Change register allocation order for ARM VFP and NEON registers to put the

callee-saved registers at the end of the lists.  Also prefer to avoid using
the low registers that are in register subclasses required by certain
instructions, so that those registers will more likely be available when needed.
This change makes a huge improvement in spilling in some cases.  Thanks to
Jakob for helping me realize the problem.

Most of this patch is fixing the testsuite.  There are quite a few places
where we're checking for specific registers.  I changed those to wildcards
in places where that doesn't weaken the tests.  The spill-q.ll and
thumb2-spill-q.ll tests stopped spilling with this change, so I added a bunch
of live values to force spills on those tests.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116055 91177308-0d34-0410-b5e6-96231b3b80d8

20 files changed, 235 insertions, 111 deletions

diff --git a/lib/Target/ARM/ARMRegisterInfo.td b/lib/Target/ARM/ARMRegisterInfo.td
index 305b232..42222f5 100644
--- a/lib/Target/ARM/ARMRegisterInfo.td
+++ b/lib/Target/ARM/ARMRegisterInfo.td
@@ -387,16 +387,18 @@ def DPR : RegisterClass<"ARM", [f64, v8i8, v4i16, v2i32, v1i64, v2f32], 64,
       ARM::D4,  ARM::D5,  ARM::D6,  ARM::D7,
       ARM::D8,  ARM::D9,  ARM::D10, ARM::D11,
       ARM::D12, ARM::D13, ARM::D14, ARM::D15 };
-    // VFP3
+    // VFP3: D8-D15 are callee saved and should be allocated last.
+    // Save other low registers for use as DPR_VFP2 and DPR_8 classes.
     static const unsigned ARM_DPR_VFP3[] = {
-      ARM::D0,  ARM::D1,  ARM::D2,  ARM::D3,
-      ARM::D4,  ARM::D5,  ARM::D6,  ARM::D7,
-      ARM::D8,  ARM::D9,  ARM::D10, ARM::D11,
-      ARM::D12, ARM::D13, ARM::D14, ARM::D15,
       ARM::D16, ARM::D17, ARM::D18, ARM::D19,
       ARM::D20, ARM::D21, ARM::D22, ARM::D23,
       ARM::D24, ARM::D25, ARM::D26, ARM::D27,
-      ARM::D28, ARM::D29, ARM::D30, ARM::D31 };
+      ARM::D28, ARM::D29, ARM::D30, ARM::D31,
+      ARM::D0,  ARM::D1,  ARM::D2,  ARM::D3,
+      ARM::D4,  ARM::D5,  ARM::D6,  ARM::D7,
+      ARM::D8,  ARM::D9,  ARM::D10, ARM::D11,
+      ARM::D12, ARM::D13, ARM::D14, ARM::D15 };
+
     DPRClass::iterator
     DPRClass::allocation_order_begin(const MachineFunction &MF) const {
       const TargetMachine &TM = MF.getTarget();
@@ -438,6 +440,29 @@ def QPR : RegisterClass<"ARM", [v16i8, v8i16, v4i32, v2i64, v4f32, v2f64], 128,
                         [Q0,  Q1,  Q2,  Q3,  Q4,  Q5,  Q6,  Q7,
                          Q8,  Q9,  Q10, Q11, Q12, Q13, Q14, Q15]> {
   let SubRegClasses = [(DPR dsub_0, dsub_1)];
+  let MethodProtos = [{
+    iterator allocation_order_begin(const MachineFunction &MF) const;
+    iterator allocation_order_end(const MachineFunction &MF) const;
+  }];
+  let MethodBodies = [{
+    // Q4-Q7 are callee saved and should be allocated last.
+    // Save other low registers for use as QPR_VFP2 and QPR_8 classes.
+    static const unsigned ARM_QPR[] = {
+      ARM::Q8,  ARM::Q9,  ARM::Q10, ARM::Q11,
+      ARM::Q12, ARM::Q13, ARM::Q14, ARM::Q15,
+      ARM::Q0,  ARM::Q1,  ARM::Q2,  ARM::Q3,
+      ARM::Q4,  ARM::Q5,  ARM::Q6,  ARM::Q7 };
+
+    QPRClass::iterator
+    QPRClass::allocation_order_begin(const MachineFunction &MF) const {
+      return ARM_QPR;
+    }
+
+    QPRClass::iterator
+    QPRClass::allocation_order_end(const MachineFunction &MF) const {
+      return ARM_QPR + (sizeof(ARM_QPR)/sizeof(unsigned));
+    }
+  }];
 }
 
 // Subset of QPR that have 32-bit SPR subregs.
@@ -463,6 +488,27 @@ def QQPR : RegisterClass<"ARM", [v4i64],
                          [QQ0, QQ1, QQ2, QQ3, QQ4, QQ5, QQ6, QQ7]> {
   let SubRegClasses = [(DPR dsub_0, dsub_1, dsub_2, dsub_3),
                        (QPR qsub_0, qsub_1)];
+  let MethodProtos = [{
+    iterator allocation_order_begin(const MachineFunction &MF) const;
+    iterator allocation_order_end(const MachineFunction &MF) const;
+  }];
+  let MethodBodies = [{
+    // QQ2-QQ3 are callee saved and should be allocated last.
+    // Save other low registers for use as QPR_VFP2 and QPR_8 classes.
+    static const unsigned ARM_QQPR[] = {
+      ARM::QQ4, ARM::QQ5, ARM::QQ6, ARM::QQ7,
+      ARM::QQ0, ARM::QQ1, ARM::QQ2, ARM::QQ3 };
+
+    QQPRClass::iterator
+    QQPRClass::allocation_order_begin(const MachineFunction &MF) const {
+      return ARM_QQPR;
+    }
+
+    QQPRClass::iterator
+    QQPRClass::allocation_order_end(const MachineFunction &MF) const {
+      return ARM_QQPR + (sizeof(ARM_QQPR)/sizeof(unsigned));
+    }
+  }];
 }
 
