CF81, %CF83
-  %L13 = load i8* %0
+  %L13 = load i8, i8* %0
   store i8 0, i8* %0
   %E14 = extractelement <2 x i64> zeroinitializer, i32 0
   %Shuff15 = shufflevector <4 x i64> <i64 -1, i64 -1, i64 -1, i64 -1>, <4 x i64> <i64 -1, i64 -1, i64 -1, i64 -1>, <4 x i32> <i32 3, i32 5, i32 7, i32 undef>
@@ -52,7 +52,7 @@ CF80:                                             ; preds = %CF80, %CF
   br i1 %Cmp19, label %CF80, label %CF81
 
 CF81:                                             ; preds = %CF80
-  %L20 = load i8* %0
+  %L20 = load i8, i8* %0
   store i8 85, i8* %0
   %E21 = extractelement <1 x i8> <i8 -1>, i32 0
   %Shuff22 = shufflevector <1 x i8> <i8 -1>, <1 x i8> %Shuff, <1 x i32> zeroinitializer
@@ -60,7 +60,7 @@ CF81:                                             ; preds = %CF80
   %FC24 = fptoui <4 x float> %FC to <4 x i16>
   %Sl25 = select i1 %Cmp, <2 x i32> zeroinitializer, <2 x i32> <i32 -1, i32 -1>
   %Cmp26 = icmp ult <4 x i64> %I16, %Shuff15
-  %L27 = load i8* %0
+  %L27 = load i8, i8* %0
   store i8 %L, i8* %0
   %E28 = extractelement <1 x i8> <i8 -1>, i32 0
   %Shuff29 = shufflevector <8 x i16> zeroinitializer, <8 x i16> zeroinitializer, <8 x i32> <i32 11, i32 undef, i32 15, i32 1, i32 3, i32 5, i32 undef, i32 9>
@@ -68,7 +68,7 @@ CF81:                                             ; preds = %CF80
   %B31 = mul i8 %E28, 85
   %PC = bitcast i32* %A3 to i32*
   %Sl32 = select i1 %Cmp12, float %FC10, float 0x4712BFE680000000
-  %L33 = load i32* %PC
+  %L33 = load i32, i32* %PC
   store i32 %L33, i32* %PC
   %E34 = extractelement <2 x i16> zeroinitializer, i32 1
   %Shuff35 = shufflevector <1 x i8> %Shuff, <1 x i8> <i8 -1>, <1 x i32> zeroinitializer
@@ -79,7 +79,7 @@ CF81:                                             ; preds = %CF80
   br i1 %Cmp39, label %CF, label %CF77
 
 CF77:                                             ; preds = %CF77, %CF81
-  %L40 = load i32* %PC
+  %L40 = load i32, i32* %PC
   store i32 %3, i32* %PC
   %E41 = extractelement <2 x i32> zeroinitializer, i32 0
   %Shuff42 = shufflevector <2 x i32> <i32 -1, i32 -1>, <2 x i32> zeroinitializer, <2 x i32> <i32 1, i32 3>
@@ -88,7 +88,7 @@ CF77:                                             ; preds = %CF77, %CF81
   %Se = sext i32 %3 to i64
   %Sl45 = select i1 true, <1 x i8> %Shuff, <1 x i8> %I43
   %Cmp46 = icmp sge <1 x i8> %I36, %Shuff
-  %L47 = load i32* %PC
+  %L47 = load i32, i32* %PC
   store i32 %L33, i32* %PC
   %E48 = extractelement <2 x i16> zeroinitializer, i32 0
   %Shuff49 = shufflevector <1 x i8> <i8 -1>, <1 x i8> <i8 -1>, <1 x i32> <i32 1>
@@ -100,7 +100,7 @@ CF77:                                             ; preds = %CF77, %CF81
   br i1 %Cmp54, label %CF77, label %CF78
 
 CF78:                                             ; preds = %CF78, %CF77
-  %L55 = load i32* %PC
+  %L55 = load i32, i32* %PC
   store i32 %L33, i32* %PC
   %E56 = extractelement <8 x i16> %Shuff29, i32 4
   %Shuff57 = shufflevector <1 x i8> <i8 -1>, <1 x i8> <i8 -1>, <1 x i32> <i32 1>
@@ -111,7 +111,7 @@ CF78:                                             ; preds = %CF78, %CF77
   br i1 %Cmp60, label %CF78, label %CF79
 
 CF79:                                             ; preds = %CF79, %CF78
-  %L61 = load i32* %PC
+  %L61 = load i32, i32* %PC
   store i32 %L33, i32* %A3
   %E62 = extractelement <4 x i64> %Shuff15, i32 1
   %Shuff63 = shufflevector <8 x i16> %Shuff29, <8 x i16> %Shuff29, <8 x i32> <i32 undef, i32 10, i32 12, i32 undef, i32 undef, i32 undef, i32 4, i32 6>
@@ -123,7 +123,7 @@ CF79:                                             ; preds = %CF79, %CF78
   br i1 %Cmp68, label %CF79, label %CF82
 
 CF82:                                             ; preds = %CF79
-  %L69 = load i32* %PC
+  %L69 = load i32, i32* %PC
   store i32 %L33, i32* %PC
   %E70 = extractelement <8 x i16> zeroinitializer, i32 3
   %Shuff71 = shufflevector <4 x i64> %Shuff15, <4 x i64> <i64 -1, i64 -1, i64 -1, i64 -1>, <4 x i32> <i32 6, i32 undef, i32 2, i32 4>
diff --git a/test/CodeGen/Mips/msa/llvm-stress-s3861334421.ll b/test/CodeGen/Mips/msa/llvm-stress-s3861334421.ll
index 1a03e55d9d54..e3cf7964497f 100644
--- a/test/CodeGen/Mips/msa/llvm-stress-s3861334421.ll
+++ b/test/CodeGen/Mips/msa/llvm-stress-s3861334421.ll
@@ -14,7 +14,7 @@ BB:
   %A2 = alloca i64
   %A1 = alloca i64
   %A = alloca double
-  %L = load i8* %0
+  %L = load i8, i8* %0
   store i8 -101, i8* %0
   %E = extractelement <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, i32 0
   %Shuff = shufflevector <8 x i64> zeroinitializer, <8 x i64> zeroinitializer, <8 x i32> <i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 undef, i32 1>
@@ -22,7 +22,7 @@ BB:
   %B = and i64 116376, 57247
   %FC = uitofp i8 7 to double
   %Sl = select i1 false, <8 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>, <8 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
-  %L5 = load i8* %0
+  %L5 = load i8, i8* %0
   store i8 %L, i8* %0
   %E6 = extractelement <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, i32 3
   %Shuff7 = shufflevector <4 x i64> zeroinitializer, <4 x i64> zeroinitializer, <4 x i32> <i32 2, i32 4, i32 6, i32 0>
@@ -33,7 +33,7 @@ BB:
   br label %CF
 
 CF:                                               ; preds = %CF, %BB
-  %L11 = load i8* %0
+  %L11 = load i8, i8* %0
   store i8 -87, i8* %0
   %E12 = extractelement <4 x i64> zeroinitializer, i32 0
   %Shuff13 = shufflevector <8 x i64> zeroinitializer, <8 x i64> zeroinitializer, <8 x i32> <i32 7, i32 9, i32 11, i32 13, i32 undef, i32 1, i32 3, i32 5>
@@ -45,7 +45,7 @@ CF:                                               ; preds = %CF, %BB
   br i1 %Cmp18, label %CF, label %CF80
 
 CF80:                                             ; preds = %CF80, %CF88, %CF
-  %L19 = load i8* %0
+  %L19 = load i8, i8* %0
   store i8 -101, i8* %0
   %E20 = extractelement <4 x i64> zeroinitializer, i32 0
   %Shuff21 = shufflevector <4 x i64> zeroinitializer, <4 x i64> %Shuff7, <4 x i32> <i32 7, i32 1, i32 3, i32 5>
@@ -56,7 +56,7 @@ CF80:                                             ; preds = %CF80, %CF88, %CF
   br i1 %Cmp25, label %CF80, label %CF83
 
 CF83:                                             ; preds = %CF83, %CF80
-  %L26 = load i8* %0
+  %L26 = load i8, i8* %0
   store i8 -87, i8* %0
   %E27 = extractelement <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, i32 0
   %Shuff28 = shufflevector <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, <4 x i32> <i32 7, i32 1, i32 3, i32 5>
@@ -68,7 +68,7 @@ CF83:                                             ; preds = %CF83, %CF80
   br i1 %Cmp33, label %CF83, label %CF88
 
 CF88:                                             ; preds = %CF83
-  %L34 = load i8* %0
+  %L34 = load i8, i8* %0
   store i8 -87, i8* %0
   %E35 = extractelement <8 x i64> %Shuff, i32 7
   %Shuff36 = shufflevector <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, <4 x i32> %Shuff28, <4 x i32> <i32 2, i32 undef, i32 undef, i32 0>
@@ -80,7 +80,7 @@ CF88:                                             ; preds = %CF83
   br i1 %Cmp40, label %CF80, label %CF81
 
 CF81:                                             ; preds = %CF81, %CF85, %CF87, %CF88
-  %L41 = load i8* %0
+  %L41 = load i8, i8* %0
   store i8 %L34, i8* %0
   %E42 = extractelement <8 x i64> %Shuff13, i32 6
   %Shuff43 = shufflevector <4 x i64> zeroinitializer, <4 x i64> zeroinitializer, <4 x i32> <i32 undef, i32 undef, i32 undef, i32 7>
@@ -92,7 +92,7 @@ CF81:                                             ; preds = %CF81, %CF85, %CF87,
   br i1 %Cmp47, label %CF81, label %CF85
 
 CF85:                                             ; preds = %CF81
-  %L48 = load i8* %0
+  %L48 = load i8, i8* %0
   store i8 -101, i8* %0
   %E49 = extractelement <8 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>, i32 2
   %Shuff50 = shufflevector <4 x i64> zeroinitializer, <4 x i64> zeroinitializer, <4 x i32> <i32 5, i32 7, i32 1, i32 3>
@@ -101,7 +101,7 @@ CF85:                                             ; preds = %CF81
   %FC53 = uitofp i8 %L48 to double
   %Sl54 = select i1 %Cmp47, i32 %3, i32 %Sl24
   %Cmp55 = icmp ne <8 x i64> %Shuff13, zeroinitializer
-  %L56 = load i8* %0
+  %L56 = load i8, i8* %0
   store i8 %L11, i8* %0
   %E57 = extractelement <4 x i64> %Shuff21, i32 1
   %Shuff58 = shufflevector <8 x i64> %Shuff, <8 x i64> zeroinitializer, <8 x i32> <i32 4, i32 6, i32 undef, i32 10, i32 12, i32 undef, i32 0, i32 2>
@@ -113,7 +113,7 @@ CF85:                                             ; preds = %CF81
 CF84:                                             ; preds = %CF84, %CF85
   %Sl62 = select i1 false, i8 %L, i8 %L48
   %Cmp63 = icmp ne <8 x i64> %I, zeroinitializer
-  %L64 = load i8* %0
+  %L64 = load i8, i8* %0
   store i8 %5, i8* %0
   %E65 = extractelement <8 x i1> %Cmp55, i32 0
   br i1 %E65, label %CF84, label %CF87
@@ -125,7 +125,7 @@ CF87:                                             ; preds = %CF84
   %ZE69 = zext <8 x i8> %Sl32 to <8 x i64>
   %Sl70 = select i1 %Tr61, i64 %E20, i64 %E12
   %Cmp71 = icmp slt <8 x i64> %I, %Shuff
-  %L72 = load i8* %0
+  %L72 = load i8, i8* %0
   store i8 %L72, i8* %0
   %E73 = extractelement <8 x i1> %Cmp55, i32 6
   br i1 %E73, label %CF81, label %CF82
diff --git a/test/CodeGen/Mips/msa/llvm-stress-s3926023935.ll b/test/CodeGen/Mips/msa/llvm-stress-s3926023935.ll
index 96547d90cb40..6f338107825e 100644
--- a/test/CodeGen/Mips/msa/llvm-stress-s3926023935.ll
+++ b/test/CodeGen/Mips/msa/llvm-stress-s3926023935.ll
@@ -14,7 +14,7 @@ BB:
   %A2 = alloca double
   %A1 = alloca float
   %A = alloca double
-  %L = load i8* %0
+  %L = load i8, i8* %0
   store i8 -123, i8* %0
   %E = extractelement <4 x i64> zeroinitializer, i32 1
   %Shuff = shufflevector <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, <4 x i32> zeroinitializer, <4 x i32> <i32 0, i32 2, i32 4, i32 6>
@@ -22,7 +22,7 @@ BB:
   %BC = bitcast i64 181325 to double
   %Sl = select i1 false, <2 x i32> zeroinitializer, <2 x i32> zeroinitializer
   %Cmp = icmp ne <4 x i64> zeroinitializer, zeroinitializer
-  %L5 = load i8* %0
+  %L5 = load i8, i8* %0
   store i8 %L, i8* %0
   %E6 = extractelement <4 x i64> zeroinitializer, i32 3
   %Shuff7 = shufflevector <2 x i16> zeroinitializer, <2 x i16> zeroinitializer, <2 x i32> <i32 2, i32 0>
@@ -33,7 +33,7 @@ BB:
   br label %CF80
 
 CF80:                                             ; preds = %BB
-  %L11 = load i8* %0
+  %L11 = load i8, i8* %0
   store i8 -123, i8* %0
   %E12 = extractelement <2 x i16> zeroinitializer, i32 1
   %Shuff13 = shufflevector <4 x i64> zeroinitializer, <4 x i64> zeroinitializer, <4 x i32> <i32 0, i32 2, i32 4, i32 6>
@@ -42,7 +42,7 @@ CF80:                                             ; preds = %BB
   %PC = bitcast i1* %A4 to i64*
   %Sl16 = select i1 %Cmp10, <4 x i32> zeroinitializer, <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>
   %Cmp17 = icmp ule <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, %Sl16
-  %L18 = load double* %A2
+  %L18 = load double, double* %A2
   store i64 498254, i64* %PC
   %E19 = extractelement <4 x i64> zeroinitializer, i32 0
   %Shuff20 = shufflevector <2 x i1> zeroinitializer, <2 x i1> %I, <2 x i32> <i32 3, i32 1>
@@ -51,7 +51,7 @@ CF80:                                             ; preds = %BB
   %ZE = zext <2 x i1> %Shuff20 to <2 x i32>
   %Sl23 = select i1 %Cmp10, <2 x i1> %Shuff20, <2 x i1> zeroinitializer
   %Cmp24 = icmp ult <2 x i32> zeroinitializer, zeroinitializer
-  %L25 = load i8* %0
+  %L25 = load i8, i8* %0
   store i8 %L25, i8* %0
   %E26 = extractelement <4 x i8> <i8 -1, i8 -1, i8 -1, i8 -1>, i32 3
   %Shuff27 = shufflevector <4 x i32> %Shuff, <4 x i32> %I14, <4 x i32> <i32 6, i32 0, i32 undef, i32 4>
@@ -63,7 +63,7 @@ CF80:                                             ; preds = %BB
 CF79:                                             ; preds = %CF80
   %Sl30 = select i1 false, i8 %B29, i8 -123
   %Cmp31 = icmp sge <2 x i1> %I, %I
-  %L32 = load i64* %PC
+  %L32 = load i64, i64* %PC
   store i8 -123, i8* %0
   %E33 = extractelement <8 x i64> <i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1>, i32 2
   %Shuff34 = shufflevector <4 x i64> zeroinitializer, <4 x i64> %Shuff13, <4 x i32> <i32 5, i32 7, i32 1, i32 3>
@@ -75,7 +75,7 @@ CF79:                                             ; preds = %CF80
   br label %CF
 
 CF:                                               ; preds = %CF, %CF79
-  %L40 = load double* %A
+  %L40 = load double, double* %A
   store i1 %Cmp39, i1* %PC37
   %E41 = extractelement <4 x i64> zeroinitializer, i32 3
   %Shuff42 = shufflevector <2 x i32> zeroinitializer, <2 x i32> %ZE, <2 x i32> <i32 2, i32 undef>
@@ -90,7 +90,7 @@ CF77:                                             ; preds = %CF77, %CF
   br i1 %Cmp46, label %CF77, label %CF78
 
 CF78:                                             ; preds = %CF78, %CF83, %CF82, %CF77
-  %L47 = load i64* %PC
+  %L47 = load i64, i64* %PC
   store i8 -123, i8* %0
   %E48 = extractelement <4 x i64> zeroinitializer, i32 3
   %Shuff49 = shufflevector <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, <4 x i32> zeroinitializer, <4 x i32> <i32 2, i32 4, i32 6, i32 undef>
@@ -105,7 +105,7 @@ CF83:                                             ; preds = %CF78
   br i1 %Cmp54, label %CF78, label %CF82
 
 CF82:                                             ; preds = %CF83
-  %L55 = load i64* %PC
+  %L55 = load i64, i64* %PC
   store i64 %L32, i64* %PC
   %E56 = extractelement <2 x i16> %Shuff7, i32 1
   %Shuff57 = shufflevector <4 x i64> zeroinitializer, <4 x i64> zeroinitializer, <4 x i32> <i32 2, i32 4, i32 6, i32 0>
@@ -114,7 +114,7 @@ CF82:                                             ; preds = %CF83
   %FC = sitofp i64 498254 to double
   %Sl60 = select i1 false, i64 %E6, i64 -1
   %Cmp61 = icmp sgt <4 x i32> %Shuff27, %I43
-  %L62 = load i64* %PC
+  %L62 = load i64, i64* %PC
   store i64 %Sl9, i64* %PC
   %E63 = extractelement <2 x i32> %ZE, i32 0
   %Shuff64 = shufflevector <4 x i64> zeroinitializer, <4 x i64> %Shuff13, <4 x i32> <i32 1, i32 3, i32 undef, i32 7>
@@ -126,7 +126,7 @@ CF82:                                             ; preds = %CF83
 
 CF81:                                             ; preds = %CF82
   %Cmp69 = icmp ne <8 x i64> <i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1>, %B36
-  %L70 = load i8* %0
+  %L70 = load i8, i8* %0
   store i64 %L55, i64* %PC
   %E71 = extractelement <4 x i32> %Shuff49, i32 1
   %Shuff72 = shufflevector <4 x i64> zeroinitializer, <4 x i64> %Shuff34, <4 x i32> <i32 0, i32 2, i32 4, i32 6>
diff --git a/test/CodeGen/Mips/msa/llvm-stress-s3997499501.ll b/test/CodeGen/Mips/msa/llvm-stress-s3997499501.ll
index bef75f3645c8..181f72abd378 100644
--- a/test/CodeGen/Mips/msa/llvm-stress-s3997499501.ll
+++ b/test/CodeGen/Mips/msa/llvm-stress-s3997499501.ll
@@ -14,7 +14,7 @@ BB:
   %A2 = alloca float
   %A1 = alloca double
   %A = alloca double
-  %L = load i8* %0
+  %L = load i8, i8* %0
   store i8 97, i8* %0
   %E = extractelement <16 x i64> <i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1, i64 -1>, i32 14
   %Shuff = shufflevector <2 x i1> zeroinitializer, <2 x i1> zeroinitializer, <2 x i32> <i32 1, i32 3>
@@ -22,7 +22,7 @@ BB:
   %Tr = trunc <1 x i64> zeroinitializer to <1 x i8>
   %Sl = select i1 false, double* %A1, double* %A
   %Cmp = icmp ne <2 x i64> zeroinitializer, zeroinitializer
-  %L5 = load double* %Sl
+  %L5 = load double, double* %Sl
   store float -4.374162e+06, float* %A2
   %E6 = extractelement <4 x i64> zeroinitializer, i32 3
   %Shuff7 = shufflevector <4 x i64> zeroinitializer, <4 x i64> %I, <4 x i32> <i32 2, i32 4, i32 6, i32 undef>
@@ -34,7 +34,7 @@ BB:
   br label %CF72
 
 CF72:                                             ; preds = %CF72, %CF80, %CF78, %BB
-  %L11 = load double* %Sl
+  %L11 = load double, double* %Sl
   store double 0.000000e+00, double* %Sl
   %E12 = extractelement <2 x i1> zeroinitializer, i32 0
   br i1 %E12, label %CF72, label %CF80
@@ -49,7 +49,7 @@ CF80:                                             ; preds = %CF72
   br i1 %Cmp17, label %CF72, label %CF77
 
 CF77:                                             ; preds = %CF77, %CF80
-  %L18 = load double* %Sl
+  %L18 = load double, double* %Sl
   store double 0.000000e+00, double* %Sl
   %E19 = extractelement <2 x i1> zeroinitializer, i32 0
   br i1 %E19, label %CF77, label %CF78
@@ -60,7 +60,7 @@ CF78:                                             ; preds = %CF77
   %B22 = sdiv <4 x i64> %Shuff7, zeroinitializer
   %FC = uitofp i8 97 to double
   %Sl23 = select i1 %Cmp10, <2 x i1> zeroinitializer, <2 x i1> zeroinitializer
-  %L24 = load double* %Sl
+  %L24 = load double, double* %Sl
   store float %Sl16, float* %PC
   %E25 = extractelement <2 x i1> %Shuff, i32 1
   br i1 %E25, label %CF72, label %CF76
@@ -71,7 +71,7 @@ CF76:                                             ; preds = %CF78
   %B28 = mul <4 x i64> %I27, zeroinitializer
   %ZE = zext <8 x i1> zeroinitializer to <8 x i64>
   %Sl29 = select i1 %Cmp17, float -4.374162e+06, float -4.374162e+06
-  %L30 = load i8* %0
+  %L30 = load i8, i8* %0
   store double %L5, double* %Sl
   %E31 = extractelement <8 x i1> zeroinitializer, i32 5
   br label %CF
@@ -85,7 +85,7 @@ CF:                                               ; preds = %CF, %CF81, %CF76
   br i1 %Cmp36, label %CF, label %CF74
 
 CF74:                                             ; preds = %CF74, %CF
-  %L37 = load float* %PC
+  %L37 = load float, float* %PC
   store double 0.000000e+00, double* %Sl
   %E38 = extractelement <2 x i1> %Sl23, i32 1
   br i1 %E38, label %CF74, label %CF75
@@ -95,7 +95,7 @@ CF75:                                             ; preds = %CF75, %CF82, %CF74
   %I40 = insertelement <4 x i64> zeroinitializer, i64 %4, i32 2
   %Sl41 = select i1 %Cmp10, i32 0, i32 %3
   %Cmp42 = icmp ne <1 x i64> zeroinitializer, zeroinitializer
-  %L43 = load double* %Sl
+  %L43 = load double, double* %Sl
   store i64 %4, i64* %2
   %E44 = extractelement <2 x i1> %Shuff20, i32 1
   br i1 %E44, label %CF75, label %CF82
@@ -109,7 +109,7 @@ CF82:                                             ; preds = %CF75
   br i1 %Cmp49, label %CF75, label %CF81
 
 CF81:                                             ; preds = %CF82
-  %L50 = load i8* %0
+  %L50 = load i8, i8* %0
   store double %L43, double* %Sl
   %E51 = extractelement <4 x i64> %Shuff7, i32 3
   %Shuff52 = shufflevector <4 x float> %BC34, <4 x float> %BC34, <4 x i32> <i32 2, i32 4, i32 6, i32 0>
@@ -117,7 +117,7 @@ CF81:                                             ; preds = %CF82
   %B54 = fdiv double %L24, %L43
   %BC55 = bitcast <4 x i64> zeroinitializer to <4 x double>
   %Sl56 = select i1 false, i8 %5, i8 97
-  %L57 = load i8* %0
+  %L57 = load i8, i8* %0
   store i8 %L50, i8* %0
   %E58 = extractelement <2 x i1> %Shuff20, i32 1
   br i1 %E58, label %CF, label %CF73
@@ -129,7 +129,7 @@ CF73:                                             ; preds = %CF73, %CF81
   %PC62 = bitcast double* %A3 to float*
   %Sl63 = select i1 %Cmp10, <1 x i64> zeroinitializer, <1 x i64> zeroinitializer
   %Cmp64 = icmp ne <2 x i1> %Cmp, %Shuff
-  %L65 = load double* %A1
+  %L65 = load double, double* %A1
   store float -4.374162e+06, float* %PC62
   %E66 = extractelement <8 x i1> %I21, i32 3
   br i1 %E66, label %CF73, label %CF79
diff --git a/test/CodeGen/Mips/msa/llvm-stress-s525530439.ll b/test/CodeGen/Mips/msa/llvm-stress-s525530439.ll
index 697871df797d..c0bc90563484 100644
--- a/test/CodeGen/Mips/msa/llvm-stress-s525530439.ll
+++ b/test/CodeGen/Mips/msa/llvm-stress-s525530439.ll
@@ -14,7 +14,7 @@ BB:
   %A2 = alloca <1 x double>
   %A1 = alloca <8 x double>
   %A = alloca i64
-  %L = load i8* %0
+  %L = load i8, i8* %0
   store i64 33695, i64* %A
   %E = extractelement <4 x i32> zeroinitializer, i32 3
   %Shuff = shufflevector <2 x i32> <i32 -1, i32 -1>, <2 x i32> <i32 -1, i32 -1>, <2 x i32> <i32 2, i32 0>
@@ -22,7 +22,7 @@ BB:
   %B = lshr <8 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>
   %ZE = fpext float 0x3B64A2B880000000 to double
   %Sl = select i1 true, i16 -1, i16 -11642
-  %L5 = load i8* %0
+  %L5 = load i8, i8* %0
   store i8 0, i8* %0
   %E6 = extractelement <4 x i32> zeroinitializer, i32 2
   %Shuff7 = shufflevector <8 x i1> zeroinitializer, <8 x i1> zeroinitializer, <8 x i32> <i32 undef, i32 7, i32 9, i32 11, i32 13, i32 15, i32 1, i32 undef>
@@ -31,7 +31,7 @@ BB:
   %BC = bitcast <2 x i32> <i32 -1, i32 -1> to <2 x float>
   %Sl10 = select i1 true, i32* %1, i32* %1
   %Cmp = icmp sge <8 x i64> zeroinitializer, zeroinitializer
-  %L11 = load i32* %Sl10
+  %L11 = load i32, i32* %Sl10
   store <1 x double> zeroinitializer, <1 x double>* %A2
   %E12 = extractelement <4 x i16> zeroinitializer, i32 0
   %Shuff13 = shufflevector <1 x i64> zeroinitializer, <1 x i64> zeroinitializer, <1 x i32> undef
@@ -43,7 +43,7 @@ BB:
   br label %CF75
 
 CF75:                                             ; preds = %CF75, %BB
-  %L19 = load i32* %Sl10
+  %L19 = load i32, i32* %Sl10
   store i32 %L11, i32* %Sl10
   %E20 = extractelement <4 x i32> zeroinitializer, i32 1
   %Shuff21 = shufflevector <4 x i32> zeroinitializer, <4 x i32> %I8, <4 x i32> <i32 undef, i32 2, i32 4, i32 6>
@@ -55,7 +55,7 @@ CF75:                                             ; preds = %CF75, %BB
   br i1 %Cmp26, label %CF75, label %CF76
 
 CF76:                                             ; preds = %CF75
-  %L27 = load i32* %Sl10
+  %L27 = load i32, i32* %Sl10
   store i32 439732, i32* %Sl10
   %E28 = extractelement <4 x i32> %Shuff21, i32 3
   %Shuff29 = shufflevector <8 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, <8 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, <8 x i32> <i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14, i32 0>
@@ -65,7 +65,7 @@ CF76:                                             ; preds = %CF75
   br label %CF74
 
 CF74:                                             ; preds = %CF74, %CF80, %CF78, %CF76
-  %L33 = load i64* %2
+  %L33 = load i64, i64* %2
   store i32 71140, i32* %Sl10
   %E34 = extractelement <4 x i32> zeroinitializer, i32 1
   %Shuff35 = shufflevector <1 x i16> zeroinitializer, <1 x i16> zeroinitializer, <1 x i32> undef
@@ -76,7 +76,7 @@ CF74:                                             ; preds = %CF74, %CF80, %CF78,
   br i1 %Cmp39, label %CF74, label %CF80
 
 CF80:                                             ; preds = %CF74
-  %L40 = load i8* %0
+  %L40 = load i8, i8* %0
   store i32 0, i32* %Sl10
   %E41 = extractelement <8 x i64> zeroinitializer, i32 1
   %Shuff42 = shufflevector <1 x i16> %I14, <1 x i16> %I14, <1 x i32> undef
@@ -86,7 +86,7 @@ CF80:                                             ; preds = %CF74
   br i1 %Sl44, label %CF74, label %CF78
 
 CF78:                                             ; preds = %CF80
-  %L45 = load i32* %Sl10
+  %L45 = load i32, i32* %Sl10
   store i8 %L5, i8* %0
   %E46 = extractelement <8 x i1> %Shuff7, i32 2
   br i1 %E46, label %CF74, label %CF77
@@ -101,7 +101,7 @@ CF77:                                             ; preds = %CF77, %CF78
   br i1 %Cmp52, label %CF77, label %CF79
 
 CF79:                                             ; preds = %CF77
-  %L53 = load i32* %Sl10
+  %L53 = load i32, i32* %Sl10
   store i8 %L40, i8* %0
   %E54 = extractelement <4 x i32> zeroinitializer, i32 1
   %Shuff55 = shufflevector <4 x i32> %Shuff21, <4 x i32> %I8, <4 x i32> <i32 4, i32 6, i32 undef, i32 2>
@@ -109,7 +109,7 @@ CF79:                                             ; preds = %CF77
   %Tr = trunc <1 x i64> %Shuff13 to <1 x i16>
   %Sl57 = select i1 %Cmp18, <2 x i32> <i32 -1, i32 -1>, <2 x i32> <i32 -1, i32 -1>
   %Cmp58 = icmp uge <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, %I56
-  %L59 = load i8* %0
+  %L59 = load i8, i8* %0
   store <1 x double> zeroinitializer, <1 x double>* %A2
   %E60 = extractelement <4 x i32> zeroinitializer, i32 0
   %Shuff61 = shufflevector <4 x i32> %I8, <4 x i32> %I8, <4 x i32> <i32 undef, i32 1, i32 undef, i32 undef>
@@ -121,7 +121,7 @@ CF79:                                             ; preds = %CF77
   br label %CF
 
 CF:                                               ; preds = %CF79
-  %L66 = load i32* %Sl10
+  %L66 = load i32, i32* %Sl10
   store i32 %E6, i32* %PC
   %E67 = extractelement <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, i32 2
   %Shuff68 = shufflevector <4 x i32> %Sl64, <4 x i32> %I8, <4 x i32> <i32 5, i32 undef, i32 1, i32 undef>
diff --git a/test/CodeGen/Mips/msa/llvm-stress-s997348632.ll b/test/CodeGen/Mips/msa/llvm-stress-s997348632.ll
index dc4200ad4285..a3150e9a67d5 100644
--- a/test/CodeGen/Mips/msa/llvm-stress-s997348632.ll
+++ b/test/CodeGen/Mips/msa/llvm-stress-s997348632.ll
@@ -14,14 +14,14 @@ BB:
   %A2 = alloca <4 x i1>
   %A1 = alloca <4 x i16>
   %A = alloca <2 x i32>
-  %L = load i8* %0
+  %L = load i8, i8* %0
   store i8 %L, i8* %0
   %E = extractelement <4 x i32> zeroinitializer, i32 0
   %Shuff = shufflevector <4 x i64> zeroinitializer, <4 x i64> zeroinitializer, <4 x i32> <i32 undef, i32 1, i32 3, i32 5>
   %I = insertelement <2 x i1> zeroinitializer, i1 false, i32 1
   %FC = sitofp <4 x i32> zeroinitializer to <4 x double>
   %Sl = select i1 false, <4 x i64> %Shuff, <4 x i64> %Shuff
-  %L5 = load i8* %0
+  %L5 = load i8, i8* %0
   store i8 %5, i8* %0
   %E6 = extractelement <1 x i16> zeroinitializer, i32 0
   %Shuff7 = shufflevector <2 x i1> %I, <2 x i1> %I, <2 x i32> <i32 1, i32 undef>
@@ -30,7 +30,7 @@ BB:
   %FC9 = fptoui float 0x406DB70180000000 to i64
   %Sl10 = select i1 false, <8 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, <8 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>
   %Cmp = icmp ult <4 x i64> zeroinitializer, zeroinitializer
-  %L11 = load i8* %0
+  %L11 = load i8, i8* %0
   store i8 %L, i8* %0
   %E12 = extractelement <4 x i64> zeroinitializer, i32 2
   %Shuff13 = shufflevector <4 x i32> zeroinitializer, <4 x i32> zeroinitializer, <4 x i32> <i32 5, i32 7, i32 undef, i32 3>
@@ -42,7 +42,7 @@ BB:
   br label %CF
 
 CF:                                               ; preds = %CF, %CF79, %CF84, %BB
-  %L18 = load i8* %0
+  %L18 = load i8, i8* %0
   store i8 %L, i8* %0
   %E19 = extractelement <4 x i64> %Sl, i32 3
   %Shuff20 = shufflevector <2 x i1> %Shuff7, <2 x i1> %I, <2 x i32> <i32 2, i32 0>
@@ -54,7 +54,7 @@ CF:                                               ; preds = %CF, %CF79, %CF84, %
   br i1 %Cmp25, label %CF, label %CF79
 
 CF79:                                             ; preds = %CF
-  %L26 = load i8* %0
+  %L26 = load i8, i8* %0
   store i8 %L26, i8* %0
   %E27 = extractelement <1 x i16> zeroinitializer, i32 0
   %Shuff28 = shufflevector <16 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, <16 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, <16 x i32> <i32 13, i32 15, i32 17, i32 19, i32 21, i32 23, i32 25, i32 27, i32 29, i32 31, i32 1, i32 3, i32 5, i32 7, i32 9, i32 11>
@@ -65,7 +65,7 @@ CF79:                                             ; preds = %CF
   br i1 %Cmp32, label %CF, label %CF78
 
 CF78:                                             ; preds = %CF78, %CF79
-  %L33 = load i8* %0
+  %L33 = load i8, i8* %0
   store i8 %L, i8* %0
   %E34 = extractelement <16 x i32> %Shuff28, i32 1
   %Shuff35 = shufflevector <4 x i64> zeroinitializer, <4 x i64> %I21, <4 x i32> <i32 undef, i32 6, i32 0, i32 2>
@@ -76,7 +76,7 @@ CF78:                                             ; preds = %CF78, %CF79
   br i1 %Cmp38, label %CF78, label %CF80
 
 CF80:                                             ; preds = %CF80, %CF82, %CF78
-  %L39 = load i8* %0
+  %L39 = load i8, i8* %0
   store i8 %L, i8* %0
   %E40 = extractelement <2 x i1> %Shuff20, i32 1
   br i1 %E40, label %CF80, label %CF82
@@ -87,7 +87,7 @@ CF82:                                             ; preds = %CF80
   %B43 = sub i32 %E, 0
   %Sl44 = select i1 %Cmp32, <16 x i32> %Shuff28, <16 x i32> %Shuff28
   %Cmp45 = icmp sgt <4 x i64> zeroinitializer, %I21
-  %L46 = load i8* %0
+  %L46 = load i8, i8* %0
   store i8 %L11, i8* %0
   %E47 = extractelement <8 x i32> %Sl16, i32 4
   %Shuff48 = shufflevector <2 x i1> zeroinitializer, <2 x i1> %Shuff7, <2 x i32> <i32 undef, i32 1>
@@ -99,7 +99,7 @@ CF82:                                             ; preds = %CF80
 CF81:                                             ; preds = %CF81, %CF82
   %Sl52 = select i1 false, float -6.749110e+06, float 0x406DB70180000000
   %Cmp53 = icmp uge <2 x i32> <i32 -1, i32 -1>, <i32 -1, i32 -1>
-  %L54 = load i8* %0
+  %L54 = load i8, i8* %0
   store i8 %L5, i8* %0
   %E55 = extractelement <8 x i32> zeroinitializer, i32 7
   %Shuff56 = shufflevector <4 x i64> zeroinitializer, <4 x i64> zeroinitializer, <4 x i32> <i32 undef, i32 4, i32 6, i32 0>
@@ -108,7 +108,7 @@ CF81:                                             ; preds = %CF81, %CF82
   %FC59 = fptoui <4 x double> %I36 to <4 x i16>
   %Sl60 = select i1 %Cmp17, <2 x i1> %I, <2 x i1> %I57
   %Cmp61 = icmp ule <8 x i32> %B50, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>
-  %L62 = load i8* %0
+  %L62 = load i8, i8* %0
   store i8 %L33, i8* %0
   %E63 = extractelement <4 x i64> %Shuff, i32 2
   %Shuff64 = shufflevector <4 x i64> %Shuff56, <4 x i64> %Shuff56, <4 x i32> <i32 5, i32 7, i32 1, i32 undef>
@@ -126,7 +126,7 @@ CF84:                                             ; preds = %CF83
   br i1 %Cmp69, label %CF, label %CF77
 
 CF77:                                             ; preds = %CF84
-  %L70 = load i8* %0
+  %L70 = load i8, i8* %0
   store i8 %L, i8* %0
   %E71 = extractelement <4 x i64> %Shuff, i32 0
   %Shuff72 = shufflevector <2 x i1> zeroinitializer, <2 x i1> %I, <2 x i32> <i32 3, i32 1>
diff --git a/test/CodeGen/Mips/msa/shuffle.ll b/test/CodeGen/Mips/msa/shuffle.ll
index faeec5d58dd4..aadff7d5a5c8 100644
--- a/test/CodeGen/Mips/msa/shuffle.ll
+++ b/test/CodeGen/Mips/msa/shuffle.ll
@@ -2,9 +2,9 @@
 ; RUN: llc -march=mipsel -mattr=+msa,+fp64 < %s | FileCheck %s
 
 define void @vshf_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
-  ; CHECK: vshf_v16i8_0:
+  ; CHECK-LABEL: vshf_v16i8_0:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <16 x i8> %1, <16 x i8> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
   ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($
@@ -14,13 +14,12 @@ define void @vshf_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind
   ; CHECK-DAG: st.b [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v16i8_0
 }
 
 define void @vshf_v16i8_1(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
-  ; CHECK: vshf_v16i8_1:
+  ; CHECK-LABEL: vshf_v16i8_1:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <16 x i8> %1, <16 x i8> undef, <16 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
   ; CHECK-DAG: splati.b [[R3:\$w[0-9]+]], [[R1]][1]
@@ -28,14 +27,13 @@ define void @vshf_v16i8_1(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind
   ; CHECK-DAG: st.b [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v16i8_1
 }
 
 define void @vshf_v16i8_2(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
-  ; CHECK: vshf_v16i8_2:
+  ; CHECK-LABEL: vshf_v16i8_2:
 