 // Subset of QQPR that have 32-bit SPR subregs.
@@ -483,6 +529,26 @@ def QQQQPR : RegisterClass<"ARM", [v8i64],
   let SubRegClasses = [(DPR dsub_0, dsub_1, dsub_2, dsub_3,
                             dsub_4, dsub_5, dsub_6, dsub_7),
                        (QPR qsub_0, qsub_1, qsub_2, qsub_3)];
+  let MethodProtos = [{
+    iterator allocation_order_begin(const MachineFunction &MF) const;
+    iterator allocation_order_end(const MachineFunction &MF) const;
+  }];
+  let MethodBodies = [{
+    // QQQQ1 is callee saved and should be allocated last.
+    // Save QQQQ0 for use as QPR_VFP2 and QPR_8 classes.
+    static const unsigned ARM_QQQQPR[] = {
+      ARM::QQQQ2, ARM::QQQQ3, ARM::QQQQ0, ARM::QQQQ1 };
+
+    QQQQPRClass::iterator
+    QQQQPRClass::allocation_order_begin(const MachineFunction &MF) const {
+      return ARM_QQQQPR;
+    }
+
+    QQQQPRClass::iterator
+    QQQQPRClass::allocation_order_end(const MachineFunction &MF) const {
+      return ARM_QQQQPR + (sizeof(ARM_QQQQPR)/sizeof(unsigned));
+    }
+  }];
 }
 
 // Condition code registers.
diff --git a/test/CodeGen/ARM/2010-06-29-PartialRedefFastAlloc.ll b/test/CodeGen/ARM/2010-06-29-PartialRedefFastAlloc.ll
index ffc47eb..b9d5600 100644
--- a/test/CodeGen/ARM/2010-06-29-PartialRedefFastAlloc.ll
+++ b/test/CodeGen/ARM/2010-06-29-PartialRedefFastAlloc.ll
@@ -10,9 +10,9 @@ target triple = "thumbv7-apple-darwin10"
 ; %reg1028 gets allocated %Q0, and if %reg1030 is reloaded for the partial
 ; redef, it cannot also get %Q0.
 
-; CHECK: vld1.64 {d0, d1}, [r{{.}}]
-; CHECK-NOT: vld1.64 {d0, d1}
-; CHECK: vmov.f64 d3, d0
+; CHECK: vld1.64 {d16, d17}, [r{{.}}]
+; CHECK-NOT: vld1.64 {d16, d17}
+; CHECK: vmov.f64 d19, d16
 
 define i32 @test(i8* %arg) nounwind {
 entry:
diff --git a/test/CodeGen/ARM/fpconsts.ll b/test/CodeGen/ARM/fpconsts.ll
index f1d6a16..9e7a8ae 100644
--- a/test/CodeGen/ARM/fpconsts.ll
+++ b/test/CodeGen/ARM/fpconsts.ll
@@ -11,7 +11,7 @@ entry:
 define double @t2(double %x) nounwind readnone optsize {
 entry:
 ; CHECK: t2:
-; CHECK: vmov.f64 d1, #3.000000e+00
+; CHECK: vmov.f64 d{{.*}}, #3.000000e+00
   %0 = fadd double %x, 3.000000e+00
   ret double %0
 }
@@ -19,7 +19,7 @@ entry:
 define double @t3(double %x) nounwind readnone optsize {
 entry:
 ; CHECK: t3:
-; CHECK: vmov.f64 d1, #-1.300000e+01
+; CHECK: vmov.f64 d{{.*}}, #-1.300000e+01
   %0 = fmul double %x, -1.300000e+01
   ret double %0
 }
diff --git a/test/CodeGen/ARM/inlineasm3.ll b/test/CodeGen/ARM/inlineasm3.ll
index 687e138..9f77ad1 100644
--- a/test/CodeGen/ARM/inlineasm3.ll
+++ b/test/CodeGen/ARM/inlineasm3.ll
@@ -7,7 +7,7 @@ define void @t() nounwind {
 entry:
 ; CHECK: vmov.I64 q15, #0
 ; CHECK: vmov.32 d30[0], r0
-; CHECK: vmov q0, q15
+; CHECK: vmov q8, q15
   %tmp = alloca %struct.int32x4_t, align 16
   call void asm sideeffect "vmov.I64 q15, #0\0Avmov.32 d30[0], $1\0Avmov ${0:q}, q15\0A", "=*w,r,~{d31},~{d30}"(%struct.int32x4_t* %tmp, i32 8192) nounwind
   ret void
@@ -18,7 +18,7 @@ entry:
 
 define void @t2() nounwind {
 entry:
-; CHECK: vmov d30, d0
+; CHECK: vmov d30, d16
 ; CHECK: vmov.32 r0, d30[0]
   %asmtmp2 = tail call i32 asm sideeffect "vmov d30, $1\0Avmov.32 $0, d30[0]\0A", "=r,w,~{d30}"(<2 x i32> undef) nounwind
   ret void
diff --git a/test/CodeGen/ARM/reg_sequence.ll b/test/CodeGen/ARM/reg_sequence.ll
index 2e4f10d..b96762a 100644
--- a/test/CodeGen/ARM/reg_sequence.ll
+++ b/test/CodeGen/ARM/reg_sequence.ll
@@ -122,9 +122,9 @@ return1:
 return2:
 ; CHECK:        %return2
 ; CHECK:        vadd.i32
-; CHECK:        vmov q1, q3
+; CHECK:        vmov q9, q11
 ; CHECK-NOT:    vmov
-; CHECK:        vst2.32 {d0, d1, d2, d3}
+; CHECK:        vst2.32 {d16, d17, d18, d19}
   %tmp100 = extractvalue %struct.__neon_int32x4x2_t %tmp2, 0 ; <<4 x i32>> [#uses=1]
   %tmp101 = extractvalue %struct.__neon_int32x4x2_t %tmp5, 1 ; <<4 x i32>> [#uses=1]
   %tmp102 = add <4 x i32> %tmp100, %tmp101              ; <<4 x i32>> [#uses=1]
@@ -136,9 +136,9 @@ return2:
 define <8 x i16> @t5(i16* %A, <8 x i16>* %B) nounwind {
 ; CHECK:        t5:
 ; CHECK:        vldmia
-; CHECK:        vmov q1, q0
+; CHECK:        vmov q9, q8
 ; CHECK-NOT:    vmov
-; CHECK:        vld2.16 {d0[1], d2[1]}, [r0]
+; CHECK:        vld2.16 {d16[1], d18[1]}, [r0]
 ; CHECK-NOT:    vmov
 ; CHECK:        vadd.i16
   %tmp0 = bitcast i16* %A to i8*                  ; <i8*> [#uses=1]
@@ -153,8 +153,8 @@ define <8 x i16> @t5(i16* %A, <8 x i16>* %B) nounwind {
 define <8 x i8> @t6(i8* %A, <8 x i8>* %B) nounwind {
 ; CHECK:        t6:
 ; CHECK:        vldr.64
-; CHECK:        vmov d1, d0
-; CHECK-NEXT:   vld2.8 {d0[1], d1[1]}
+; CHECK:        vmov d17, d16
+; CHECK-NEXT:   vld2.8 {d16[1], d17[1]}
   %tmp1 = load <8 x i8>* %B                       ; <<8 x i8>> [#uses=2]
   %tmp2 = call %struct.__neon_int8x8x2_t @llvm.arm.neon.vld2lane.v8i8(i8* %A, <8 x i8> %tmp1, <8 x i8> %tmp1, i32 1, i32 1) ; <%struct.__neon_int8x8x2_t> [#uses=2]
   %tmp3 = extractvalue %struct.__neon_int8x8x2_t %tmp2, 0 ; <<8 x i8>> [#uses=1]
@@ -168,10 +168,10 @@ entry:
 ; CHECK:        t7:
 ; CHECK:        vld2.32
 ; CHECK:        vst2.32
-; CHECK:        vld1.32 {d0, d1},
-; CHECK:        vmov q1, q0
+; CHECK:        vld1.32 {d16, d17},
+; CHECK:        vmov q9, q8
 ; CHECK-NOT:    vmov
-; CHECK:        vuzp.32 q0, q1
+; CHECK:        vuzp.32 q8, q9
 ; CHECK:        vst1.32
   %0 = bitcast i32* %iptr to i8*                  ; <i8*> [#uses=2]
   %1 = tail call %struct.__neon_int32x4x2_t @llvm.arm.neon.vld2.v4i32(i8* %0, i32 1) ; <%struct.__neon_int32x4x2_t> [#uses=2]
@@ -188,7 +188,7 @@ entry:
 ; PR7156
 define arm_aapcs_vfpcc i32 @t8() nounwind {
 ; CHECK: t8:
-; CHECK: vrsqrte.f32 q0, q0
+; CHECK: vrsqrte.f32 q8, q8
 bb.nph55.bb.nph55.split_crit_edge:
   br label %bb3
 
@@ -238,10 +238,10 @@ bb14:                                             ; preds = %bb6
 define arm_aapcs_vfpcc float @t9(%0* nocapture, %3* nocapture) nounwind {
 ; CHECK:        t9:
 ; CHECK:        vldr.64
-; CHECK-NOT:    vmov d{{.*}}, d0
-; CHECK:        vmov.i32 d1
-; CHECK-NEXT:   vstmia r0, {d0, d1}
-; CHECK-NEXT:   vstmia r0, {d0, d1}
+; CHECK-NOT:    vmov d{{.*}}, d16
+; CHECK:        vmov.i32 d17
+; CHECK-NEXT:   vstmia r0, {d16, d17}
+; CHECK-NEXT:   vstmia r0, {d16, d17}
   %3 = bitcast double 0.000000e+00 to <2 x float> ; <<2 x float>> [#uses=2]
   %4 = shufflevector <2 x float> %3, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; <<4 x float>> [#uses=1]
   store <4 x float> %4, <4 x float>* undef, align 16
@@ -269,9 +269,9 @@ define arm_aapcs_vfpcc float @t9(%0* nocapture, %3* nocapture) nounwind {
 define arm_aapcs_vfpcc i32 @t10() nounwind {
 entry:
 ; CHECK: t10:
-; CHECK: vmov.i32 q1, #0x3F000000
-; CHECK: vmov d0, d1
-; CHECK: vmla.f32 q0, q0, d0[0]
+; CHECK: vmov.i32 q9, #0x3F000000
+; CHECK: vmov d0, d17
+; CHECK: vmla.f32 q8, q8, d0[0]
   %0 = shufflevector <4 x float> zeroinitializer, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
   %1 = insertelement <4 x float> %0, float undef, i32 1 ; <<4 x float>> [#uses=1]
   %2 = insertelement <4 x float> %1, float undef, i32 2 ; <<4 x float>> [#uses=1]
diff --git a/test/CodeGen/ARM/spill-q.ll b/test/CodeGen/ARM/spill-q.ll
index ae1ba2f..dc83e25 100644
--- a/test/CodeGen/ARM/spill-q.ll
+++ b/test/CodeGen/ARM/spill-q.ll
@@ -20,6 +20,26 @@ entry:
   %1 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind ; <<4 x float>> [#uses=1]
   store float 0.000000e+00, float* undef, align 4
   %2 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind ; <<4 x float>> [#uses=1]
+  %ld3 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld4 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld5 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld6 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld7 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld8 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld9 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld10 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld11 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld12 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
   %val173 = load <4 x float>* undef               ; <<4 x float>> [#uses=1]
   br label %bb4
 
@@ -44,7 +64,16 @@ bb4:                                              ; preds = %bb193, %entry
   %18 = fmul <4 x float> %17, %val173             ; <<4 x float>> [#uses=1]
   %19 = shufflevector <4 x float> %18, <4 x float> undef, <2 x i32> <i32 2, i32 3> ; <<2 x float>> [#uses=1]
   %20 = shufflevector <2 x float> %19, <2 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
-  %21 = fadd <4 x float> zeroinitializer, %20     ; <<4 x float>> [#uses=2]
+  %tmp1 = fadd <4 x float> %20, %ld3
+  %tmp2 = fadd <4 x float> %tmp1, %ld4
+  %tmp3 = fadd <4 x float> %tmp2, %ld5
+  %tmp4 = fadd <4 x float> %tmp3, %ld6
+  %tmp5 = fadd <4 x float> %tmp4, %ld7
+  %tmp6 = fadd <4 x float> %tmp5, %ld8
+  %tmp7 = fadd <4 x float> %tmp6, %ld9
+  %tmp8 = fadd <4 x float> %tmp7, %ld10
+  %tmp9 = fadd <4 x float> %tmp8, %ld11
+  %21 = fadd <4 x float> %tmp9, %ld12
   %22 = fcmp ogt <4 x float> %besterror.0.2264, %21 ; <<4 x i1>> [#uses=0]
   %tmp = extractelement <4 x i1> %22, i32 0
   br i1 %tmp, label %bb193, label %bb186
diff --git a/test/CodeGen/ARM/vcgt.ll b/test/CodeGen/ARM/vcgt.ll
index 194093c..e3318cd 100644
--- a/test/CodeGen/ARM/vcgt.ll
+++ b/test/CodeGen/ARM/vcgt.ll
@@ -161,9 +161,9 @@ define <4 x i32> @vacgtQf32(<4 x float>* %A, <4 x float>* %B) nounwind {
 ; rdar://7923010
 define <4 x i32> @vcgt_zext(<4 x float>* %A, <4 x float>* %B) nounwind {
 ;CHECK: vcgt_zext:
-;CHECK: vcgt.f32 q0
-;CHECK: vmov.i32 q1, #0x1
-;CHECK: vand q0, q0, q1
+;CHECK: vcgt.f32 q8
+;CHECK: vmov.i32 q9, #0x1
+;CHECK: vand q8, q8, q9
 	%tmp1 = load <4 x float>* %A
 	%tmp2 = load <4 x float>* %B
 	%tmp3 = fcmp ogt <4 x float> %tmp1, %tmp2
diff --git a/test/CodeGen/ARM/vget_lane.ll b/test/CodeGen/ARM/vget_lane.ll
index 05e7f50..0fd1843 100644
--- a/test/CodeGen/ARM/vget_lane.ll
+++ b/test/CodeGen/ARM/vget_lane.ll
@@ -96,7 +96,7 @@ define i32 @vgetQ_lanei32(<4 x i32>* %A) nounwind {
 