-  %1 = load <16 x i8>* %a
-  %2 = load <16 x i8>* %b
+  %1 = load <16 x i8>, <16 x i8>* %a
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <16 x i8> %1, <16 x i8> %2, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 16>
   ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($
@@ -45,15 +43,14 @@ define void @vshf_v16i8_2(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind
   ; CHECK-DAG: st.b [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v16i8_2
 }
 
 define void @vshf_v16i8_3(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
-  ; CHECK: vshf_v16i8_3:
+  ; CHECK-LABEL: vshf_v16i8_3:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <16 x i8> %1, <16 x i8> %2, <16 x i32> <i32 17, i32 24, i32 25, i32 18, i32 19, i32 20, i32 28, i32 19, i32 1, i32 8, i32 9, i32 2, i32 3, i32 4, i32 12, i32 3>
   ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($
@@ -65,13 +62,12 @@ define void @vshf_v16i8_3(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind
   ; CHECK-DAG: st.b [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v16i8_3
 }
 
 define void @vshf_v16i8_4(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
-  ; CHECK: vshf_v16i8_4:
+  ; CHECK-LABEL: vshf_v16i8_4:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <16 x i8> %1, <16 x i8> %1, <16 x i32> <i32 1, i32 17, i32 1, i32 17, i32 1, i32 17, i32 1, i32 17, i32 1, i32 17, i32 1, i32 17, i32 1, i32 17, i32 1, i32 17>
   ; CHECK-DAG: splati.b [[R3:\$w[0-9]+]], [[R1]][1]
@@ -79,13 +75,12 @@ define void @vshf_v16i8_4(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind
   ; CHECK-DAG: st.b [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v16i8_4
 }
 
 define void @vshf_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
-  ; CHECK: vshf_v8i16_0:
+  ; CHECK-LABEL: vshf_v8i16_0:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <8 x i16> %1, <8 x i16> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
   ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($
@@ -95,13 +90,12 @@ define void @vshf_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind
   ; CHECK-DAG: st.h [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v8i16_0
 }
 
 define void @vshf_v8i16_1(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
-  ; CHECK: vshf_v8i16_1:
+  ; CHECK-LABEL: vshf_v8i16_1:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <8 x i16> %1, <8 x i16> undef, <8 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
   ; CHECK-DAG: splati.h [[R3:\$w[0-9]+]], [[R1]][1]
@@ -109,14 +103,13 @@ define void @vshf_v8i16_1(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind
   ; CHECK-DAG: st.h [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v8i16_1
 }
 
 define void @vshf_v8i16_2(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
-  ; CHECK: vshf_v8i16_2:
+  ; CHECK-LABEL: vshf_v8i16_2:
 
-  %1 = load <8 x i16>* %a
-  %2 = load <8 x i16>* %b
+  %1 = load <8 x i16>, <8 x i16>* %a
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 8>
   ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($
@@ -126,15 +119,14 @@ define void @vshf_v8i16_2(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind
   ; CHECK-DAG: st.h [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v8i16_2
 }
 
 define void @vshf_v8i16_3(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
-  ; CHECK: vshf_v8i16_3:
+  ; CHECK-LABEL: vshf_v8i16_3:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 1, i32 8, i32 9, i32 2, i32 3, i32 4, i32 12, i32 3>
   ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($
@@ -146,13 +138,12 @@ define void @vshf_v8i16_3(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind
   ; CHECK-DAG: st.h [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v8i16_3
 }
 
 define void @vshf_v8i16_4(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
-  ; CHECK: vshf_v8i16_4:
+  ; CHECK-LABEL: vshf_v8i16_4:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <8 x i16> %1, <8 x i16> %1, <8 x i32> <i32 1, i32 9, i32 1, i32 9, i32 1, i32 9, i32 1, i32 9>
   ; CHECK-DAG: splati.h [[R3:\$w[0-9]+]], [[R1]][1]
@@ -160,16 +151,15 @@ define void @vshf_v8i16_4(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind
   ; CHECK-DAG: st.h [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v8i16_4
 }
 
 ; Note: v4i32 only has one 4-element set so it's impossible to get a vshf.w
 ; instruction when using a single vector.
 
 define void @vshf_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
-  ; CHECK: vshf_v4i32_0:
+  ; CHECK-LABEL: vshf_v4i32_0:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
   ; CHECK-DAG: shf.w [[R3:\$w[0-9]+]], [[R1]], 27
@@ -177,28 +167,26 @@ define void @vshf_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind
   ; CHECK-DAG: st.w [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v4i32_0
 }
 
 define void @vshf_v4i32_1(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
-  ; CHECK: vshf_v4i32_1:
+  ; CHECK-LABEL: vshf_v4i32_1:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
-  ; CHECK-DAG: shf.w [[R3:\$w[0-9]+]], [[R1]], 85
+  ; CHECK-DAG: splati.w [[R3:\$w[0-9]+]], [[R1]][1]
   store <4 x i32> %2, <4 x i32>* %c
   ; CHECK-DAG: st.w [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v4i32_1
 }
 
 define void @vshf_v4i32_2(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
-  ; CHECK: vshf_v4i32_2:
+  ; CHECK-LABEL: vshf_v4i32_2:
 
-  %1 = load <4 x i32>* %a
-  %2 = load <4 x i32>* %b
+  %1 = load <4 x i32>, <4 x i32>* %a
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 4, i32 5, i32 6, i32 4>
   ; CHECK-DAG: shf.w [[R3:\$w[0-9]+]], [[R2]], 36
@@ -206,15 +194,14 @@ define void @vshf_v4i32_2(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind
   ; CHECK-DAG: st.w [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v4i32_2
 }
 
 define void @vshf_v4i32_3(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
-  ; CHECK: vshf_v4i32_3:
+  ; CHECK-LABEL: vshf_v4i32_3:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 1, i32 5, i32 6, i32 4>
   ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($
@@ -226,27 +213,26 @@ define void @vshf_v4i32_3(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind
   ; CHECK-DAG: st.w [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v4i32_3
 }
 
 define void @vshf_v4i32_4(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
-  ; CHECK: vshf_v4i32_4:
+  ; CHECK-LABEL: vshf_v4i32_4:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <4 x i32> %1, <4 x i32> %1, <4 x i32> <i32 1, i32 5, i32 5, i32 1>
-  ; CHECK-DAG: shf.w [[R3:\$w[0-9]+]], [[R1]], 85
+  ; The two operand vectors are the same so element 1 and 5 are equivalent.
+  ; CHECK-DAG: splati.w [[R3:\$w[0-9]+]], [[R1]][1]
   store <4 x i32> %2, <4 x i32>* %c
   ; CHECK-DAG: st.w [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v4i32_4
 }
 
 define void @vshf_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
-  ; CHECK: vshf_v2i64_0:
+  ; CHECK-LABEL: vshf_v2i64_0:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <2 x i64> %1, <2 x i64> undef, <2 x i32> <i32 1, i32 0>
   ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($
@@ -256,13 +242,12 @@ define void @vshf_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind
   ; CHECK-DAG: st.d [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v2i64_0
 }
 
 define void @vshf_v2i64_1(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
-  ; CHECK: vshf_v2i64_1:
+  ; CHECK-LABEL: vshf_v2i64_1:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <2 x i64> %1, <2 x i64> undef, <2 x i32> <i32 1, i32 1>
   ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][1]
@@ -270,14 +255,13 @@ define void @vshf_v2i64_1(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind
   ; CHECK-DAG: st.d [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v2i64_1
 }
 
 define void @vshf_v2i64_2(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
-  ; CHECK: vshf_v2i64_2:
+  ; CHECK-LABEL: vshf_v2i64_2:
 
-  %1 = load <2 x i64>* %a
-  %2 = load <2 x i64>* %b
+  %1 = load <2 x i64>, <2 x i64>* %a
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 3, i32 2>
   ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($
@@ -287,15 +271,14 @@ define void @vshf_v2i64_2(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind
   ; CHECK-DAG: st.d [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v2i64_2
 }
 
 define void @vshf_v2i64_3(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
-  ; CHECK: vshf_v2i64_3:
+  ; CHECK-LABEL: vshf_v2i64_3:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 1, i32 2>
   ; CHECK-DAG: addiu [[PTR_A:\$[0-9]+]], {{.*}}, %lo($
@@ -307,13 +290,12 @@ define void @vshf_v2i64_3(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind
   ; CHECK-DAG: st.d [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v2i64_3
 }
 
 define void @vshf_v2i64_4(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
-  ; CHECK: vshf_v2i64_4:
+  ; CHECK-LABEL: vshf_v2i64_4:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <2 x i64> %1, <2 x i64> %1, <2 x i32> <i32 1, i32 3>
   ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][1]
@@ -321,13 +303,12 @@ define void @vshf_v2i64_4(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind
   ; CHECK-DAG: st.d [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size vshf_v2i64_4
 }
 
 define void @shf_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
-  ; CHECK: shf_v16i8_0:
+  ; CHECK-LABEL: shf_v16i8_0:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <16 x i8> %1, <16 x i8> undef, <16 x i32> <i32 1, i32 3, i32 2, i32 0, i32 5, i32 7, i32 6, i32 4, i32 9, i32 11, i32 10, i32 8, i32 13, i32 15, i32 14, i32 12>
   ; CHECK-DAG: shf.b [[R3:\$w[0-9]+]], [[R1]], 45
@@ -335,13 +316,12 @@ define void @shf_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
   ; CHECK-DAG: st.b [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size shf_v16i8_0
 }
 
 define void @shf_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
-  ; CHECK: shf_v8i16_0:
+  ; CHECK-LABEL: shf_v8i16_0:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <8 x i16> %1, <8 x i16> undef, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4>
   ; CHECK-DAG: shf.h [[R3:\$w[0-9]+]], [[R1]], 27
@@ -349,13 +329,12 @@ define void @shf_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
   ; CHECK-DAG: st.h [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size shf_v8i16_0
 }
 
 define void @shf_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
-  ; CHECK: shf_v4i32_0:
+  ; CHECK-LABEL: shf_v4i32_0:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
   ; CHECK-DAG: shf.w [[R3:\$w[0-9]+]], [[R1]], 27
@@ -363,82 +342,194 @@ define void @shf_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
   ; CHECK-DAG: st.w [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size shf_v4i32_0
 }
 
 ; shf.d does not exist
 
 define void @ilvev_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
-  ; CHECK: ilvev_v16i8_0:
+  ; CHECK-LABEL: ilvev_v16i8_0:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
                      <16 x i32> <i32 0, i32 16, i32 2, i32 18, i32 4, i32 20, i32 6, i32 22, i32 8, i32 24, i32 10, i32 26, i32 12, i32 28, i32 14, i32 30>
-  ; CHECK-DAG: ilvev.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; CHECK-DAG: ilvev.b [[R3:\$w[0-9]+]], [[R2]], [[R1]]
   store <16 x i8> %3, <16 x i8>* %c
   ; CHECK-DAG: st.b [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size ilvev_v16i8_0
 }
 
 define void @ilvev_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
-  ; CHECK: ilvev_v8i16_0:
+  ; CHECK-LABEL: ilvev_v8i16_0:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 0, i32 8, i32 2, i32 10, i32 4, i32 12, i32 6, i32 14>
-  ; CHECK-DAG: ilvev.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; CHECK-DAG: ilvev.h [[R3:\$w[0-9]+]], [[R2]], [[R1]]
   store <8 x i16> %3, <8 x i16>* %c
   ; CHECK-DAG: st.h [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size ilvev_v8i16_0
 }
 
 define void @ilvev_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
-  ; CHECK: ilvev_v4i32_0:
+  ; CHECK-LABEL: ilvev_v4i32_0:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 0, i32 4, i32 2, i32 6>
-  ; CHECK-DAG: ilvev.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; CHECK-DAG: ilvev.w [[R3:\$w[0-9]+]], [[R2]], [[R1]]
   store <4 x i32> %3, <4 x i32>* %c
   ; CHECK-DAG: st.w [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size ilvev_v4i32_0
 }
 
 define void @ilvev_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
-  ; CHECK: ilvev_v2i64_0:
+  ; CHECK-LABEL: ilvev_v2i64_0:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 0, i32 2>
-  ; CHECK-DAG: ilvev.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; CHECK-DAG: ilvev.d [[R3:\$w[0-9]+]], [[R2]], [[R1]]
+  store <2 x i64> %3, <2 x i64>* %c
+  ; CHECK-DAG: st.d [[R3]], 0($4)
+
+  ret void
+}
+
+; Interleaving one operand with itself.
+define void @ilvev_v16i8_1(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
+  ; CHECK-LABEL: ilvev_v16i8_1:
+
+  %1 = load <16 x i8>, <16 x i8>* %a
+  ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <16 x i8>, <16 x i8>* %b
+  %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
+                     <16 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6, i32 8, i32 8, i32 10, i32 10, i32 12, i32 12, i32 14, i32 14>
+  ; CHECK-DAG: ilvev.b [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <16 x i8> %3, <16 x i8>* %c
+  ; CHECK-DAG: st.b [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvev_v8i16_1(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
+  ; CHECK-LABEL: ilvev_v8i16_1:
+
+  %1 = load <8 x i16>, <8 x i16>* %a
+  ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <8 x i16>, <8 x i16>* %b
+  %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6>
+  ; CHECK-DAG: ilvev.h [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <8 x i16> %3, <8 x i16>* %c
+  ; CHECK-DAG: st.h [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvev_v4i32_1(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
+  ; CHECK-LABEL: ilvev_v4i32_1:
+
+  %1 = load <4 x i32>, <4 x i32>* %a
+  ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <4 x i32>, <4 x i32>* %b
+  %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
+  ; CHECK-DAG: ilvev.w [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <4 x i32> %3, <4 x i32>* %c
+  ; CHECK-DAG: st.w [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvev_v2i64_1(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
+  ; CHECK-LABEL: ilvev_v2i64_1:
+
+  %1 = load <2 x i64>, <2 x i64>* %a
+  ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <2 x i64>, <2 x i64>* %b
+  %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 0, i32 0>
+  ; ilvev.d with two identical operands is equivalent to splati.d
+  ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][0]
+  store <2 x i64> %3, <2 x i64>* %c
+  ; CHECK-DAG: st.d [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvev_v16i8_2(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
+  ; CHECK-LABEL: ilvev_v16i8_2:
+
+  %1 = load <16 x i8>, <16 x i8>* %a
+  %2 = load <16 x i8>, <16 x i8>* %b
+  ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
+                     <16 x i32> <i32 16, i32 16, i32 18, i32 18, i32 20, i32 20, i32 22, i32 22, i32 24, i32 24, i32 26, i32 26, i32 28, i32 28, i32 30, i32 30>
+  ; CHECK-DAG: ilvev.b [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <16 x i8> %3, <16 x i8>* %c
+  ; CHECK-DAG: st.b [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvev_v8i16_2(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
+  ; CHECK-LABEL: ilvev_v8i16_2:
+
+  %1 = load <8 x i16>, <8 x i16>* %a
+  %2 = load <8 x i16>, <8 x i16>* %b
+  ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 8, i32 8, i32 10, i32 10, i32 12, i32 12, i32 14, i32 14>
+  ; CHECK-DAG: ilvev.h [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <8 x i16> %3, <8 x i16>* %c
+  ; CHECK-DAG: st.h [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvev_v4i32_2(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
+  ; CHECK-LABEL: ilvev_v4i32_2:
+
+  %1 = load <4 x i32>, <4 x i32>* %a
+  %2 = load <4 x i32>, <4 x i32>* %b
+  ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 4, i32 4, i32 6, i32 6>
+  ; CHECK-DAG: ilvev.w [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <4 x i32> %3, <4 x i32>* %c
+  ; CHECK-DAG: st.w [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvev_v2i64_2(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
+  ; CHECK-LABEL: ilvev_v2i64_2:
+
+  %1 = load <2 x i64>, <2 x i64>* %a
+  %2 = load <2 x i64>, <2 x i64>* %b
+  ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 2, i32 2>
+  ; ilvev.d with two identical operands is equivalent to splati.d
+  ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R2]][0]
   store <2 x i64> %3, <2 x i64>* %c
   ; CHECK-DAG: st.d [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size ilvev_v2i64_0
 }
 
 define void @ilvod_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
-  ; CHECK: ilvod_v16i8_0:
+  ; CHECK-LABEL: ilvod_v16i8_0:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
                      <16 x i32> <i32 1, i32 17, i32 3, i32 19, i32 5, i32 21, i32 7, i32 23, i32 9, i32 25, i32 11, i32 27, i32 13, i32 29, i32 15, i32 31>
@@ -447,15 +538,14 @@ define void @ilvod_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind
   ; CHECK-DAG: st.b [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size ilvod_v16i8_0
 }
 
 define void @ilvod_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
-  ; CHECK: ilvod_v8i16_0:
+  ; CHECK-LABEL: ilvod_v8i16_0:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 1, i32 9, i32 3, i32 11, i32 5, i32 13, i32 7, i32 15>
   ; CHECK-DAG: ilvod.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -463,15 +553,14 @@ define void @ilvod_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind
   ; CHECK-DAG: st.h [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size ilvod_v8i16_0
 }
 
 define void @ilvod_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
-  ; CHECK: ilvod_v4i32_0:
+  ; CHECK-LABEL: ilvod_v4i32_0:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 1, i32 5, i32 3, i32 7>
   ; CHECK-DAG: ilvod.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -479,15 +568,14 @@ define void @ilvod_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind
   ; CHECK-DAG: st.w [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size ilvod_v4i32_0
 }
 
 define void @ilvod_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
-  ; CHECK: ilvod_v2i64_0:
+  ; CHECK-LABEL: ilvod_v2i64_0:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 1, i32 3>
   ; CHECK-DAG: ilvod.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
@@ -495,262 +583,710 @@ define void @ilvod_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind
   ; CHECK-DAG: st.d [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size ilvod_v2i64_0
 }
 
-define void @ilvl_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
-  ; CHECK: ilvl_v16i8_0:
+define void @ilvod_v16i8_1(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
+  ; CHECK-LABEL: ilvod_v16i8_1:
+
+  %1 = load <16 x i8>, <16 x i8>* %a
+  ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <16 x i8>, <16 x i8>* %b
+  %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
+                     <16 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 7, i32 9, i32 9, i32 11, i32 11, i32 13, i32 13, i32 15, i32 15>
+  ; CHECK-DAG: ilvod.b [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <16 x i8> %3, <16 x i8>* %c
+  ; CHECK-DAG: st.b [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvod_v8i16_1(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
+  ; CHECK-LABEL: ilvod_v8i16_1:
+
+  %1 = load <8 x i16>, <8 x i16>* %a
+  ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <8 x i16>, <8 x i16>* %b
+  %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 7>
+  ; CHECK-DAG: ilvod.h [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <8 x i16> %3, <8 x i16>* %c
+  ; CHECK-DAG: st.h [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvod_v4i32_1(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
+  ; CHECK-LABEL: ilvod_v4i32_1:
+
+  %1 = load <4 x i32>, <4 x i32>* %a
+  ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <4 x i32>, <4 x i32>* %b
+  %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
+  ; CHECK-DAG: ilvod.w [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <4 x i32> %3, <4 x i32>* %c
+  ; CHECK-DAG: st.w [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvod_v2i64_1(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
+  ; CHECK-LABEL: ilvod_v2i64_1:
+
+  %1 = load <2 x i64>, <2 x i64>* %a
+  ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <2 x i64>, <2 x i64>* %b
+  %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 1, i32 1>
+  ; ilvod.d with two identical operands is equivalent to splati.d
+  ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][1]
+  store <2 x i64> %3, <2 x i64>* %c
+  ; CHECK-DAG: st.d [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvod_v16i8_2(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
+  ; CHECK-LABEL: ilvod_v16i8_2:
+
+  %1 = load <16 x i8>, <16 x i8>* %a
+  %2 = load <16 x i8>, <16 x i8>* %b
+  ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
+                     <16 x i32> <i32 17, i32 17, i32 19, i32 19, i32 21, i32 21, i32 23, i32 23, i32 25, i32 25, i32 27, i32 27, i32 29, i32 29, i32 31, i32 31>
+  ; CHECK-DAG: ilvod.b [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <16 x i8> %3, <16 x i8>* %c
+  ; CHECK-DAG: st.b [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvod_v8i16_2(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
+  ; CHECK-LABEL: ilvod_v8i16_2:
+
+  %1 = load <8 x i16>, <8 x i16>* %a
+  %2 = load <8 x i16>, <8 x i16>* %b
+  ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 9, i32 9, i32 11, i32 11, i32 13, i32 13, i32 15, i32 15>
+  ; CHECK-DAG: ilvod.h [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <8 x i16> %3, <8 x i16>* %c
+  ; CHECK-DAG: st.h [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvod_v4i32_2(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
+  ; CHECK-LABEL: ilvod_v4i32_2:
+
+  %1 = load <4 x i32>, <4 x i32>* %a
+  %2 = load <4 x i32>, <4 x i32>* %b
+  ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 5, i32 5, i32 7, i32 7>
+  ; CHECK-DAG: ilvod.w [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <4 x i32> %3, <4 x i32>* %c
+  ; CHECK-DAG: st.w [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvod_v2i64_2(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
+  ; CHECK-LABEL: ilvod_v2i64_2:
+
+  %1 = load <2 x i64>, <2 x i64>* %a
+  %2 = load <2 x i64>, <2 x i64>* %b
+  ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 3, i32 3>
+  ; ilvod.d with two identical operands is equivalent to splati.d
+  ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R2]][1]
+  store <2 x i64> %3, <2 x i64>* %c
+  ; CHECK-DAG: st.d [[R3]], 0($4)
 
-  %1 = load <16 x i8>* %a
+  ret void
+}
+
+define void @ilvr_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
+  ; CHECK-LABEL: ilvr_v16i8_0:
+
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
                      <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
-  ; CHECK-DAG: ilvl.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; CHECK-DAG: ilvr.b [[R3:\$w[0-9]+]], [[R2]], [[R1]]
   store <16 x i8> %3, <16 x i8>* %c
   ; CHECK-DAG: st.b [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size ilvl_v16i8_0
 }
 
-define void @ilvl_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
-  ; CHECK: ilvl_v8i16_0:
+define void @ilvr_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
+  ; CHECK-LABEL: ilvr_v8i16_0:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 2, i32 10, i32 3, i32 11>
-  ; CHECK-DAG: ilvl.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; CHECK-DAG: ilvr.h [[R3:\$w[0-9]+]], [[R2]], [[R1]]
   store <8 x i16> %3, <8 x i16>* %c
   ; CHECK-DAG: st.h [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size ilvl_v8i16_0
 }
 
-define void @ilvl_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
-  ; CHECK: ilvl_v4i32_0:
+define void @ilvr_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
+  ; CHECK-LABEL: ilvr_v4i32_0:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
-  ; CHECK-DAG: ilvl.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; CHECK-DAG: ilvr.w [[R3:\$w[0-9]+]], [[R2]], [[R1]]
   store <4 x i32> %3, <4 x i32>* %c
   ; CHECK-DAG: st.w [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size ilvl_v4i32_0
 }
 
-define void @ilvl_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
-  ; CHECK: ilvl_v2i64_0:
+define void @ilvr_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
+  ; CHECK-LABEL: ilvr_v2i64_0:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 0, i32 2>
-  ; ilvl.d and ilvev.d are equivalent for v2i64
-  ; CHECK-DAG: ilvev.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; ilvr.d and ilvev.d are equivalent for v2i64
+  ; CHECK-DAG: ilvev.d [[R3:\$w[0-9]+]], [[R2]], [[R1]]
   store <2 x i64> %3, <2 x i64>* %c
   ; CHECK-DAG: st.d [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size ilvl_v2i64_0
 }
 
-define void @ilvr_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
-  ; CHECK: ilvr_v16i8_0:
+define void @ilvr_v16i8_1(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
+  ; CHECK-LABEL: ilvr_v16i8_1:
+
+  %1 = load <16 x i8>, <16 x i8>* %a
+  %2 = load <16 x i8>, <16 x i8>* %b
+  ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
+                     <16 x i32> <i32 16, i32 16, i32 17, i32 17, i32 18, i32 18, i32 19, i32 19, i32 20, i32 20, i32 21, i32 21, i32 22, i32 22, i32 23, i32 23>
+  ; CHECK-DAG: ilvr.b [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <16 x i8> %3, <16 x i8>* %c
+  ; CHECK-DAG: st.b [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvr_v8i16_1(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
+  ; CHECK-LABEL: ilvr_v8i16_1:
+
+  %1 = load <8 x i16>, <8 x i16>* %a
+  %2 = load <8 x i16>, <8 x i16>* %b
+  ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 8, i32 8, i32 9, i32 9, i32 10, i32 10, i32 11, i32 11>
+  ; CHECK-DAG: ilvr.h [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <8 x i16> %3, <8 x i16>* %c
+  ; CHECK-DAG: st.h [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvr_v4i32_1(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
+  ; CHECK-LABEL: ilvr_v4i32_1:
+
+  %1 = load <4 x i32>, <4 x i32>* %a
+  %2 = load <4 x i32>, <4 x i32>* %b
+  ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 4, i32 4, i32 5, i32 5>
+  ; CHECK-DAG: ilvr.w [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <4 x i32> %3, <4 x i32>* %c
+  ; CHECK-DAG: st.w [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvr_v2i64_1(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
+  ; CHECK-LABEL: ilvr_v2i64_1:
+
+  %1 = load <2 x i64>, <2 x i64>* %a
+  %2 = load <2 x i64>, <2 x i64>* %b
+  ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 2, i32 2>
+  ; ilvr.d and splati.d are equivalent for v2i64
+  ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R2]][0]
+  store <2 x i64> %3, <2 x i64>* %c
+  ; CHECK-DAG: st.d [[R3]], 0($4)
 
-  %1 = load <16 x i8>* %a
+  ret void
+}
+
+define void @ilvr_v16i8_2(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
+  ; CHECK-LABEL: ilvr_v16i8_2:
+
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
+  %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
+                     <16 x i32> <i32 0, i32 0, i32 1, i32 1, i32 2, i32 2, i32 3, i32 3, i32 4, i32 4, i32 5, i32 5, i32 6, i32 6, i32 7, i32 7>
+  ; CHECK-DAG: ilvr.b [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <16 x i8> %3, <16 x i8>* %c
+  ; CHECK-DAG: st.b [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvr_v8i16_2(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
+  ; CHECK-LABEL: ilvr_v8i16_2:
+
+  %1 = load <8 x i16>, <8 x i16>* %a
+  ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <8 x i16>, <8 x i16>* %b
+  %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 0, i32 0, i32 1, i32 1, i32 2, i32 2, i32 3, i32 3>
+  ; CHECK-DAG: ilvr.h [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <8 x i16> %3, <8 x i16>* %c
+  ; CHECK-DAG: st.h [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvr_v4i32_2(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
+  ; CHECK-LABEL: ilvr_v4i32_2:
+
+  %1 = load <4 x i32>, <4 x i32>* %a
+  ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <4 x i32>, <4 x i32>* %b
+  %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 0, i32 0, i32 1, i32 1>
+  ; CHECK-DAG: ilvr.w [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <4 x i32> %3, <4 x i32>* %c
+  ; CHECK-DAG: st.w [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvr_v2i64_2(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
+  ; CHECK-LABEL: ilvr_v2i64_2:
+
+  %1 = load <2 x i64>, <2 x i64>* %a
+  ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <2 x i64>, <2 x i64>* %b
+  %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 0, i32 0>
+  ; ilvr.d and splati.d are equivalent for v2i64
+  ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][0]
+  store <2 x i64> %3, <2 x i64>* %c
+  ; CHECK-DAG: st.d [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvl_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
+  ; CHECK-LABEL: ilvl_v16i8_0:
+
+  %1 = load <16 x i8>, <16 x i8>* %a
+  ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
                      <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
-  ; CHECK-DAG: ilvr.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; CHECK-DAG: ilvl.b [[R3:\$w[0-9]+]], [[R2]], [[R1]]
   store <16 x i8> %3, <16 x i8>* %c
   ; CHECK-DAG: st.b [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size ilvr_v16i8_0
 }
 
-define void @ilvr_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
-  ; CHECK: ilvr_v8i16_0:
+define void @ilvl_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
+  ; CHECK-LABEL: ilvl_v8i16_0:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 4, i32 12, i32 5, i32 13, i32 6, i32 14, i32 7, i32 15>
-  ; CHECK-DAG: ilvr.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; CHECK-DAG: ilvl.h [[R3:\$w[0-9]+]], [[R2]], [[R1]]
   store <8 x i16> %3, <8 x i16>* %c
   ; CHECK-DAG: st.h [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size ilvr_v8i16_0
 }
 
-define void @ilvr_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
-  ; CHECK: ilvr_v4i32_0:
+define void @ilvl_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
+  ; CHECK-LABEL: ilvl_v4i32_0:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 2, i32 6, i32 3, i32 7>
-  ; CHECK-DAG: ilvr.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; CHECK-DAG: ilvl.w [[R3:\$w[0-9]+]], [[R2]], [[R1]]
   store <4 x i32> %3, <4 x i32>* %c
   ; CHECK-DAG: st.w [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size ilvr_v4i32_0
 }
 
-define void @ilvr_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
-  ; CHECK: ilvr_v2i64_0:
+define void @ilvl_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
+  ; CHECK-LABEL: ilvl_v2i64_0:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 1, i32 3>
-  ; ilvr.d and ilvod.d are equivalent for v2i64
+  ; ilvl.d and ilvod.d are equivalent for v2i64
   ; CHECK-DAG: ilvod.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
   store <2 x i64> %3, <2 x i64>* %c
   ; CHECK-DAG: st.d [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size ilvr_v2i64_0
+}
+
+define void @ilvl_v16i8_1(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
+  ; CHECK-LABEL: ilvl_v16i8_1:
+
+  %1 = load <16 x i8>, <16 x i8>* %a
+  %2 = load <16 x i8>, <16 x i8>* %b
+  ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
+                     <16 x i32> <i32 24, i32 24, i32 25, i32 25, i32 26, i32 26, i32 27, i32 27, i32 28, i32 28, i32 29, i32 29, i32 30, i32 30, i32 31, i32 31>
+  ; CHECK-DAG: ilvl.b [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <16 x i8> %3, <16 x i8>* %c
+  ; CHECK-DAG: st.b [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvl_v8i16_1(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
+  ; CHECK-LABEL: ilvl_v8i16_1:
+
+  %1 = load <8 x i16>, <8 x i16>* %a
+  %2 = load <8 x i16>, <8 x i16>* %b
+  ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 12, i32 12, i32 13, i32 13, i32 14, i32 14, i32 15, i32 15>
+  ; CHECK-DAG: ilvl.h [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <8 x i16> %3, <8 x i16>* %c
+  ; CHECK-DAG: st.h [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvl_v4i32_1(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
+  ; CHECK-LABEL: ilvl_v4i32_1:
+
+  %1 = load <4 x i32>, <4 x i32>* %a
+  %2 = load <4 x i32>, <4 x i32>* %b
+  ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 6, i32 6, i32 7, i32 7>
+  ; CHECK-DAG: ilvl.w [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <4 x i32> %3, <4 x i32>* %c
+  ; CHECK-DAG: st.w [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvl_v2i64_1(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
+  ; CHECK-LABEL: ilvl_v2i64_1:
+
+  %1 = load <2 x i64>, <2 x i64>* %a
+  %2 = load <2 x i64>, <2 x i64>* %b
+  ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 3, i32 3>
+  ; ilvl.d and splati.d are equivalent for v2i64
+  ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R2]][1]
+  store <2 x i64> %3, <2 x i64>* %c
+  ; CHECK-DAG: st.d [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvl_v16i8_2(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
+  ; CHECK-LABEL: ilvl_v16i8_2:
+
+  %1 = load <16 x i8>, <16 x i8>* %a
+  ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <16 x i8>, <16 x i8>* %b
+  %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
+                     <16 x i32> <i32 8, i32 8, i32 9, i32 9, i32 10, i32 10, i32 11, i32 11, i32 12, i32 12, i32 13, i32 13, i32 14, i32 14, i32 15, i32 15>
+  ; CHECK-DAG: ilvl.b [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <16 x i8> %3, <16 x i8>* %c
+  ; CHECK-DAG: st.b [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvl_v8i16_2(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
+  ; CHECK-LABEL: ilvl_v8i16_2:
+
+  %1 = load <8 x i16>, <8 x i16>* %a
+  ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <8 x i16>, <8 x i16>* %b
+  %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 4, i32 4, i32 5, i32 5, i32 6, i32 6, i32 7, i32 7>
+  ; CHECK-DAG: ilvl.h [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <8 x i16> %3, <8 x i16>* %c
+  ; CHECK-DAG: st.h [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvl_v4i32_2(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
+  ; CHECK-LABEL: ilvl_v4i32_2:
+
+  %1 = load <4 x i32>, <4 x i32>* %a
+  ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <4 x i32>, <4 x i32>* %b
+  %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 2, i32 2, i32 3, i32 3>
+  ; CHECK-DAG: ilvl.w [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <4 x i32> %3, <4 x i32>* %c
+  ; CHECK-DAG: st.w [[R3]], 0($4)
+
+  ret void
+}
+
+define void @ilvl_v2i64_2(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
+  ; CHECK-LABEL: ilvl_v2i64_2:
+
+  %1 = load <2 x i64>, <2 x i64>* %a
+  ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <2 x i64>, <2 x i64>* %b
+  %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 1, i32 1>
+  ; ilvl.d and splati.d are equivalent for v2i64
+  ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][1]
+  store <2 x i64> %3, <2 x i64>* %c
+  ; CHECK-DAG: st.d [[R3]], 0($4)
+
+  ret void
 }
 
 define void @pckev_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
-  ; CHECK: pckev_v16i8_0:
+  ; CHECK-LABEL: pckev_v16i8_0:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
                      <16 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14, i32 16, i32 18, i32 20, i32 22, i32 24, i32 26, i32 28, i32 30>
-  ; CHECK-DAG: pckev.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; CHECK-DAG: pckev.b [[R3:\$w[0-9]+]], [[R2]], [[R1]]
   store <16 x i8> %3, <16 x i8>* %c
   ; CHECK-DAG: st.b [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size pckev_v16i8_0
 }
 
 define void @pckev_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
-  ; CHECK: pckev_v8i16_0:
+  ; CHECK-LABEL: pckev_v8i16_0:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14>
-  ; CHECK-DAG: pckev.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; CHECK-DAG: pckev.h [[R3:\$w[0-9]+]], [[R2]], [[R1]]
   store <8 x i16> %3, <8 x i16>* %c
   ; CHECK-DAG: st.h [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size pckev_v8i16_0
 }
 
 define void @pckev_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
-  ; CHECK: pckev_v4i32_0:
+  ; CHECK-LABEL: pckev_v4i32_0:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 0, i32 2, i32 4, i32 6>
-  ; CHECK-DAG: pckev.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; CHECK-DAG: pckev.w [[R3:\$w[0-9]+]], [[R2]], [[R1]]
   store <4 x i32> %3, <4 x i32>* %c
   ; CHECK-DAG: st.w [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size pckev_v4i32_0
 }
 
 define void @pckev_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
-  ; CHECK: pckev_v2i64_0:
+  ; CHECK-LABEL: pckev_v2i64_0:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 0, i32 2>
   ; pckev.d and ilvev.d are equivalent for v2i64
-  ; CHECK-DAG: ilvev.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; CHECK-DAG: ilvev.d [[R3:\$w[0-9]+]], [[R2]], [[R1]]
+  store <2 x i64> %3, <2 x i64>* %c
+  ; CHECK-DAG: st.d [[R3]], 0($4)
+
+  ret void
+}
+
+define void @pckev_v16i8_1(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
+  ; CHECK-LABEL: pckev_v16i8_1:
+
+  %1 = load <16 x i8>, <16 x i8>* %a
+  %2 = load <16 x i8>, <16 x i8>* %b
+  ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
+                     <16 x i32> <i32 16, i32 18, i32 20, i32 22, i32 24, i32 26, i32 28, i32 30, i32 16, i32 18, i32 20, i32 22, i32 24, i32 26, i32 28, i32 30>
+  ; CHECK-DAG: pckev.b [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <16 x i8> %3, <16 x i8>* %c
+  ; CHECK-DAG: st.b [[R3]], 0($4)
+
+  ret void
+}
+
+define void @pckev_v8i16_1(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
+  ; CHECK-LABEL: pckev_v8i16_1:
+
+  %1 = load <8 x i16>, <8 x i16>* %a
+  %2 = load <8 x i16>, <8 x i16>* %b
+  ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 8, i32 10, i32 12, i32 14, i32 8, i32 10, i32 12, i32 14>
+  ; CHECK-DAG: pckev.h [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <8 x i16> %3, <8 x i16>* %c
+  ; CHECK-DAG: st.h [[R3]], 0($4)
+
+  ret void
+}
+
+define void @pckev_v4i32_1(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
+  ; CHECK-LABEL: pckev_v4i32_1:
+
+  %1 = load <4 x i32>, <4 x i32>* %a
+  %2 = load <4 x i32>, <4 x i32>* %b
+  ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 4, i32 6, i32 4, i32 6>
+  ; CHECK-DAG: pckev.w [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <4 x i32> %3, <4 x i32>* %c
+  ; CHECK-DAG: st.w [[R3]], 0($4)
+
+  ret void
+}
+
+define void @pckev_v2i64_1(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
+  ; CHECK-LABEL: pckev_v2i64_1:
+
+  %1 = load <2 x i64>, <2 x i64>* %a
+  %2 = load <2 x i64>, <2 x i64>* %b
+  ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 2, i32 2>
+  ; pckev.d and splati.d are equivalent for v2i64
+  ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R2]][0]
+  store <2 x i64> %3, <2 x i64>* %c
+  ; CHECK-DAG: st.d [[R3]], 0($4)
+
+  ret void
+}
+
+define void @pckev_v16i8_2(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
+  ; CHECK-LABEL: pckev_v16i8_2:
+
+  %1 = load <16 x i8>, <16 x i8>* %a
+  ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <16 x i8>, <16 x i8>* %b
+  %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
+                     <16 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14, i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14>
+  ; CHECK-DAG: pckev.b [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <16 x i8> %3, <16 x i8>* %c
+  ; CHECK-DAG: st.b [[R3]], 0($4)
+
+  ret void
+}
+
+define void @pckev_v8i16_2(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
+  ; CHECK-LABEL: pckev_v8i16_2:
+
+  %1 = load <8 x i16>, <8 x i16>* %a
+  ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <8 x i16>, <8 x i16>* %b
+  %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 0, i32 2, i32 4, i32 6>
+  ; CHECK-DAG: pckev.h [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <8 x i16> %3, <8 x i16>* %c
+  ; CHECK-DAG: st.h [[R3]], 0($4)
+
+  ret void
+}
+
+define void @pckev_v4i32_2(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
+  ; CHECK-LABEL: pckev_v4i32_2:
+
+  %1 = load <4 x i32>, <4 x i32>* %a
+  ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <4 x i32>, <4 x i32>* %b
+  %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 0, i32 2, i32 0, i32 2>
+  ; CHECK-DAG: pckev.w [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <4 x i32> %3, <4 x i32>* %c
+  ; CHECK-DAG: st.w [[R3]], 0($4)
+
+  ret void
+}
+
+define void @pckev_v2i64_2(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
+  ; CHECK-LABEL: pckev_v2i64_2:
+
+  %1 = load <2 x i64>, <2 x i64>* %a
+  ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <2 x i64>, <2 x i64>* %b
+  %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 0, i32 0>
+  ; pckev.d and splati.d are equivalent for v2i64
+  ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][0]
   store <2 x i64> %3, <2 x i64>* %c
   ; CHECK-DAG: st.d [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size pckev_v2i64_0
 }
 