 define arm_aapcs_vfpcc void @test_vget_laneu16() nounwind {
 entry:
-; CHECK: vmov.u16 r0, d0[1]
+; CHECK: vmov.u16 r0, d{{.*}}[1]
   %arg0_uint16x4_t = alloca <4 x i16>             ; <<4 x i16>*> [#uses=1]
   %out_uint16_t = alloca i16                      ; <i16*> [#uses=1]
   %"alloca point" = bitcast i32 0 to i32          ; <i32> [#uses=0]
@@ -111,7 +111,7 @@ return:                                           ; preds = %entry
 
 define arm_aapcs_vfpcc void @test_vget_laneu8() nounwind {
 entry:
-; CHECK: vmov.u8 r0, d0[1]
+; CHECK: vmov.u8 r0, d{{.*}}[1]
   %arg0_uint8x8_t = alloca <8 x i8>               ; <<8 x i8>*> [#uses=1]
   %out_uint8_t = alloca i8                        ; <i8*> [#uses=1]
   %"alloca point" = bitcast i32 0 to i32          ; <i32> [#uses=0]
@@ -126,7 +126,7 @@ return:                                           ; preds = %entry
 
 define arm_aapcs_vfpcc void @test_vgetQ_laneu16() nounwind {
 entry:
-; CHECK: vmov.u16 r0, d0[1]
+; CHECK: vmov.u16 r0, d{{.*}}[1]
   %arg0_uint16x8_t = alloca <8 x i16>             ; <<8 x i16>*> [#uses=1]
   %out_uint16_t = alloca i16                      ; <i16*> [#uses=1]
   %"alloca point" = bitcast i32 0 to i32          ; <i32> [#uses=0]
@@ -141,7 +141,7 @@ return:                                           ; preds = %entry
 