 define void @pckod_v16i8_0(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
-  ; CHECK: pckod_v16i8_0:
+  ; CHECK-LABEL: pckod_v16i8_0:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <16 x i8>* %b
+  %2 = load <16 x i8>, <16 x i8>* %b
   ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
                      <16 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15, i32 17, i32 19, i32 21, i32 23, i32 25, i32 27, i32 29, i32 31>
-  ; CHECK-DAG: pckod.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; CHECK-DAG: pckod.b [[R3:\$w[0-9]+]], [[R2]], [[R1]]
   store <16 x i8> %3, <16 x i8>* %c
   ; CHECK-DAG: st.b [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size pckod_v16i8_0
 }
 
 define void @pckod_v8i16_0(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
-  ; CHECK: pckod_v8i16_0:
+  ; CHECK-LABEL: pckod_v8i16_0:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <8 x i16>* %b
+  %2 = load <8 x i16>, <8 x i16>* %b
   ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15>
-  ; CHECK-DAG: pckod.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; CHECK-DAG: pckod.h [[R3:\$w[0-9]+]], [[R2]], [[R1]]
   store <8 x i16> %3, <8 x i16>* %c
   ; CHECK-DAG: st.h [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size pckod_v8i16_0
 }
 
 define void @pckod_v4i32_0(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
-  ; CHECK: pckod_v4i32_0:
+  ; CHECK-LABEL: pckod_v4i32_0:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <4 x i32>* %b
+  %2 = load <4 x i32>, <4 x i32>* %b
   ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 1, i32 3, i32 5, i32 7>
-  ; CHECK-DAG: pckod.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
+  ; CHECK-DAG: pckod.w [[R3:\$w[0-9]+]], [[R2]], [[R1]]
   store <4 x i32> %3, <4 x i32>* %c
   ; CHECK-DAG: st.w [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size pckod_v4i32_0
 }
 
 define void @pckod_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
-  ; CHECK: pckod_v2i64_0:
+  ; CHECK-LABEL: pckod_v2i64_0:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
-  %2 = load <2 x i64>* %b
+  %2 = load <2 x i64>, <2 x i64>* %b
   ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
   %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 1, i32 3>
   ; pckod.d and ilvod.d are equivalent for v2i64
@@ -759,13 +1295,128 @@ define void @pckod_v2i64_0(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind
   ; CHECK-DAG: st.d [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size pckod_v2i64_0
+}
+
+define void @pckod_v16i8_1(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
+  ; CHECK-LABEL: pckod_v16i8_1:
+
+  %1 = load <16 x i8>, <16 x i8>* %a
+  %2 = load <16 x i8>, <16 x i8>* %b
+  ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
+                     <16 x i32> <i32 17, i32 19, i32 21, i32 23, i32 25, i32 27, i32 29, i32 31, i32 17, i32 19, i32 21, i32 23, i32 25, i32 27, i32 29, i32 31>
+  ; CHECK-DAG: pckod.b [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <16 x i8> %3, <16 x i8>* %c
+  ; CHECK-DAG: st.b [[R3]], 0($4)
+
+  ret void
+}
+
+define void @pckod_v8i16_1(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
+  ; CHECK-LABEL: pckod_v8i16_1:
+
+  %1 = load <8 x i16>, <8 x i16>* %a
+  %2 = load <8 x i16>, <8 x i16>* %b
+  ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 9, i32 11, i32 13, i32 15, i32 9, i32 11, i32 13, i32 15>
+  ; CHECK-DAG: pckod.h [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <8 x i16> %3, <8 x i16>* %c
+  ; CHECK-DAG: st.h [[R3]], 0($4)
+
+  ret void
+}
+
+define void @pckod_v4i32_1(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
+  ; CHECK-LABEL: pckod_v4i32_1:
+
+  %1 = load <4 x i32>, <4 x i32>* %a
+  %2 = load <4 x i32>, <4 x i32>* %b
+  ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 5, i32 7, i32 5, i32 7>
+  ; CHECK-DAG: pckod.w [[R3:\$w[0-9]+]], [[R2]], [[R2]]
+  store <4 x i32> %3, <4 x i32>* %c
+  ; CHECK-DAG: st.w [[R3]], 0($4)
+
+  ret void
+}
+
+define void @pckod_v2i64_1(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
+  ; CHECK-LABEL: pckod_v2i64_1:
+
+  %1 = load <2 x i64>, <2 x i64>* %a
+  %2 = load <2 x i64>, <2 x i64>* %b
+  ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
+  %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 3, i32 3>
+  ; pckod.d and splati.d are equivalent for v2i64
+  ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R2]][1]
+  store <2 x i64> %3, <2 x i64>* %c
+  ; CHECK-DAG: st.d [[R3]], 0($4)
+
+  ret void
+}
+
+define void @pckod_v16i8_2(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
+  ; CHECK-LABEL: pckod_v16i8_2:
+
+  %1 = load <16 x i8>, <16 x i8>* %a
+  ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <16 x i8>, <16 x i8>* %b
+  %3 = shufflevector <16 x i8> %1, <16 x i8> %2,
+                     <16 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15, i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15>
+  ; CHECK-DAG: pckod.b [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <16 x i8> %3, <16 x i8>* %c
+  ; CHECK-DAG: st.b [[R3]], 0($4)
+
+  ret void
+}
+
+define void @pckod_v8i16_2(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
+  ; CHECK-LABEL: pckod_v8i16_2:
+
+  %1 = load <8 x i16>, <8 x i16>* %a
+  ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <8 x i16>, <8 x i16>* %b
+  %3 = shufflevector <8 x i16> %1, <8 x i16> %2, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 1, i32 3, i32 5, i32 7>
+  ; CHECK-DAG: pckod.h [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <8 x i16> %3, <8 x i16>* %c
+  ; CHECK-DAG: st.h [[R3]], 0($4)
+
+  ret void
+}
+
+define void @pckod_v4i32_2(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
+  ; CHECK-LABEL: pckod_v4i32_2:
+
+  %1 = load <4 x i32>, <4 x i32>* %a
+  ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <4 x i32>, <4 x i32>* %b
+  %3 = shufflevector <4 x i32> %1, <4 x i32> %2, <4 x i32> <i32 1, i32 3, i32 1, i32 3>
+  ; CHECK-DAG: pckod.w [[R3:\$w[0-9]+]], [[R1]], [[R1]]
+  store <4 x i32> %3, <4 x i32>* %c
+  ; CHECK-DAG: st.w [[R3]], 0($4)
+
+  ret void
+}
+
+define void @pckod_v2i64_2(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
+  ; CHECK-LABEL: pckod_v2i64_2:
+
+  %1 = load <2 x i64>, <2 x i64>* %a
+  ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
+  %2 = load <2 x i64>, <2 x i64>* %b
+  %3 = shufflevector <2 x i64> %1, <2 x i64> %2, <2 x i32> <i32 1, i32 1>
+  ; pckod.d and splati.d are equivalent for v2i64
+  ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][1]
+  store <2 x i64> %3, <2 x i64>* %c
+  ; CHECK-DAG: st.d [[R3]], 0($4)
+
+  ret void
 }
 
 define void @splati_v16i8_0(<16 x i8>* %c, <16 x i8>* %a) nounwind {
-  ; CHECK: splati_v16i8_0:
+  ; CHECK-LABEL: splati_v16i8_0:
 
-  %1 = load <16 x i8>* %a
+  %1 = load <16 x i8>, <16 x i8>* %a
   ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <16 x i8> %1, <16 x i8> undef,
                      <16 x i32> <i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4>
@@ -774,13 +1425,12 @@ define void @splati_v16i8_0(<16 x i8>* %c, <16 x i8>* %a) nounwind {
   ; CHECK-DAG: st.b [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size splati_v16i8_0
 }
 
 define void @splati_v8i16_0(<8 x i16>* %c, <8 x i16>* %a) nounwind {
-  ; CHECK: splati_v8i16_0:
+  ; CHECK-LABEL: splati_v8i16_0:
 
-  %1 = load <8 x i16>* %a
+  %1 = load <8 x i16>, <8 x i16>* %a
   ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <8 x i16> %1, <8 x i16> undef, <8 x i32> <i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4>
   ; CHECK-DAG: splati.h [[R3:\$w[0-9]+]], [[R1]][4]
@@ -788,28 +1438,25 @@ define void @splati_v8i16_0(<8 x i16>* %c, <8 x i16>* %a) nounwind {
   ; CHECK-DAG: st.h [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size splati_v8i16_0
 }
 
 define void @splati_v4i32_0(<4 x i32>* %c, <4 x i32>* %a) nounwind {
-  ; CHECK: splati_v4i32_0:
+  ; CHECK-LABEL: splati_v4i32_0:
 
-  %1 = load <4 x i32>* %a
+  %1 = load <4 x i32>, <4 x i32>* %a
   ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 3, i32 3, i32 3, i32 3>
-  ; shf.w and splati.w are equivalent
-  ; CHECK-DAG: shf.w [[R3:\$w[0-9]+]], [[R1]], 255
+  ; CHECK-DAG: splati.w [[R3:\$w[0-9]+]], [[R1]][3]
   store <4 x i32> %2, <4 x i32>* %c
   ; CHECK-DAG: st.w [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size splati_v4i32_0
 }
 
 define void @splati_v2i64_0(<2 x i64>* %c, <2 x i64>* %a) nounwind {
-  ; CHECK: splati_v2i64_0:
+  ; CHECK-LABEL: splati_v2i64_0:
 
-  %1 = load <2 x i64>* %a
+  %1 = load <2 x i64>, <2 x i64>* %a
   ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
   %2 = shufflevector <2 x i64> %1, <2 x i64> undef, <2 x i32> <i32 1, i32 1>
   ; CHECK-DAG: splati.d [[R3:\$w[0-9]+]], [[R1]][1]
@@ -817,5 +1464,4 @@ define void @splati_v2i64_0(<2 x i64>* %c, <2 x i64>* %a) nounwind {
   ; CHECK-DAG: st.d [[R3]], 0($4)
 
   ret void
-  ; CHECK: .size splati_v2i64_0
 }
diff --git a/test/CodeGen/Mips/msa/spill.ll b/test/CodeGen/Mips/msa/spill.ll
index 66f896ac4684..8c9a79912351 100644
--- a/test/CodeGen/Mips/msa/spill.ll
+++ b/test/CodeGen/Mips/msa/spill.ll
@@ -6,73 +6,73 @@
 
 define i32 @test_i8(<16 x i8>* %p0, <16 x i8>* %q1) nounwind {
 entry:
-  %p1  = getelementptr <16 x i8>* %p0, i32 1
-  %p2  = getelementptr <16 x i8>* %p0, i32 2
-  %p3  = getelementptr <16 x i8>* %p0, i32 3
-  %p4  = getelementptr <16 x i8>* %p0, i32 4
-  %p5  = getelementptr <16 x i8>* %p0, i32 5
-  %p6  = getelementptr <16 x i8>* %p0, i32 6
-  %p7  = getelementptr <16 x i8>* %p0, i32 7
-  %p8  = getelementptr <16 x i8>* %p0, i32 8
-  %p9  = getelementptr <16 x i8>* %p0, i32 9
-  %p10 = getelementptr <16 x i8>* %p0, i32 10
-  %p11 = getelementptr <16 x i8>* %p0, i32 11
-  %p12 = getelementptr <16 x i8>* %p0, i32 12
-  %p13 = getelementptr <16 x i8>* %p0, i32 13
-  %p14 = getelementptr <16 x i8>* %p0, i32 14
-  %p15 = getelementptr <16 x i8>* %p0, i32 15
-  %p16 = getelementptr <16 x i8>* %p0, i32 16
-  %p17 = getelementptr <16 x i8>* %p0, i32 17
-  %p18 = getelementptr <16 x i8>* %p0, i32 18
-  %p19 = getelementptr <16 x i8>* %p0, i32 19
-  %p20 = getelementptr <16 x i8>* %p0, i32 20
-  %p21 = getelementptr <16 x i8>* %p0, i32 21
-  %p22 = getelementptr <16 x i8>* %p0, i32 22
-  %p23 = getelementptr <16 x i8>* %p0, i32 23
-  %p24 = getelementptr <16 x i8>* %p0, i32 24
-  %p25 = getelementptr <16 x i8>* %p0, i32 25
-  %p26 = getelementptr <16 x i8>* %p0, i32 26
-  %p27 = getelementptr <16 x i8>* %p0, i32 27
-  %p28 = getelementptr <16 x i8>* %p0, i32 28
-  %p29 = getelementptr <16 x i8>* %p0, i32 29
-  %p30 = getelementptr <16 x i8>* %p0, i32 30
-  %p31 = getelementptr <16 x i8>* %p0, i32 31
-  %p32 = getelementptr <16 x i8>* %p0, i32 32
-  %p33 = getelementptr <16 x i8>* %p0, i32 33
-  %0  = load <16 x i8>* %p0, align 16
-  %1  = load <16 x i8>* %p1, align 16
-  %2  = load <16 x i8>* %p2, align 16
-  %3  = load <16 x i8>* %p3, align 16
-  %4  = load <16 x i8>* %p4, align 16
-  %5  = load <16 x i8>* %p5, align 16
-  %6  = load <16 x i8>* %p6, align 16
-  %7  = load <16 x i8>* %p7, align 16
-  %8  = load <16 x i8>* %p8, align 16
-  %9  = load <16 x i8>* %p9, align 16
-  %10 = load <16 x i8>* %p10, align 16
-  %11 = load <16 x i8>* %p11, align 16
-  %12 = load <16 x i8>* %p12, align 16
-  %13 = load <16 x i8>* %p13, align 16
-  %14 = load <16 x i8>* %p14, align 16
-  %15 = load <16 x i8>* %p15, align 16
-  %16 = load <16 x i8>* %p16, align 16
-  %17 = load <16 x i8>* %p17, align 16
-  %18 = load <16 x i8>* %p18, align 16
-  %19 = load <16 x i8>* %p19, align 16
-  %20 = load <16 x i8>* %p20, align 16
-  %21 = load <16 x i8>* %p21, align 16
-  %22 = load <16 x i8>* %p22, align 16
-  %23 = load <16 x i8>* %p23, align 16
-  %24 = load <16 x i8>* %p24, align 16
-  %25 = load <16 x i8>* %p25, align 16
-  %26 = load <16 x i8>* %p26, align 16
-  %27 = load <16 x i8>* %p27, align 16
-  %28 = load <16 x i8>* %p28, align 16
-  %29 = load <16 x i8>* %p29, align 16
-  %30 = load <16 x i8>* %p30, align 16
-  %31 = load <16 x i8>* %p31, align 16
-  %32 = load <16 x i8>* %p32, align 16
-  %33 = load <16 x i8>* %p33, align 16
+  %p1  = getelementptr <16 x i8>, <16 x i8>* %p0, i32 1
+  %p2  = getelementptr <16 x i8>, <16 x i8>* %p0, i32 2
+  %p3  = getelementptr <16 x i8>, <16 x i8>* %p0, i32 3
+  %p4  = getelementptr <16 x i8>, <16 x i8>* %p0, i32 4
+  %p5  = getelementptr <16 x i8>, <16 x i8>* %p0, i32 5
+  %p6  = getelementptr <16 x i8>, <16 x i8>* %p0, i32 6
+  %p7  = getelementptr <16 x i8>, <16 x i8>* %p0, i32 7
+  %p8  = getelementptr <16 x i8>, <16 x i8>* %p0, i32 8
+  %p9  = getelementptr <16 x i8>, <16 x i8>* %p0, i32 9
+  %p10 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 10
+  %p11 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 11
+  %p12 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 12
+  %p13 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 13
+  %p14 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 14
+  %p15 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 15
+  %p16 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 16
+  %p17 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 17
+  %p18 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 18
+  %p19 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 19
+  %p20 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 20
+  %p21 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 21
+  %p22 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 22
+  %p23 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 23
+  %p24 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 24
+  %p25 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 25
+  %p26 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 26
+  %p27 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 27
+  %p28 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 28
+  %p29 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 29
+  %p30 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 30
+  %p31 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 31
+  %p32 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 32
+  %p33 = getelementptr <16 x i8>, <16 x i8>* %p0, i32 33
+  %0  = load <16 x i8>, <16 x i8>* %p0, align 16
+  %1  = load <16 x i8>, <16 x i8>* %p1, align 16
+  %2  = load <16 x i8>, <16 x i8>* %p2, align 16
+  %3  = load <16 x i8>, <16 x i8>* %p3, align 16
+  %4  = load <16 x i8>, <16 x i8>* %p4, align 16
+  %5  = load <16 x i8>, <16 x i8>* %p5, align 16
+  %6  = load <16 x i8>, <16 x i8>* %p6, align 16
+  %7  = load <16 x i8>, <16 x i8>* %p7, align 16
+  %8  = load <16 x i8>, <16 x i8>* %p8, align 16
+  %9  = load <16 x i8>, <16 x i8>* %p9, align 16
+  %10 = load <16 x i8>, <16 x i8>* %p10, align 16
+  %11 = load <16 x i8>, <16 x i8>* %p11, align 16
+  %12 = load <16 x i8>, <16 x i8>* %p12, align 16
+  %13 = load <16 x i8>, <16 x i8>* %p13, align 16
+  %14 = load <16 x i8>, <16 x i8>* %p14, align 16
+  %15 = load <16 x i8>, <16 x i8>* %p15, align 16
+  %16 = load <16 x i8>, <16 x i8>* %p16, align 16
+  %17 = load <16 x i8>, <16 x i8>* %p17, align 16
+  %18 = load <16 x i8>, <16 x i8>* %p18, align 16
+  %19 = load <16 x i8>, <16 x i8>* %p19, align 16
+  %20 = load <16 x i8>, <16 x i8>* %p20, align 16
+  %21 = load <16 x i8>, <16 x i8>* %p21, align 16
+  %22 = load <16 x i8>, <16 x i8>* %p22, align 16
+  %23 = load <16 x i8>, <16 x i8>* %p23, align 16
+  %24 = load <16 x i8>, <16 x i8>* %p24, align 16
+  %25 = load <16 x i8>, <16 x i8>* %p25, align 16
+  %26 = load <16 x i8>, <16 x i8>* %p26, align 16
+  %27 = load <16 x i8>, <16 x i8>* %p27, align 16
+  %28 = load <16 x i8>, <16 x i8>* %p28, align 16
+  %29 = load <16 x i8>, <16 x i8>* %p29, align 16
+  %30 = load <16 x i8>, <16 x i8>* %p30, align 16
+  %31 = load <16 x i8>, <16 x i8>* %p31, align 16
+  %32 = load <16 x i8>, <16 x i8>* %p32, align 16
+  %33 = load <16 x i8>, <16 x i8>* %p33, align 16
   %r1  = call <16 x i8> @llvm.mips.addv.b(<16 x i8> %0,   <16 x i8> %1)
   %r2  = call <16 x i8> @llvm.mips.addv.b(<16 x i8> %r1,  <16 x i8> %2)
   %r3  = call <16 x i8> @llvm.mips.addv.b(<16 x i8> %r2,  <16 x i8> %3)
@@ -155,73 +155,73 @@ declare i32       @llvm.mips.copy.s.b(<16 x i8>, i32) nounwind
 
 define i32 @test_i16(<8 x i16>* %p0, <8 x i16>* %q1) nounwind {
 entry:
-  %p1  = getelementptr <8 x i16>* %p0, i32 1
-  %p2  = getelementptr <8 x i16>* %p0, i32 2
-  %p3  = getelementptr <8 x i16>* %p0, i32 3
-  %p4  = getelementptr <8 x i16>* %p0, i32 4
-  %p5  = getelementptr <8 x i16>* %p0, i32 5
-  %p6  = getelementptr <8 x i16>* %p0, i32 6
-  %p7  = getelementptr <8 x i16>* %p0, i32 7
-  %p8  = getelementptr <8 x i16>* %p0, i32 8
-  %p9  = getelementptr <8 x i16>* %p0, i32 9
-  %p10 = getelementptr <8 x i16>* %p0, i32 10
-  %p11 = getelementptr <8 x i16>* %p0, i32 11
-  %p12 = getelementptr <8 x i16>* %p0, i32 12
-  %p13 = getelementptr <8 x i16>* %p0, i32 13
-  %p14 = getelementptr <8 x i16>* %p0, i32 14
-  %p15 = getelementptr <8 x i16>* %p0, i32 15
-  %p16 = getelementptr <8 x i16>* %p0, i32 16
-  %p17 = getelementptr <8 x i16>* %p0, i32 17
-  %p18 = getelementptr <8 x i16>* %p0, i32 18
-  %p19 = getelementptr <8 x i16>* %p0, i32 19
-  %p20 = getelementptr <8 x i16>* %p0, i32 20
-  %p21 = getelementptr <8 x i16>* %p0, i32 21
-  %p22 = getelementptr <8 x i16>* %p0, i32 22
-  %p23 = getelementptr <8 x i16>* %p0, i32 23
-  %p24 = getelementptr <8 x i16>* %p0, i32 24
-  %p25 = getelementptr <8 x i16>* %p0, i32 25
-  %p26 = getelementptr <8 x i16>* %p0, i32 26
-  %p27 = getelementptr <8 x i16>* %p0, i32 27
-  %p28 = getelementptr <8 x i16>* %p0, i32 28
-  %p29 = getelementptr <8 x i16>* %p0, i32 29
-  %p30 = getelementptr <8 x i16>* %p0, i32 30
-  %p31 = getelementptr <8 x i16>* %p0, i32 31
-  %p32 = getelementptr <8 x i16>* %p0, i32 32
-  %p33 = getelementptr <8 x i16>* %p0, i32 33
-  %0  = load <8 x i16>* %p0, align 16
-  %1  = load <8 x i16>* %p1, align 16
-  %2  = load <8 x i16>* %p2, align 16
-  %3  = load <8 x i16>* %p3, align 16
-  %4  = load <8 x i16>* %p4, align 16
-  %5  = load <8 x i16>* %p5, align 16
-  %6  = load <8 x i16>* %p6, align 16
-  %7  = load <8 x i16>* %p7, align 16
-  %8  = load <8 x i16>* %p8, align 16
-  %9  = load <8 x i16>* %p9, align 16
-  %10 = load <8 x i16>* %p10, align 16
-  %11 = load <8 x i16>* %p11, align 16
-  %12 = load <8 x i16>* %p12, align 16
-  %13 = load <8 x i16>* %p13, align 16
-  %14 = load <8 x i16>* %p14, align 16
-  %15 = load <8 x i16>* %p15, align 16
-  %16 = load <8 x i16>* %p16, align 16
-  %17 = load <8 x i16>* %p17, align 16
-  %18 = load <8 x i16>* %p18, align 16
-  %19 = load <8 x i16>* %p19, align 16
-  %20 = load <8 x i16>* %p20, align 16
-  %21 = load <8 x i16>* %p21, align 16
-  %22 = load <8 x i16>* %p22, align 16
-  %23 = load <8 x i16>* %p23, align 16
-  %24 = load <8 x i16>* %p24, align 16
-  %25 = load <8 x i16>* %p25, align 16
-  %26 = load <8 x i16>* %p26, align 16
-  %27 = load <8 x i16>* %p27, align 16
-  %28 = load <8 x i16>* %p28, align 16
-  %29 = load <8 x i16>* %p29, align 16
-  %30 = load <8 x i16>* %p30, align 16
-  %31 = load <8 x i16>* %p31, align 16
-  %32 = load <8 x i16>* %p32, align 16
-  %33 = load <8 x i16>* %p33, align 16
+  %p1  = getelementptr <8 x i16>, <8 x i16>* %p0, i32 1
+  %p2  = getelementptr <8 x i16>, <8 x i16>* %p0, i32 2
+  %p3  = getelementptr <8 x i16>, <8 x i16>* %p0, i32 3
+  %p4  = getelementptr <8 x i16>, <8 x i16>* %p0, i32 4
+  %p5  = getelementptr <8 x i16>, <8 x i16>* %p0, i32 5
+  %p6  = getelementptr <8 x i16>, <8 x i16>* %p0, i32 6
+  %p7  = getelementptr <8 x i16>, <8 x i16>* %p0, i32 7
+  %p8  = getelementptr <8 x i16>, <8 x i16>* %p0, i32 8
+  %p9  = getelementptr <8 x i16>, <8 x i16>* %p0, i32 9
+  %p10 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 10
+  %p11 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 11
+  %p12 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 12
+  %p13 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 13
+  %p14 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 14
+  %p15 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 15
+  %p16 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 16
+  %p17 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 17
+  %p18 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 18
+  %p19 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 19
+  %p20 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 20
+  %p21 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 21
+  %p22 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 22
+  %p23 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 23
+  %p24 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 24
+  %p25 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 25
+  %p26 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 26
+  %p27 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 27
+  %p28 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 28
+  %p29 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 29
+  %p30 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 30
+  %p31 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 31
+  %p32 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 32
+  %p33 = getelementptr <8 x i16>, <8 x i16>* %p0, i32 33
+  %0  = load <8 x i16>, <8 x i16>* %p0, align 16
+  %1  = load <8 x i16>, <8 x i16>* %p1, align 16
+  %2  = load <8 x i16>, <8 x i16>* %p2, align 16
+  %3  = load <8 x i16>, <8 x i16>* %p3, align 16
+  %4  = load <8 x i16>, <8 x i16>* %p4, align 16
+  %5  = load <8 x i16>, <8 x i16>* %p5, align 16
+  %6  = load <8 x i16>, <8 x i16>* %p6, align 16
+  %7  = load <8 x i16>, <8 x i16>* %p7, align 16
+  %8  = load <8 x i16>, <8 x i16>* %p8, align 16
+  %9  = load <8 x i16>, <8 x i16>* %p9, align 16
+  %10 = load <8 x i16>, <8 x i16>* %p10, align 16
+  %11 = load <8 x i16>, <8 x i16>* %p11, align 16
+  %12 = load <8 x i16>, <8 x i16>* %p12, align 16
+  %13 = load <8 x i16>, <8 x i16>* %p13, align 16
+  %14 = load <8 x i16>, <8 x i16>* %p14, align 16
+  %15 = load <8 x i16>, <8 x i16>* %p15, align 16
+  %16 = load <8 x i16>, <8 x i16>* %p16, align 16
+  %17 = load <8 x i16>, <8 x i16>* %p17, align 16
+  %18 = load <8 x i16>, <8 x i16>* %p18, align 16
+  %19 = load <8 x i16>, <8 x i16>* %p19, align 16
+  %20 = load <8 x i16>, <8 x i16>* %p20, align 16
+  %21 = load <8 x i16>, <8 x i16>* %p21, align 16
+  %22 = load <8 x i16>, <8 x i16>* %p22, align 16
+  %23 = load <8 x i16>, <8 x i16>* %p23, align 16
+  %24 = load <8 x i16>, <8 x i16>* %p24, align 16
+  %25 = load <8 x i16>, <8 x i16>* %p25, align 16
+  %26 = load <8 x i16>, <8 x i16>* %p26, align 16
+  %27 = load <8 x i16>, <8 x i16>* %p27, align 16
+  %28 = load <8 x i16>, <8 x i16>* %p28, align 16
+  %29 = load <8 x i16>, <8 x i16>* %p29, align 16
+  %30 = load <8 x i16>, <8 x i16>* %p30, align 16
+  %31 = load <8 x i16>, <8 x i16>* %p31, align 16
+  %32 = load <8 x i16>, <8 x i16>* %p32, align 16
+  %33 = load <8 x i16>, <8 x i16>* %p33, align 16
   %r1  = call <8 x i16> @llvm.mips.addv.h(<8 x i16> %0,   <8 x i16> %1)
   %r2  = call <8 x i16> @llvm.mips.addv.h(<8 x i16> %r1,  <8 x i16> %2)
   %r3  = call <8 x i16> @llvm.mips.addv.h(<8 x i16> %r2,  <8 x i16> %3)
@@ -304,73 +304,73 @@ declare i32       @llvm.mips.copy.s.h(<8 x i16>, i32) nounwind
 
 define i32 @test_i32(<4 x i32>* %p0, <4 x i32>* %q1) nounwind {
 entry:
-  %p1  = getelementptr <4 x i32>* %p0, i32 1
-  %p2  = getelementptr <4 x i32>* %p0, i32 2
-  %p3  = getelementptr <4 x i32>* %p0, i32 3
-  %p4  = getelementptr <4 x i32>* %p0, i32 4
-  %p5  = getelementptr <4 x i32>* %p0, i32 5
-  %p6  = getelementptr <4 x i32>* %p0, i32 6
-  %p7  = getelementptr <4 x i32>* %p0, i32 7
-  %p8  = getelementptr <4 x i32>* %p0, i32 8
-  %p9  = getelementptr <4 x i32>* %p0, i32 9
-  %p10 = getelementptr <4 x i32>* %p0, i32 10
-  %p11 = getelementptr <4 x i32>* %p0, i32 11
-  %p12 = getelementptr <4 x i32>* %p0, i32 12
-  %p13 = getelementptr <4 x i32>* %p0, i32 13
-  %p14 = getelementptr <4 x i32>* %p0, i32 14
-  %p15 = getelementptr <4 x i32>* %p0, i32 15
-  %p16 = getelementptr <4 x i32>* %p0, i32 16
-  %p17 = getelementptr <4 x i32>* %p0, i32 17
-  %p18 = getelementptr <4 x i32>* %p0, i32 18
-  %p19 = getelementptr <4 x i32>* %p0, i32 19
-  %p20 = getelementptr <4 x i32>* %p0, i32 20
-  %p21 = getelementptr <4 x i32>* %p0, i32 21
-  %p22 = getelementptr <4 x i32>* %p0, i32 22
-  %p23 = getelementptr <4 x i32>* %p0, i32 23
-  %p24 = getelementptr <4 x i32>* %p0, i32 24
-  %p25 = getelementptr <4 x i32>* %p0, i32 25
-  %p26 = getelementptr <4 x i32>* %p0, i32 26
-  %p27 = getelementptr <4 x i32>* %p0, i32 27
-  %p28 = getelementptr <4 x i32>* %p0, i32 28
-  %p29 = getelementptr <4 x i32>* %p0, i32 29
-  %p30 = getelementptr <4 x i32>* %p0, i32 30
-  %p31 = getelementptr <4 x i32>* %p0, i32 31
-  %p32 = getelementptr <4 x i32>* %p0, i32 32
-  %p33 = getelementptr <4 x i32>* %p0, i32 33
-  %0  = load <4 x i32>* %p0, align 16
-  %1  = load <4 x i32>* %p1, align 16
-  %2  = load <4 x i32>* %p2, align 16
-  %3  = load <4 x i32>* %p3, align 16
-  %4  = load <4 x i32>* %p4, align 16
-  %5  = load <4 x i32>* %p5, align 16
-  %6  = load <4 x i32>* %p6, align 16
-  %7  = load <4 x i32>* %p7, align 16
-  %8  = load <4 x i32>* %p8, align 16
-  %9  = load <4 x i32>* %p9, align 16
-  %10 = load <4 x i32>* %p10, align 16
-  %11 = load <4 x i32>* %p11, align 16
-  %12 = load <4 x i32>* %p12, align 16
-  %13 = load <4 x i32>* %p13, align 16
-  %14 = load <4 x i32>* %p14, align 16
-  %15 = load <4 x i32>* %p15, align 16
-  %16 = load <4 x i32>* %p16, align 16
-  %17 = load <4 x i32>* %p17, align 16
-  %18 = load <4 x i32>* %p18, align 16
-  %19 = load <4 x i32>* %p19, align 16
-  %20 = load <4 x i32>* %p20, align 16
-  %21 = load <4 x i32>* %p21, align 16
-  %22 = load <4 x i32>* %p22, align 16
-  %23 = load <4 x i32>* %p23, align 16
-  %24 = load <4 x i32>* %p24, align 16
-  %25 = load <4 x i32>* %p25, align 16
-  %26 = load <4 x i32>* %p26, align 16
-  %27 = load <4 x i32>* %p27, align 16
-  %28 = load <4 x i32>* %p28, align 16
-  %29 = load <4 x i32>* %p29, align 16
-  %30 = load <4 x i32>* %p30, align 16
-  %31 = load <4 x i32>* %p31, align 16
-  %32 = load <4 x i32>* %p32, align 16
-  %33 = load <4 x i32>* %p33, align 16
+  %p1  = getelementptr <4 x i32>, <4 x i32>* %p0, i32 1
+  %p2  = getelementptr <4 x i32>, <4 x i32>* %p0, i32 2
+  %p3  = getelementptr <4 x i32>, <4 x i32>* %p0, i32 3
+  %p4  = getelementptr <4 x i32>, <4 x i32>* %p0, i32 4
+  %p5  = getelementptr <4 x i32>, <4 x i32>* %p0, i32 5
+  %p6  = getelementptr <4 x i32>, <4 x i32>* %p0, i32 6
+  %p7  = getelementptr <4 x i32>, <4 x i32>* %p0, i32 7
+  %p8  = getelementptr <4 x i32>, <4 x i32>* %p0, i32 8
+  %p9  = getelementptr <4 x i32>, <4 x i32>* %p0, i32 9
+  %p10 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 10
+  %p11 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 11
+  %p12 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 12
+  %p13 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 13
+  %p14 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 14
+  %p15 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 15
+  %p16 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 16
+  %p17 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 17
+  %p18 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 18
+  %p19 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 19
+  %p20 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 20
+  %p21 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 21
+  %p22 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 22
+  %p23 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 23
+  %p24 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 24
+  %p25 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 25
+  %p26 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 26
+  %p27 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 27
+  %p28 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 28
+  %p29 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 29
+  %p30 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 30
+  %p31 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 31
+  %p32 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 32
+  %p33 = getelementptr <4 x i32>, <4 x i32>* %p0, i32 33
+  %0  = load <4 x i32>, <4 x i32>* %p0, align 16
+  %1  = load <4 x i32>, <4 x i32>* %p1, align 16
+  %2  = load <4 x i32>, <4 x i32>* %p2, align 16
+  %3  = load <4 x i32>, <4 x i32>* %p3, align 16
+  %4  = load <4 x i32>, <4 x i32>* %p4, align 16
+  %5  = load <4 x i32>, <4 x i32>* %p5, align 16
+  %6  = load <4 x i32>, <4 x i32>* %p6, align 16
+  %7  = load <4 x i32>, <4 x i32>* %p7, align 16
+  %8  = load <4 x i32>, <4 x i32>* %p8, align 16
+  %9  = load <4 x i32>, <4 x i32>* %p9, align 16
+  %10 = load <4 x i32>, <4 x i32>* %p10, align 16
+  %11 = load <4 x i32>, <4 x i32>* %p11, align 16
+  %12 = load <4 x i32>, <4 x i32>* %p12, align 16
+  %13 = load <4 x i32>, <4 x i32>* %p13, align 16
+  %14 = load <4 x i32>, <4 x i32>* %p14, align 16
+  %15 = load <4 x i32>, <4 x i32>* %p15, align 16
+  %16 = load <4 x i32>, <4 x i32>* %p16, align 16
+  %17 = load <4 x i32>, <4 x i32>* %p17, align 16
+  %18 = load <4 x i32>, <4 x i32>* %p18, align 16
+  %19 = load <4 x i32>, <4 x i32>* %p19, align 16
+  %20 = load <4 x i32>, <4 x i32>* %p20, align 16
+  %21 = load <4 x i32>, <4 x i32>* %p21, align 16
+  %22 = load <4 x i32>, <4 x i32>* %p22, align 16
+  %23 = load <4 x i32>, <4 x i32>* %p23, align 16
+  %24 = load <4 x i32>, <4 x i32>* %p24, align 16
+  %25 = load <4 x i32>, <4 x i32>* %p25, align 16
+  %26 = load <4 x i32>, <4 x i32>* %p26, align 16
+  %27 = load <4 x i32>, <4 x i32>* %p27, align 16
+  %28 = load <4 x i32>, <4 x i32>* %p28, align 16
+  %29 = load <4 x i32>, <4 x i32>* %p29, align 16
+  %30 = load <4 x i32>, <4 x i32>* %p30, align 16
+  %31 = load <4 x i32>, <4 x i32>* %p31, align 16
+  %32 = load <4 x i32>, <4 x i32>* %p32, align 16
+  %33 = load <4 x i32>, <4 x i32>* %p33, align 16
   %r1 = call <4 x i32> @llvm.mips.addv.w(<4 x i32> %0, <4 x i32> %1)
   %r2 = call <4 x i32> @llvm.mips.addv.w(<4 x i32> %r1, <4 x i32> %2)
   %r3 = call <4 x i32> @llvm.mips.addv.w(<4 x i32> %r2, <4 x i32> %3)
@@ -453,73 +453,73 @@ declare i32       @llvm.mips.copy.s.w(<4 x i32>, i32) nounwind
 