 define arm_aapcs_vfpcc void @test_vgetQ_laneu8() nounwind {
 entry:
-; CHECK: vmov.u8 r0, d0[1]
+; CHECK: vmov.u8 r0, d{{.*}}[1]
   %arg0_uint8x16_t = alloca <16 x i8>             ; <<16 x i8>*> [#uses=1]
   %out_uint8_t = alloca i8                        ; <i8*> [#uses=1]
   %"alloca point" = bitcast i32 0 to i32          ; <i32> [#uses=0]
diff --git a/test/CodeGen/ARM/vld1.ll b/test/CodeGen/ARM/vld1.ll
index 2800475..16bd832 100644
--- a/test/CodeGen/ARM/vld1.ll
+++ b/test/CodeGen/ARM/vld1.ll
@@ -3,7 +3,7 @@
 define <8 x i8> @vld1i8(i8* %A) nounwind {
 ;CHECK: vld1i8:
 ;Check the alignment value.  Max for this instruction is 64 bits:
-;CHECK: vld1.8 {d0}, [r0, :64]
+;CHECK: vld1.8 {d16}, [r0, :64]
 	%tmp1 = call <8 x i8> @llvm.arm.neon.vld1.v8i8(i8* %A, i32 16)
 	ret <8 x i8> %tmp1
 }
@@ -43,7 +43,7 @@ define <1 x i64> @vld1i64(i64* %A) nounwind {
 define <16 x i8> @vld1Qi8(i8* %A) nounwind {
 ;CHECK: vld1Qi8:
 ;Check the alignment value.  Max for this instruction is 128 bits:
-;CHECK: vld1.8 {d0, d1}, [r0, :64]
+;CHECK: vld1.8 {d16, d17}, [r0, :64]
 	%tmp1 = call <16 x i8> @llvm.arm.neon.vld1.v16i8(i8* %A, i32 8)
 	ret <16 x i8> %tmp1
 }
@@ -51,7 +51,7 @@ define <16 x i8> @vld1Qi8(i8* %A) nounwind {
 define <8 x i16> @vld1Qi16(i16* %A) nounwind {
 ;CHECK: vld1Qi16:
 ;Check the alignment value.  Max for this instruction is 128 bits:
-;CHECK: vld1.16 {d0, d1}, [r0, :128]
+;CHECK: vld1.16 {d16, d17}, [r0, :128]
 	%tmp0 = bitcast i16* %A to i8*
 	%tmp1 = call <8 x i16> @llvm.arm.neon.vld1.v8i16(i8* %tmp0, i32 32)
 	ret <8 x i16> %tmp1
diff --git a/test/CodeGen/ARM/vld2.ll b/test/CodeGen/ARM/vld2.ll
index fe99dd5..3fdd7b7 100644
--- a/test/CodeGen/ARM/vld2.ll
+++ b/test/CodeGen/ARM/vld2.ll
@@ -14,7 +14,7 @@
 define <8 x i8> @vld2i8(i8* %A) nounwind {
 ;CHECK: vld2i8:
 ;Check the alignment value.  Max for this instruction is 128 bits:
-;CHECK: vld2.8 {d0, d1}, [r0, :64]
+;CHECK: vld2.8 {d16, d17}, [r0, :64]
 	%tmp1 = call %struct.__neon_int8x8x2_t @llvm.arm.neon.vld2.v8i8(i8* %A, i32 8)
         %tmp2 = extractvalue %struct.__neon_int8x8x2_t %tmp1, 0
         %tmp3 = extractvalue %struct.__neon_int8x8x2_t %tmp1, 1
@@ -25,7 +25,7 @@ define <8 x i8> @vld2i8(i8* %A) nounwind {
 define <4 x i16> @vld2i16(i16* %A) nounwind {
 ;CHECK: vld2i16:
 ;Check the alignment value.  Max for this instruction is 128 bits:
-;CHECK: vld2.16 {d0, d1}, [r0, :128]
+;CHECK: vld2.16 {d16, d17}, [r0, :128]
 	%tmp0 = bitcast i16* %A to i8*
 	%tmp1 = call %struct.__neon_int16x4x2_t @llvm.arm.neon.vld2.v4i16(i8* %tmp0, i32 32)
         %tmp2 = extractvalue %struct.__neon_int16x4x2_t %tmp1, 0
@@ -59,7 +59,7 @@ define <2 x float> @vld2f(float* %A) nounwind {
 define <1 x i64> @vld2i64(i64* %A) nounwind {
 ;CHECK: vld2i64:
 ;Check the alignment value.  Max for this instruction is 128 bits:
-;CHECK: vld1.64 {d0, d1}, [r0, :128]
+;CHECK: vld1.64 {d16, d17}, [r0, :128]
 	%tmp0 = bitcast i64* %A to i8*
 	%tmp1 = call %struct.__neon_int64x1x2_t @llvm.arm.neon.vld2.v1i64(i8* %tmp0, i32 32)
         %tmp2 = extractvalue %struct.__neon_int64x1x2_t %tmp1, 0
@@ -71,7 +71,7 @@ define <1 x i64> @vld2i64(i64* %A) nounwind {
 define <16 x i8> @vld2Qi8(i8* %A) nounwind {
 ;CHECK: vld2Qi8:
 ;Check the alignment value.  Max for this instruction is 256 bits:
-;CHECK: vld2.8 {d0, d1, d2, d3}, [r0, :64]
+;CHECK: vld2.8 {d16, d17, d18, d19}, [r0, :64]
 	%tmp1 = call %struct.__neon_int8x16x2_t @llvm.arm.neon.vld2.v16i8(i8* %A, i32 8)
         %tmp2 = extractvalue %struct.__neon_int8x16x2_t %tmp1, 0
         %tmp3 = extractvalue %struct.__neon_int8x16x2_t %tmp1, 1
@@ -82,7 +82,7 @@ define <16 x i8> @vld2Qi8(i8* %A) nounwind {
 define <8 x i16> @vld2Qi16(i16* %A) nounwind {
 ;CHECK: vld2Qi16:
 ;Check the alignment value.  Max for this instruction is 256 bits:
-;CHECK: vld2.16 {d0, d1, d2, d3}, [r0, :128]
+;CHECK: vld2.16 {d16, d17, d18, d19}, [r0, :128]
 	%tmp0 = bitcast i16* %A to i8*
 	%tmp1 = call %struct.__neon_int16x8x2_t @llvm.arm.neon.vld2.v8i16(i8* %tmp0, i32 16)
         %tmp2 = extractvalue %struct.__neon_int16x8x2_t %tmp1, 0
@@ -94,7 +94,7 @@ define <8 x i16> @vld2Qi16(i16* %A) nounwind {
 define <4 x i32> @vld2Qi32(i32* %A) nounwind {
 ;CHECK: vld2Qi32:
 ;Check the alignment value.  Max for this instruction is 256 bits:
-;CHECK: vld2.32 {d0, d1, d2, d3}, [r0, :256]
+;CHECK: vld2.32 {d16, d17, d18, d19}, [r0, :256]
 	%tmp0 = bitcast i32* %A to i8*
 	%tmp1 = call %struct.__neon_int32x4x2_t @llvm.arm.neon.vld2.v4i32(i8* %tmp0, i32 64)
         %tmp2 = extractvalue %struct.__neon_int32x4x2_t %tmp1, 0
diff --git a/test/CodeGen/ARM/vld3.ll b/test/CodeGen/ARM/vld3.ll
index 7c4b8e1..0e541bb 100644
--- a/test/CodeGen/ARM/vld3.ll
+++ b/test/CodeGen/ARM/vld3.ll
@@ -14,7 +14,7 @@
 define <8 x i8> @vld3i8(i8* %A) nounwind {
 ;CHECK: vld3i8:
 ;Check the alignment value.  Max for this instruction is 64 bits:
-;CHECK: vld3.8 {d0, d1, d2}, [r0, :64]
+;CHECK: vld3.8 {d16, d17, d18}, [r0, :64]
 	%tmp1 = call %struct.__neon_int8x8x3_t @llvm.arm.neon.vld3.v8i8(i8* %A, i32 32)
         %tmp2 = extractvalue %struct.__neon_int8x8x3_t %tmp1, 0
         %tmp3 = extractvalue %struct.__neon_int8x8x3_t %tmp1, 2
@@ -58,7 +58,7 @@ define <2 x float> @vld3f(float* %A) nounwind {
 define <1 x i64> @vld3i64(i64* %A) nounwind {
 ;CHECK: vld3i64:
 ;Check the alignment value.  Max for this instruction is 64 bits:
-;CHECK: vld1.64 {d0, d1, d2}, [r0, :64]
+;CHECK: vld1.64 {d16, d17, d18}, [r0, :64]
 	%tmp0 = bitcast i64* %A to i8*
 	%tmp1 = call %struct.__neon_int64x1x3_t @llvm.arm.neon.vld3.v1i64(i8* %tmp0, i32 16)
         %tmp2 = extractvalue %struct.__neon_int64x1x3_t %tmp1, 0
@@ -70,8 +70,8 @@ define <1 x i64> @vld3i64(i64* %A) nounwind {
 define <16 x i8> @vld3Qi8(i8* %A) nounwind {
 ;CHECK: vld3Qi8:
 ;Check the alignment value.  Max for this instruction is 64 bits:
-;CHECK: vld3.8 {d0, d2, d4}, [r0, :64]!
-;CHECK: vld3.8 {d1, d3, d5}, [r0, :64]
+;CHECK: vld3.8 {d16, d18, d20}, [r0, :64]!
+;CHECK: vld3.8 {d17, d19, d21}, [r0, :64]
 	%tmp1 = call %struct.__neon_int8x16x3_t @llvm.arm.neon.vld3.v16i8(i8* %A, i32 32)
         %tmp2 = extractvalue %struct.__neon_int8x16x3_t %tmp1, 0
         %tmp3 = extractvalue %struct.__neon_int8x16x3_t %tmp1, 2
diff --git a/test/CodeGen/ARM/vld4.ll b/test/CodeGen/ARM/vld4.ll
index 6330d4f..a616a98 100644
--- a/test/CodeGen/ARM/vld4.ll
+++ b/test/CodeGen/ARM/vld4.ll
@@ -14,7 +14,7 @@
 define <8 x i8> @vld4i8(i8* %A) nounwind {
 ;CHECK: vld4i8:
 ;Check the alignment value.  Max for this instruction is 256 bits:
-;CHECK: vld4.8 {d0, d1, d2, d3}, [r0, :64]
+;CHECK: vld4.8 {d16, d17, d18, d19}, [r0, :64]
 	%tmp1 = call %struct.__neon_int8x8x4_t @llvm.arm.neon.vld4.v8i8(i8* %A, i32 8)
         %tmp2 = extractvalue %struct.__neon_int8x8x4_t %tmp1, 0
         %tmp3 = extractvalue %struct.__neon_int8x8x4_t %tmp1, 2
@@ -25,7 +25,7 @@ define <8 x i8> @vld4i8(i8* %A) nounwind {
 define <4 x i16> @vld4i16(i16* %A) nounwind {
 ;CHECK: vld4i16:
 ;Check the alignment value.  Max for this instruction is 256 bits:
-;CHECK: vld4.16 {d0, d1, d2, d3}, [r0, :128]
+;CHECK: vld4.16 {d16, d17, d18, d19}, [r0, :128]
 	%tmp0 = bitcast i16* %A to i8*
 	%tmp1 = call %struct.__neon_int16x4x4_t @llvm.arm.neon.vld4.v4i16(i8* %tmp0, i32 16)
         %tmp2 = extractvalue %struct.__neon_int16x4x4_t %tmp1, 0
@@ -37,7 +37,7 @@ define <4 x i16> @vld4i16(i16* %A) nounwind {
 define <2 x i32> @vld4i32(i32* %A) nounwind {
 ;CHECK: vld4i32:
 ;Check the alignment value.  Max for this instruction is 256 bits:
-;CHECK: vld4.32 {d0, d1, d2, d3}, [r0, :256]
+;CHECK: vld4.32 {d16, d17, d18, d19}, [r0, :256]
 	%tmp0 = bitcast i32* %A to i8*
 	%tmp1 = call %struct.__neon_int32x2x4_t @llvm.arm.neon.vld4.v2i32(i8* %tmp0, i32 32)
         %tmp2 = extractvalue %struct.__neon_int32x2x4_t %tmp1, 0
@@ -60,7 +60,7 @@ define <2 x float> @vld4f(float* %A) nounwind {
 define <1 x i64> @vld4i64(i64* %A) nounwind {
 ;CHECK: vld4i64:
 ;Check the alignment value.  Max for this instruction is 256 bits:
-;CHECK: vld1.64 {d0, d1, d2, d3}, [r0, :256]
+;CHECK: vld1.64 {d16, d17, d18, d19}, [r0, :256]
 	%tmp0 = bitcast i64* %A to i8*
 	%tmp1 = call %struct.__neon_int64x1x4_t @llvm.arm.neon.vld4.v1i64(i8* %tmp0, i32 64)
         %tmp2 = extractvalue %struct.__neon_int64x1x4_t %tmp1, 0
@@ -72,8 +72,8 @@ define <1 x i64> @vld4i64(i64* %A) nounwind {
 define <16 x i8> @vld4Qi8(i8* %A) nounwind {
 ;CHECK: vld4Qi8:
 ;Check the alignment value.  Max for this instruction is 256 bits:
-;CHECK: vld4.8 {d0, d2, d4, d6}, [r0, :256]!
-;CHECK: vld4.8 {d1, d3, d5, d7}, [r0, :256]
+;CHECK: vld4.8 {d16, d18, d20, d22}, [r0, :256]!
+;CHECK: vld4.8 {d17, d19, d21, d23}, [r0, :256]
 	%tmp1 = call %struct.__neon_int8x16x4_t @llvm.arm.neon.vld4.v16i8(i8* %A, i32 64)
         %tmp2 = extractvalue %struct.__neon_int8x16x4_t %tmp1, 0
         %tmp3 = extractvalue %struct.__neon_int8x16x4_t %tmp1, 2
@@ -84,8 +84,8 @@ define <16 x i8> @vld4Qi8(i8* %A) nounwind {
 define <8 x i16> @vld4Qi16(i16* %A) nounwind {
 ;CHECK: vld4Qi16:
 ;Check for no alignment specifier.
-;CHECK: vld4.16 {d0, d2, d4, d6}, [r0]!
-;CHECK: vld4.16 {d1, d3, d5, d7}, [r0]
+;CHECK: vld4.16 {d16, d18, d20, d22}, [r0]!
+;CHECK: vld4.16 {d17, d19, d21, d23}, [r0]
 	%tmp0 = bitcast i16* %A to i8*
 	%tmp1 = call %struct.__neon_int16x8x4_t @llvm.arm.neon.vld4.v8i16(i8* %tmp0, i32 1)
         %tmp2 = extractvalue %struct.__neon_int16x8x4_t %tmp1, 0
diff --git a/test/CodeGen/ARM/vmov.ll b/test/CodeGen/ARM/vmov.ll
index 8cd9457..4d654d7 100644
--- a/test/CodeGen/ARM/vmov.ll
+++ b/test/CodeGen/ARM/vmov.ll
@@ -2,169 +2,169 @@
 