 define i32 @test_i64(<2 x i64>* %p0, <2 x i64>* %q1) nounwind {
 entry:
-  %p1  = getelementptr <2 x i64>* %p0, i32 1
-  %p2  = getelementptr <2 x i64>* %p0, i32 2
-  %p3  = getelementptr <2 x i64>* %p0, i32 3
-  %p4  = getelementptr <2 x i64>* %p0, i32 4
-  %p5  = getelementptr <2 x i64>* %p0, i32 5
-  %p6  = getelementptr <2 x i64>* %p0, i32 6
-  %p7  = getelementptr <2 x i64>* %p0, i32 7
-  %p8  = getelementptr <2 x i64>* %p0, i32 8
-  %p9  = getelementptr <2 x i64>* %p0, i32 9
-  %p10 = getelementptr <2 x i64>* %p0, i32 10
-  %p11 = getelementptr <2 x i64>* %p0, i32 11
-  %p12 = getelementptr <2 x i64>* %p0, i32 12
-  %p13 = getelementptr <2 x i64>* %p0, i32 13
-  %p14 = getelementptr <2 x i64>* %p0, i32 14
-  %p15 = getelementptr <2 x i64>* %p0, i32 15
-  %p16 = getelementptr <2 x i64>* %p0, i32 16
-  %p17 = getelementptr <2 x i64>* %p0, i32 17
-  %p18 = getelementptr <2 x i64>* %p0, i32 18
-  %p19 = getelementptr <2 x i64>* %p0, i32 19
-  %p20 = getelementptr <2 x i64>* %p0, i32 20
-  %p21 = getelementptr <2 x i64>* %p0, i32 21
-  %p22 = getelementptr <2 x i64>* %p0, i32 22
-  %p23 = getelementptr <2 x i64>* %p0, i32 23
-  %p24 = getelementptr <2 x i64>* %p0, i32 24
-  %p25 = getelementptr <2 x i64>* %p0, i32 25
-  %p26 = getelementptr <2 x i64>* %p0, i32 26
-  %p27 = getelementptr <2 x i64>* %p0, i32 27
-  %p28 = getelementptr <2 x i64>* %p0, i32 28
-  %p29 = getelementptr <2 x i64>* %p0, i32 29
-  %p30 = getelementptr <2 x i64>* %p0, i32 30
-  %p31 = getelementptr <2 x i64>* %p0, i32 31
-  %p32 = getelementptr <2 x i64>* %p0, i32 32
-  %p33 = getelementptr <2 x i64>* %p0, i32 33
-  %0  = load <2 x i64>* %p0, align 16
-  %1  = load <2 x i64>* %p1, align 16
-  %2  = load <2 x i64>* %p2, align 16
-  %3  = load <2 x i64>* %p3, align 16
-  %4  = load <2 x i64>* %p4, align 16
-  %5  = load <2 x i64>* %p5, align 16
-  %6  = load <2 x i64>* %p6, align 16
-  %7  = load <2 x i64>* %p7, align 16
-  %8  = load <2 x i64>* %p8, align 16
-  %9  = load <2 x i64>* %p9, align 16
-  %10 = load <2 x i64>* %p10, align 16
-  %11 = load <2 x i64>* %p11, align 16
-  %12 = load <2 x i64>* %p12, align 16
-  %13 = load <2 x i64>* %p13, align 16
-  %14 = load <2 x i64>* %p14, align 16
-  %15 = load <2 x i64>* %p15, align 16
-  %16 = load <2 x i64>* %p16, align 16
-  %17 = load <2 x i64>* %p17, align 16
-  %18 = load <2 x i64>* %p18, align 16
-  %19 = load <2 x i64>* %p19, align 16
-  %20 = load <2 x i64>* %p20, align 16
-  %21 = load <2 x i64>* %p21, align 16
-  %22 = load <2 x i64>* %p22, align 16
-  %23 = load <2 x i64>* %p23, align 16
-  %24 = load <2 x i64>* %p24, align 16
-  %25 = load <2 x i64>* %p25, align 16
-  %26 = load <2 x i64>* %p26, align 16
-  %27 = load <2 x i64>* %p27, align 16
-  %28 = load <2 x i64>* %p28, align 16
-  %29 = load <2 x i64>* %p29, align 16
-  %30 = load <2 x i64>* %p30, align 16
-  %31 = load <2 x i64>* %p31, align 16
-  %32 = load <2 x i64>* %p32, align 16
-  %33 = load <2 x i64>* %p33, align 16
+  %p1  = getelementptr <2 x i64>, <2 x i64>* %p0, i32 1
+  %p2  = getelementptr <2 x i64>, <2 x i64>* %p0, i32 2
+  %p3  = getelementptr <2 x i64>, <2 x i64>* %p0, i32 3
+  %p4  = getelementptr <2 x i64>, <2 x i64>* %p0, i32 4
+  %p5  = getelementptr <2 x i64>, <2 x i64>* %p0, i32 5
+  %p6  = getelementptr <2 x i64>, <2 x i64>* %p0, i32 6
+  %p7  = getelementptr <2 x i64>, <2 x i64>* %p0, i32 7
+  %p8  = getelementptr <2 x i64>, <2 x i64>* %p0, i32 8
+  %p9  = getelementptr <2 x i64>, <2 x i64>* %p0, i32 9
+  %p10 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 10
+  %p11 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 11
+  %p12 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 12
+  %p13 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 13
+  %p14 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 14
+  %p15 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 15
+  %p16 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 16
+  %p17 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 17
+  %p18 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 18
+  %p19 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 19
+  %p20 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 20
+  %p21 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 21
+  %p22 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 22
+  %p23 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 23
+  %p24 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 24
+  %p25 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 25
+  %p26 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 26
+  %p27 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 27
+  %p28 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 28
+  %p29 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 29
+  %p30 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 30
+  %p31 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 31
+  %p32 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 32
+  %p33 = getelementptr <2 x i64>, <2 x i64>* %p0, i32 33
+  %0  = load <2 x i64>, <2 x i64>* %p0, align 16
+  %1  = load <2 x i64>, <2 x i64>* %p1, align 16
+  %2  = load <2 x i64>, <2 x i64>* %p2, align 16
+  %3  = load <2 x i64>, <2 x i64>* %p3, align 16
+  %4  = load <2 x i64>, <2 x i64>* %p4, align 16
+  %5  = load <2 x i64>, <2 x i64>* %p5, align 16
+  %6  = load <2 x i64>, <2 x i64>* %p6, align 16
+  %7  = load <2 x i64>, <2 x i64>* %p7, align 16
+  %8  = load <2 x i64>, <2 x i64>* %p8, align 16
+  %9  = load <2 x i64>, <2 x i64>* %p9, align 16
+  %10 = load <2 x i64>, <2 x i64>* %p10, align 16
+  %11 = load <2 x i64>, <2 x i64>* %p11, align 16
+  %12 = load <2 x i64>, <2 x i64>* %p12, align 16
+  %13 = load <2 x i64>, <2 x i64>* %p13, align 16
+  %14 = load <2 x i64>, <2 x i64>* %p14, align 16
+  %15 = load <2 x i64>, <2 x i64>* %p15, align 16
+  %16 = load <2 x i64>, <2 x i64>* %p16, align 16
+  %17 = load <2 x i64>, <2 x i64>* %p17, align 16
+  %18 = load <2 x i64>, <2 x i64>* %p18, align 16
+  %19 = load <2 x i64>, <2 x i64>* %p19, align 16
+  %20 = load <2 x i64>, <2 x i64>* %p20, align 16
+  %21 = load <2 x i64>, <2 x i64>* %p21, align 16
+  %22 = load <2 x i64>, <2 x i64>* %p22, align 16
+  %23 = load <2 x i64>, <2 x i64>* %p23, align 16
+  %24 = load <2 x i64>, <2 x i64>* %p24, align 16
+  %25 = load <2 x i64>, <2 x i64>* %p25, align 16
+  %26 = load <2 x i64>, <2 x i64>* %p26, align 16
+  %27 = load <2 x i64>, <2 x i64>* %p27, align 16
+  %28 = load <2 x i64>, <2 x i64>* %p28, align 16
+  %29 = load <2 x i64>, <2 x i64>* %p29, align 16
+  %30 = load <2 x i64>, <2 x i64>* %p30, align 16
+  %31 = load <2 x i64>, <2 x i64>* %p31, align 16
+  %32 = load <2 x i64>, <2 x i64>* %p32, align 16
+  %33 = load <2 x i64>, <2 x i64>* %p33, align 16
   %r1  = call <2 x i64> @llvm.mips.addv.d(<2 x i64> %0,   <2 x i64> %1)
   %r2  = call <2 x i64> @llvm.mips.addv.d(<2 x i64> %r1,  <2 x i64> %2)
   %r3  = call <2 x i64> @llvm.mips.addv.d(<2 x i64> %r2,  <2 x i64> %3)
diff --git a/test/CodeGen/Mips/msa/vec.ll b/test/CodeGen/Mips/msa/vec.ll
index d5b97f52fb83..8790923ce727 100644
--- a/test/CodeGen/Mips/msa/vec.ll
+++ b/test/CodeGen/Mips/msa/vec.ll
@@ -9,8 +9,8 @@
 
 define void @llvm_mips_and_v_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_and_v_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_and_v_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_and_v_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_and_v_b_ARG2
   %2 = bitcast <16 x i8> %0 to <16 x i8>
   %3 = bitcast <16 x i8> %1 to <16 x i8>
   %4 = tail call <16 x i8> @llvm.mips.and.v(<16 x i8> %2, <16 x i8> %3)
@@ -32,8 +32,8 @@ entry:
 
 define void @llvm_mips_and_v_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_and_v_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_and_v_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_and_v_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_and_v_h_ARG2
   %2 = bitcast <8 x i16> %0 to <16 x i8>
   %3 = bitcast <8 x i16> %1 to <16 x i8>
   %4 = tail call <16 x i8> @llvm.mips.and.v(<16 x i8> %2, <16 x i8> %3)
@@ -55,8 +55,8 @@ entry:
 
 define void @llvm_mips_and_v_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_and_v_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_and_v_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_and_v_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_and_v_w_ARG2
   %2 = bitcast <4 x i32> %0 to <16 x i8>
   %3 = bitcast <4 x i32> %1 to <16 x i8>
   %4 = tail call <16 x i8> @llvm.mips.and.v(<16 x i8> %2, <16 x i8> %3)
@@ -78,8 +78,8 @@ entry:
 
 define void @llvm_mips_and_v_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_and_v_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_and_v_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_and_v_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_and_v_d_ARG2
   %2 = bitcast <2 x i64> %0 to <16 x i8>
   %3 = bitcast <2 x i64> %1 to <16 x i8>
   %4 = tail call <16 x i8> @llvm.mips.and.v(<16 x i8> %2, <16 x i8> %3)
@@ -97,8 +97,8 @@ entry:
 ;
 define void @and_v_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_and_v_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_and_v_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_and_v_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_and_v_b_ARG2
   %2 = and <16 x i8> %0, %1
   store <16 x i8> %2, <16 x i8>* @llvm_mips_and_v_b_RES
   ret void
@@ -113,8 +113,8 @@ entry:
 ;
 define void @and_v_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_and_v_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_and_v_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_and_v_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_and_v_h_ARG2
   %2 = and <8 x i16> %0, %1
   store <8 x i16> %2, <8 x i16>* @llvm_mips_and_v_h_RES
   ret void
@@ -130,8 +130,8 @@ entry:
 
 define void @and_v_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_and_v_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_and_v_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_and_v_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_and_v_w_ARG2
   %2 = and <4 x i32> %0, %1
   store <4 x i32> %2, <4 x i32>* @llvm_mips_and_v_w_RES
   ret void
@@ -147,8 +147,8 @@ entry:
 
 define void @and_v_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_and_v_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_and_v_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_and_v_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_and_v_d_ARG2
   %2 = and <2 x i64> %0, %1
   store <2 x i64> %2, <2 x i64>* @llvm_mips_and_v_d_RES
   ret void
@@ -168,9 +168,9 @@ entry:
 
 define void @llvm_mips_bmnz_v_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_bmnz_v_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_bmnz_v_b_ARG2
-  %2 = load <16 x i8>* @llvm_mips_bmnz_v_b_ARG3
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bmnz_v_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bmnz_v_b_ARG2
+  %2 = load <16 x i8>, <16 x i8>* @llvm_mips_bmnz_v_b_ARG3
   %3 = bitcast <16 x i8> %0 to <16 x i8>
   %4 = bitcast <16 x i8> %1 to <16 x i8>
   %5 = bitcast <16 x i8> %2 to <16 x i8>
@@ -198,9 +198,9 @@ entry:
 
 define void @llvm_mips_bmnz_v_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_bmnz_v_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_bmnz_v_h_ARG2
-  %2 = load <8 x i16>* @llvm_mips_bmnz_v_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bmnz_v_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_bmnz_v_h_ARG2
+  %2 = load <8 x i16>, <8 x i16>* @llvm_mips_bmnz_v_h_ARG3
   %3 = bitcast <8 x i16> %0 to <16 x i8>
   %4 = bitcast <8 x i16> %1 to <16 x i8>
   %5 = bitcast <8 x i16> %2 to <16 x i8>
@@ -228,9 +228,9 @@ entry:
 
 define void @llvm_mips_bmnz_v_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_bmnz_v_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_bmnz_v_w_ARG2
-  %2 = load <4 x i32>* @llvm_mips_bmnz_v_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bmnz_v_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_bmnz_v_w_ARG2
+  %2 = load <4 x i32>, <4 x i32>* @llvm_mips_bmnz_v_w_ARG3
   %3 = bitcast <4 x i32> %0 to <16 x i8>
   %4 = bitcast <4 x i32> %1 to <16 x i8>
   %5 = bitcast <4 x i32> %2 to <16 x i8>
@@ -258,9 +258,9 @@ entry:
 
 define void @llvm_mips_bmnz_v_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_bmnz_v_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_bmnz_v_d_ARG2
-  %2 = load <2 x i64>* @llvm_mips_bmnz_v_d_ARG3
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bmnz_v_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_bmnz_v_d_ARG2
+  %2 = load <2 x i64>, <2 x i64>* @llvm_mips_bmnz_v_d_ARG3
   %3 = bitcast <2 x i64> %0 to <16 x i8>
   %4 = bitcast <2 x i64> %1 to <16 x i8>
   %5 = bitcast <2 x i64> %2 to <16 x i8>
@@ -288,9 +288,9 @@ entry:
 
 define void @llvm_mips_bmz_v_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_bmz_v_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_bmz_v_b_ARG2
-  %2 = load <16 x i8>* @llvm_mips_bmz_v_b_ARG3
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bmz_v_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bmz_v_b_ARG2
+  %2 = load <16 x i8>, <16 x i8>* @llvm_mips_bmz_v_b_ARG3
   %3 = bitcast <16 x i8> %0 to <16 x i8>
   %4 = bitcast <16 x i8> %1 to <16 x i8>
   %5 = bitcast <16 x i8> %2 to <16 x i8>
@@ -319,9 +319,9 @@ entry:
 
 define void @llvm_mips_bmz_v_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_bmz_v_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_bmz_v_h_ARG2
-  %2 = load <8 x i16>* @llvm_mips_bmz_v_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bmz_v_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_bmz_v_h_ARG2
+  %2 = load <8 x i16>, <8 x i16>* @llvm_mips_bmz_v_h_ARG3
   %3 = bitcast <8 x i16> %0 to <16 x i8>
   %4 = bitcast <8 x i16> %1 to <16 x i8>
   %5 = bitcast <8 x i16> %2 to <16 x i8>
@@ -350,9 +350,9 @@ entry:
 
 define void @llvm_mips_bmz_v_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_bmz_v_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_bmz_v_w_ARG2
-  %2 = load <4 x i32>* @llvm_mips_bmz_v_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bmz_v_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_bmz_v_w_ARG2
+  %2 = load <4 x i32>, <4 x i32>* @llvm_mips_bmz_v_w_ARG3
   %3 = bitcast <4 x i32> %0 to <16 x i8>
   %4 = bitcast <4 x i32> %1 to <16 x i8>
   %5 = bitcast <4 x i32> %2 to <16 x i8>
@@ -381,9 +381,9 @@ entry:
 
 define void @llvm_mips_bmz_v_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_bmz_v_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_bmz_v_d_ARG2
-  %2 = load <2 x i64>* @llvm_mips_bmz_v_d_ARG3
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bmz_v_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_bmz_v_d_ARG2
+  %2 = load <2 x i64>, <2 x i64>* @llvm_mips_bmz_v_d_ARG3
   %3 = bitcast <2 x i64> %0 to <16 x i8>
   %4 = bitcast <2 x i64> %1 to <16 x i8>
   %5 = bitcast <2 x i64> %2 to <16 x i8>
@@ -412,9 +412,9 @@ entry:
 
 define void @llvm_mips_bsel_v_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_bsel_v_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_bsel_v_b_ARG2
-  %2 = load <16 x i8>* @llvm_mips_bsel_v_b_ARG3
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bsel_v_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_bsel_v_b_ARG2
+  %2 = load <16 x i8>, <16 x i8>* @llvm_mips_bsel_v_b_ARG3
   %3 = bitcast <16 x i8> %0 to <16 x i8>
   %4 = bitcast <16 x i8> %1 to <16 x i8>
   %5 = bitcast <16 x i8> %2 to <16 x i8>
@@ -443,9 +443,9 @@ entry:
 
 define void @llvm_mips_bsel_v_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_bsel_v_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_bsel_v_h_ARG2
-  %2 = load <8 x i16>* @llvm_mips_bsel_v_h_ARG3
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_bsel_v_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_bsel_v_h_ARG2
+  %2 = load <8 x i16>, <8 x i16>* @llvm_mips_bsel_v_h_ARG3
   %3 = bitcast <8 x i16> %0 to <16 x i8>
   %4 = bitcast <8 x i16> %1 to <16 x i8>
   %5 = bitcast <8 x i16> %2 to <16 x i8>
@@ -474,9 +474,9 @@ entry:
 
 define void @llvm_mips_bsel_v_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_bsel_v_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_bsel_v_w_ARG2
-  %2 = load <4 x i32>* @llvm_mips_bsel_v_w_ARG3
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_bsel_v_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_bsel_v_w_ARG2
+  %2 = load <4 x i32>, <4 x i32>* @llvm_mips_bsel_v_w_ARG3
   %3 = bitcast <4 x i32> %0 to <16 x i8>
   %4 = bitcast <4 x i32> %1 to <16 x i8>
   %5 = bitcast <4 x i32> %2 to <16 x i8>
@@ -505,9 +505,9 @@ entry:
 
 define void @llvm_mips_bsel_v_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_bsel_v_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_bsel_v_d_ARG2
-  %2 = load <2 x i64>* @llvm_mips_bsel_v_d_ARG3
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_bsel_v_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_bsel_v_d_ARG2
+  %2 = load <2 x i64>, <2 x i64>* @llvm_mips_bsel_v_d_ARG3
   %3 = bitcast <2 x i64> %0 to <16 x i8>
   %4 = bitcast <2 x i64> %1 to <16 x i8>
   %5 = bitcast <2 x i64> %2 to <16 x i8>
@@ -535,8 +535,8 @@ entry:
 
 define void @llvm_mips_nor_v_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_nor_v_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_nor_v_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_nor_v_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_nor_v_b_ARG2
   %2 = bitcast <16 x i8> %0 to <16 x i8>
   %3 = bitcast <16 x i8> %1 to <16 x i8>
   %4 = tail call <16 x i8> @llvm.mips.nor.v(<16 x i8> %2, <16 x i8> %3)
@@ -558,8 +558,8 @@ entry:
 
 define void @llvm_mips_nor_v_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_nor_v_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_nor_v_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_nor_v_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_nor_v_h_ARG2
   %2 = bitcast <8 x i16> %0 to <16 x i8>
   %3 = bitcast <8 x i16> %1 to <16 x i8>
   %4 = tail call <16 x i8> @llvm.mips.nor.v(<16 x i8> %2, <16 x i8> %3)
@@ -581,8 +581,8 @@ entry:
 
 define void @llvm_mips_nor_v_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_nor_v_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_nor_v_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_nor_v_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_nor_v_w_ARG2
   %2 = bitcast <4 x i32> %0 to <16 x i8>
   %3 = bitcast <4 x i32> %1 to <16 x i8>
   %4 = tail call <16 x i8> @llvm.mips.nor.v(<16 x i8> %2, <16 x i8> %3)
@@ -604,8 +604,8 @@ entry:
 
 define void @llvm_mips_nor_v_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_nor_v_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_nor_v_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_nor_v_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_nor_v_d_ARG2
   %2 = bitcast <2 x i64> %0 to <16 x i8>
   %3 = bitcast <2 x i64> %1 to <16 x i8>
   %4 = tail call <16 x i8> @llvm.mips.nor.v(<16 x i8> %2, <16 x i8> %3)
@@ -627,8 +627,8 @@ entry:
 
 define void @llvm_mips_or_v_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_or_v_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_or_v_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_or_v_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_or_v_b_ARG2
   %2 = bitcast <16 x i8> %0 to <16 x i8>
   %3 = bitcast <16 x i8> %1 to <16 x i8>
   %4 = tail call <16 x i8> @llvm.mips.or.v(<16 x i8> %2, <16 x i8> %3)
@@ -650,8 +650,8 @@ entry:
 
 define void @llvm_mips_or_v_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_or_v_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_or_v_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_or_v_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_or_v_h_ARG2
   %2 = bitcast <8 x i16> %0 to <16 x i8>
   %3 = bitcast <8 x i16> %1 to <16 x i8>
   %4 = tail call <16 x i8> @llvm.mips.or.v(<16 x i8> %2, <16 x i8> %3)
@@ -673,8 +673,8 @@ entry:
 
 define void @llvm_mips_or_v_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_or_v_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_or_v_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_or_v_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_or_v_w_ARG2
   %2 = bitcast <4 x i32> %0 to <16 x i8>
   %3 = bitcast <4 x i32> %1 to <16 x i8>
   %4 = tail call <16 x i8> @llvm.mips.or.v(<16 x i8> %2, <16 x i8> %3)
@@ -696,8 +696,8 @@ entry:
 
 define void @llvm_mips_or_v_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_or_v_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_or_v_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_or_v_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_or_v_d_ARG2
   %2 = bitcast <2 x i64> %0 to <16 x i8>
   %3 = bitcast <2 x i64> %1 to <16 x i8>
   %4 = tail call <16 x i8> @llvm.mips.or.v(<16 x i8> %2, <16 x i8> %3)
@@ -715,8 +715,8 @@ entry:
 ;
 define void @or_v_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_or_v_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_or_v_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_or_v_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_or_v_b_ARG2
   %2 = or <16 x i8> %0, %1
   store <16 x i8> %2, <16 x i8>* @llvm_mips_or_v_b_RES
   ret void
@@ -731,8 +731,8 @@ entry:
 ;
 define void @or_v_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_or_v_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_or_v_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_or_v_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_or_v_h_ARG2
   %2 = or <8 x i16> %0, %1
   store <8 x i16> %2, <8 x i16>* @llvm_mips_or_v_h_RES
   ret void
@@ -748,8 +748,8 @@ entry:
 
 define void @or_v_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_or_v_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_or_v_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_or_v_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_or_v_w_ARG2
   %2 = or <4 x i32> %0, %1
   store <4 x i32> %2, <4 x i32>* @llvm_mips_or_v_w_RES
   ret void
@@ -765,8 +765,8 @@ entry:
 
 define void @or_v_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_or_v_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_or_v_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_or_v_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_or_v_d_ARG2
   %2 = or <2 x i64> %0, %1
   store <2 x i64> %2, <2 x i64>* @llvm_mips_or_v_d_RES
   ret void
@@ -785,8 +785,8 @@ entry:
 
 define void @llvm_mips_xor_v_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_xor_v_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_xor_v_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_xor_v_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_xor_v_b_ARG2
   %2 = bitcast <16 x i8> %0 to <16 x i8>
   %3 = bitcast <16 x i8> %1 to <16 x i8>
   %4 = tail call <16 x i8> @llvm.mips.xor.v(<16 x i8> %2, <16 x i8> %3)
@@ -808,8 +808,8 @@ entry:
 
 define void @llvm_mips_xor_v_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_xor_v_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_xor_v_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_xor_v_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_xor_v_h_ARG2
   %2 = bitcast <8 x i16> %0 to <16 x i8>
   %3 = bitcast <8 x i16> %1 to <16 x i8>
   %4 = tail call <16 x i8> @llvm.mips.xor.v(<16 x i8> %2, <16 x i8> %3)
@@ -831,8 +831,8 @@ entry:
 
 define void @llvm_mips_xor_v_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_xor_v_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_xor_v_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_xor_v_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_xor_v_w_ARG2
   %2 = bitcast <4 x i32> %0 to <16 x i8>
   %3 = bitcast <4 x i32> %1 to <16 x i8>
   %4 = tail call <16 x i8> @llvm.mips.xor.v(<16 x i8> %2, <16 x i8> %3)
@@ -854,8 +854,8 @@ entry:
 
 define void @llvm_mips_xor_v_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_xor_v_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_xor_v_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_xor_v_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_xor_v_d_ARG2
   %2 = bitcast <2 x i64> %0 to <16 x i8>
   %3 = bitcast <2 x i64> %1 to <16 x i8>
   %4 = tail call <16 x i8> @llvm.mips.xor.v(<16 x i8> %2, <16 x i8> %3)
@@ -873,8 +873,8 @@ entry:
 ;
 define void @xor_v_b_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_xor_v_b_ARG1
-  %1 = load <16 x i8>* @llvm_mips_xor_v_b_ARG2
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_xor_v_b_ARG1
+  %1 = load <16 x i8>, <16 x i8>* @llvm_mips_xor_v_b_ARG2
   %2 = xor <16 x i8> %0, %1
   store <16 x i8> %2, <16 x i8>* @llvm_mips_xor_v_b_RES
   ret void
@@ -889,8 +889,8 @@ entry:
 ;
 define void @xor_v_h_test() nounwind {
 entry:
-  %0 = load <8 x i16>* @llvm_mips_xor_v_h_ARG1
-  %1 = load <8 x i16>* @llvm_mips_xor_v_h_ARG2
+  %0 = load <8 x i16>, <8 x i16>* @llvm_mips_xor_v_h_ARG1
+  %1 = load <8 x i16>, <8 x i16>* @llvm_mips_xor_v_h_ARG2
   %2 = xor <8 x i16> %0, %1
   store <8 x i16> %2, <8 x i16>* @llvm_mips_xor_v_h_RES
   ret void
@@ -906,8 +906,8 @@ entry:
 
 define void @xor_v_w_test() nounwind {
 entry:
-  %0 = load <4 x i32>* @llvm_mips_xor_v_w_ARG1
-  %1 = load <4 x i32>* @llvm_mips_xor_v_w_ARG2
+  %0 = load <4 x i32>, <4 x i32>* @llvm_mips_xor_v_w_ARG1
+  %1 = load <4 x i32>, <4 x i32>* @llvm_mips_xor_v_w_ARG2
   %2 = xor <4 x i32> %0, %1
   store <4 x i32> %2, <4 x i32>* @llvm_mips_xor_v_w_RES
   ret void
@@ -923,8 +923,8 @@ entry:
 
 define void @xor_v_d_test() nounwind {
 entry:
-  %0 = load <2 x i64>* @llvm_mips_xor_v_d_ARG1
-  %1 = load <2 x i64>* @llvm_mips_xor_v_d_ARG2
+  %0 = load <2 x i64>, <2 x i64>* @llvm_mips_xor_v_d_ARG1
+  %1 = load <2 x i64>, <2 x i64>* @llvm_mips_xor_v_d_ARG2
   %2 = xor <2 x i64> %0, %1
   store <2 x i64> %2, <2 x i64>* @llvm_mips_xor_v_d_RES
   ret void
diff --git a/test/CodeGen/Mips/msa/vecs10.ll b/test/CodeGen/Mips/msa/vecs10.ll
index e22e0755ef00..f442f7727447 100644
--- a/test/CodeGen/Mips/msa/vecs10.ll
+++ b/test/CodeGen/Mips/msa/vecs10.ll
@@ -7,7 +7,7 @@
 
 define i32 @llvm_mips_bnz_v_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_bnz_v_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bnz_v_ARG1
   %1 = tail call i32 @llvm.mips.bnz.v(<16 x i8> %0)
   %2 = icmp eq i32 %1, 0
   br i1 %2, label %true, label %false
@@ -28,7 +28,7 @@ declare i32 @llvm.mips.bnz.v(<16 x i8>) nounwind
 
 define i32 @llvm_mips_bz_v_test() nounwind {
 entry:
-  %0 = load <16 x i8>* @llvm_mips_bz_v_ARG1
+  %0 = load <16 x i8>, <16 x i8>* @llvm_mips_bz_v_ARG1
   %1 = tail call i32 @llvm.mips.bz.v(<16 x i8> %0)
   %2 = icmp eq i32 %1, 0
   br i1 %2, label %true, label %false
diff --git a/test/CodeGen/Mips/mul.ll b/test/CodeGen/Mips/mul.ll
index 4ce801b1c9f4..3231f9cac38e 100644
--- a/test/CodeGen/Mips/mul.ll
+++ b/test/CodeGen/Mips/mul.ll
@@ -6,8 +6,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @iiii, align 4
-  %1 = load i32* @jjjj, align 4
+  %0 = load i32, i32* @iiii, align 4
+  %1 = load i32, i32* @jjjj, align 4
   %mul = mul nsw i32 %1, %0
 ; 16:	mult	${{[0-9]+}}, ${{[0-9]+}}
 ; 16: 	mflo	${{[0-9]+}}
diff --git a/test/CodeGen/Mips/mulll.ll b/test/CodeGen/Mips/mulll.ll
index e37b9197df82..6e5ba647b8bf 100644
--- a/test/CodeGen/Mips/mulll.ll
+++ b/test/CodeGen/Mips/mulll.ll
@@ -6,8 +6,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i64* @iiii, align 8
-  %1 = load i64* @jjjj, align 8
+  %0 = load i64, i64* @iiii, align 8
+  %1 = load i64, i64* @jjjj, align 8
   %mul = mul nsw i64 %1, %0
   store i64 %mul, i64* @kkkk, align 8
 ; 16:	multu	${{[0-9]+}}, ${{[0-9]+}}
diff --git a/test/CodeGen/Mips/mulull.ll b/test/CodeGen/Mips/mulull.ll
index 4d23c693184b..c1334484fb66 100644
--- a/test/CodeGen/Mips/mulull.ll
+++ b/test/CodeGen/Mips/mulull.ll
@@ -7,8 +7,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i64* @iiii, align 8
-  %1 = load i64* @jjjj, align 8
+  %0 = load i64, i64* @iiii, align 8
+  %1 = load i64, i64* @jjjj, align 8
   %mul = mul nsw i64 %1, %0
   store i64 %mul, i64* @kkkk, align 8
 ; 16:	multu	${{[0-9]+}}, ${{[0-9]+}}
diff --git a/test/CodeGen/Mips/nacl-align.ll b/test/CodeGen/Mips/nacl-align.ll
index e61b8347760e..ec8f3f06afdf 100644
--- a/test/CodeGen/Mips/nacl-align.ll
+++ b/test/CodeGen/Mips/nacl-align.ll
@@ -67,8 +67,8 @@ default:
 
 define i32 @test2(i32 %i) {
 entry:
-  %elementptr = getelementptr inbounds [2 x i8*]* @bb_array, i32 0, i32 %i
-  %0 = load i8** %elementptr, align 4
+  %elementptr = getelementptr inbounds [2 x i8*], [2 x i8*]* @bb_array, i32 0, i32 %i
+  %0 = load i8*, i8** %elementptr, align 4
   indirectbr i8* %0, [label %bb1, label %bb2]
 
 bb1:
diff --git a/test/CodeGen/Mips/nacl-branch-delay.ll b/test/CodeGen/Mips/nacl-branch-delay.ll
index d251eee07526..2927f39a416c 100644
--- a/test/CodeGen/Mips/nacl-branch-delay.ll
+++ b/test/CodeGen/Mips/nacl-branch-delay.ll
@@ -10,7 +10,7 @@ declare void @f2()
 
 
 define void @test1() {
-  %1 = load i32* @x, align 4
+  %1 = load i32, i32* @x, align 4
   call void @f1(i32 %1)
   ret void
 
diff --git a/test/CodeGen/Mips/nacl-reserved-regs.ll b/test/CodeGen/Mips/nacl-reserved-regs.ll
index ae21283b1fb7..efe2a663a3c5 100644
--- a/test/CodeGen/Mips/nacl-reserved-regs.ll
+++ b/test/CodeGen/Mips/nacl-reserved-regs.ll
@@ -5,22 +5,22 @@
 @var = external global i32
 
 define void @f() {
-  %val1 = load volatile i32* @var
-  %val2 = load volatile i32* @var
-  %val3 = load volatile i32* @var
-  %val4 = load volatile i32* @var
-  %val5 = load volatile i32* @var
-  %val6 = load volatile i32* @var
-  %val7 = load volatile i32* @var
-  %val8 = load volatile i32* @var
-  %val9 = load volatile i32* @var
-  %val10 = load volatile i32* @var
-  %val11 = load volatile i32* @var
-  %val12 = load volatile i32* @var
-  %val13 = load volatile i32* @var
-  %val14 = load volatile i32* @var
-  %val15 = load volatile i32* @var
-  %val16 = load volatile i32* @var
+  %val1 = load volatile i32, i32* @var
+  %val2 = load volatile i32, i32* @var
+  %val3 = load volatile i32, i32* @var
+  %val4 = load volatile i32, i32* @var
+  %val5 = load volatile i32, i32* @var
+  %val6 = load volatile i32, i32* @var
+  %val7 = load volatile i32, i32* @var
+  %val8 = load volatile i32, i32* @var
+  %val9 = load volatile i32, i32* @var
+  %val10 = load volatile i32, i32* @var
+  %val11 = load volatile i32, i32* @var
+  %val12 = load volatile i32, i32* @var
+  %val13 = load volatile i32, i32* @var
+  %val14 = load volatile i32, i32* @var
+  %val15 = load volatile i32, i32* @var
+  %val16 = load volatile i32, i32* @var
   store volatile i32 %val1, i32* @var
   store volatile i32 %val2, i32* @var
   store volatile i32 %val3, i32* @var
diff --git a/test/CodeGen/Mips/named-register-n32.ll b/test/CodeGen/Mips/named-register-n32.ll
index 1e5f53ac5c9c..b15e9283fc43 100644
--- a/test/CodeGen/Mips/named-register-n32.ll
+++ b/test/CodeGen/Mips/named-register-n32.ll
@@ -1,4 +1,4 @@
-; RUN: llc -march=mips64 -relocation-model=static -mattr=+noabicalls,-n64,+n32 < %s | FileCheck %s
+; RUN: llc -march=mips64 -relocation-model=static -mattr=+noabicalls -target-abi n32 < %s | FileCheck %s
 
 define i32* @get_gp() {
 entry:
diff --git a/test/CodeGen/Mips/neg1.ll b/test/CodeGen/Mips/neg1.ll
index 281e62621565..36275a2991f6 100644
--- a/test/CodeGen/Mips/neg1.ll
+++ b/test/CodeGen/Mips/neg1.ll
@@ -5,10 +5,10 @@
 
 define i32 @main() nounwind {
 entry:
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %sub = sub nsw i32 0, %0
 ; 16:	neg	${{[0-9]+}}, ${{[0-9]+}}
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i32 0, i32 0), i32 %sub)
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([5 x i8], [5 x i8]* @.str, i32 0, i32 0), i32 %sub)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/no-odd-spreg-msa.ll b/test/CodeGen/Mips/no-odd-spreg-msa.ll
index 30dd1ff82d73..cf79557cc97f 100644
--- a/test/CodeGen/Mips/no-odd-spreg-msa.ll
+++ b/test/CodeGen/Mips/no-odd-spreg-msa.ll
@@ -8,7 +8,7 @@ entry:
   ; Force the float into an odd-numbered register using named registers and
   ; load the vector.
   %b = call float asm sideeffect "mov.s $0, $1", "={$f13},{$f12}" (float %a)
-  %0 = load volatile <4 x float>* @v4f32
+  %0 = load volatile <4 x float>, <4 x float>* @v4f32
 
   ; Clobber all except $f12/$w12 and $f13
   ;
@@ -42,7 +42,7 @@ entry:
   ; Force the float into an odd-numbered register using named registers and
   ; load the vector.
   %b = call float asm sideeffect "mov.s $0, $1", "={$f13},{$f12}" (float %a)
-  %0 = load volatile <4 x float>* @v4f32
+  %0 = load volatile <4 x float>, <4 x float>* @v4f32
 
   ; Clobber all except $f12/$w12 and $f13
   ;
@@ -73,7 +73,7 @@ entry:
 
 define float @msa_extract_0() {
 entry:
-  %0 = load volatile <4 x float>* @v4f32
+  %0 = load volatile <4 x float>, <4 x float>* @v4f32
   %1 = call <4 x float> asm sideeffect "move.v $0, $1", "={$w13},{$w12}" (<4 x float> %0)
 
   ; Clobber all except $f12, and $f13
@@ -101,7 +101,7 @@ entry:
 
 define float @msa_extract_1() {
 entry:
-  %0 = load volatile <4 x float>* @v4f32
+  %0 = load volatile <4 x float>, <4 x float>* @v4f32
   %1 = call <4 x float> asm sideeffect "move.v $0, $1", "={$w13},{$w12}" (<4 x float> %0)
 
   ; Clobber all except $f13
diff --git a/test/CodeGen/Mips/nomips16.ll b/test/CodeGen/Mips/nomips16.ll
index 5f7d74e41979..418d8ead2c39 100644
--- a/test/CodeGen/Mips/nomips16.ll
+++ b/test/CodeGen/Mips/nomips16.ll
@@ -6,7 +6,7 @@
 ; Function Attrs: nounwind
 define void @foo() #0 {
 entry:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   %conv = fpext float %0 to double
   %add = fadd double %conv, 1.500000e+00
   %conv1 = fptrunc double %add to float
@@ -20,7 +20,7 @@ entry:
 ; Function Attrs: nounwind
 define void @nofoo() #1 {
 entry:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   %conv = fpext float %0 to double
   %add = fadd double %conv, 3.900000e+00
   %conv1 = fptrunc double %add to float
diff --git a/test/CodeGen/Mips/not1.ll b/test/CodeGen/Mips/not1.ll
index 2163b236c56f..f5ec5b60e421 100644
--- a/test/CodeGen/Mips/not1.ll
+++ b/test/CodeGen/Mips/not1.ll
@@ -6,10 +6,10 @@
 
 define i32 @main() nounwind {
 entry:
-  %0 = load i32* @x, align 4
+  %0 = load i32, i32* @x, align 4
   %neg = xor i32 %0, -1
 ; 16:	not	${{[0-9]+}}, ${{[0-9]+}}
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([7 x i8]* @.str, i32 0, i32 0), i32 %neg)
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([7 x i8], [7 x i8]* @.str, i32 0, i32 0), i32 %neg)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/o32_cc_byval.ll b/test/CodeGen/Mips/o32_cc_byval.ll
index 5db47acc5a85..108c663ab1cd 100644
--- a/test/CodeGen/Mips/o32_cc_byval.ll
+++ b/test/CodeGen/Mips/o32_cc_byval.ll
@@ -30,7 +30,7 @@ entry:
   %agg.tmp10 = alloca %struct.S3, align 4
   call void @callee1(float 2.000000e+01, %struct.S1* byval bitcast (%0* @f1.s1 to %struct.S1*)) nounwind
   call void @callee2(%struct.S2* byval @f1.s2) nounwind
-  %tmp11 = getelementptr inbounds %struct.S3* %agg.tmp10, i32 0, i32 0
+  %tmp11 = getelementptr inbounds %struct.S3, %struct.S3* %agg.tmp10, i32 0, i32 0
   store i8 11, i8* %tmp11, align 4
   call void @callee3(float 2.100000e+01, %struct.S3* byval %agg.tmp10, %struct.S1* byval bitcast (%0* @f1.s1 to %struct.S1*)) nounwind
   ret void
@@ -61,18 +61,18 @@ entry:
 ; CHECK: sw  $[[R3]], 16($sp)
 ; CHECK: mfc1 $6, $f[[F0]]
 
-  %i2 = getelementptr inbounds %struct.S1* %s1, i32 0, i32 5
-  %tmp = load i32* %i2, align 4
-  %d = getelementptr inbounds %struct.S1* %s1, i32 0, i32 4
-  %tmp1 = load double* %d, align 8
-  %ll = getelementptr inbounds %struct.S1* %s1, i32 0, i32 3
-  %tmp2 = load i64* %ll, align 8
-  %i = getelementptr inbounds %struct.S1* %s1, i32 0, i32 2
-  %tmp3 = load i32* %i, align 4
-  %s = getelementptr inbounds %struct.S1* %s1, i32 0, i32 1
-  %tmp4 = load i16* %s, align 2
-  %c = getelementptr inbounds %struct.S1* %s1, i32 0, i32 0
-  %tmp5 = load i8* %c, align 1
+  %i2 = getelementptr inbounds %struct.S1, %struct.S1* %s1, i32 0, i32 5
+  %tmp = load i32, i32* %i2, align 4
+  %d = getelementptr inbounds %struct.S1, %struct.S1* %s1, i32 0, i32 4
+  %tmp1 = load double, double* %d, align 8
+  %ll = getelementptr inbounds %struct.S1, %struct.S1* %s1, i32 0, i32 3
+  %tmp2 = load i64, i64* %ll, align 8
+  %i = getelementptr inbounds %struct.S1, %struct.S1* %s1, i32 0, i32 2
+  %tmp3 = load i32, i32* %i, align 4
+  %s = getelementptr inbounds %struct.S1, %struct.S1* %s1, i32 0, i32 1
+  %tmp4 = load i16, i16* %s, align 2
+  %c = getelementptr inbounds %struct.S1, %struct.S1* %s1, i32 0, i32 0
+  %tmp5 = load i8, i8* %c, align 1
   tail call void @callee4(i32 %tmp, double %tmp1, i64 %tmp2, i32 %tmp3, i16 signext %tmp4, i8 signext %tmp5, float %f) nounwind
   ret void
 }
@@ -90,10 +90,10 @@ entry:
 ; CHECK: lw  $[[R0:[0-9]+]], 60($sp)
 ; CHECK: sw  $[[R0]], 24($sp)
 
-  %arrayidx = getelementptr inbounds %struct.S2* %s2, i32 0, i32 0, i32 0
-  %tmp = load i32* %arrayidx, align 4
-  %arrayidx2 = getelementptr inbounds %struct.S2* %s2, i32 0, i32 0, i32 3
-  %tmp3 = load i32* %arrayidx2, align 4
+  %arrayidx = getelementptr inbounds %struct.S2, %struct.S2* %s2, i32 0, i32 0, i32 0
+  %tmp = load i32, i32* %arrayidx, align 4
+  %arrayidx2 = getelementptr inbounds %struct.S2, %struct.S2* %s2, i32 0, i32 0, i32 3
+  %tmp3 = load i32, i32* %arrayidx2, align 4
   tail call void @callee4(i32 %tmp, double 2.000000e+00, i64 3, i32 %tmp3, i16 signext 4, i8 signext 5, float 6.000000e+00) nounwind
   ret void
 }
@@ -110,12 +110,12 @@ entry:
 ; CHECK: sw  $[[R0]], 32($sp)
 ; CHECK: sw  $[[R1]], 24($sp)
 
-  %i = getelementptr inbounds %struct.S1* %s1, i32 0, i32 2
-  %tmp = load i32* %i, align 4
-  %i2 = getelementptr inbounds %struct.S1* %s1, i32 0, i32 5
-  %tmp1 = load i32* %i2, align 4
-  %c = getelementptr inbounds %struct.S3* %s3, i32 0, i32 0
-  %tmp2 = load i8* %c, align 1
+  %i = getelementptr inbounds %struct.S1, %struct.S1* %s1, i32 0, i32 2
+  %tmp = load i32, i32* %i, align 4
+  %i2 = getelementptr inbounds %struct.S1, %struct.S1* %s1, i32 0, i32 5
+  %tmp1 = load i32, i32* %i2, align 4
+  %c = getelementptr inbounds %struct.S3, %struct.S3* %s3, i32 0, i32 0
+  %tmp2 = load i8, i8* %c, align 1
   tail call void @callee4(i32 %tmp, double 2.000000e+00, i64 3, i32 %tmp1, i16 signext 4, i8 signext %tmp2, float 6.000000e+00) nounwind
   ret void
 }
diff --git a/test/CodeGen/Mips/o32_cc_vararg.ll b/test/CodeGen/Mips/o32_cc_vararg.ll
index 10972e884ac3..b4597a3214e2 100644
--- a/test/CodeGen/Mips/o32_cc_vararg.ll
+++ b/test/CodeGen/Mips/o32_cc_vararg.ll
@@ -24,7 +24,7 @@ entry:
   store i32 %0, i32* %b, align 4
   %ap2 = bitcast i8** %ap to i8*
   call void @llvm.va_end(i8* %ap2)
-  %tmp = load i32* %b, align 4
+  %tmp = load i32, i32* %b, align 4
   ret i32 %tmp
 
 ; CHECK-LABEL: va1:
@@ -50,7 +50,7 @@ entry:
   store double %0, double* %b, align 8
   %ap2 = bitcast i8** %ap to i8*
   call void @llvm.va_end(i8* %ap2)
-  %tmp = load double* %b, align 8
+  %tmp = load double, double* %b, align 8
   ret double %tmp
 
 ; CHECK-LABEL: va2:
@@ -78,7 +78,7 @@ entry:
   store i32 %0, i32* %b, align 4
   %ap2 = bitcast i8** %ap to i8*
   call void @llvm.va_end(i8* %ap2)
-  %tmp = load i32* %b, align 4
+  %tmp = load i32, i32* %b, align 4
   ret i32 %tmp
 
 ; CHECK-LABEL: va3:
@@ -101,7 +101,7 @@ entry:
   store double %0, double* %b, align 8
   %ap2 = bitcast i8** %ap to i8*
   call void @llvm.va_end(i8* %ap2)
-  %tmp = load double* %b, align 8
+  %tmp = load double, double* %b, align 8
   ret double %tmp
 
 ; CHECK-LABEL: va4:
@@ -129,7 +129,7 @@ entry:
   store i32 %0, i32* %d, align 4
   %ap2 = bitcast i8** %ap to i8*
   call void @llvm.va_end(i8* %ap2)
-  %tmp = load i32* %d, align 4
+  %tmp = load i32, i32* %d, align 4
   ret i32 %tmp
 
 ; CHECK-LABEL: va5:
@@ -155,7 +155,7 @@ entry:
   store double %0, double* %d, align 8
   %ap2 = bitcast i8** %ap to i8*
   call void @llvm.va_end(i8* %ap2)
-  %tmp = load double* %d, align 8
+  %tmp = load double, double* %d, align 8
   ret double %tmp
 
 ; CHECK-LABEL: va6:
@@ -183,7 +183,7 @@ entry:
   store i32 %0, i32* %c, align 4
   %ap2 = bitcast i8** %ap to i8*
   call void @llvm.va_end(i8* %ap2)
-  %tmp = load i32* %c, align 4
+  %tmp = load i32, i32* %c, align 4
   ret i32 %tmp
 
 ; CHECK-LABEL: va7:
@@ -206,7 +206,7 @@ entry:
   store double %0, double* %c, align 8
   %ap2 = bitcast i8** %ap to i8*
   call void @llvm.va_end(i8* %ap2)
-  %tmp = load double* %c, align 8
+  %tmp = load double, double* %c, align 8
   ret double %tmp
 
 ; CHECK-LABEL: va8:
@@ -232,7 +232,7 @@ entry:
   store i32 %0, i32* %d, align 4
   %ap2 = bitcast i8** %ap to i8*
   call void @llvm.va_end(i8* %ap2)
-  %tmp = load i32* %d, align 4
+  %tmp = load i32, i32* %d, align 4
   ret i32 %tmp
 
 ; CHECK-LABEL: va9:
@@ -257,7 +257,7 @@ entry:
   store double %0, double* %d, align 8
   %ap2 = bitcast i8** %ap to i8*
   call void @llvm.va_end(i8* %ap2)
-  %tmp = load double* %d, align 8
+  %tmp = load double, double* %d, align 8
   ret double %tmp
 
 ; CHECK-LABEL: va10:
diff --git a/test/CodeGen/Mips/octeon.ll b/test/CodeGen/Mips/octeon.ll
index 9d82b74f5b7e..499ce3c1ddbf 100644
--- a/test/CodeGen/Mips/octeon.ll
+++ b/test/CodeGen/Mips/octeon.ll
@@ -1,15 +1,14 @@
-; RUN: llc -O1 < %s -march=mips64 -mcpu=octeon | FileCheck %s -check-prefix=OCTEON
-; RUN: llc -O1 < %s -march=mips64 -mcpu=mips64 | FileCheck %s -check-prefix=MIPS64
+; RUN: llc -O1 < %s -march=mips64 -mcpu=octeon | FileCheck %s -check-prefix=ALL -check-prefix=OCTEON
+; RUN: llc -O1 < %s -march=mips64 -mcpu=mips64 | FileCheck %s -check-prefix=ALL -check-prefix=MIPS64
 
 define i64 @addi64(i64 %a, i64 %b) nounwind {
 entry:
-; OCTEON-LABEL: addi64:
+; ALL-LABEL: addi64:
 ; OCTEON: jr      $ra
 ; OCTEON: baddu   $2, $4, $5
-; MIPS64-LABEL: addi64:
-; MIPS64: daddu
-; MIPS64: jr
-; MIPS64: andi
+; MIPS64: daddu   $[[T0:[0-9]+]], $4, $5
+; MIPS64: jr      $ra
+; MIPS64: andi    $2, $[[T0]], 255
   %add = add i64 %a, %b
   %and = and i64 %add, 255
   ret i64 %and
@@ -17,28 +16,26 @@ entry:
 
 define i64 @mul(i64 %a, i64 %b) nounwind {
 entry:
-; OCTEON-LABEL: mul:
+; ALL-LABEL: mul:
 ; OCTEON: jr    $ra
 ; OCTEON: dmul  $2, $4, $5
-; MIPS64-LABEL: mul:
-; MIPS64: dmult
-; MIPS64: jr
-; MIPS64: mflo
+; MIPS64: dmult $4, $5
+; MIPS64: jr    $ra
+; MIPS64: mflo  $2
   %res = mul i64 %a, %b
   ret i64 %res
 }
 
 define i64 @cmpeq(i64 %a, i64 %b) nounwind {
 entry:
-; OCTEON-LABEL: cmpeq:
+; ALL-LABEL: cmpeq:
 ; OCTEON: jr     $ra
 ; OCTEON: seq    $2, $4, $5
-; MIPS64-LABEL: cmpeq:
-; MIPS64: xor    $1, $4, $5
-; MIPS64: sltiu  $1, $1, 1
-; MIPS64: dsll   $1, $1, 32
+; MIPS64: xor    $[[T0:[0-9]+]], $4, $5
+; MIPS64: sltiu  $[[T1:[0-9]+]], $[[T0]], 1
+; MIPS64: dsll   $[[T2:[0-9]+]], $[[T1]], 32
 ; MIPS64: jr     $ra
-; MIPS64: dsrl   $2, $1, 32
+; MIPS64: dsrl   $2, $[[T2]], 32
   %res = icmp eq i64 %a, %b
   %res2 = zext i1 %res to i64
   ret i64 %res2
@@ -46,16 +43,15 @@ entry:
 
 define i64 @cmpeqi(i64 %a) nounwind {
 entry:
-; OCTEON-LABEL: cmpeqi:
+; ALL-LABEL: cmpeqi:
 ; OCTEON: jr     $ra
 ; OCTEON: seqi   $2, $4, 42
-; MIPS64-LABEL: cmpeqi:
-; MIPS64: daddiu $1, $zero, 42
-; MIPS64: xor    $1, $4, $1
-; MIPS64: sltiu  $1, $1, 1
-; MIPS64: dsll   $1, $1, 32
+; MIPS64: daddiu $[[T0:[0-9]+]], $zero, 42
+; MIPS64: xor    $[[T1:[0-9]+]], $4, $[[T0]]
+; MIPS64: sltiu  $[[T2:[0-9]+]], $[[T1]], 1
+; MIPS64: dsll   $[[T3:[0-9]+]], $[[T2]], 32
 ; MIPS64: jr     $ra
-; MIPS64: dsrl   $2, $1, 32
+; MIPS64: dsrl   $2, $[[T3]], 32
   %res = icmp eq i64 %a, 42
   %res2 = zext i1 %res to i64
   ret i64 %res2
@@ -63,15 +59,14 @@ entry:
 
 define i64 @cmpne(i64 %a, i64 %b) nounwind {
 entry:
-; OCTEON-LABEL: cmpne:
+; ALL-LABEL: cmpne:
 ; OCTEON: jr     $ra
 ; OCTEON: sne    $2, $4, $5
-; MIPS64-LABEL: cmpne:
-; MIPS64: xor    $1, $4, $5
-; MIPS64: sltu   $1, $zero, $1
-; MIPS64: dsll   $1, $1, 32
+; MIPS64: xor    $[[T0:[0-9]+]], $4, $5
+; MIPS64: sltu   $[[T1:[0-9]+]], $zero, $[[T0]]
+; MIPS64: dsll   $[[T2:[0-9]+]], $[[T1]], 32
 ; MIPS64: jr     $ra
-; MIPS64: dsrl   $2, $1, 32
+; MIPS64: dsrl   $2, $[[T2]], 32
   %res = icmp ne i64 %a, %b
   %res2 = zext i1 %res to i64
   ret i64 %res2
@@ -79,17 +74,84 @@ entry:
 
 define i64 @cmpnei(i64 %a) nounwind {
 entry:
-; OCTEON-LABEL: cmpnei:
+; ALL-LABEL: cmpnei:
 ; OCTEON: jr     $ra
 ; OCTEON: snei   $2, $4, 42
-; MIPS64-LABEL: cmpnei:
-; MIPS64: daddiu $1, $zero, 42
-; MIPS64: xor    $1, $4, $1
-; MIPS64: sltu   $1, $zero, $1
-; MIPS64: dsll   $1, $1, 32
+; MIPS64: daddiu $[[T0:[0-9]+]], $zero, 42
+; MIPS64: xor    $[[T1:[0-9]+]], $4, $[[T0]]
+; MIPS64: sltu   $[[T2:[0-9]+]], $zero, $[[T1]]
+; MIPS64: dsll   $[[T3:[0-9]+]], $[[T2]], 32
 ; MIPS64: jr     $ra
-; MIPS64: dsrl   $2, $1, 32
+; MIPS64: dsrl   $2, $[[T3]], 32
   %res = icmp ne i64 %a, 42
   %res2 = zext i1 %res to i64
   ret i64 %res2
 }
+
+define i64 @bbit0(i64 %a) nounwind {
+entry:
+; ALL-LABEL: bbit0:
+; OCTEON: bbit0   $4, 3, $[[BB0:BB[0-9_]+]]
+; MIPS64: andi  $[[T0:[0-9]+]], $4, 8
+; MIPS64: bnez  $[[T0]], $[[BB0:BB[0-9_]+]]
+  %bit = and i64 %a, 8
+  %res = icmp eq i64 %bit, 0
+  br i1 %res, label %endif, label %if
+if:
+  ret i64 48
+
+endif:
+  ret i64 12
+}
+
+define i64 @bbit032(i64 %a) nounwind {
+entry:
+; ALL-LABEL: bbit032:
+; OCTEON: bbit032 $4, 3, $[[BB0:BB[0-9_]+]]
+; MIPS64: daddiu  $[[T0:[0-9]+]], $zero, 1
+; MIPS64: dsll    $[[T1:[0-9]+]], $[[T0]], 35
+; MIPS64: and     $[[T2:[0-9]+]], $4, $[[T1]]
+; MIPS64: bnez    $[[T2]], $[[BB0:BB[0-9_]+]]
+  %bit = and i64 %a, 34359738368
+  %res = icmp eq i64 %bit, 0
+  br i1 %res, label %endif, label %if
+if:
+  ret i64 48
+
+endif:
+  ret i64 12
+}
+
+define i64 @bbit1(i64 %a) nounwind {
+entry:
+; ALL-LABEL: bbit1:
+; OCTEON: bbit1 $4, 3, $[[BB0:BB[0-9_]+]]
+; MIPS64: andi  $[[T0:[0-9]+]], $4, 8
+; MIPS64: beqz  $[[T0]], $[[BB0:BB[0-9_]+]]
+  %bit = and i64 %a, 8
+  %res = icmp ne i64 %bit, 0
+  br i1 %res, label %endif, label %if
+if:
+  ret i64 48
+
+endif:
+  ret i64 12
+}
+
+define i64 @bbit132(i64 %a) nounwind {
+entry:
+; ALL-LABEL: bbit132:
+; OCTEON: bbit132 $4, 3, $[[BB0:BB[0-9_]+]]
+; MIPS64: daddiu  $[[T0:[0-9]+]], $zero, 1
+; MIPS64: dsll    $[[T1:[0-9]+]], $[[T0]], 35
+; MIPS64: and     $[[T2:[0-9]+]], $4, $[[T1]]
+; MIPS64: beqz    $[[T2]], $[[BB0:BB[0-9_]+]]
+  %bit = and i64 %a, 34359738368
+  %res = icmp ne i64 %bit, 0
+  br i1 %res, label %endif, label %if
+if:
+  ret i64 48
+
+endif:
+  ret i64 12
+}
diff --git a/test/CodeGen/Mips/optimize-pic-o0.ll b/test/CodeGen/Mips/optimize-pic-o0.ll
index 554d49e728c7..454bc851484d 100644
--- a/test/CodeGen/Mips/optimize-pic-o0.ll
+++ b/test/CodeGen/Mips/optimize-pic-o0.ll
@@ -10,7 +10,7 @@ entry:
   br label %for.cond
 
 for.cond:                                         ; preds = %for.inc, %entry
-  %0 = load i32* %i, align 4
+  %0 = load i32, i32* %i, align 4
   %cmp = icmp slt i32 %0, 10
   br i1 %cmp, label %for.body, label %for.end
 
@@ -20,13 +20,13 @@ for.body:                                         ; preds = %for.cond
   br label %for.inc
 
 for.inc:                                          ; preds = %for.body
-  %1 = load i32* %i, align 4
+  %1 = load i32, i32* %i, align 4
   %inc = add nsw i32 %1, 1
   store i32 %inc, i32* %i, align 4
   br label %for.cond
 
 for.end:                                          ; preds = %for.cond
-  %2 = load i32* %retval
+  %2 = load i32, i32* %retval
   ret i32 %2
 }
 
diff --git a/test/CodeGen/Mips/or1.ll b/test/CodeGen/Mips/or1.ll
index b1c36961f92b..51b6ebfe8e3b 100644
--- a/test/CodeGen/Mips/or1.ll
+++ b/test/CodeGen/Mips/or1.ll
@@ -6,11 +6,11 @@
 
 define i32 @main() nounwind {
 entry:
-  %0 = load i32* @x, align 4
-  %1 = load i32* @y, align 4
+  %0 = load i32, i32* @x, align 4
+  %1 = load i32, i32* @y, align 4
   %or = or i32 %0, %1
 ; 16:	or	${{[0-9]+}}, ${{[0-9]+}}
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([7 x i8]* @.str, i32 0, i32 0), i32 %or)
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([7 x i8], [7 x i8]* @.str, i32 0, i32 0), i32 %or)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/prevent-hoisting.ll b/test/CodeGen/Mips/prevent-hoisting.ll
index 210fe3b0f6d8..81b14d7441b3 100644
--- a/test/CodeGen/Mips/prevent-hoisting.ll
+++ b/test/CodeGen/Mips/prevent-hoisting.ll
@@ -46,7 +46,7 @@
 
 define void @readLumaCoeff8x8_CABAC(%struct.img_par* %img, i32 %b8) {
 
-  %1 = load i32* undef, align 4
+  %1 = load i32, i32* undef, align 4
   br i1 false, label %2, label %3
 
 ; <label>:2                                       ; preds = %0
@@ -65,7 +65,7 @@ switch.lookup:                                    ; preds = %3
   br label %6
 
 ; <label>:6                                       ; preds = %5, %4
-  %7 = phi [2 x i8]* [ getelementptr inbounds ([64 x [2 x i8]]* @FIELD_SCAN8x8, i32 0, i32 0), %4 ], [ null, %5 ]
+  %7 = phi [2 x i8]* [ getelementptr inbounds ([64 x [2 x i8]], [64 x [2 x i8]]* @FIELD_SCAN8x8, i32 0, i32 0), %4 ], [ null, %5 ]
   br i1 undef, label %switch.lookup6, label %8
 
 switch.lookup6:                                   ; preds = %6
@@ -77,7 +77,7 @@ switch.lookup6:                                   ; preds = %6
 ; <label>:9                                       ; preds = %8
   %10 = and i32 %b8, 1
   %11 = shl nuw nsw i32 %10, 3
-  %12 = getelementptr inbounds %struct.Slice* null, i32 0, i32 9
+  %12 = getelementptr inbounds %struct.Slice, %struct.Slice* null, i32 0, i32 9
   br i1 undef, label %.preheader, label %.preheader11
 
 .preheader11:                                     ; preds = %21, %9
@@ -92,19 +92,19 @@ switch.lookup6:                                   ; preds = %6
   br label %15
 
 ; <label>:15                                      ; preds = %14, %13
-  %16 = getelementptr inbounds [0 x [20 x i32]]* @assignSE2partition, i32 0, i32 %1, i32 undef
-  %17 = load i32* %16, align 4
-  %18 = getelementptr inbounds %struct.datapartition* null, i32 %17, i32 2
-  %19 = load i32 (%struct.syntaxelement*, %struct.img_par*, %struct.datapartition*)** %18, align 4
+  %16 = getelementptr inbounds [0 x [20 x i32]], [0 x [20 x i32]]* @assignSE2partition, i32 0, i32 %1, i32 undef
+  %17 = load i32, i32* %16, align 4
+  %18 = getelementptr inbounds %struct.datapartition, %struct.datapartition* null, i32 %17, i32 2
+  %19 = load i32 (%struct.syntaxelement*, %struct.img_par*, %struct.datapartition*)*, i32 (%struct.syntaxelement*, %struct.img_par*, %struct.datapartition*)** %18, align 4
   %20 = call i32 %19(%struct.syntaxelement* undef, %struct.img_par* %img, %struct.datapartition* undef)
   br i1 false, label %.loopexit, label %21
 
 ; <label>:21                                      ; preds = %15
   %22 = add i32 %coef_ctr.013, 1
   %23 = add i32 %22, 0
-  %24 = getelementptr inbounds [2 x i8]* %7, i32 %23, i32 0
+  %24 = getelementptr inbounds [2 x i8], [2 x i8]* %7, i32 %23, i32 0
   %25 = add nsw i32 0, %11
-  %26 = getelementptr inbounds %struct.img_par* %img, i32 0, i32 27, i32 undef, i32 %25
+  %26 = getelementptr inbounds %struct.img_par, %struct.img_par* %img, i32 0, i32 27, i32 undef, i32 %25
   store i32 0, i32* %26, align 4
   %27 = add nsw i32 %k.014, 1
   %28 = icmp slt i32 %27, 65
@@ -122,21 +122,21 @@ switch.lookup6:                                   ; preds = %6
   br label %31
 
 ; <label>:31                                      ; preds = %30, %29
-  %32 = getelementptr inbounds [0 x [20 x i32]]* @assignSE2partition, i32 0, i32 %1, i32 undef
-  %33 = load i32* %32, align 4
-  %34 = getelementptr inbounds %struct.datapartition* null, i32 %33
+  %32 = getelementptr inbounds [0 x [20 x i32]], [0 x [20 x i32]]* @assignSE2partition, i32 0, i32 %1, i32 undef
+  %33 = load i32, i32* %32, align 4
+  %34 = getelementptr inbounds %struct.datapartition, %struct.datapartition* null, i32 %33
   %35 = call i32 undef(%struct.syntaxelement* undef, %struct.img_par* %img, %struct.datapartition* %34)
   br i1 false, label %.loopexit, label %36
 
 ; <label>:36                                      ; preds = %31
-  %37 = load i32* undef, align 4
+  %37 = load i32, i32* undef, align 4
   %38 = add i32 %coef_ctr.29, 1
   %39 = add i32 %38, %37
-  %40 = getelementptr inbounds [2 x i8]* %7, i32 %39, i32 0
-  %41 = load i8* %40, align 1
+  %40 = getelementptr inbounds [2 x i8], [2 x i8]* %7, i32 %39, i32 0
+  %41 = load i8, i8* %40, align 1
   %42 = zext i8 %41 to i32
   %43 = add nsw i32 %42, %11
-  %44 = getelementptr inbounds %struct.img_par* %img, i32 0, i32 27, i32 undef, i32 %43
+  %44 = getelementptr inbounds %struct.img_par, %struct.img_par* %img, i32 0, i32 27, i32 undef, i32 %43
   store i32 0, i32* %44, align 4
   %45 = add nsw i32 %k.110, 1
   %46 = icmp slt i32 %45, 65
diff --git a/test/CodeGen/Mips/private.ll b/test/CodeGen/Mips/private.ll
index 058db0bb977a..5907dbd644ae 100644
--- a/test/CodeGen/Mips/private.ll
+++ b/test/CodeGen/Mips/private.ll
@@ -15,6 +15,6 @@ define i32 @bar() {
 ; CHECK: lw $[[R0:[0-9]+]], %got($baz)($
 ; CHECK: lw ${{[0-9]+}}, %lo($baz)($[[R0]])
   call void @foo()
-  %1 = load i32* @baz, align 4
+  %1 = load i32, i32* @baz, align 4
   ret i32 %1
 }
diff --git a/test/CodeGen/Mips/ra-allocatable.ll b/test/CodeGen/Mips/ra-allocatable.ll
index afc5cb0c2556..048d4325a411 100644
--- a/test/CodeGen/Mips/ra-allocatable.ll
+++ b/test/CodeGen/Mips/ra-allocatable.ll
@@ -98,186 +98,186 @@ entry:
 ; CHECK: lw  $ra, {{[0-9]+}}($sp)            # 4-byte Folded Reload
 ; CHECK: jr  $ra
 
-  %0 = load i32* @a0, align 4
-  %1 = load i32** @b0, align 4
+  %0 = load i32, i32* @a0, align 4
+  %1 = load i32*, i32** @b0, align 4
   store i32 %0, i32* %1, align 4
-  %2 = load i32* @a1, align 4
-  %3 = load i32** @b1, align 4
+  %2 = load i32, i32* @a1, align 4
+  %3 = load i32*, i32** @b1, align 4
   store i32 %2, i32* %3, align 4
-  %4 = load i32* @a2, align 4
-  %5 = load i32** @b2, align 4
+  %4 = load i32, i32* @a2, align 4
+  %5 = load i32*, i32** @b2, align 4
   store i32 %4, i32* %5, align 4
-  %6 = load i32* @a3, align 4
-  %7 = load i32** @b3, align 4
+  %6 = load i32, i32* @a3, align 4
+  %7 = load i32*, i32** @b3, align 4
   store i32 %6, i32* %7, align 4
-  %8 = load i32* @a4, align 4
-  %9 = load i32** @b4, align 4
+  %8 = load i32, i32* @a4, align 4
+  %9 = load i32*, i32** @b4, align 4
   store i32 %8, i32* %9, align 4
-  %10 = load i32* @a5, align 4
-  %11 = load i32** @b5, align 4
+  %10 = load i32, i32* @a5, align 4
+  %11 = load i32*, i32** @b5, align 4
   store i32 %10, i32* %11, align 4
-  %12 = load i32* @a6, align 4
-  %13 = load i32** @b6, align 4
+  %12 = load i32, i32* @a6, align 4
+  %13 = load i32*, i32** @b6, align 4
   store i32 %12, i32* %13, align 4
-  %14 = load i32* @a7, align 4
-  %15 = load i32** @b7, align 4
+  %14 = load i32, i32* @a7, align 4
+  %15 = load i32*, i32** @b7, align 4
   store i32 %14, i32* %15, align 4
-  %16 = load i32* @a8, align 4
-  %17 = load i32** @b8, align 4
+  %16 = load i32, i32* @a8, align 4
+  %17 = load i32*, i32** @b8, align 4
   store i32 %16, i32* %17, align 4
-  %18 = load i32* @a9, align 4
-  %19 = load i32** @b9, align 4
+  %18 = load i32, i32* @a9, align 4
+  %19 = load i32*, i32** @b9, align 4
   store i32 %18, i32* %19, align 4
-  %20 = load i32* @a10, align 4
-  %21 = load i32** @b10, align 4
+  %20 = load i32, i32* @a10, align 4
+  %21 = load i32*, i32** @b10, align 4
   store i32 %20, i32* %21, align 4
-  %22 = load i32* @a11, align 4
-  %23 = load i32** @b11, align 4
+  %22 = load i32, i32* @a11, align 4
+  %23 = load i32*, i32** @b11, align 4
   store i32 %22, i32* %23, align 4
-  %24 = load i32* @a12, align 4
-  %25 = load i32** @b12, align 4
+  %24 = load i32, i32* @a12, align 4
+  %25 = load i32*, i32** @b12, align 4
   store i32 %24, i32* %25, align 4
-  %26 = load i32* @a13, align 4
-  %27 = load i32** @b13, align 4
+  %26 = load i32, i32* @a13, align 4
+  %27 = load i32*, i32** @b13, align 4
   store i32 %26, i32* %27, align 4
-  %28 = load i32* @a14, align 4
-  %29 = load i32** @b14, align 4
+  %28 = load i32, i32* @a14, align 4
+  %29 = load i32*, i32** @b14, align 4
   store i32 %28, i32* %29, align 4
-  %30 = load i32* @a15, align 4
-  %31 = load i32** @b15, align 4
+  %30 = load i32, i32* @a15, align 4
+  %31 = load i32*, i32** @b15, align 4
   store i32 %30, i32* %31, align 4
-  %32 = load i32* @a16, align 4
-  %33 = load i32** @b16, align 4
+  %32 = load i32, i32* @a16, align 4
+  %33 = load i32*, i32** @b16, align 4
   store i32 %32, i32* %33, align 4
-  %34 = load i32* @a17, align 4
-  %35 = load i32** @b17, align 4
+  %34 = load i32, i32* @a17, align 4
+  %35 = load i32*, i32** @b17, align 4
   store i32 %34, i32* %35, align 4
-  %36 = load i32* @a18, align 4
-  %37 = load i32** @b18, align 4
+  %36 = load i32, i32* @a18, align 4
+  %37 = load i32*, i32** @b18, align 4
   store i32 %36, i32* %37, align 4
-  %38 = load i32* @a19, align 4
-  %39 = load i32** @b19, align 4
+  %38 = load i32, i32* @a19, align 4
+  %39 = load i32*, i32** @b19, align 4
   store i32 %38, i32* %39, align 4
-  %40 = load i32* @a20, align 4
-  %41 = load i32** @b20, align 4
+  %40 = load i32, i32* @a20, align 4
+  %41 = load i32*, i32** @b20, align 4
   store i32 %40, i32* %41, align 4
-  %42 = load i32* @a21, align 4
-  %43 = load i32** @b21, align 4
+  %42 = load i32, i32* @a21, align 4
+  %43 = load i32*, i32** @b21, align 4
   store i32 %42, i32* %43, align 4
-  %44 = load i32* @a22, align 4
-  %45 = load i32** @b22, align 4
+  %44 = load i32, i32* @a22, align 4
+  %45 = load i32*, i32** @b22, align 4
   store i32 %44, i32* %45, align 4
-  %46 = load i32* @a23, align 4
-  %47 = load i32** @b23, align 4
+  %46 = load i32, i32* @a23, align 4
+  %47 = load i32*, i32** @b23, align 4
   store i32 %46, i32* %47, align 4
-  %48 = load i32* @a24, align 4
-  %49 = load i32** @b24, align 4
+  %48 = load i32, i32* @a24, align 4
+  %49 = load i32*, i32** @b24, align 4
   store i32 %48, i32* %49, align 4
-  %50 = load i32* @a25, align 4
-  %51 = load i32** @b25, align 4
+  %50 = load i32, i32* @a25, align 4
+  %51 = load i32*, i32** @b25, align 4
   store i32 %50, i32* %51, align 4
-  %52 = load i32* @a26, align 4
-  %53 = load i32** @b26, align 4
+  %52 = load i32, i32* @a26, align 4
+  %53 = load i32*, i32** @b26, align 4
   store i32 %52, i32* %53, align 4
-  %54 = load i32* @a27, align 4
-  %55 = load i32** @b27, align 4
+  %54 = load i32, i32* @a27, align 4
+  %55 = load i32*, i32** @b27, align 4
   store i32 %54, i32* %55, align 4
-  %56 = load i32* @a28, align 4
-  %57 = load i32** @b28, align 4
+  %56 = load i32, i32* @a28, align 4
+  %57 = load i32*, i32** @b28, align 4
   store i32 %56, i32* %57, align 4
-  %58 = load i32* @a29, align 4
-  %59 = load i32** @b29, align 4
+  %58 = load i32, i32* @a29, align 4
+  %59 = load i32*, i32** @b29, align 4
   store i32 %58, i32* %59, align 4
-  %60 = load i32* @a0, align 4
-  %61 = load i32** @c0, align 4
+  %60 = load i32, i32* @a0, align 4
+  %61 = load i32*, i32** @c0, align 4
   store i32 %60, i32* %61, align 4
-  %62 = load i32* @a1, align 4
-  %63 = load i32** @c1, align 4
+  %62 = load i32, i32* @a1, align 4
+  %63 = load i32*, i32** @c1, align 4
   store i32 %62, i32* %63, align 4
-  %64 = load i32* @a2, align 4
-  %65 = load i32** @c2, align 4
+  %64 = load i32, i32* @a2, align 4
+  %65 = load i32*, i32** @c2, align 4
   store i32 %64, i32* %65, align 4
-  %66 = load i32* @a3, align 4
-  %67 = load i32** @c3, align 4
+  %66 = load i32, i32* @a3, align 4
+  %67 = load i32*, i32** @c3, align 4
   store i32 %66, i32* %67, align 4
-  %68 = load i32* @a4, align 4
-  %69 = load i32** @c4, align 4
+  %68 = load i32, i32* @a4, align 4
+  %69 = load i32*, i32** @c4, align 4
   store i32 %68, i32* %69, align 4
-  %70 = load i32* @a5, align 4
-  %71 = load i32** @c5, align 4
+  %70 = load i32, i32* @a5, align 4
+  %71 = load i32*, i32** @c5, align 4
   store i32 %70, i32* %71, align 4
-  %72 = load i32* @a6, align 4
-  %73 = load i32** @c6, align 4
+  %72 = load i32, i32* @a6, align 4
+  %73 = load i32*, i32** @c6, align 4
   store i32 %72, i32* %73, align 4
-  %74 = load i32* @a7, align 4
-  %75 = load i32** @c7, align 4
+  %74 = load i32, i32* @a7, align 4
+  %75 = load i32*, i32** @c7, align 4
   store i32 %74, i32* %75, align 4
-  %76 = load i32* @a8, align 4
-  %77 = load i32** @c8, align 4
+  %76 = load i32, i32* @a8, align 4
+  %77 = load i32*, i32** @c8, align 4
   store i32 %76, i32* %77, align 4
-  %78 = load i32* @a9, align 4
-  %79 = load i32** @c9, align 4
+  %78 = load i32, i32* @a9, align 4
+  %79 = load i32*, i32** @c9, align 4
   store i32 %78, i32* %79, align 4
-  %80 = load i32* @a10, align 4
-  %81 = load i32** @c10, align 4
+  %80 = load i32, i32* @a10, align 4
+  %81 = load i32*, i32** @c10, align 4
   store i32 %80, i32* %81, align 4
-  %82 = load i32* @a11, align 4
-  %83 = load i32** @c11, align 4
+  %82 = load i32, i32* @a11, align 4
+  %83 = load i32*, i32** @c11, align 4
   store i32 %82, i32* %83, align 4
-  %84 = load i32* @a12, align 4
-  %85 = load i32** @c12, align 4
+  %84 = load i32, i32* @a12, align 4
+  %85 = load i32*, i32** @c12, align 4
   store i32 %84, i32* %85, align 4
-  %86 = load i32* @a13, align 4
-  %87 = load i32** @c13, align 4
+  %86 = load i32, i32* @a13, align 4
+  %87 = load i32*, i32** @c13, align 4
   store i32 %86, i32* %87, align 4
-  %88 = load i32* @a14, align 4
-  %89 = load i32** @c14, align 4
+  %88 = load i32, i32* @a14, align 4
+  %89 = load i32*, i32** @c14, align 4
   store i32 %88, i32* %89, align 4
-  %90 = load i32* @a15, align 4
-  %91 = load i32** @c15, align 4
+  %90 = load i32, i32* @a15, align 4
+  %91 = load i32*, i32** @c15, align 4
   store i32 %90, i32* %91, align 4
-  %92 = load i32* @a16, align 4
-  %93 = load i32** @c16, align 4
+  %92 = load i32, i32* @a16, align 4
+  %93 = load i32*, i32** @c16, align 4
   store i32 %92, i32* %93, align 4
-  %94 = load i32* @a17, align 4
-  %95 = load i32** @c17, align 4
+  %94 = load i32, i32* @a17, align 4
+  %95 = load i32*, i32** @c17, align 4
   store i32 %94, i32* %95, align 4
-  %96 = load i32* @a18, align 4
-  %97 = load i32** @c18, align 4
+  %96 = load i32, i32* @a18, align 4
+  %97 = load i32*, i32** @c18, align 4
   store i32 %96, i32* %97, align 4
-  %98 = load i32* @a19, align 4
-  %99 = load i32** @c19, align 4
+  %98 = load i32, i32* @a19, align 4
+  %99 = load i32*, i32** @c19, align 4
   store i32 %98, i32* %99, align 4
-  %100 = load i32* @a20, align 4
-  %101 = load i32** @c20, align 4
+  %100 = load i32, i32* @a20, align 4
+  %101 = load i32*, i32** @c20, align 4
   store i32 %100, i32* %101, align 4
-  %102 = load i32* @a21, align 4
-  %103 = load i32** @c21, align 4
+  %102 = load i32, i32* @a21, align 4
+  %103 = load i32*, i32** @c21, align 4
   store i32 %102, i32* %103, align 4
-  %104 = load i32* @a22, align 4
-  %105 = load i32** @c22, align 4
+  %104 = load i32, i32* @a22, align 4
+  %105 = load i32*, i32** @c22, align 4
   store i32 %104, i32* %105, align 4
-  %106 = load i32* @a23, align 4
-  %107 = load i32** @c23, align 4
+  %106 = load i32, i32* @a23, align 4
+  %107 = load i32*, i32** @c23, align 4
   store i32 %106, i32* %107, align 4
-  %108 = load i32* @a24, align 4
-  %109 = load i32** @c24, align 4
+  %108 = load i32, i32* @a24, align 4
+  %109 = load i32*, i32** @c24, align 4
   store i32 %108, i32* %109, align 4
-  %110 = load i32* @a25, align 4
-  %111 = load i32** @c25, align 4
+  %110 = load i32, i32* @a25, align 4
+  %111 = load i32*, i32** @c25, align 4
   store i32 %110, i32* %111, align 4
-  %112 = load i32* @a26, align 4
-  %113 = load i32** @c26, align 4
+  %112 = load i32, i32* @a26, align 4
+  %113 = load i32*, i32** @c26, align 4
   store i32 %112, i32* %113, align 4
-  %114 = load i32* @a27, align 4
-  %115 = load i32** @c27, align 4
+  %114 = load i32, i32* @a27, align 4
+  %115 = load i32*, i32** @c27, align 4
   store i32 %114, i32* %115, align 4
-  %116 = load i32* @a28, align 4
-  %117 = load i32** @c28, align 4
+  %116 = load i32, i32* @a28, align 4
+  %117 = load i32*, i32** @c28, align 4
   store i32 %116, i32* %117, align 4
-  %118 = load i32* @a29, align 4
-  %119 = load i32** @c29, align 4
+  %118 = load i32, i32* @a29, align 4
+  %119 = load i32*, i32** @c29, align 4
   store i32 %118, i32* %119, align 4
-  %120 = load i32* @a0, align 4
+  %120 = load i32, i32* @a0, align 4
   ret i32 %120
 }
diff --git a/test/CodeGen/Mips/rdhwr-directives.ll b/test/CodeGen/Mips/rdhwr-directives.ll
index 27010d4d3699..ebc91ea4459e 100644
--- a/test/CodeGen/Mips/rdhwr-directives.ll
+++ b/test/CodeGen/Mips/rdhwr-directives.ll
@@ -9,7 +9,7 @@ entry:
 ; CHECK: rdhwr 
 ; CHECK: .set  pop
 