 define <8 x i8> @v_movi8() nounwind {
 ;CHECK: v_movi8:
-;CHECK: vmov.i8 d0, #0x8
+;CHECK: vmov.i8 d{{.*}}, #0x8
 	ret <8 x i8> < i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8 >
 }
 
 define <4 x i16> @v_movi16a() nounwind {
 ;CHECK: v_movi16a:
-;CHECK: vmov.i16 d0, #0x10
+;CHECK: vmov.i16 d{{.*}}, #0x10
 	ret <4 x i16> < i16 16, i16 16, i16 16, i16 16 >
 }
 
 define <4 x i16> @v_movi16b() nounwind {
 ;CHECK: v_movi16b:
-;CHECK: vmov.i16 d0, #0x1000
+;CHECK: vmov.i16 d{{.*}}, #0x1000
 	ret <4 x i16> < i16 4096, i16 4096, i16 4096, i16 4096 >
 }
 
 define <4 x i16> @v_mvni16a() nounwind {
 ;CHECK: v_mvni16a:
-;CHECK: vmvn.i16 d0, #0x10
+;CHECK: vmvn.i16 d{{.*}}, #0x10
 	ret <4 x i16> < i16 65519, i16 65519, i16 65519, i16 65519 >
 }
 
 define <4 x i16> @v_mvni16b() nounwind {
 ;CHECK: v_mvni16b:
-;CHECK: vmvn.i16 d0, #0x1000
+;CHECK: vmvn.i16 d{{.*}}, #0x1000
 	ret <4 x i16> < i16 61439, i16 61439, i16 61439, i16 61439 >
 }
 
 define <2 x i32> @v_movi32a() nounwind {
 ;CHECK: v_movi32a:
-;CHECK: vmov.i32 d0, #0x20
+;CHECK: vmov.i32 d{{.*}}, #0x20
 	ret <2 x i32> < i32 32, i32 32 >
 }
 