-  %0 = load i32* @a, align 4
+  %0 = load i32, i32* @a, align 4
   ret i32 %0
 }
 
diff --git a/test/CodeGen/Mips/rem.ll b/test/CodeGen/Mips/rem.ll
index b18f85dcbecf..70f957ce15f6 100644
--- a/test/CodeGen/Mips/rem.ll
+++ b/test/CodeGen/Mips/rem.ll
@@ -7,8 +7,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @iiii, align 4
-  %1 = load i32* @jjjj, align 4
+  %0 = load i32, i32* @iiii, align 4
+  %1 = load i32, i32* @jjjj, align 4
   %rem = srem i32 %0, %1
 ; 16:	div	$zero, ${{[0-9]+}}, ${{[0-9]+}}
 ; 16: 	mfhi	${{[0-9]+}}
diff --git a/test/CodeGen/Mips/remat-immed-load.ll b/test/CodeGen/Mips/remat-immed-load.ll
index b53b156e9eec..3d37b43bbc63 100644
--- a/test/CodeGen/Mips/remat-immed-load.ll
+++ b/test/CodeGen/Mips/remat-immed-load.ll
@@ -1,6 +1,6 @@
 ; RUN: llc -march=mipsel < %s | FileCheck %s -check-prefix=32
-; RUN: llc -march=mips64el -mcpu=mips4 -mattr=n64 < %s | FileCheck %s -check-prefix=64
-; RUN: llc -march=mips64el -mcpu=mips64 -mattr=n64 < %s | FileCheck %s -check-prefix=64
+; RUN: llc -march=mips64el -mcpu=mips4 -target-abi=n64 < %s | FileCheck %s -check-prefix=64
+; RUN: llc -march=mips64el -mcpu=mips64 -target-abi=n64 < %s | FileCheck %s -check-prefix=64
 
 define void @f0() nounwind {
 entry:
diff --git a/test/CodeGen/Mips/remu.ll b/test/CodeGen/Mips/remu.ll
index 472503c38403..12679727952f 100644
--- a/test/CodeGen/Mips/remu.ll
+++ b/test/CodeGen/Mips/remu.ll
@@ -7,8 +7,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @iiii, align 4
-  %1 = load i32* @jjjj, align 4
+  %0 = load i32, i32* @iiii, align 4
+  %1 = load i32, i32* @jjjj, align 4
   %rem = urem i32 %0, %1
 ; 16:	divu	$zero, ${{[0-9]+}}, ${{[0-9]+}}
 ; 16: 	mfhi	${{[0-9]+}}
diff --git a/test/CodeGen/Mips/return-vector.ll b/test/CodeGen/Mips/return-vector.ll
index 0e0d51587541..3870fe092156 100644
--- a/test/CodeGen/Mips/return-vector.ll
+++ b/test/CodeGen/Mips/return-vector.ll
@@ -12,7 +12,7 @@ declare <4 x double> @d4(...)
 
 define i32 @call_i8() {
 entry:
-  %call = call <8 x i32> (...)* @i8()
+  %call = call <8 x i32> (...) @i8()
   %v0 = extractelement <8 x i32> %call, i32 0
   %v1 = extractelement <8 x i32> %call, i32 1
   %v2 = extractelement <8 x i32> %call, i32 2
@@ -46,7 +46,7 @@ entry:
 
 define float @call_f4() {
 entry:
-  %call = call <4 x float> (...)* @f4()
+  %call = call <4 x float> (...) @f4()
   %v0 = extractelement <4 x float> %call, i32 0
   %v1 = extractelement <4 x float> %call, i32 1
   %v2 = extractelement <4 x float> %call, i32 2
@@ -68,7 +68,7 @@ entry:
 
 define double @call_d4() {
 entry:
-  %call = call <4 x double> (...)* @d4()
+  %call = call <4 x double> (...) @d4()
   %v0 = extractelement <4 x double> %call, i32 0
   %v1 = extractelement <4 x double> %call, i32 1
   %v2 = extractelement <4 x double> %call, i32 2
@@ -99,7 +99,7 @@ declare <2 x double> @d2(...)
 
 define i32 @call_i4() {
 entry:
-  %call = call <4 x i32> (...)* @i4()
+  %call = call <4 x i32> (...) @i4()
   %v0 = extractelement <4 x i32> %call, i32 0
   %v1 = extractelement <4 x i32> %call, i32 1
   %v2 = extractelement <4 x i32> %call, i32 2
@@ -120,7 +120,7 @@ entry:
 
 define float @call_f2() {
 entry:
-  %call = call <2 x float> (...)* @f2()
+  %call = call <2 x float> (...) @f2()
   %v0 = extractelement <2 x float> %call, i32 0
   %v1 = extractelement <2 x float> %call, i32 1
   %add1 = fadd float %v0, %v1
@@ -135,7 +135,7 @@ entry:
 
 define double @call_d2() {
 entry:
-  %call = call <2 x double> (...)* @d2()
+  %call = call <2 x double> (...) @d2()
   %v0 = extractelement <2 x double> %call, i32 0
   %v1 = extractelement <2 x double> %call, i32 1
   %add1 = fadd double %v0, %v1
diff --git a/test/CodeGen/Mips/s2rem.ll b/test/CodeGen/Mips/s2rem.ll
index 9edb5be2771e..65e48fe57c92 100644
--- a/test/CodeGen/Mips/s2rem.ll
+++ b/test/CodeGen/Mips/s2rem.ll
@@ -56,7 +56,7 @@ declare double @d() #1
 ; Function Attrs: nounwind
 define void @fft() #0 {
 entry:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   %call = call float @ff(float %0)
   store float %call, float* @x, align 4
   ret void
@@ -71,7 +71,7 @@ declare float @ff(float) #1
 ; Function Attrs: nounwind
 define void @vft() #0 {
 entry:
-  %0 = load float* @x, align 4
+  %0 = load float, float* @x, align 4
   call void @vf(float %0)
   ret void
 ; PIC: 	.ent	vft
diff --git a/test/CodeGen/Mips/sb1.ll b/test/CodeGen/Mips/sb1.ll
index e1a28d459548..d2e8510024e5 100644
--- a/test/CodeGen/Mips/sb1.ll
+++ b/test/CodeGen/Mips/sb1.ll
@@ -6,14 +6,14 @@
 
 define i32 @main() nounwind {
 entry:
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %conv = trunc i32 %0 to i8
   store i8 %conv, i8* @c, align 1
-  %1 = load i32* @i, align 4
-  %2 = load i8* @c, align 1
+  %1 = load i32, i32* @i, align 4
+  %2 = load i8, i8* @c, align 1
   %conv1 = sext i8 %2 to i32
 ; 16:	sb	${{[0-9]+}}, 0(${{[0-9]+}})
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([8 x i8]* @.str, i32 0, i32 0), i32 %1, i32 %conv1)
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([8 x i8], [8 x i8]* @.str, i32 0, i32 0), i32 %1, i32 %conv1)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/sel1c.ll b/test/CodeGen/Mips/sel1c.ll
index edd2e3e43b79..6753af106e0f 100644
--- a/test/CodeGen/Mips/sel1c.ll
+++ b/test/CodeGen/Mips/sel1c.ll
@@ -1,4 +1,4 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=cond-b-short
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=cond-b-short
 
 @i = global i32 1, align 4
 @j = global i32 2, align 4
@@ -7,8 +7,8 @@
 ; Function Attrs: nounwind optsize
 define void @t() #0 {
 entry:
-  %0 = load i32* @i, align 4
-  %1 = load i32* @j, align 4
+  %0 = load i32, i32* @i, align 4
+  %1 = load i32, i32* @j, align 4
   %cmp = icmp eq i32 %0, %1
   %cond = select i1 %cmp, i32 1, i32 3
   store i32 %cond, i32* @k, align 4
diff --git a/test/CodeGen/Mips/sel2c.ll b/test/CodeGen/Mips/sel2c.ll
index 4b211245f46e..987cccad5bf4 100644
--- a/test/CodeGen/Mips/sel2c.ll
+++ b/test/CodeGen/Mips/sel2c.ll
@@ -1,4 +1,4 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=cond-b-short
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=pic -mips16-constant-islands   < %s | FileCheck %s -check-prefix=cond-b-short
 
 @i = global i32 1, align 4
 @j = global i32 2, align 4
@@ -7,8 +7,8 @@
 ; Function Attrs: nounwind optsize
 define void @t() #0 {
 entry:
-  %0 = load i32* @i, align 4
-  %1 = load i32* @j, align 4
+  %0 = load i32, i32* @i, align 4
+  %1 = load i32, i32* @j, align 4
   %cmp = icmp ne i32 %0, %1
   %cond = select i1 %cmp, i32 1, i32 3
   store i32 %cond, i32* @k, align 4
diff --git a/test/CodeGen/Mips/selTBteqzCmpi.ll b/test/CodeGen/Mips/selTBteqzCmpi.ll
index 9cb8227f9d2b..5a72ea01073c 100644
--- a/test/CodeGen/Mips/selTBteqzCmpi.ll
+++ b/test/CodeGen/Mips/selTBteqzCmpi.ll
@@ -8,10 +8,10 @@
 
 define void @t() nounwind "target-cpu"="mips16" "target-features"="+mips16,+o32" {
 entry:
-  %0 = load i32* @a, align 4
+  %0 = load i32, i32* @a, align 4
   %cmp = icmp eq i32 %0, 10
-  %1 = load i32* @i, align 4
-  %2 = load i32* @j, align 4
+  %1 = load i32, i32* @i, align 4
+  %2 = load i32, i32* @j, align 4
   %cond = select i1 %cmp, i32 %1, i32 %2
   store i32 %cond, i32* @i, align 4
   ret void
diff --git a/test/CodeGen/Mips/selTBtnezCmpi.ll b/test/CodeGen/Mips/selTBtnezCmpi.ll
index bd334f59d33b..b6407e67f27a 100644
--- a/test/CodeGen/Mips/selTBtnezCmpi.ll
+++ b/test/CodeGen/Mips/selTBtnezCmpi.ll
@@ -8,10 +8,10 @@
 
 define void @t() nounwind "target-cpu"="mips16" "target-features"="+mips16,+o32" {
 entry:
-  %0 = load i32* @a, align 4
+  %0 = load i32, i32* @a, align 4
   %cmp = icmp ne i32 %0, 10
-  %1 = load i32* @i, align 4
-  %2 = load i32* @j, align 4
+  %1 = load i32, i32* @i, align 4
+  %2 = load i32, i32* @j, align 4
   %cond = select i1 %cmp, i32 %1, i32 %2
   store i32 %cond, i32* @i, align 4
   ret void
diff --git a/test/CodeGen/Mips/selTBtnezSlti.ll b/test/CodeGen/Mips/selTBtnezSlti.ll
index 593f6f274eb3..2f1cdb866294 100644
--- a/test/CodeGen/Mips/selTBtnezSlti.ll
+++ b/test/CodeGen/Mips/selTBtnezSlti.ll
@@ -8,10 +8,10 @@
 
 define void @t() nounwind "target-cpu"="mips16" "target-features"="+mips16,+o32" {
 entry:
-  %0 = load i32* @a, align 4
+  %0 = load i32, i32* @a, align 4
   %cmp = icmp slt i32 %0, 10
-  %1 = load i32* @j, align 4
-  %2 = load i32* @i, align 4
+  %1 = load i32, i32* @j, align 4
+  %2 = load i32, i32* @i, align 4
   %cond = select i1 %cmp, i32 %1, i32 %2
   store i32 %cond, i32* @i, align 4
   ret void
diff --git a/test/CodeGen/Mips/select.ll b/test/CodeGen/Mips/select.ll
index d6e1826c30c8..96bd3782c058 100644
--- a/test/CodeGen/Mips/select.ll
+++ b/test/CodeGen/Mips/select.ll
@@ -700,8 +700,8 @@ entry:
 ; 64R6:          selnez $[[NE:[0-9]+]], $4, $[[CCGPR]]
 ; 64R6:          or $2, $[[NE]], $[[EQ]]
 
-  %tmp = load double* @d2, align 8
-  %tmp1 = load double* @d3, align 8
+  %tmp = load double, double* @d2, align 8
+  %tmp1 = load double, double* @d3, align 8
   %cmp = fcmp oeq double %tmp, %tmp1
   %cond = select i1 %cmp, i32 %f0, i32 %f1
   ret i32 %cond
@@ -777,8 +777,8 @@ entry:
 ; 64R6:          selnez $[[NE:[0-9]+]], $4, $[[CCGPR]]
 ; 64R6:          or $2, $[[NE]], $[[EQ]]
 
-  %tmp = load double* @d2, align 8
-  %tmp1 = load double* @d3, align 8
+  %tmp = load double, double* @d2, align 8
+  %tmp1 = load double, double* @d3, align 8
   %cmp = fcmp olt double %tmp, %tmp1
   %cond = select i1 %cmp, i32 %f0, i32 %f1
   ret i32 %cond
@@ -854,8 +854,8 @@ entry:
 ; 64R6:          selnez $[[NE:[0-9]+]], $4, $[[CCGPR]]
 ; 64R6:          or $2, $[[NE]], $[[EQ]]
 
-  %tmp = load double* @d2, align 8
-  %tmp1 = load double* @d3, align 8
+  %tmp = load double, double* @d2, align 8
+  %tmp1 = load double, double* @d3, align 8
   %cmp = fcmp ogt double %tmp, %tmp1
   %cond = select i1 %cmp, i32 %f0, i32 %f1
   ret i32 %cond
diff --git a/test/CodeGen/Mips/seleq.ll b/test/CodeGen/Mips/seleq.ll
index 9af422fa1bdb..bd25358fb9e6 100644
--- a/test/CodeGen/Mips/seleq.ll
+++ b/test/CodeGen/Mips/seleq.ll
@@ -12,65 +12,65 @@
 
 define void @calc_seleq() nounwind {
 entry:
-  %0 = load i32* @a, align 4
-  %1 = load i32* @b, align 4
+  %0 = load i32, i32* @a, align 4
+  %1 = load i32, i32* @b, align 4
   %cmp = icmp eq i32 %0, %1
   br i1 %cmp, label %cond.true, label %cond.false
 
 cond.true:                                        ; preds = %entry
-  %2 = load i32* @f, align 4
+  %2 = load i32, i32* @f, align 4
   br label %cond.end
 
 cond.false:                                       ; preds = %entry
-  %3 = load i32* @t, align 4
+  %3 = load i32, i32* @t, align 4
   br label %cond.end
 
 cond.end:                                         ; preds = %cond.false, %cond.true
   %cond = phi i32 [ %2, %cond.true ], [ %3, %cond.false ]
   store i32 %cond, i32* @z1, align 4
-  %4 = load i32* @b, align 4
-  %5 = load i32* @a, align 4
+  %4 = load i32, i32* @b, align 4
+  %5 = load i32, i32* @a, align 4
   %cmp1 = icmp eq i32 %4, %5
   br i1 %cmp1, label %cond.true2, label %cond.false3
 
 cond.true2:                                       ; preds = %cond.end
-  %6 = load i32* @f, align 4
+  %6 = load i32, i32* @f, align 4
   br label %cond.end4
 
 cond.false3:                                      ; preds = %cond.end
-  %7 = load i32* @t, align 4
+  %7 = load i32, i32* @t, align 4
   br label %cond.end4
 
 cond.end4:                                        ; preds = %cond.false3, %cond.true2
   %cond5 = phi i32 [ %6, %cond.true2 ], [ %7, %cond.false3 ]
   store i32 %cond5, i32* @z2, align 4
-  %8 = load i32* @c, align 4
-  %9 = load i32* @a, align 4
+  %8 = load i32, i32* @c, align 4
+  %9 = load i32, i32* @a, align 4
   %cmp6 = icmp eq i32 %8, %9
   br i1 %cmp6, label %cond.true7, label %cond.false8
 
 cond.true7:                                       ; preds = %cond.end4
-  %10 = load i32* @t, align 4
+  %10 = load i32, i32* @t, align 4
   br label %cond.end9
 
 cond.false8:                                      ; preds = %cond.end4
-  %11 = load i32* @f, align 4
+  %11 = load i32, i32* @f, align 4
   br label %cond.end9
 
 cond.end9:                                        ; preds = %cond.false8, %cond.true7
   %cond10 = phi i32 [ %10, %cond.true7 ], [ %11, %cond.false8 ]
   store i32 %cond10, i32* @z3, align 4
-  %12 = load i32* @a, align 4
-  %13 = load i32* @c, align 4
+  %12 = load i32, i32* @a, align 4
+  %13 = load i32, i32* @c, align 4
   %cmp11 = icmp eq i32 %12, %13
   br i1 %cmp11, label %cond.true12, label %cond.false13
 
 cond.true12:                                      ; preds = %cond.end9
-  %14 = load i32* @t, align 4
+  %14 = load i32, i32* @t, align 4
   br label %cond.end14
 
 cond.false13:                                     ; preds = %cond.end9
-  %15 = load i32* @f, align 4
+  %15 = load i32, i32* @f, align 4
   br label %cond.end14
 
 cond.end14:                                       ; preds = %cond.false13, %cond.true12
diff --git a/test/CodeGen/Mips/seleqk.ll b/test/CodeGen/Mips/seleqk.ll
index 3ca622d5d8fe..2eeaa9e33738 100644
--- a/test/CodeGen/Mips/seleqk.ll
+++ b/test/CodeGen/Mips/seleqk.ll
@@ -12,61 +12,61 @@
 
 define void @calc_seleqk() nounwind "target-cpu"="mips16" "target-features"="+mips16,+o32" {
 entry:
-  %0 = load i32* @a, align 4
+  %0 = load i32, i32* @a, align 4
   %cmp = icmp eq i32 %0, 1
   br i1 %cmp, label %cond.true, label %cond.false
 
 cond.true:                                        ; preds = %entry
-  %1 = load i32* @t, align 4
+  %1 = load i32, i32* @t, align 4
   br label %cond.end
 
 cond.false:                                       ; preds = %entry
-  %2 = load i32* @f, align 4
+  %2 = load i32, i32* @f, align 4
   br label %cond.end
 
 cond.end:                                         ; preds = %cond.false, %cond.true
   %cond = phi i32 [ %1, %cond.true ], [ %2, %cond.false ]
   store i32 %cond, i32* @z1, align 4
-  %3 = load i32* @a, align 4
+  %3 = load i32, i32* @a, align 4
   %cmp1 = icmp eq i32 %3, 1000
   br i1 %cmp1, label %cond.true2, label %cond.false3
 
 cond.true2:                                       ; preds = %cond.end
-  %4 = load i32* @f, align 4
+  %4 = load i32, i32* @f, align 4
   br label %cond.end4
 
 cond.false3:                                      ; preds = %cond.end
-  %5 = load i32* @t, align 4
+  %5 = load i32, i32* @t, align 4
   br label %cond.end4
 
 cond.end4:                                        ; preds = %cond.false3, %cond.true2
   %cond5 = phi i32 [ %4, %cond.true2 ], [ %5, %cond.false3 ]
   store i32 %cond5, i32* @z2, align 4
-  %6 = load i32* @b, align 4
+  %6 = load i32, i32* @b, align 4
   %cmp6 = icmp eq i32 %6, 3
   br i1 %cmp6, label %cond.true7, label %cond.false8
 
 cond.true7:                                       ; preds = %cond.end4
-  %7 = load i32* @f, align 4
+  %7 = load i32, i32* @f, align 4
   br label %cond.end9
 
 cond.false8:                                      ; preds = %cond.end4
-  %8 = load i32* @t, align 4
+  %8 = load i32, i32* @t, align 4
   br label %cond.end9
 
 cond.end9:                                        ; preds = %cond.false8, %cond.true7
   %cond10 = phi i32 [ %7, %cond.true7 ], [ %8, %cond.false8 ]
   store i32 %cond10, i32* @z3, align 4
-  %9 = load i32* @b, align 4
+  %9 = load i32, i32* @b, align 4
   %cmp11 = icmp eq i32 %9, 1000
   br i1 %cmp11, label %cond.true12, label %cond.false13
 
 cond.true12:                                      ; preds = %cond.end9
-  %10 = load i32* @t, align 4
+  %10 = load i32, i32* @t, align 4
   br label %cond.end14
 
 cond.false13:                                     ; preds = %cond.end9
-  %11 = load i32* @f, align 4
+  %11 = load i32, i32* @f, align 4
   br label %cond.end14
 
 cond.end14:                                       ; preds = %cond.false13, %cond.true12
diff --git a/test/CodeGen/Mips/selgek.ll b/test/CodeGen/Mips/selgek.ll
index 8ab4046e92cb..38ad95ee01a9 100644
--- a/test/CodeGen/Mips/selgek.ll
+++ b/test/CodeGen/Mips/selgek.ll
@@ -13,61 +13,61 @@
 
 define void @calc_z() nounwind "target-cpu"="mips16" "target-features"="+mips16,+o32" {
 entry:
-  %0 = load i32* @a, align 4
+  %0 = load i32, i32* @a, align 4
   %cmp = icmp sge i32 %0, 1000
   br i1 %cmp, label %cond.true, label %cond.false
 
 cond.true:                                        ; preds = %entry
-  %1 = load i32* @f, align 4
+  %1 = load i32, i32* @f, align 4
   br label %cond.end
 
 cond.false:                                       ; preds = %entry
-  %2 = load i32* @t, align 4
+  %2 = load i32, i32* @t, align 4
   br label %cond.end
 
 cond.end:                                         ; preds = %cond.false, %cond.true
   %cond = phi i32 [ %1, %cond.true ], [ %2, %cond.false ]
   store i32 %cond, i32* @z1, align 4
-  %3 = load i32* @b, align 4
+  %3 = load i32, i32* @b, align 4
   %cmp1 = icmp sge i32 %3, 1
   br i1 %cmp1, label %cond.true2, label %cond.false3
 
 cond.true2:                                       ; preds = %cond.end
-  %4 = load i32* @t, align 4
+  %4 = load i32, i32* @t, align 4
   br label %cond.end4
 
 cond.false3:                                      ; preds = %cond.end
-  %5 = load i32* @f, align 4
+  %5 = load i32, i32* @f, align 4
   br label %cond.end4
 
 cond.end4:                                        ; preds = %cond.false3, %cond.true2
   %cond5 = phi i32 [ %4, %cond.true2 ], [ %5, %cond.false3 ]
   store i32 %cond5, i32* @z2, align 4
-  %6 = load i32* @c, align 4
+  %6 = load i32, i32* @c, align 4
   %cmp6 = icmp sge i32 %6, 2
   br i1 %cmp6, label %cond.true7, label %cond.false8
 
 cond.true7:                                       ; preds = %cond.end4
-  %7 = load i32* @t, align 4
+  %7 = load i32, i32* @t, align 4
   br label %cond.end9
 
 cond.false8:                                      ; preds = %cond.end4
-  %8 = load i32* @f, align 4
+  %8 = load i32, i32* @f, align 4
   br label %cond.end9
 
 cond.end9:                                        ; preds = %cond.false8, %cond.true7
   %cond10 = phi i32 [ %7, %cond.true7 ], [ %8, %cond.false8 ]
   store i32 %cond10, i32* @z3, align 4
-  %9 = load i32* @a, align 4
+  %9 = load i32, i32* @a, align 4
   %cmp11 = icmp sge i32 %9, 2
   br i1 %cmp11, label %cond.true12, label %cond.false13
 
 cond.true12:                                      ; preds = %cond.end9
-  %10 = load i32* @t, align 4
+  %10 = load i32, i32* @t, align 4
   br label %cond.end14
 
 cond.false13:                                     ; preds = %cond.end9
-  %11 = load i32* @f, align 4
+  %11 = load i32, i32* @f, align 4
   br label %cond.end14
 
 cond.end14:                                       ; preds = %cond.false13, %cond.true12
diff --git a/test/CodeGen/Mips/selgt.ll b/test/CodeGen/Mips/selgt.ll
index 67b9b498709b..a2e1e39e742f 100644
--- a/test/CodeGen/Mips/selgt.ll
+++ b/test/CodeGen/Mips/selgt.ll
@@ -14,71 +14,71 @@
 define i32 @calc_z() nounwind "target-cpu"="mips16" "target-features"="+mips16,+o32" {
 entry:
   %retval = alloca i32, align 4
-  %0 = load i32* @a, align 4
-  %1 = load i32* @b, align 4
+  %0 = load i32, i32* @a, align 4
+  %1 = load i32, i32* @b, align 4
   %cmp = icmp sgt i32 %0, %1
   br i1 %cmp, label %cond.true, label %cond.false
 
 cond.true:                                        ; preds = %entry
-  %2 = load i32* @f, align 4
+  %2 = load i32, i32* @f, align 4
   br label %cond.end
 
 cond.false:                                       ; preds = %entry
-  %3 = load i32* @t, align 4
+  %3 = load i32, i32* @t, align 4
   br label %cond.end
 
 cond.end:                                         ; preds = %cond.false, %cond.true
   %cond = phi i32 [ %2, %cond.true ], [ %3, %cond.false ]
   store i32 %cond, i32* @z1, align 4
-  %4 = load i32* @b, align 4
-  %5 = load i32* @a, align 4
+  %4 = load i32, i32* @b, align 4
+  %5 = load i32, i32* @a, align 4
   %cmp1 = icmp sgt i32 %4, %5
   br i1 %cmp1, label %cond.true2, label %cond.false3
 
 cond.true2:                                       ; preds = %cond.end
-  %6 = load i32* @t, align 4
+  %6 = load i32, i32* @t, align 4
   br label %cond.end4
 
 cond.false3:                                      ; preds = %cond.end
-  %7 = load i32* @f, align 4
+  %7 = load i32, i32* @f, align 4
   br label %cond.end4
 
 cond.end4:                                        ; preds = %cond.false3, %cond.true2
   %cond5 = phi i32 [ %6, %cond.true2 ], [ %7, %cond.false3 ]
   store i32 %cond5, i32* @z2, align 4
-  %8 = load i32* @c, align 4
-  %9 = load i32* @a, align 4
+  %8 = load i32, i32* @c, align 4
+  %9 = load i32, i32* @a, align 4
   %cmp6 = icmp sgt i32 %8, %9
   br i1 %cmp6, label %cond.true7, label %cond.false8
 
 cond.true7:                                       ; preds = %cond.end4
-  %10 = load i32* @f, align 4
+  %10 = load i32, i32* @f, align 4
   br label %cond.end9
 
 cond.false8:                                      ; preds = %cond.end4
-  %11 = load i32* @t, align 4
+  %11 = load i32, i32* @t, align 4
   br label %cond.end9
 
 cond.end9:                                        ; preds = %cond.false8, %cond.true7
   %cond10 = phi i32 [ %10, %cond.true7 ], [ %11, %cond.false8 ]
   store i32 %cond10, i32* @z3, align 4
-  %12 = load i32* @a, align 4
-  %13 = load i32* @c, align 4
+  %12 = load i32, i32* @a, align 4
+  %13 = load i32, i32* @c, align 4
   %cmp11 = icmp sgt i32 %12, %13
   br i1 %cmp11, label %cond.true12, label %cond.false13
 
 cond.true12:                                      ; preds = %cond.end9
-  %14 = load i32* @f, align 4
+  %14 = load i32, i32* @f, align 4
   br label %cond.end14
 
 cond.false13:                                     ; preds = %cond.end9
-  %15 = load i32* @t, align 4
+  %15 = load i32, i32* @t, align 4
   br label %cond.end14
 
 cond.end14:                                       ; preds = %cond.false13, %cond.true12
   %cond15 = phi i32 [ %14, %cond.true12 ], [ %15, %cond.false13 ]
   store i32 %cond15, i32* @z4, align 4
-  %16 = load i32* %retval
+  %16 = load i32, i32* %retval
   ret i32 %16
 }
 
diff --git a/test/CodeGen/Mips/selle.ll b/test/CodeGen/Mips/selle.ll
index b27df45e6739..1adefb7846e4 100644
--- a/test/CodeGen/Mips/selle.ll
+++ b/test/CodeGen/Mips/selle.ll
@@ -13,65 +13,65 @@
 
 define void @calc_z() nounwind "target-cpu"="mips16" "target-features"="+mips16,+o32" {
 entry:
-  %0 = load i32* @a, align 4
-  %1 = load i32* @b, align 4
+  %0 = load i32, i32* @a, align 4
+  %1 = load i32, i32* @b, align 4
   %cmp = icmp sle i32 %0, %1
   br i1 %cmp, label %cond.true, label %cond.false
 
 cond.true:                                        ; preds = %entry
-  %2 = load i32* @t, align 4
+  %2 = load i32, i32* @t, align 4
   br label %cond.end
 
 cond.false:                                       ; preds = %entry
-  %3 = load i32* @f, align 4
+  %3 = load i32, i32* @f, align 4
   br label %cond.end
 
 cond.end:                                         ; preds = %cond.false, %cond.true
   %cond = phi i32 [ %2, %cond.true ], [ %3, %cond.false ]
   store i32 %cond, i32* @z1, align 4
-  %4 = load i32* @b, align 4
-  %5 = load i32* @a, align 4
+  %4 = load i32, i32* @b, align 4
+  %5 = load i32, i32* @a, align 4
   %cmp1 = icmp sle i32 %4, %5
   br i1 %cmp1, label %cond.true2, label %cond.false3
 
 cond.true2:                                       ; preds = %cond.end
-  %6 = load i32* @f, align 4
+  %6 = load i32, i32* @f, align 4
   br label %cond.end4
 
 cond.false3:                                      ; preds = %cond.end
-  %7 = load i32* @t, align 4
+  %7 = load i32, i32* @t, align 4
   br label %cond.end4
 
 cond.end4:                                        ; preds = %cond.false3, %cond.true2
   %cond5 = phi i32 [ %6, %cond.true2 ], [ %7, %cond.false3 ]
   store i32 %cond5, i32* @z2, align 4
-  %8 = load i32* @c, align 4
-  %9 = load i32* @a, align 4
+  %8 = load i32, i32* @c, align 4
+  %9 = load i32, i32* @a, align 4
   %cmp6 = icmp sle i32 %8, %9
   br i1 %cmp6, label %cond.true7, label %cond.false8
 
 cond.true7:                                       ; preds = %cond.end4
-  %10 = load i32* @t, align 4
+  %10 = load i32, i32* @t, align 4
   br label %cond.end9
 
 cond.false8:                                      ; preds = %cond.end4
-  %11 = load i32* @f, align 4
+  %11 = load i32, i32* @f, align 4
   br label %cond.end9
 
 cond.end9:                                        ; preds = %cond.false8, %cond.true7
   %cond10 = phi i32 [ %10, %cond.true7 ], [ %11, %cond.false8 ]
   store i32 %cond10, i32* @z3, align 4
-  %12 = load i32* @a, align 4
-  %13 = load i32* @c, align 4
+  %12 = load i32, i32* @a, align 4
+  %13 = load i32, i32* @c, align 4
   %cmp11 = icmp sle i32 %12, %13
   br i1 %cmp11, label %cond.true12, label %cond.false13
 
 cond.true12:                                      ; preds = %cond.end9
-  %14 = load i32* @t, align 4
+  %14 = load i32, i32* @t, align 4
   br label %cond.end14
 
 cond.false13:                                     ; preds = %cond.end9
-  %15 = load i32* @f, align 4
+  %15 = load i32, i32* @f, align 4
   br label %cond.end14
 
 cond.end14:                                       ; preds = %cond.false13, %cond.true12
diff --git a/test/CodeGen/Mips/selltk.ll b/test/CodeGen/Mips/selltk.ll
index 1471b892c92a..db9f8c171b78 100644
--- a/test/CodeGen/Mips/selltk.ll
+++ b/test/CodeGen/Mips/selltk.ll
@@ -13,61 +13,61 @@
 
 define void @calc_selltk() nounwind "target-cpu"="mips16" "target-features"="+mips16,+o32" {
 entry:
-  %0 = load i32* @a, align 4
+  %0 = load i32, i32* @a, align 4
   %cmp = icmp slt i32 %0, 1000
   br i1 %cmp, label %cond.true, label %cond.false
 
 cond.true:                                        ; preds = %entry
-  %1 = load i32* @t, align 4
+  %1 = load i32, i32* @t, align 4
   br label %cond.end
 
 cond.false:                                       ; preds = %entry
-  %2 = load i32* @f, align 4
+  %2 = load i32, i32* @f, align 4
   br label %cond.end
 
 cond.end:                                         ; preds = %cond.false, %cond.true
   %cond = phi i32 [ %1, %cond.true ], [ %2, %cond.false ]
   store i32 %cond, i32* @z1, align 4
-  %3 = load i32* @b, align 4
+  %3 = load i32, i32* @b, align 4
   %cmp1 = icmp slt i32 %3, 2
   br i1 %cmp1, label %cond.true2, label %cond.false3
 
 cond.true2:                                       ; preds = %cond.end
-  %4 = load i32* @f, align 4
+  %4 = load i32, i32* @f, align 4
   br label %cond.end4
 
 cond.false3:                                      ; preds = %cond.end
-  %5 = load i32* @t, align 4
+  %5 = load i32, i32* @t, align 4
   br label %cond.end4
 
 cond.end4:                                        ; preds = %cond.false3, %cond.true2
   %cond5 = phi i32 [ %4, %cond.true2 ], [ %5, %cond.false3 ]
   store i32 %cond5, i32* @z2, align 4
-  %6 = load i32* @c, align 4
+  %6 = load i32, i32* @c, align 4
   %cmp6 = icmp sgt i32 %6, 2
   br i1 %cmp6, label %cond.true7, label %cond.false8
 
 cond.true7:                                       ; preds = %cond.end4
-  %7 = load i32* @f, align 4
+  %7 = load i32, i32* @f, align 4
   br label %cond.end9
 
 cond.false8:                                      ; preds = %cond.end4
-  %8 = load i32* @t, align 4
+  %8 = load i32, i32* @t, align 4
   br label %cond.end9
 
 cond.end9:                                        ; preds = %cond.false8, %cond.true7
   %cond10 = phi i32 [ %7, %cond.true7 ], [ %8, %cond.false8 ]
   store i32 %cond10, i32* @z3, align 4
-  %9 = load i32* @a, align 4
+  %9 = load i32, i32* @a, align 4
   %cmp11 = icmp sgt i32 %9, 2
   br i1 %cmp11, label %cond.true12, label %cond.false13
 
 cond.true12:                                      ; preds = %cond.end9
-  %10 = load i32* @f, align 4
+  %10 = load i32, i32* @f, align 4
   br label %cond.end14
 
 cond.false13:                                     ; preds = %cond.end9
-  %11 = load i32* @t, align 4
+  %11 = load i32, i32* @t, align 4
   br label %cond.end14
 
 cond.end14:                                       ; preds = %cond.false13, %cond.true12
diff --git a/test/CodeGen/Mips/selne.ll b/test/CodeGen/Mips/selne.ll
index e3d82b8cf5d0..9be99d669475 100644
--- a/test/CodeGen/Mips/selne.ll
+++ b/test/CodeGen/Mips/selne.ll
@@ -13,65 +13,65 @@
 
 define void @calc_seleq() nounwind "target-cpu"="mips16" "target-features"="+mips16,+o32" {
 entry:
-  %0 = load i32* @a, align 4
-  %1 = load i32* @b, align 4
+  %0 = load i32, i32* @a, align 4
+  %1 = load i32, i32* @b, align 4
   %cmp = icmp ne i32 %0, %1
   br i1 %cmp, label %cond.true, label %cond.false
 
 cond.true:                                        ; preds = %entry
-  %2 = load i32* @f, align 4
+  %2 = load i32, i32* @f, align 4
   br label %cond.end
 
 cond.false:                                       ; preds = %entry
-  %3 = load i32* @t, align 4
+  %3 = load i32, i32* @t, align 4
   br label %cond.end
 
 cond.end:                                         ; preds = %cond.false, %cond.true
   %cond = phi i32 [ %2, %cond.true ], [ %3, %cond.false ]
   store i32 %cond, i32* @z1, align 4
-  %4 = load i32* @b, align 4
-  %5 = load i32* @a, align 4
+  %4 = load i32, i32* @b, align 4
+  %5 = load i32, i32* @a, align 4
   %cmp1 = icmp ne i32 %4, %5
   br i1 %cmp1, label %cond.true2, label %cond.false3
 
 cond.true2:                                       ; preds = %cond.end
-  %6 = load i32* @f, align 4
+  %6 = load i32, i32* @f, align 4
   br label %cond.end4
 
 cond.false3:                                      ; preds = %cond.end
-  %7 = load i32* @t, align 4
+  %7 = load i32, i32* @t, align 4
   br label %cond.end4
 
 cond.end4:                                        ; preds = %cond.false3, %cond.true2
   %cond5 = phi i32 [ %6, %cond.true2 ], [ %7, %cond.false3 ]
   store i32 %cond5, i32* @z2, align 4
-  %8 = load i32* @c, align 4
-  %9 = load i32* @a, align 4
+  %8 = load i32, i32* @c, align 4
+  %9 = load i32, i32* @a, align 4
   %cmp6 = icmp ne i32 %8, %9
   br i1 %cmp6, label %cond.true7, label %cond.false8
 
 cond.true7:                                       ; preds = %cond.end4
-  %10 = load i32* @t, align 4
+  %10 = load i32, i32* @t, align 4
   br label %cond.end9
 
 cond.false8:                                      ; preds = %cond.end4
-  %11 = load i32* @f, align 4
+  %11 = load i32, i32* @f, align 4
   br label %cond.end9
 
 cond.end9:                                        ; preds = %cond.false8, %cond.true7
   %cond10 = phi i32 [ %10, %cond.true7 ], [ %11, %cond.false8 ]
   store i32 %cond10, i32* @z3, align 4
-  %12 = load i32* @a, align 4
-  %13 = load i32* @c, align 4
+  %12 = load i32, i32* @a, align 4
+  %13 = load i32, i32* @c, align 4
   %cmp11 = icmp ne i32 %12, %13
   br i1 %cmp11, label %cond.true12, label %cond.false13
 
 cond.true12:                                      ; preds = %cond.end9
-  %14 = load i32* @t, align 4
+  %14 = load i32, i32* @t, align 4
   br label %cond.end14
 
 cond.false13:                                     ; preds = %cond.end9
-  %15 = load i32* @f, align 4
+  %15 = load i32, i32* @f, align 4
   br label %cond.end14
 
 cond.end14:                                       ; preds = %cond.false13, %cond.true12
diff --git a/test/CodeGen/Mips/selnek.ll b/test/CodeGen/Mips/selnek.ll
index 64834b256fe5..5b6aa2afa1af 100644
--- a/test/CodeGen/Mips/selnek.ll
+++ b/test/CodeGen/Mips/selnek.ll
@@ -12,61 +12,61 @@
 
 define void @calc_z() nounwind "target-cpu"="mips16" "target-features"="+mips16,+o32" {
 entry:
-  %0 = load i32* @a, align 4
+  %0 = load i32, i32* @a, align 4
   %cmp = icmp ne i32 %0, 1
   br i1 %cmp, label %cond.true, label %cond.false
 
 cond.true:                                        ; preds = %entry
-  %1 = load i32* @f, align 4
+  %1 = load i32, i32* @f, align 4
   br label %cond.end
 
 cond.false:                                       ; preds = %entry
-  %2 = load i32* @t, align 4
+  %2 = load i32, i32* @t, align 4
   br label %cond.end
 
 cond.end:                                         ; preds = %cond.false, %cond.true
   %cond = phi i32 [ %1, %cond.true ], [ %2, %cond.false ]
   store i32 %cond, i32* @z1, align 4
-  %3 = load i32* @a, align 4
+  %3 = load i32, i32* @a, align 4
   %cmp1 = icmp ne i32 %3, 1000
   br i1 %cmp1, label %cond.true2, label %cond.false3
 
 cond.true2:                                       ; preds = %cond.end
-  %4 = load i32* @t, align 4
+  %4 = load i32, i32* @t, align 4
   br label %cond.end4
 
 cond.false3:                                      ; preds = %cond.end
-  %5 = load i32* @f, align 4
+  %5 = load i32, i32* @f, align 4
   br label %cond.end4
 
 cond.end4:                                        ; preds = %cond.false3, %cond.true2
   %cond5 = phi i32 [ %4, %cond.true2 ], [ %5, %cond.false3 ]
   store i32 %cond5, i32* @z2, align 4
-  %6 = load i32* @b, align 4
+  %6 = load i32, i32* @b, align 4
   %cmp6 = icmp ne i32 %6, 3
   br i1 %cmp6, label %cond.true7, label %cond.false8
 
 cond.true7:                                       ; preds = %cond.end4
-  %7 = load i32* @t, align 4
+  %7 = load i32, i32* @t, align 4
   br label %cond.end9
 
 cond.false8:                                      ; preds = %cond.end4
-  %8 = load i32* @f, align 4
+  %8 = load i32, i32* @f, align 4
   br label %cond.end9
 
 cond.end9:                                        ; preds = %cond.false8, %cond.true7
   %cond10 = phi i32 [ %7, %cond.true7 ], [ %8, %cond.false8 ]
   store i32 %cond10, i32* @z3, align 4
-  %9 = load i32* @b, align 4
+  %9 = load i32, i32* @b, align 4
   %cmp11 = icmp ne i32 %9, 1000
   br i1 %cmp11, label %cond.true12, label %cond.false13
 
 cond.true12:                                      ; preds = %cond.end9
-  %10 = load i32* @f, align 4
+  %10 = load i32, i32* @f, align 4
   br label %cond.end14
 
 cond.false13:                                     ; preds = %cond.end9
-  %11 = load i32* @t, align 4
+  %11 = load i32, i32* @t, align 4
   br label %cond.end14
 
 cond.end14:                                       ; preds = %cond.false13, %cond.true12
@@ -78,14 +78,14 @@ cond.end14:                                       ; preds = %cond.false13, %cond
 define i32 @main() nounwind "target-cpu"="mips16" "target-features"="+mips16,+o32" {
 entry:
   call void @calc_z() "target-cpu"="mips16" "target-features"="+mips16,+o32"
-  %0 = load i32* @z1, align 4
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i32 0, i32 0), i32 %0) "target-cpu"="mips16" "target-features"="+mips16,+o32"
-  %1 = load i32* @z2, align 4
-  %call1 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i32 0, i32 0), i32 %1) "target-cpu"="mips16" "target-features"="+mips16,+o32"
-  %2 = load i32* @z3, align 4
-  %call2 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i32 0, i32 0), i32 %2) "target-cpu"="mips16" "target-features"="+mips16,+o32"
-  %3 = load i32* @z4, align 4
-  %call3 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i32 0, i32 0), i32 %3) "target-cpu"="mips16" "target-features"="+mips16,+o32"
+  %0 = load i32, i32* @z1, align 4
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([5 x i8], [5 x i8]* @.str, i32 0, i32 0), i32 %0) "target-cpu"="mips16" "target-features"="+mips16,+o32"
+  %1 = load i32, i32* @z2, align 4
+  %call1 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([5 x i8], [5 x i8]* @.str, i32 0, i32 0), i32 %1) "target-cpu"="mips16" "target-features"="+mips16,+o32"
+  %2 = load i32, i32* @z3, align 4
+  %call2 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([5 x i8], [5 x i8]* @.str, i32 0, i32 0), i32 %2) "target-cpu"="mips16" "target-features"="+mips16,+o32"
+  %3 = load i32, i32* @z4, align 4
+  %call3 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([5 x i8], [5 x i8]* @.str, i32 0, i32 0), i32 %3) "target-cpu"="mips16" "target-features"="+mips16,+o32"
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/selpat.ll b/test/CodeGen/Mips/selpat.ll
index 8eda8de45e08..c682d8182a46 100644
--- a/test/CodeGen/Mips/selpat.ll
+++ b/test/CodeGen/Mips/selpat.ll
@@ -12,18 +12,18 @@
 
 define void @calc_seleq() nounwind {
 entry:
-  %0 = load i32* @a, align 4
-  %1 = load i32* @b, align 4
+  %0 = load i32, i32* @a, align 4
+  %1 = load i32, i32* @b, align 4
   %cmp = icmp eq i32 %0, %1
-  %2 = load i32* @f, align 4
-  %3 = load i32* @t, align 4
+  %2 = load i32, i32* @f, align 4
+  %3 = load i32, i32* @t, align 4
   %cond = select i1 %cmp, i32 %2, i32 %3
   store i32 %cond, i32* @z1, align 4
 ; 16:	cmp	${{[0-9]+}}, ${{[0-9]+}}
 ; 16:	bteqz	$BB{{[0-9]+}}_{{[0-9]}}
 ; 16: 	move    ${{[0-9]+}}, ${{[0-9]+}}
   store i32 %cond, i32* @z2, align 4
-  %4 = load i32* @c, align 4
+  %4 = load i32, i32* @c, align 4
   %cmp6 = icmp eq i32 %4, %0
   %cond10 = select i1 %cmp6, i32 %3, i32 %2
   store i32 %cond10, i32* @z3, align 4
@@ -34,10 +34,10 @@ entry:
 
 define void @calc_seleqk() nounwind {
 entry:
-  %0 = load i32* @a, align 4
+  %0 = load i32, i32* @a, align 4
   %cmp = icmp eq i32 %0, 1
-  %1 = load i32* @t, align 4
-  %2 = load i32* @f, align 4
+  %1 = load i32, i32* @t, align 4
+  %2 = load i32, i32* @f, align 4
   %cond = select i1 %cmp, i32 %1, i32 %2
   store i32 %cond, i32* @z1, align 4
 ; 16:	cmpi	${{[0-9]+}}, 1
@@ -46,7 +46,7 @@ entry:
   %cmp1 = icmp eq i32 %0, 10
   %cond5 = select i1 %cmp1, i32 %2, i32 %1
   store i32 %cond5, i32* @z2, align 4
-  %3 = load i32* @b, align 4
+  %3 = load i32, i32* @b, align 4
   %cmp6 = icmp eq i32 %3, 3
   %cond10 = select i1 %cmp6, i32 %2, i32 %1
   store i32 %cond10, i32* @z3, align 4
@@ -61,19 +61,19 @@ entry:
 
 define void @calc_seleqz() nounwind {
 entry:
-  %0 = load i32* @a, align 4
+  %0 = load i32, i32* @a, align 4
   %cmp = icmp eq i32 %0, 0
-  %1 = load i32* @t, align 4
-  %2 = load i32* @f, align 4
+  %1 = load i32, i32* @t, align 4
+  %2 = load i32, i32* @f, align 4
   %cond = select i1 %cmp, i32 %1, i32 %2
   store i32 %cond, i32* @z1, align 4
 ; 16:	beqz	${{[0-9]+}}, $BB{{[0-9]+}}_{{[0-9]}}
 ; 16: 	move    ${{[0-9]+}}, ${{[0-9]+}}
-  %3 = load i32* @b, align 4
+  %3 = load i32, i32* @b, align 4
   %cmp1 = icmp eq i32 %3, 0
   %cond5 = select i1 %cmp1, i32 %2, i32 %1
   store i32 %cond5, i32* @z2, align 4
-  %4 = load i32* @c, align 4
+  %4 = load i32, i32* @c, align 4
   %cmp6 = icmp eq i32 %4, 0
   %cond10 = select i1 %cmp6, i32 %1, i32 %2
   store i32 %cond10, i32* @z3, align 4
@@ -83,11 +83,11 @@ entry:
 
 define void @calc_selge() nounwind {
 entry:
-  %0 = load i32* @a, align 4
-  %1 = load i32* @b, align 4
+  %0 = load i32, i32* @a, align 4
+  %1 = load i32, i32* @b, align 4
   %cmp = icmp sge i32 %0, %1
-  %2 = load i32* @f, align 4
-  %3 = load i32* @t, align 4
+  %2 = load i32, i32* @f, align 4
+  %3 = load i32, i32* @t, align 4
   %cond = select i1 %cmp, i32 %2, i32 %3
   store i32 %cond, i32* @z1, align 4
 ; 16:	slt	${{[0-9]+}}, ${{[0-9]+}}
@@ -96,7 +96,7 @@ entry:
   %cmp1 = icmp sge i32 %1, %0
   %cond5 = select i1 %cmp1, i32 %3, i32 %2
   store i32 %cond5, i32* @z2, align 4
-  %4 = load i32* @c, align 4
+  %4 = load i32, i32* @c, align 4
   %cmp6 = icmp sge i32 %4, %0
   %cond10 = select i1 %cmp6, i32 %3, i32 %2
   store i32 %cond10, i32* @z3, align 4
@@ -108,20 +108,20 @@ entry:
 
 define i32 @calc_selgt() nounwind {
 entry:
-  %0 = load i32* @a, align 4
-  %1 = load i32* @b, align 4
+  %0 = load i32, i32* @a, align 4
+  %1 = load i32, i32* @b, align 4
   %cmp = icmp sgt i32 %0, %1
 ; 16:	slt	${{[0-9]+}}, ${{[0-9]+}}
 ; 16:	btnez	$BB{{[0-9]+}}_{{[0-9]}}
 ; 16: 	move    ${{[0-9]+}}, ${{[0-9]+}}
-  %2 = load i32* @f, align 4
-  %3 = load i32* @t, align 4
+  %2 = load i32, i32* @f, align 4
+  %3 = load i32, i32* @t, align 4
   %cond = select i1 %cmp, i32 %2, i32 %3
   store i32 %cond, i32* @z1, align 4
   %cmp1 = icmp sgt i32 %1, %0
   %cond5 = select i1 %cmp1, i32 %3, i32 %2
   store i32 %cond5, i32* @z2, align 4
-  %4 = load i32* @c, align 4
+  %4 = load i32, i32* @c, align 4
   %cmp6 = icmp sgt i32 %4, %0
   %cond10 = select i1 %cmp6, i32 %2, i32 %3
   store i32 %cond10, i32* @z3, align 4
@@ -133,11 +133,11 @@ entry:
 
 define void @calc_selle() nounwind {
 entry:
-  %0 = load i32* @a, align 4
-  %1 = load i32* @b, align 4
+  %0 = load i32, i32* @a, align 4
+  %1 = load i32, i32* @b, align 4
   %cmp = icmp sle i32 %0, %1
-  %2 = load i32* @t, align 4
-  %3 = load i32* @f, align 4
+  %2 = load i32, i32* @t, align 4
+  %3 = load i32, i32* @f, align 4
   %cond = select i1 %cmp, i32 %2, i32 %3
   store i32 %cond, i32* @z1, align 4
 ; 16:	slt	${{[0-9]+}}, ${{[0-9]+}}
@@ -146,7 +146,7 @@ entry:
   %cmp1 = icmp sle i32 %1, %0
   %cond5 = select i1 %cmp1, i32 %3, i32 %2
   store i32 %cond5, i32* @z2, align 4
-  %4 = load i32* @c, align 4
+  %4 = load i32, i32* @c, align 4
   %cmp6 = icmp sle i32 %4, %0
   %cond10 = select i1 %cmp6, i32 %2, i32 %3
   store i32 %cond10, i32* @z3, align 4
@@ -158,20 +158,20 @@ entry:
 
 define void @calc_selltk() nounwind {
 entry:
-  %0 = load i32* @a, align 4
+  %0 = load i32, i32* @a, align 4
   %cmp = icmp slt i32 %0, 10
-  %1 = load i32* @t, align 4
-  %2 = load i32* @f, align 4
+  %1 = load i32, i32* @t, align 4
+  %2 = load i32, i32* @f, align 4
   %cond = select i1 %cmp, i32 %1, i32 %2
   store i32 %cond, i32* @z1, align 4
 ; 16:	slti	${{[0-9]+}}, {{[0-9]+}}
 ; 16:	btnez	$BB{{[0-9]+}}_{{[0-9]}}
 ; 16: 	move    ${{[0-9]+}}, ${{[0-9]+}}
-  %3 = load i32* @b, align 4
+  %3 = load i32, i32* @b, align 4
   %cmp1 = icmp slt i32 %3, 2
   %cond5 = select i1 %cmp1, i32 %2, i32 %1
   store i32 %cond5, i32* @z2, align 4
-  %4 = load i32* @c, align 4
+  %4 = load i32, i32* @c, align 4
   %cmp6 = icmp sgt i32 %4, 2
   %cond10 = select i1 %cmp6, i32 %2, i32 %1
   store i32 %cond10, i32* @z3, align 4
@@ -184,18 +184,18 @@ entry:
 
 define void @calc_selne() nounwind {
 entry:
-  %0 = load i32* @a, align 4
-  %1 = load i32* @b, align 4
+  %0 = load i32, i32* @a, align 4
+  %1 = load i32, i32* @b, align 4
   %cmp = icmp ne i32 %0, %1
-  %2 = load i32* @t, align 4
-  %3 = load i32* @f, align 4
+  %2 = load i32, i32* @t, align 4
+  %3 = load i32, i32* @f, align 4
   %cond = select i1 %cmp, i32 %2, i32 %3
   store i32 %cond, i32* @z1, align 4
 ; 16:	cmp	${{[0-9]+}}, ${{[0-9]+}}
 ; 16:	btnez	$BB{{[0-9]+}}_{{[0-9]}}
 ; 16: 	move    ${{[0-9]+}}, ${{[0-9]+}}
   store i32 %cond, i32* @z2, align 4
-  %4 = load i32* @c, align 4
+  %4 = load i32, i32* @c, align 4
   %cmp6 = icmp ne i32 %4, %0
   %cond10 = select i1 %cmp6, i32 %3, i32 %2
   store i32 %cond10, i32* @z3, align 4
@@ -205,10 +205,10 @@ entry:
 
 define void @calc_selnek() nounwind {
 entry:
-  %0 = load i32* @a, align 4
+  %0 = load i32, i32* @a, align 4
   %cmp = icmp ne i32 %0, 1
-  %1 = load i32* @f, align 4
-  %2 = load i32* @t, align 4
+  %1 = load i32, i32* @f, align 4
+  %2 = load i32, i32* @t, align 4
   %cond = select i1 %cmp, i32 %1, i32 %2
   store i32 %cond, i32* @z1, align 4
 ; 16:	cmpi	${{[0-9]+}}, 1
@@ -217,7 +217,7 @@ entry:
   %cmp1 = icmp ne i32 %0, 10
   %cond5 = select i1 %cmp1, i32 %2, i32 %1
   store i32 %cond5, i32* @z2, align 4
-  %3 = load i32* @b, align 4
+  %3 = load i32, i32* @b, align 4
   %cmp6 = icmp ne i32 %3, 3
   %cond10 = select i1 %cmp6, i32 %2, i32 %1
   store i32 %cond10, i32* @z3, align 4
@@ -232,19 +232,19 @@ entry:
 
 define void @calc_selnez() nounwind {
 entry:
-  %0 = load i32* @a, align 4
+  %0 = load i32, i32* @a, align 4
   %cmp = icmp ne i32 %0, 0
-  %1 = load i32* @f, align 4
-  %2 = load i32* @t, align 4
+  %1 = load i32, i32* @f, align 4
+  %2 = load i32, i32* @t, align 4
   %cond = select i1 %cmp, i32 %1, i32 %2
   store i32 %cond, i32* @z1, align 4
 ; 16:	bnez	${{[0-9]+}}, $BB{{[0-9]+}}_{{[0-9]}}
 ; 16: 	move    ${{[0-9]+}}, ${{[0-9]+}}
-  %3 = load i32* @b, align 4
+  %3 = load i32, i32* @b, align 4
   %cmp1 = icmp ne i32 %3, 0
   %cond5 = select i1 %cmp1, i32 %2, i32 %1
   store i32 %cond5, i32* @z2, align 4
-  %4 = load i32* @c, align 4
+  %4 = load i32, i32* @c, align 4
   %cmp6 = icmp ne i32 %4, 0
   %cond10 = select i1 %cmp6, i32 %1, i32 %2
   store i32 %cond10, i32* @z3, align 4
@@ -254,19 +254,19 @@ entry:
 
 define void @calc_selnez2() nounwind {
 entry:
-  %0 = load i32* @a, align 4
+  %0 = load i32, i32* @a, align 4
   %tobool = icmp ne i32 %0, 0
-  %1 = load i32* @f, align 4
-  %2 = load i32* @t, align 4
+  %1 = load i32, i32* @f, align 4
+  %2 = load i32, i32* @t, align 4
   %cond = select i1 %tobool, i32 %1, i32 %2
   store i32 %cond, i32* @z1, align 4
 ; 16:	bnez	${{[0-9]+}}, $BB{{[0-9]+}}_{{[0-9]}}
 ; 16: 	move    ${{[0-9]+}}, ${{[0-9]+}}
-  %3 = load i32* @b, align 4
+  %3 = load i32, i32* @b, align 4
   %tobool1 = icmp ne i32 %3, 0
   %cond5 = select i1 %tobool1, i32 %2, i32 %1
   store i32 %cond5, i32* @z2, align 4
-  %4 = load i32* @c, align 4
+  %4 = load i32, i32* @c, align 4
   %tobool6 = icmp ne i32 %4, 0
   %cond10 = select i1 %tobool6, i32 %1, i32 %2
   store i32 %cond10, i32* @z3, align 4
@@ -276,11 +276,11 @@ entry:
 
 define void @calc_seluge() nounwind {
 entry:
-  %0 = load i32* @a, align 4
-  %1 = load i32* @b, align 4
+  %0 = load i32, i32* @a, align 4
+  %1 = load i32, i32* @b, align 4
   %cmp = icmp uge i32 %0, %1
-  %2 = load i32* @f, align 4
-  %3 = load i32* @t, align 4
+  %2 = load i32, i32* @f, align 4
+  %3 = load i32, i32* @t, align 4
   %cond = select i1 %cmp, i32 %2, i32 %3
   store i32 %cond, i32* @z1, align 4
 ; 16:	sltu	${{[0-9]+}}, ${{[0-9]+}}
@@ -289,7 +289,7 @@ entry:
   %cmp1 = icmp uge i32 %1, %0
   %cond5 = select i1 %cmp1, i32 %3, i32 %2
   store i32 %cond5, i32* @z2, align 4
-  %4 = load i32* @c, align 4
+  %4 = load i32, i32* @c, align 4
   %cmp6 = icmp uge i32 %4, %0
   %cond10 = select i1 %cmp6, i32 %3, i32 %2
   store i32 %cond10, i32* @z3, align 4
@@ -301,11 +301,11 @@ entry:
 
 define void @calc_selugt() nounwind {
 entry:
-  %0 = load i32* @a, align 4
-  %1 = load i32* @b, align 4
+  %0 = load i32, i32* @a, align 4
+  %1 = load i32, i32* @b, align 4
   %cmp = icmp ugt i32 %0, %1
-  %2 = load i32* @f, align 4
-  %3 = load i32* @t, align 4
+  %2 = load i32, i32* @f, align 4
+  %3 = load i32, i32* @t, align 4
   %cond = select i1 %cmp, i32 %2, i32 %3
   store i32 %cond, i32* @z1, align 4
 ; 16:	sltu	${{[0-9]+}}, ${{[0-9]+}}
@@ -314,7 +314,7 @@ entry:
   %cmp1 = icmp ugt i32 %1, %0
   %cond5 = select i1 %cmp1, i32 %3, i32 %2
   store i32 %cond5, i32* @z2, align 4
-  %4 = load i32* @c, align 4
+  %4 = load i32, i32* @c, align 4
   %cmp6 = icmp ugt i32 %4, %0
   %cond10 = select i1 %cmp6, i32 %2, i32 %3
   store i32 %cond10, i32* @z3, align 4
@@ -326,11 +326,11 @@ entry:
 
 define void @calc_selule() nounwind {
 entry:
-  %0 = load i32* @a, align 4
-  %1 = load i32* @b, align 4
+  %0 = load i32, i32* @a, align 4
+  %1 = load i32, i32* @b, align 4
   %cmp = icmp ule i32 %0, %1
-  %2 = load i32* @t, align 4
-  %3 = load i32* @f, align 4
+  %2 = load i32, i32* @t, align 4
+  %3 = load i32, i32* @f, align 4
   %cond = select i1 %cmp, i32 %2, i32 %3
   store i32 %cond, i32* @z1, align 4
 ; 16:	sltu	${{[0-9]+}}, ${{[0-9]+}}
@@ -339,7 +339,7 @@ entry:
   %cmp1 = icmp ule i32 %1, %0
   %cond5 = select i1 %cmp1, i32 %3, i32 %2
   store i32 %cond5, i32* @z2, align 4
-  %4 = load i32* @c, align 4
+  %4 = load i32, i32* @c, align 4
   %cmp6 = icmp ule i32 %4, %0
   %cond10 = select i1 %cmp6, i32 %2, i32 %3
   store i32 %cond10, i32* @z3, align 4
diff --git a/test/CodeGen/Mips/seteq.ll b/test/CodeGen/Mips/seteq.ll
index 5fadf78d57a0..8fad6122bdbe 100644
--- a/test/CodeGen/Mips/seteq.ll
+++ b/test/CodeGen/Mips/seteq.ll
@@ -8,8 +8,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @i, align 4
-  %1 = load i32* @k, align 4
+  %0 = load i32, i32* @i, align 4
+  %1 = load i32, i32* @k, align 4
   %cmp = icmp eq i32 %0, %1
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @r1, align 4
diff --git a/test/CodeGen/Mips/seteqz.ll b/test/CodeGen/Mips/seteqz.ll
index 80dc3120a6a1..8e9a4beac75b 100644
--- a/test/CodeGen/Mips/seteqz.ll
+++ b/test/CodeGen/Mips/seteqz.ll
@@ -7,13 +7,13 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %cmp = icmp eq i32 %0, 0
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @r1, align 4
 ; 16:	sltiu	${{[0-9]+}}, 1
 ; 16:	move	${{[0-9]+}}, $24
-  %1 = load i32* @j, align 4
+  %1 = load i32, i32* @j, align 4
   %cmp1 = icmp eq i32 %1, 99
   %conv2 = zext i1 %cmp1 to i32
   store i32 %conv2, i32* @r2, align 4
diff --git a/test/CodeGen/Mips/setge.ll b/test/CodeGen/Mips/setge.ll
index 8869eb8fc547..8fb729964cf5 100644
--- a/test/CodeGen/Mips/setge.ll
+++ b/test/CodeGen/Mips/setge.ll
@@ -11,15 +11,15 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @k, align 4
-  %1 = load i32* @j, align 4
+  %0 = load i32, i32* @k, align 4
+  %1 = load i32, i32* @j, align 4
   %cmp = icmp sge i32 %0, %1
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @r1, align 4
 ; 16:	slt	${{[0-9]+}}, ${{[0-9]+}}
 ; 16:	move	$[[REGISTER:[0-9]+]], $24
 ; 16:	xor	$[[REGISTER]], ${{[0-9]+}}
-  %2 = load i32* @m, align 4
+  %2 = load i32, i32* @m, align 4
   %cmp1 = icmp sge i32 %0, %2
   %conv2 = zext i1 %cmp1 to i32
   store i32 %conv2, i32* @r2, align 4
diff --git a/test/CodeGen/Mips/setgek.ll b/test/CodeGen/Mips/setgek.ll
index 18a0fcf62130..1148d1b67bda 100644
--- a/test/CodeGen/Mips/setgek.ll
+++ b/test/CodeGen/Mips/setgek.ll
@@ -7,7 +7,7 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @k, align 4
+  %0 = load i32, i32* @k, align 4
   %cmp = icmp sgt i32 %0, -32769
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @r1, align 4
diff --git a/test/CodeGen/Mips/setle.ll b/test/CodeGen/Mips/setle.ll
index 2df6774c1fad..fe4a2c37eb54 100644
--- a/test/CodeGen/Mips/setle.ll
+++ b/test/CodeGen/Mips/setle.ll
@@ -10,15 +10,15 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @j, align 4
-  %1 = load i32* @k, align 4
+  %0 = load i32, i32* @j, align 4
+  %1 = load i32, i32* @k, align 4
   %cmp = icmp sle i32 %0, %1
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @r1, align 4
 ; 16:	slt	${{[0-9]+}}, ${{[0-9]+}}
 ; 16:	move	$[[REGISTER:[0-9]+]], $24
 ; 16:	xor	$[[REGISTER]], ${{[0-9]+}}
-  %2 = load i32* @m, align 4
+  %2 = load i32, i32* @m, align 4
   %cmp1 = icmp sle i32 %2, %1
   %conv2 = zext i1 %cmp1 to i32
   store i32 %conv2, i32* @r2, align 4
diff --git a/test/CodeGen/Mips/setlt.ll b/test/CodeGen/Mips/setlt.ll
index 3dac74bf2e01..c4211e6dd696 100644
--- a/test/CodeGen/Mips/setlt.ll
+++ b/test/CodeGen/Mips/setlt.ll
@@ -10,8 +10,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @j, align 4
-  %1 = load i32* @k, align 4
+  %0 = load i32, i32* @j, align 4
+  %1 = load i32, i32* @k, align 4
   %cmp = icmp slt i32 %0, %1
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @r1, align 4
diff --git a/test/CodeGen/Mips/setltk.ll b/test/CodeGen/Mips/setltk.ll
index ecebc7e578e1..8c0041111270 100644
--- a/test/CodeGen/Mips/setltk.ll
+++ b/test/CodeGen/Mips/setltk.ll
@@ -10,7 +10,7 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @j, align 4
+  %0 = load i32, i32* @j, align 4
   %cmp = icmp slt i32 %0, 10
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @r1, align 4
diff --git a/test/CodeGen/Mips/setne.ll b/test/CodeGen/Mips/setne.ll
index 9e66901e32b5..484674e5da32 100644
--- a/test/CodeGen/Mips/setne.ll
+++ b/test/CodeGen/Mips/setne.ll
@@ -8,8 +8,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @i, align 4
-  %1 = load i32* @k, align 4
+  %0 = load i32, i32* @i, align 4
+  %1 = load i32, i32* @k, align 4
   %cmp = icmp ne i32 %0, %1
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @r1, align 4
diff --git a/test/CodeGen/Mips/setuge.ll b/test/CodeGen/Mips/setuge.ll
index 1c9b5bbe8114..025b4dcefd76 100644
--- a/test/CodeGen/Mips/setuge.ll
+++ b/test/CodeGen/Mips/setuge.ll
@@ -10,15 +10,15 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @k, align 4
-  %1 = load i32* @j, align 4
+  %0 = load i32, i32* @k, align 4
+  %1 = load i32, i32* @j, align 4
   %cmp = icmp uge i32 %0, %1
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @r1, align 4
 ; 16:	sltu	${{[0-9]+}}, ${{[0-9]+}}
 ; 16:	move    $[[REGISTER:[0-9]+]], $24
 ; 16:	xor	$[[REGISTER]], ${{[0-9]+}}
-  %2 = load i32* @m, align 4
+  %2 = load i32, i32* @m, align 4
   %cmp1 = icmp uge i32 %0, %2
   %conv2 = zext i1 %cmp1 to i32
   store i32 %conv2, i32* @r2, align 4
diff --git a/test/CodeGen/Mips/setugt.ll b/test/CodeGen/Mips/setugt.ll
index f10b47ae7178..0ce317e0df9e 100644
--- a/test/CodeGen/Mips/setugt.ll
+++ b/test/CodeGen/Mips/setugt.ll
@@ -10,8 +10,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @k, align 4
-  %1 = load i32* @j, align 4
+  %0 = load i32, i32* @k, align 4
+  %1 = load i32, i32* @j, align 4
   %cmp = icmp ugt i32 %0, %1
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @r1, align 4
diff --git a/test/CodeGen/Mips/setule.ll b/test/CodeGen/Mips/setule.ll
index a6d6bf064052..4255fd27c5cd 100644
--- a/test/CodeGen/Mips/setule.ll
+++ b/test/CodeGen/Mips/setule.ll
@@ -10,15 +10,15 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @j, align 4
-  %1 = load i32* @k, align 4
+  %0 = load i32, i32* @j, align 4
+  %1 = load i32, i32* @k, align 4
   %cmp = icmp ule i32 %0, %1
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @r1, align 4
 ; 16:	sltu	${{[0-9]+}}, ${{[0-9]+}}
 ; 16:	move	$[[REGISTER:[0-9]+]], $24
 ; 16:	xor	$[[REGISTER]], ${{[0-9]+}}
-  %2 = load i32* @m, align 4
+  %2 = load i32, i32* @m, align 4
   %cmp1 = icmp ule i32 %2, %1
   %conv2 = zext i1 %cmp1 to i32
   store i32 %conv2, i32* @r2, align 4
diff --git a/test/CodeGen/Mips/setult.ll b/test/CodeGen/Mips/setult.ll
index 00ee437a2ffe..d30107e54dd0 100644
--- a/test/CodeGen/Mips/setult.ll
+++ b/test/CodeGen/Mips/setult.ll
@@ -10,8 +10,8 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @j, align 4
-  %1 = load i32* @k, align 4
+  %0 = load i32, i32* @j, align 4
+  %1 = load i32, i32* @k, align 4
   %cmp = icmp ult i32 %0, %1
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @r1, align 4
diff --git a/test/CodeGen/Mips/setultk.ll b/test/CodeGen/Mips/setultk.ll
index eb9edbaad7f8..1b79f103bed7 100644
--- a/test/CodeGen/Mips/setultk.ll
+++ b/test/CodeGen/Mips/setultk.ll
@@ -10,7 +10,7 @@
 
 define void @test() nounwind {
 entry:
-  %0 = load i32* @j, align 4
+  %0 = load i32, i32* @j, align 4
   %cmp = icmp ult i32 %0, 10
   %conv = zext i1 %cmp to i32
   store i32 %conv, i32* @r1, align 4
diff --git a/test/CodeGen/Mips/sh1.ll b/test/CodeGen/Mips/sh1.ll
index 1746ae284f2a..3f70b9bc6e68 100644
--- a/test/CodeGen/Mips/sh1.ll
+++ b/test/CodeGen/Mips/sh1.ll
@@ -6,14 +6,14 @@
 
 define i32 @main() nounwind {
 entry:
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %conv = trunc i32 %0 to i16
   store i16 %conv, i16* @s, align 2
-  %1 = load i32* @i, align 4
-  %2 = load i16* @s, align 2
+  %1 = load i32, i32* @i, align 4
+  %2 = load i16, i16* @s, align 2
   %conv1 = sext i16 %2 to i32
 ; 16:	sh	${{[0-9]+}}, 0(${{[0-9]+}})
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([9 x i8]* @.str, i32 0, i32 0), i32 %1, i32 %conv1)
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([9 x i8], [9 x i8]* @.str, i32 0, i32 0), i32 %1, i32 %conv1)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/simplebr.ll b/test/CodeGen/Mips/simplebr.ll
index a1d63671b4ee..2aeacc903fbe 100644
--- a/test/CodeGen/Mips/simplebr.ll
+++ b/test/CodeGen/Mips/simplebr.ll
@@ -1,4 +1,4 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mips16-hard-float -soft-float -relocation-model=static < %s | FileCheck %s -check-prefix=CHECK-STATIC16
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mips16-hard-float -mattr=+soft-float -relocation-model=static < %s | FileCheck %s -check-prefix=CHECK-STATIC16
 