 define <2 x i32> @v_movi32b() nounwind {
 ;CHECK: v_movi32b:
-;CHECK: vmov.i32 d0, #0x2000
+;CHECK: vmov.i32 d{{.*}}, #0x2000
 	ret <2 x i32> < i32 8192, i32 8192 >
 }
 
 define <2 x i32> @v_movi32c() nounwind {
 ;CHECK: v_movi32c:
-;CHECK: vmov.i32 d0, #0x200000
+;CHECK: vmov.i32 d{{.*}}, #0x200000
 	ret <2 x i32> < i32 2097152, i32 2097152 >
 }
 
 define <2 x i32> @v_movi32d() nounwind {
 ;CHECK: v_movi32d:
-;CHECK: vmov.i32 d0, #0x20000000
+;CHECK: vmov.i32 d{{.*}}, #0x20000000
 	ret <2 x i32> < i32 536870912, i32 536870912 >
 }
 
 define <2 x i32> @v_movi32e() nounwind {
 ;CHECK: v_movi32e:
-;CHECK: vmov.i32 d0, #0x20FF
+;CHECK: vmov.i32 d{{.*}}, #0x20FF
 	ret <2 x i32> < i32 8447, i32 8447 >
 }
 
 define <2 x i32> @v_movi32f() nounwind {
 ;CHECK: v_movi32f:
-;CHECK: vmov.i32 d0, #0x20FFFF
+;CHECK: vmov.i32 d{{.*}}, #0x20FFFF
 	ret <2 x i32> < i32 2162687, i32 2162687 >
 }
 
 define <2 x i32> @v_mvni32a() nounwind {
 ;CHECK: v_mvni32a:
-;CHECK: vmvn.i32 d0, #0x20
+;CHECK: vmvn.i32 d{{.*}}, #0x20
 	ret <2 x i32> < i32 4294967263, i32 4294967263 >
 }
 
 define <2 x i32> @v_mvni32b() nounwind {
 ;CHECK: v_mvni32b:
-;CHECK: vmvn.i32 d0, #0x2000
+;CHECK: vmvn.i32 d{{.*}}, #0x2000
 	ret <2 x i32> < i32 4294959103, i32 4294959103 >
 }
 
 define <2 x i32> @v_mvni32c() nounwind {
 ;CHECK: v_mvni32c:
-;CHECK: vmvn.i32 d0, #0x200000
+;CHECK: vmvn.i32 d{{.*}}, #0x200000
 	ret <2 x i32> < i32 4292870143, i32 4292870143 >
 }
 
 define <2 x i32> @v_mvni32d() nounwind {
 ;CHECK: v_mvni32d:
-;CHECK: vmvn.i32 d0, #0x20000000
+;CHECK: vmvn.i32 d{{.*}}, #0x20000000
 	ret <2 x i32> < i32 3758096383, i32 3758096383 >
 }
 
 define <2 x i32> @v_mvni32e() nounwind {
 ;CHECK: v_mvni32e:
-;CHECK: vmvn.i32 d0, #0x20FF
+;CHECK: vmvn.i32 d{{.*}}, #0x20FF
 	ret <2 x i32> < i32 4294958848, i32 4294958848 >
 }
 
 define <2 x i32> @v_mvni32f() nounwind {
 ;CHECK: v_mvni32f:
-;CHECK: vmvn.i32 d0, #0x20FFFF
+;CHECK: vmvn.i32 d{{.*}}, #0x20FFFF
 	ret <2 x i32> < i32 4292804608, i32 4292804608 >
 }
 
 define <1 x i64> @v_movi64() nounwind {
 ;CHECK: v_movi64:
-;CHECK: vmov.i64 d0, #0xFF0000FF0000FFFF
+;CHECK: vmov.i64 d{{.*}}, #0xFF0000FF0000FFFF
 	ret <1 x i64> < i64 18374687574888349695 >
 }
 
 define <16 x i8> @v_movQi8() nounwind {
 ;CHECK: v_movQi8:
-;CHECK: vmov.i8 q0, #0x8
+;CHECK: vmov.i8 q{{.*}}, #0x8
 	ret <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 >
 }
 
 define <8 x i16> @v_movQi16a() nounwind {
 ;CHECK: v_movQi16a:
-;CHECK: vmov.i16 q0, #0x10
+;CHECK: vmov.i16 q{{.*}}, #0x10
 	ret <8 x i16> < i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16, i16 16 >
 }
 
 define <8 x i16> @v_movQi16b() nounwind {
 ;CHECK: v_movQi16b:
-;CHECK: vmov.i16 q0, #0x1000
+;CHECK: vmov.i16 q{{.*}}, #0x1000
 	ret <8 x i16> < i16 4096, i16 4096, i16 4096, i16 4096, i16 4096, i16 4096, i16 4096, i16 4096 >
 }
 
 define <4 x i32> @v_movQi32a() nounwind {
 ;CHECK: v_movQi32a:
-;CHECK: vmov.i32 q0, #0x20
+;CHECK: vmov.i32 q{{.*}}, #0x20
 	ret <4 x i32> < i32 32, i32 32, i32 32, i32 32 >
 }
 
 define <4 x i32> @v_movQi32b() nounwind {
 ;CHECK: v_movQi32b:
-;CHECK: vmov.i32 q0, #0x2000
+;CHECK: vmov.i32 q{{.*}}, #0x2000
 	ret <4 x i32> < i32 8192, i32 8192, i32 8192, i32 8192 >
 }
 
 define <4 x i32> @v_movQi32c() nounwind {
 ;CHECK: v_movQi32c:
-;CHECK: vmov.i32 q0, #0x200000
+;CHECK: vmov.i32 q{{.*}}, #0x200000
 	ret <4 x i32> < i32 2097152, i32 2097152, i32 2097152, i32 2097152 >
 }
 
 define <4 x i32> @v_movQi32d() nounwind {
 ;CHECK: v_movQi32d:
-;CHECK: vmov.i32 q0, #0x20000000
+;CHECK: vmov.i32 q{{.*}}, #0x20000000
 	ret <4 x i32> < i32 536870912, i32 536870912, i32 536870912, i32 536870912 >
 }
 
 define <4 x i32> @v_movQi32e() nounwind {
 ;CHECK: v_movQi32e:
-;CHECK: vmov.i32 q0, #0x20FF
+;CHECK: vmov.i32 q{{.*}}, #0x20FF
 	ret <4 x i32> < i32 8447, i32 8447, i32 8447, i32 8447 >
 }
 
 define <4 x i32> @v_movQi32f() nounwind {
 ;CHECK: v_movQi32f:
-;CHECK: vmov.i32 q0, #0x20FFFF
+;CHECK: vmov.i32 q{{.*}}, #0x20FFFF
 	ret <4 x i32> < i32 2162687, i32 2162687, i32 2162687, i32 2162687 >
 }
 
 define <2 x i64> @v_movQi64() nounwind {
 ;CHECK: v_movQi64:
-;CHECK: vmov.i64 q0, #0xFF0000FF0000FFFF
+;CHECK: vmov.i64 q{{.*}}, #0xFF0000FF0000FFFF
 	ret <2 x i64> < i64 18374687574888349695, i64 18374687574888349695 >
 }
 