 ; ModuleID = 'simplebr.c'
 target datalayout = "E-p:32:32:32-i1:8:8-i8:8:32-i16:16:32-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-n32-S64"
@@ -9,7 +9,7 @@ target triple = "mips--linux-gnu"
 ; Function Attrs: nounwind
 define void @foo() #0 {
 entry:
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %tobool = icmp ne i32 %0, 0
   br i1 %tobool, label %if.then, label %if.else
 
diff --git a/test/CodeGen/Mips/sitofp-selectcc-opt.ll b/test/CodeGen/Mips/sitofp-selectcc-opt.ll
index 576cbd8e9637..c60fceb1a04c 100644
--- a/test/CodeGen/Mips/sitofp-selectcc-opt.ll
+++ b/test/CodeGen/Mips/sitofp-selectcc-opt.ll
@@ -14,7 +14,7 @@ entry:
   %tobool1. = or i1 %tobool1, %not.tobool
   %lor.ext = zext i1 %tobool1. to i32
   %conv = sitofp i32 %lor.ext to double
-  %1 = load double* @foo12.d4, align 8
+  %1 = load double, double* @foo12.d4, align 8
   %add = fadd double %conv, %1
   store double %add, double* @foo12.d4, align 8
   ret double %add
diff --git a/test/CodeGen/Mips/sll1.ll b/test/CodeGen/Mips/sll1.ll
index fdcd38c84b3a..4d35b64e0b58 100644
--- a/test/CodeGen/Mips/sll1.ll
+++ b/test/CodeGen/Mips/sll1.ll
@@ -7,12 +7,12 @@
 define i32 @main() nounwind {
 entry:
 ; 16:	sll	${{[0-9]+}}, ${{[0-9]+}}, {{[0-9]+}}
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %shl = shl i32 %0, 4
 ; 16:	sll	${{[0-9]+}}, ${{[0-9]+}}, {{[0-9]+}}
   store i32 %shl, i32* @j, align 4
-  %1 = load i32* @j, align 4
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i32 0, i32 0), i32 %1)
+  %1 = load i32, i32* @j, align 4
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([5 x i8], [5 x i8]* @.str, i32 0, i32 0), i32 %1)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/sll2.ll b/test/CodeGen/Mips/sll2.ll
index c2af454cc853..dc2236b10ccf 100644
--- a/test/CodeGen/Mips/sll2.ll
+++ b/test/CodeGen/Mips/sll2.ll
@@ -6,13 +6,13 @@
 
 define i32 @main() nounwind {
 entry:
-  %0 = load i32* @i, align 4
-  %1 = load i32* @j, align 4
+  %0 = load i32, i32* @i, align 4
+  %1 = load i32, i32* @j, align 4
   %shl = shl i32 %0, %1
 ; 16:	sllv	${{[0-9]+}}, ${{[0-9]+}}
   store i32 %shl, i32* @i, align 4
-  %2 = load i32* @j, align 4
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i32 0, i32 0), i32 %2)
+  %2 = load i32, i32* @j, align 4
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([5 x i8], [5 x i8]* @.str, i32 0, i32 0), i32 %2)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/small-section-reserve-gp.ll b/test/CodeGen/Mips/small-section-reserve-gp.ll
index cbf0681c78e5..c4e37665aaf6 100644
--- a/test/CodeGen/Mips/small-section-reserve-gp.ll
+++ b/test/CodeGen/Mips/small-section-reserve-gp.ll
@@ -6,7 +6,7 @@
 define i32 @geti() nounwind readonly {
 entry:
 ; CHECK: lw ${{[0-9]+}}, %gp_rel(i)($gp)
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   ret i32 %0
 }
 
diff --git a/test/CodeGen/Mips/spill-copy-acreg.ll b/test/CodeGen/Mips/spill-copy-acreg.ll
index 6563a5cffd91..fd160b67cf24 100644
--- a/test/CodeGen/Mips/spill-copy-acreg.ll
+++ b/test/CodeGen/Mips/spill-copy-acreg.ll
@@ -6,7 +6,7 @@
 
 define i64 @test_acreg_copy(i32 %a0, i32 %a1, i32 %a2, i32 %a3) {
 entry:
-  %0 = load i64* @g1, align 8
+  %0 = load i64, i64* @g1, align 8
   %1 = tail call i64 @llvm.mips.maddu(i64 %0, i32 %a0, i32 %a1)
   %2 = tail call i64 @llvm.mips.maddu(i64 %0, i32 %a2, i32 %a3)
   store i64 %1, i64* @g1, align 8
@@ -32,8 +32,8 @@ entry:
   %sext = sext <2 x i1> %cmp3 to <2 x i16>
   store <2 x i16> %sext, <2 x i16>* @g4, align 4
   tail call void @foo1()
-  %2 = load <2 x i16>* @g5, align 4
-  %3 = load <2 x i16>* @g6, align 4
+  %2 = load <2 x i16>, <2 x i16>* @g5, align 4
+  %3 = load <2 x i16>, <2 x i16>* @g6, align 4
   %or = select <2 x i1> %cmp3, <2 x i16> %2, <2 x i16> %3
   %4 = bitcast <2 x i16> %or to i32
   %.fca.0.insert = insertvalue { i32 } undef, i32 %4, 0
diff --git a/test/CodeGen/Mips/sr1.ll b/test/CodeGen/Mips/sr1.ll
index 610693d58b3f..69655f7b842c 100644
--- a/test/CodeGen/Mips/sr1.ll
+++ b/test/CodeGen/Mips/sr1.ll
@@ -8,9 +8,9 @@
 define void @foo1() #0 {
 entry:
   %c = alloca [10 x i8], align 1
-  %arraydecay = getelementptr inbounds [10 x i8]* %c, i32 0, i32 0
+  %arraydecay = getelementptr inbounds [10 x i8], [10 x i8]* %c, i32 0, i32 0
   call void @x(i8* %arraydecay)
-  %arraydecay1 = getelementptr inbounds [10 x i8]* %c, i32 0, i32 0
+  %arraydecay1 = getelementptr inbounds [10 x i8], [10 x i8]* %c, i32 0, i32 0
   call void @x(i8* %arraydecay1)
   ret void
 ; CHECK: 	.ent	foo1
@@ -25,9 +25,9 @@ declare void @x(i8*) #1
 define void @foo2() #0 {
 entry:
   %c = alloca [150 x i8], align 1
-  %arraydecay = getelementptr inbounds [150 x i8]* %c, i32 0, i32 0
+  %arraydecay = getelementptr inbounds [150 x i8], [150 x i8]* %c, i32 0, i32 0
   call void @x(i8* %arraydecay)
-  %arraydecay1 = getelementptr inbounds [150 x i8]* %c, i32 0, i32 0
+  %arraydecay1 = getelementptr inbounds [150 x i8], [150 x i8]* %c, i32 0, i32 0
   call void @x(i8* %arraydecay1)
   ret void
 ; CHECK: 	.ent	foo2
diff --git a/test/CodeGen/Mips/sra1.ll b/test/CodeGen/Mips/sra1.ll
index 15bf8d644ea3..1c7d417cb13a 100644
--- a/test/CodeGen/Mips/sra1.ll
+++ b/test/CodeGen/Mips/sra1.ll
@@ -5,10 +5,10 @@
 
 define i32 @main() nounwind {
 entry:
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %shr = ashr i32 %0, 3
 ; 16:	sra	${{[0-9]+}}, ${{[0-9]+}}, {{[0-9]+}}
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i32 0, i32 0), i32 %shr)
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([5 x i8], [5 x i8]* @.str, i32 0, i32 0), i32 %shr)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/sra2.ll b/test/CodeGen/Mips/sra2.ll
index 26bf19d44020..771d0f4a79e3 100644
--- a/test/CodeGen/Mips/sra2.ll
+++ b/test/CodeGen/Mips/sra2.ll
@@ -6,11 +6,11 @@
 
 define i32 @main() nounwind {
 entry:
-  %0 = load i32* @i, align 4
-  %1 = load i32* @j, align 4
+  %0 = load i32, i32* @i, align 4
+  %1 = load i32, i32* @j, align 4
   %shr = ashr i32 %0, %1
 ; 16:	srav	${{[0-9]+}}, ${{[0-9]+}}
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i32 0, i32 0), i32 %shr)
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([5 x i8], [5 x i8]* @.str, i32 0, i32 0), i32 %shr)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/srl1.ll b/test/CodeGen/Mips/srl1.ll
index 3474283faef9..a748eabb066f 100644
--- a/test/CodeGen/Mips/srl1.ll
+++ b/test/CodeGen/Mips/srl1.ll
@@ -6,12 +6,12 @@
 
 define i32 @main() nounwind {
 entry:
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %shr = lshr i32 %0, 4
 ; 16:	srl	${{[0-9]+}}, ${{[0-9]+}}, {{[0-9]+}}
   store i32 %shr, i32* @j, align 4
-  %1 = load i32* @j, align 4
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i32 0, i32 0), i32 %1)
+  %1 = load i32, i32* @j, align 4
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([5 x i8], [5 x i8]* @.str, i32 0, i32 0), i32 %1)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/srl2.ll b/test/CodeGen/Mips/srl2.ll
index 26ec0927a559..6e338b39350f 100644
--- a/test/CodeGen/Mips/srl2.ll
+++ b/test/CodeGen/Mips/srl2.ll
@@ -7,13 +7,13 @@
 
 define i32 @main() nounwind {
 entry:
-  %0 = load i32* @i, align 4
-  %1 = load i32* @k, align 4
+  %0 = load i32, i32* @i, align 4
+  %1 = load i32, i32* @k, align 4
   %shr = lshr i32 %0, %1
 ; 16:	srlv	${{[0-9]+}}, ${{[0-9]+}}
   store i32 %shr, i32* @j, align 4
-  %2 = load i32* @j, align 4
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i32 0, i32 0), i32 %2)
+  %2 = load i32, i32* @j, align 4
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([5 x i8], [5 x i8]* @.str, i32 0, i32 0), i32 %2)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/stackcoloring.ll b/test/CodeGen/Mips/stackcoloring.ll
index 4987dad5338b..5516b5a3c023 100644
--- a/test/CodeGen/Mips/stackcoloring.ll
+++ b/test/CodeGen/Mips/stackcoloring.ll
@@ -12,15 +12,15 @@ entry:
   %b = alloca [16 x i32], align 4
   %0 = bitcast [16 x i32]* %b to i8*
   call void @llvm.lifetime.start(i64 64, i8* %0)
-  %arraydecay = getelementptr inbounds [16 x i32]* %b, i32 0, i32 0
+  %arraydecay = getelementptr inbounds [16 x i32], [16 x i32]* %b, i32 0, i32 0
   br label %for.body
 
 for.body:                                         ; preds = %for.body, %entry
   %i.05 = phi i32 [ 0, %entry ], [ %inc, %for.body ]
   %v.04 = phi i32 [ 0, %entry ], [ %add, %for.body ]
-  %1 = load i32** @g1, align 4
-  %arrayidx = getelementptr inbounds i32* %1, i32 %i.05
-  %2 = load i32* %arrayidx, align 4
+  %1 = load i32*, i32** @g1, align 4
+  %arrayidx = getelementptr inbounds i32, i32* %1, i32 %i.05
+  %2 = load i32, i32* %arrayidx, align 4
   %call = call i32 @foo2(i32 %2, i32* %arraydecay)
   %add = add nsw i32 %call, %v.04
   %inc = add nsw i32 %i.05, 1
diff --git a/test/CodeGen/Mips/start-asm-file.ll b/test/CodeGen/Mips/start-asm-file.ll
index 9dc501ce10b4..60c047a4e8cc 100644
--- a/test/CodeGen/Mips/start-asm-file.ll
+++ b/test/CodeGen/Mips/start-asm-file.ll
@@ -19,36 +19,36 @@
 
 ; ### N32 ABI ###
 ; RUN: llc -filetype=asm -mtriple mips64-unknown-linux -mcpu=mips64 \
-; RUN: -relocation-model=static -mattr=-n64,+n32 %s -o - | \
+; RUN: -relocation-model=static -target-abi n32 %s -o - | \
 ; RUN:   FileCheck -check-prefix=CHECK-STATIC-N32 -check-prefix=CHECK-STATIC-N32-NLEGACY %s
 
 ; RUN: llc -filetype=asm -mtriple mips64-unknown-linux -mcpu=mips64 \
-; RUN: -relocation-model=pic -mattr=-n64,+n32 %s -o - | \
+; RUN: -relocation-model=pic -target-abi n32 %s -o - | \
 ; RUN:   FileCheck -check-prefix=CHECK-PIC-N32 -check-prefix=CHECK-PIC-N32-NLEGACY %s
 
 ; RUN: llc -filetype=asm -mtriple mips64-unknown-linux -mcpu=mips64 \
-; RUN: -relocation-model=static -mattr=-n64,+n32,+nan2008 %s -o - | \
+; RUN: -relocation-model=static -target-abi n32 -mattr=+nan2008 %s -o - | \
 ; RUN:   FileCheck -check-prefix=CHECK-STATIC-N32 -check-prefix=CHECK-STATIC-N32-N2008 %s
 
 ; RUN: llc -filetype=asm -mtriple mips64-unknown-linux -mcpu=mips64 \
-; RUN: -relocation-model=pic -mattr=-n64,+n32,+nan2008 %s -o - | \
+; RUN: -relocation-model=pic -target-abi n32 -mattr=+nan2008 %s -o - | \
 ; RUN:   FileCheck -check-prefix=CHECK-PIC-N32 -check-prefix=CHECK-PIC-N32-N2008 %s
 
 ; ### N64 ABI ###
 ; RUN: llc -filetype=asm -mtriple mips64-unknown-linux -mcpu=mips64 \
-; RUN: -relocation-model=static -mattr=+n64 %s -o - | \
+; RUN: -relocation-model=static -target-abi n64 %s -o - | \
 ; RUN:   FileCheck -check-prefix=CHECK-STATIC-N64 -check-prefix=CHECK-STATIC-N64-NLEGACY %s
 
 ; RUN: llc -filetype=asm -mtriple mips64-unknown-linux -mcpu=mips64 \
-; RUN: -relocation-model=pic -mattr=+n64 %s -o - | \
+; RUN: -relocation-model=pic -target-abi n64 %s -o - | \
 ; RUN:   FileCheck -check-prefix=CHECK-PIC-N64 -check-prefix=CHECK-PIC-N64-NLEGACY %s
 
 ; RUN: llc -filetype=asm -mtriple mips64-unknown-linux -mcpu=mips64 \
-; RUN: -relocation-model=static -mattr=+n64,+nan2008 %s -o - | \
+; RUN: -relocation-model=static -target-abi n64 -mattr=+nan2008 %s -o - | \
 ; RUN:   FileCheck -check-prefix=CHECK-STATIC-N64 -check-prefix=CHECK-STATIC-N64-N2008 %s
 
 ; RUN: llc -filetype=asm -mtriple mips64-unknown-linux -mcpu=mips64 \
-; RUN: -relocation-model=pic -mattr=+n64,+nan2008 %s -o - | \
+; RUN: -relocation-model=pic -target-abi n64 -mattr=+nan2008 %s -o - | \
 ; RUN:   FileCheck -check-prefix=CHECK-PIC-N64 -check-prefix=CHECK-PIC-N64-N2008 %s
 
 ; CHECK-STATIC-O32: .abicalls
diff --git a/test/CodeGen/Mips/stchar.ll b/test/CodeGen/Mips/stchar.ll
index 12eae3487ff1..6bc4889931a7 100644
--- a/test/CodeGen/Mips/stchar.ll
+++ b/test/CodeGen/Mips/stchar.ll
@@ -9,7 +9,7 @@ define void @p1(i16 signext %s, i8 signext %c) nounwind {
 entry:
   %conv = sext i16 %s to i32
   %conv1 = sext i8 %c to i32
-  %call = tail call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([9 x i8]* @.str, i32 0, i32 0), i32 %conv, i32 %conv1) nounwind
+  %call = tail call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([9 x i8], [9 x i8]* @.str, i32 0, i32 0), i32 %conv, i32 %conv1) nounwind
   ret void
 }
 
@@ -17,16 +17,16 @@ declare i32 @printf(i8* nocapture, ...) nounwind
 
 define void @p2() nounwind {
 entry:
-  %0 = load i16** @sp, align 4
-  %1 = load i16* %0, align 2
-  %2 = load i8** @cp, align 4
-  %3 = load i8* %2, align 1
+  %0 = load i16*, i16** @sp, align 4
+  %1 = load i16, i16* %0, align 2
+  %2 = load i8*, i8** @cp, align 4
+  %3 = load i8, i8* %2, align 1
   %conv.i = sext i16 %1 to i32
   %conv1.i = sext i8 %3 to i32
-  %call.i = tail call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([9 x i8]* @.str, i32 0, i32 0), i32 %conv.i, i32 %conv1.i) nounwind
-  %4 = load i16** @sp, align 4
+  %call.i = tail call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([9 x i8], [9 x i8]* @.str, i32 0, i32 0), i32 %conv.i, i32 %conv1.i) nounwind
+  %4 = load i16*, i16** @sp, align 4
   store i16 32, i16* %4, align 2
-  %5 = load i8** @cp, align 4
+  %5 = load i8*, i8** @cp, align 4
   store i8 97, i8* %5, align 1
   ret void
 }
@@ -39,16 +39,16 @@ entry:
   store i8 99, i8* %c, align 4
   store i16* %s, i16** @sp, align 4
   store i8* %c, i8** @cp, align 4
-  %call.i.i = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([9 x i8]* @.str, i32 0, i32 0), i32 16, i32 99) nounwind
-  %0 = load i16** @sp, align 4
+  %call.i.i = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([9 x i8], [9 x i8]* @.str, i32 0, i32 0), i32 16, i32 99) nounwind
+  %0 = load i16*, i16** @sp, align 4
   store i16 32, i16* %0, align 2
-  %1 = load i8** @cp, align 4
+  %1 = load i8*, i8** @cp, align 4
   store i8 97, i8* %1, align 1
-  %2 = load i16* %s, align 4
-  %3 = load i8* %c, align 4
+  %2 = load i16, i16* %s, align 4
+  %3 = load i8, i8* %c, align 4
   %conv.i = sext i16 %2 to i32
   %conv1.i = sext i8 %3 to i32
-  %call.i = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([9 x i8]* @.str, i32 0, i32 0), i32 %conv.i, i32 %conv1.i) nounwind
+  %call.i = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([9 x i8], [9 x i8]* @.str, i32 0, i32 0), i32 %conv.i, i32 %conv1.i) nounwind
   ret void
 ; 16_b-LABEL: test:
 ; 16_h-LABEL: test:
@@ -69,16 +69,16 @@ entry:
   store i8 99, i8* %c.i, align 4
   store i16* %s.i, i16** @sp, align 4
   store i8* %c.i, i8** @cp, align 4
-  %call.i.i.i = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([9 x i8]* @.str, i32 0, i32 0), i32 16, i32 99) nounwind
-  %1 = load i16** @sp, align 4
+  %call.i.i.i = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([9 x i8], [9 x i8]* @.str, i32 0, i32 0), i32 16, i32 99) nounwind
+  %1 = load i16*, i16** @sp, align 4
   store i16 32, i16* %1, align 2
-  %2 = load i8** @cp, align 4
+  %2 = load i8*, i8** @cp, align 4
   store i8 97, i8* %2, align 1
-  %3 = load i16* %s.i, align 4
-  %4 = load i8* %c.i, align 4
+  %3 = load i16, i16* %s.i, align 4
+  %4 = load i8, i8* %c.i, align 4
   %conv.i.i = sext i16 %3 to i32
   %conv1.i.i = sext i8 %4 to i32
-  %call.i.i = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([9 x i8]* @.str, i32 0, i32 0), i32 %conv.i.i, i32 %conv1.i.i) nounwind
+  %call.i.i = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([9 x i8], [9 x i8]* @.str, i32 0, i32 0), i32 %conv.i.i, i32 %conv1.i.i) nounwind
   call void @llvm.lifetime.end(i64 -1, i8* %0) nounwind
   call void @llvm.lifetime.end(i64 -1, i8* %c.i) nounwind
   ret i32 0
diff --git a/test/CodeGen/Mips/stldst.ll b/test/CodeGen/Mips/stldst.ll
index 4182b9e76d63..4eef5ece0589 100644
--- a/test/CodeGen/Mips/stldst.ll
+++ b/test/CodeGen/Mips/stldst.ll
@@ -12,25 +12,25 @@
 
 define i32 @main() nounwind {
 entry:
-  %0 = load i32* @kkkk, align 4
-  %1 = load i32* @llll, align 4
+  %0 = load i32, i32* @kkkk, align 4
+  %1 = load i32, i32* @llll, align 4
   %add = add nsw i32 %0, 10
   %add1 = add nsw i32 %1, 10
-  %2 = load i32* @mmmm, align 4
+  %2 = load i32, i32* @mmmm, align 4
   %sub = add nsw i32 %2, -3
-  %3 = load i32* @nnnn, align 4
+  %3 = load i32, i32* @nnnn, align 4
   %add2 = add nsw i32 %3, 10
-  %4 = load i32* @oooo, align 4
+  %4 = load i32, i32* @oooo, align 4
   %add3 = add nsw i32 %4, 4
-  %5 = load i32* @pppp, align 4
+  %5 = load i32, i32* @pppp, align 4
   %sub4 = add nsw i32 %5, -5
-  %6 = load i32* @qqqq, align 4
+  %6 = load i32, i32* @qqqq, align 4
   %sub5 = add nsw i32 %6, -10
-  %7 = load i32* @rrrr, align 4
+  %7 = load i32, i32* @rrrr, align 4
   %add6 = add nsw i32 %7, 6
 
-  %call = tail call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([32 x i8]* @.str, i32 0, i32 0), i32 %sub5, i32 %add6, i32 %0, i32 %1, i32 %2, i32 %3, i32 %4, i32 %5, i32 %6, i32 %7) nounwind
-  %call7 = tail call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([32 x i8]* @.str, i32 0, i32 0), i32 %0, i32 %1, i32 %add, i32 %add1, i32 %sub, i32 %add2, i32 %add3, i32 %sub4, i32 %sub5, i32 %add6) nounwind
+  %call = tail call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([32 x i8], [32 x i8]* @.str, i32 0, i32 0), i32 %sub5, i32 %add6, i32 %0, i32 %1, i32 %2, i32 %3, i32 %4, i32 %5, i32 %6, i32 %7) nounwind
+  %call7 = tail call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([32 x i8], [32 x i8]* @.str, i32 0, i32 0), i32 %0, i32 %1, i32 %add, i32 %add1, i32 %sub, i32 %add2, i32 %add3, i32 %sub4, i32 %sub5, i32 %add6) nounwind
   ret i32 0
 }
 ; 16:	sw	${{[0-9]+}}, {{[0-9]+}} ( $sp );         # 4-byte Folded Spill
diff --git a/test/CodeGen/Mips/sub1.ll b/test/CodeGen/Mips/sub1.ll
index 195750b805d6..636ab8f2c5f3 100644
--- a/test/CodeGen/Mips/sub1.ll
+++ b/test/CodeGen/Mips/sub1.ll
@@ -5,10 +5,10 @@
 
 define i32 @main() nounwind {
 entry:
-  %0 = load i32* @i, align 4
+  %0 = load i32, i32* @i, align 4
   %sub = sub nsw i32 %0, 5
 ; 16:	addiu	${{[0-9]+}}, -{{[0-9]+}}
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([4 x i8]* @.str, i32 0, i32 0), i32 %sub)
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([4 x i8], [4 x i8]* @.str, i32 0, i32 0), i32 %sub)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/sub2.ll b/test/CodeGen/Mips/sub2.ll
index 4f6bfccec401..a97f5e947ca9 100644
--- a/test/CodeGen/Mips/sub2.ll
+++ b/test/CodeGen/Mips/sub2.ll
@@ -6,11 +6,11 @@
 
 define i32 @main() nounwind {
 entry:
-  %0 = load i32* @j, align 4
-  %1 = load i32* @i, align 4
+  %0 = load i32, i32* @j, align 4
+  %1 = load i32, i32* @i, align 4
   %sub = sub nsw i32 %0, %1
 ; 16:	subu	${{[0-9]+}}, ${{[0-9]+}}, ${{[0-9]+}}
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([4 x i8]* @.str, i32 0, i32 0), i32 %sub)
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([4 x i8], [4 x i8]* @.str, i32 0, i32 0), i32 %sub)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/swzero.ll b/test/CodeGen/Mips/swzero.ll
index 9f91a3902d7d..9aaee1509806 100644
--- a/test/CodeGen/Mips/swzero.ll
+++ b/test/CodeGen/Mips/swzero.ll
@@ -6,7 +6,7 @@ define void @zero_u(%struct.unaligned* nocapture %p) nounwind {
 entry:
 ; CHECK: swl $zero
 ; CHECK: swr $zero
-  %x = getelementptr inbounds %struct.unaligned* %p, i32 0, i32 0
+  %x = getelementptr inbounds %struct.unaligned, %struct.unaligned* %p, i32 0, i32 0
   store i32 0, i32* %x, align 1
   ret void
 }
diff --git a/test/CodeGen/Mips/tail16.ll b/test/CodeGen/Mips/tail16.ll
index 4e62e557478a..13f27fcc513b 100644
--- a/test/CodeGen/Mips/tail16.ll
+++ b/test/CodeGen/Mips/tail16.ll
@@ -1,4 +1,4 @@
-; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -soft-float -mips16-hard-float -relocation-model=pic   < %s | FileCheck %s 
+; RUN: llc -mtriple=mipsel-linux-gnu -march=mipsel -mcpu=mips16 -mattr=+soft-float -mips16-hard-float -relocation-model=pic   < %s | FileCheck %s 
 
 ; Function Attrs: nounwind optsize
 define float @h()  {
diff --git a/test/CodeGen/Mips/tailcall.ll b/test/CodeGen/Mips/tailcall.ll
index 30f47abc06cb..6a0d64b7eed8 100644
--- a/test/CodeGen/Mips/tailcall.ll
+++ b/test/CodeGen/Mips/tailcall.ll
@@ -85,16 +85,16 @@ entry:
 ; PIC16: jalrc
 ; PIC16: .end caller5
 
-  %0 = load i32* @g0, align 4
-  %1 = load i32* @g1, align 4
-  %2 = load i32* @g2, align 4
-  %3 = load i32* @g3, align 4
-  %4 = load i32* @g4, align 4
-  %5 = load i32* @g5, align 4
-  %6 = load i32* @g6, align 4
-  %7 = load i32* @g7, align 4
-  %8 = load i32* @g8, align 4
-  %9 = load i32* @g9, align 4
+  %0 = load i32, i32* @g0, align 4
+  %1 = load i32, i32* @g1, align 4
+  %2 = load i32, i32* @g2, align 4
+  %3 = load i32, i32* @g3, align 4
+  %4 = load i32, i32* @g4, align 4
+  %5 = load i32, i32* @g5, align 4
+  %6 = load i32, i32* @g6, align 4
+  %7 = load i32, i32* @g7, align 4
+  %8 = load i32, i32* @g8, align 4
+  %9 = load i32, i32* @g9, align 4
   %call = tail call fastcc i32 @callee5(i32 %0, i32 %1, i32 %2, i32 %3, i32 %4, i32 %5, i32 %6, i32 %7, i32 %8, i32 %9)
   ret i32 %call
 }
@@ -136,7 +136,7 @@ entry:
 ; PIC16: jalrc
 ; PIC16: .end caller8_1
 
-  %call = tail call i32 (i32, ...)* @callee8(i32 2, i32 1) nounwind
+  %call = tail call i32 (i32, ...) @callee8(i32 2, i32 1) nounwind
   ret i32 %call
 }
 
@@ -239,7 +239,7 @@ entry:
 ; PIC16: .ent caller13
 ; PIC16: jalrc
 
-  %call = tail call i32 (i32, ...)* @callee13(i32 1, i32 2) nounwind
+  %call = tail call i32 (i32, ...) @callee13(i32 1, i32 2) nounwind
   ret i32 %call
 }
 
diff --git a/test/CodeGen/Mips/tls.ll b/test/CodeGen/Mips/tls.ll
index b14ad5ba452b..97e270fc59a6 100644
--- a/test/CodeGen/Mips/tls.ll
+++ b/test/CodeGen/Mips/tls.ll
@@ -10,7 +10,7 @@
 
 define i32 @f1() nounwind {
 entry:
-  %tmp = load i32* @t1, align 4
+  %tmp = load i32, i32* @t1, align 4
   ret i32 %tmp
 
 ; PIC-LABEL:       f1:
@@ -33,7 +33,7 @@ entry:
 
 define i32 @f2() nounwind {
 entry:
-  %tmp = load i32* @t2, align 4
+  %tmp = load i32, i32* @t2, align 4
   ret i32 %tmp
 
 ; PIC-LABEL:       f2:
@@ -69,7 +69,7 @@ entry:
 ; PIC:   addu    $[[R1:[0-9]+]], $[[R0]], $2
 ; PIC:   lw      ${{[0-9]+}}, %dtprel_lo(f3.i)($[[R1]])
 
-  %0 = load i32* @f3.i, align 4
+  %0 = load i32, i32* @f3.i, align 4
   %inc = add nsw i32 %0, 1
   store i32 %inc, i32* @f3.i, align 4
   ret i32 %inc
diff --git a/test/CodeGen/Mips/tls16.ll b/test/CodeGen/Mips/tls16.ll
index 861864bcfe0f..3d324d7ed1e8 100644
--- a/test/CodeGen/Mips/tls16.ll
+++ b/test/CodeGen/Mips/tls16.ll
@@ -4,7 +4,7 @@
 
 define i32 @foo() nounwind readonly {
 entry:
-  %0 = load i32* @a, align 4
+  %0 = load i32, i32* @a, align 4
 ; PIC16:	lw	${{[0-9]+}}, %call16(__tls_get_addr)(${{[0-9]+}})
 ; PIC16:	addiu	${{[0-9]+}}, %tlsgd(a)
   ret i32 %0
diff --git a/test/CodeGen/Mips/tls16_2.ll b/test/CodeGen/Mips/tls16_2.ll
index b33e3c3766b6..0a6a4123e116 100644
--- a/test/CodeGen/Mips/tls16_2.ll
+++ b/test/CodeGen/Mips/tls16_2.ll
@@ -4,7 +4,7 @@
 
 define i8* @f(i8* nocapture %a) nounwind {
 entry:
-  %0 = load i32* @f.i, align 4
+  %0 = load i32, i32* @f.i, align 4
   %inc = add nsw i32 %0, 1
   store i32 %inc, i32* @f.i, align 4
   %1 = inttoptr i32 %inc to i8*
diff --git a/test/CodeGen/Mips/uitofp.ll b/test/CodeGen/Mips/uitofp.ll
index aff70c24f07c..83c2069f9661 100644
--- a/test/CodeGen/Mips/uitofp.ll
+++ b/test/CodeGen/Mips/uitofp.ll
@@ -5,7 +5,7 @@ entry:
   %b = alloca i32, align 4
   %a = alloca float, align 4
   store volatile i32 1, i32* %b, align 4
-  %0 = load volatile i32* %b, align 4
+  %0 = load volatile i32, i32* %b, align 4
   %conv = uitofp i32 %0 to float
   store float %conv, float* %a, align 4
   ret void
diff --git a/test/CodeGen/Mips/ul1.ll b/test/CodeGen/Mips/ul1.ll
index 7e64ff4d90fd..ad0992954631 100644
--- a/test/CodeGen/Mips/ul1.ll
+++ b/test/CodeGen/Mips/ul1.ll
@@ -5,7 +5,7 @@
 
 define i32 @main() nounwind {
 entry:
-  store i32 10, i32* getelementptr inbounds (%struct.ua* @foo, i32 0, i32 1), align 1
+  store i32 10, i32* getelementptr inbounds (%struct.ua, %struct.ua* @foo, i32 0, i32 1), align 1
 ; 16:   sb  ${{[0-9]+}}, {{[0-9]+}}(${{[0-9]+}})
 ; 16:   sb  ${{[0-9]+}}, {{[0-9]+}}(${{[0-9]+}})
 ; 16:   sb  ${{[0-9]+}}, {{[0-9]+}}(${{[0-9]+}})
diff --git a/test/CodeGen/Mips/unalignedload.ll b/test/CodeGen/Mips/unalignedload.ll
index 2002b1c60abe..9e453a6e794b 100644
--- a/test/CodeGen/Mips/unalignedload.ll
+++ b/test/CodeGen/Mips/unalignedload.ll
@@ -30,7 +30,7 @@ entry:
 
 ; MIPS32R6-DAG:  lhu $[[PART1:[0-9]+]], 2($[[R0]])
 
-  tail call void @foo2(%struct.S1* byval getelementptr inbounds (%struct.S2* @s2, i32 0, i32 1)) nounwind
+  tail call void @foo2(%struct.S1* byval getelementptr inbounds (%struct.S2, %struct.S2* @s2, i32 0, i32 1)) nounwind
   ret void
 }
 
diff --git a/test/CodeGen/Mips/vector-load-store.ll b/test/CodeGen/Mips/vector-load-store.ll
index d88996309908..61cbc5a6dee1 100644
--- a/test/CodeGen/Mips/vector-load-store.ll
+++ b/test/CodeGen/Mips/vector-load-store.ll
@@ -10,7 +10,7 @@ entry:
 ; CHECK: lw
 ; CHECK: sw
 
-  %0 = load <2 x i16>* @g1, align 4
+  %0 = load <2 x i16>, <2 x i16>* @g1, align 4
   store <2 x i16> %0, <2 x i16>* @g0, align 4
   ret void
 }
@@ -20,7 +20,7 @@ entry:
 ; CHECK: lw
 ; CHECK: sw
 
-  %0 = load <4 x i8>* @g3, align 4
+  %0 = load <4 x i8>, <4 x i8>* @g3, align 4
   store <4 x i8> %0, <4 x i8>* @g2, align 4
   ret void
 }
diff --git a/test/CodeGen/Mips/vector-setcc.ll b/test/CodeGen/Mips/vector-setcc.ll
index aeff4918c8bb..64b84e40513e 100644
--- a/test/CodeGen/Mips/vector-setcc.ll
+++ b/test/CodeGen/Mips/vector-setcc.ll
@@ -6,8 +6,8 @@
 
 define void @foo0() nounwind {
 entry:
-  %0 = load <4 x i32>* @a, align 16
-  %1 = load <4 x i32>* @b, align 16
+  %0 = load <4 x i32>, <4 x i32>* @a, align 16
+  %1 = load <4 x i32>, <4 x i32>* @b, align 16
   %cmp = icmp slt <4 x i32> %0, %1
   %sext = sext <4 x i1> %cmp to <4 x i32>
   store <4 x i32> %sext, <4 x i32>* @g0, align 16
diff --git a/test/CodeGen/Mips/xor1.ll b/test/CodeGen/Mips/xor1.ll
index f2c13169cf7a..dd51f143bb6c 100644
--- a/test/CodeGen/Mips/xor1.ll
+++ b/test/CodeGen/Mips/xor1.ll
@@ -6,11 +6,11 @@
 
 define i32 @main() nounwind {
 entry:
-  %0 = load i32* @x, align 4
-  %1 = load i32* @y, align 4
+  %0 = load i32, i32* @x, align 4
+  %1 = load i32, i32* @y, align 4
   %xor = xor i32 %0, %1
 ; 16:	xor	${{[0-9]+}}, ${{[0-9]+}}
-  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([7 x i8]* @.str, i32 0, i32 0), i32 %xor)
+  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([7 x i8], [7 x i8]* @.str, i32 0, i32 0), i32 %xor)
   ret i32 0
 }
 
diff --git a/test/CodeGen/Mips/zeroreg.ll b/test/CodeGen/Mips/zeroreg.ll
index c766d3b3cc2a..6baf9d4fbff2 100644
--- a/test/CodeGen/Mips/zeroreg.ll
+++ b/test/CodeGen/Mips/zeroreg.ll
@@ -25,7 +25,7 @@ entry:
 ; 64R6:          seleqz $2, $[[R0]], $4
 
   %tobool = icmp ne i32 %s, 0
-  %0 = load i32* @g1, align 4
+  %0 = load i32, i32* @g1, align 4
   %cond = select i1 %tobool, i32 0, i32 %0
   ret i32 %cond
 }
@@ -47,7 +47,7 @@ entry:
 ; 64R6:          selnez $2, $[[R0]], $4
 
   %tobool = icmp ne i32 %s, 0
-  %0 = load i32* @g1, align 4
+  %0 = load i32, i32* @g1, align 4
   %cond = select i1 %tobool, i32 %0, i32 0
   ret i32 %cond
 }
@@ -76,7 +76,7 @@ entry:
 ; 64R6:          seleqz $2, $[[R0]], $4
 
   %tobool = icmp ne i64 %s, 0
-  %0 = load i64* @g2, align 4
+  %0 = load i64, i64* @g2, align 4
   %cond = select i1 %tobool, i64 0, i64 %0
   ret i64 %cond
 }
@@ -103,7 +103,7 @@ entry:
 ; 64R6:          selnez $2, $[[R0]], $4
 
   %tobool = icmp ne i64 %s, 0
-  %0 = load i64* @g2, align 4
+  %0 = load i64, i64* @g2, align 4
   %cond = select i1 %tobool, i64 %0, i64 0
   ret i64 %cond
 }