@@ -173,7 +173,7 @@ define <2 x i64> @v_movQi64() nounwind {
 define void @vdupn128(%struct.int8x8_t* noalias nocapture sret %agg.result) nounwind {
 entry:
 ;CHECK: vdupn128:
-;CHECK: vmov.i8 d0, #0x80
+;CHECK: vmov.i8 d{{.*}}, #0x80
   %0 = getelementptr inbounds %struct.int8x8_t* %agg.result, i32 0, i32 0 ; <<8 x i8>*> [#uses=1]
   store <8 x i8> <i8 -128, i8 -128, i8 -128, i8 -128, i8 -128, i8 -128, i8 -128, i8 -128>, <8 x i8>* %0, align 8
   ret void
@@ -182,7 +182,7 @@ entry:
 define void @vdupnneg75(%struct.int8x8_t* noalias nocapture sret %agg.result) nounwind {
 entry:
 ;CHECK: vdupnneg75:
-;CHECK: vmov.i8 d0, #0xB5
+;CHECK: vmov.i8 d{{.*}}, #0xB5
   %0 = getelementptr inbounds %struct.int8x8_t* %agg.result, i32 0, i32 0 ; <<8 x i8>*> [#uses=1]
   store <8 x i8> <i8 -75, i8 -75, i8 -75, i8 -75, i8 -75, i8 -75, i8 -75, i8 -75>, <8 x i8>* %0, align 8
   ret void
diff --git a/test/CodeGen/ARM/vst1.ll b/test/CodeGen/ARM/vst1.ll
index c43b8aa..70f3a4c 100644
--- a/test/CodeGen/ARM/vst1.ll
+++ b/test/CodeGen/ARM/vst1.ll
@@ -3,7 +3,7 @@
 define void @vst1i8(i8* %A, <8 x i8>* %B) nounwind {
 ;CHECK: vst1i8:
 ;Check the alignment value.  Max for this instruction is 64 bits:
-;CHECK: vst1.8 {d0}, [r0, :64]
+;CHECK: vst1.8 {d16}, [r0, :64]
 	%tmp1 = load <8 x i8>* %B
 	call void @llvm.arm.neon.vst1.v8i8(i8* %A, <8 x i8> %tmp1, i32 16)
 	ret void
@@ -48,7 +48,7 @@ define void @vst1i64(i64* %A, <1 x i64>* %B) nounwind {
 define void @vst1Qi8(i8* %A, <16 x i8>* %B) nounwind {
 ;CHECK: vst1Qi8:
 ;Check the alignment value.  Max for this instruction is 128 bits:
-;CHECK: vst1.8 {d0, d1}, [r0, :64]
+;CHECK: vst1.8 {d16, d17}, [r0, :64]
 	%tmp1 = load <16 x i8>* %B
 	call void @llvm.arm.neon.vst1.v16i8(i8* %A, <16 x i8> %tmp1, i32 8)
 	ret void
@@ -57,7 +57,7 @@ define void @vst1Qi8(i8* %A, <16 x i8>* %B) nounwind {
 define void @vst1Qi16(i16* %A, <8 x i16>* %B) nounwind {
 ;CHECK: vst1Qi16:
 ;Check the alignment value.  Max for this instruction is 128 bits:
-;CHECK: vst1.16 {d0, d1}, [r0, :128]
+;CHECK: vst1.16 {d16, d17}, [r0, :128]
 	%tmp0 = bitcast i16* %A to i8*
 	%tmp1 = load <8 x i16>* %B
 	call void @llvm.arm.neon.vst1.v8i16(i8* %tmp0, <8 x i16> %tmp1, i32 32)
diff --git a/test/CodeGen/ARM/vst2.ll b/test/CodeGen/ARM/vst2.ll
index 1124640..ed2498b 100644
--- a/test/CodeGen/ARM/vst2.ll
+++ b/test/CodeGen/ARM/vst2.ll
@@ -3,7 +3,7 @@
 define void @vst2i8(i8* %A, <8 x i8>* %B) nounwind {
 ;CHECK: vst2i8:
 ;Check the alignment value.  Max for this instruction is 128 bits:
-;CHECK: vst2.8 {d0, d1}, [r0, :64]
+;CHECK: vst2.8 {d16, d17}, [r0, :64]
 	%tmp1 = load <8 x i8>* %B
 	call void @llvm.arm.neon.vst2.v8i8(i8* %A, <8 x i8> %tmp1, <8 x i8> %tmp1, i32 8)
 	ret void
@@ -12,7 +12,7 @@ define void @vst2i8(i8* %A, <8 x i8>* %B) nounwind {
 define void @vst2i16(i16* %A, <4 x i16>* %B) nounwind {
 ;CHECK: vst2i16:
 ;Check the alignment value.  Max for this instruction is 128 bits:
-;CHECK: vst2.16 {d0, d1}, [r0, :128]
+;CHECK: vst2.16 {d16, d17}, [r0, :128]
 	%tmp0 = bitcast i16* %A to i8*
 	%tmp1 = load <4 x i16>* %B
 	call void @llvm.arm.neon.vst2.v4i16(i8* %tmp0, <4 x i16> %tmp1, <4 x i16> %tmp1, i32 32)
@@ -40,7 +40,7 @@ define void @vst2f(float* %A, <2 x float>* %B) nounwind {
 define void @vst2i64(i64* %A, <1 x i64>* %B) nounwind {
 ;CHECK: vst2i64:
 ;Check the alignment value.  Max for this instruction is 128 bits:
-;CHECK: vst1.64 {d0, d1}, [r0, :128]
+;CHECK: vst1.64 {d16, d17}, [r0, :128]
 	%tmp0 = bitcast i64* %A to i8*
 	%tmp1 = load <1 x i64>* %B
 	call void @llvm.arm.neon.vst2.v1i64(i8* %tmp0, <1 x i64> %tmp1, <1 x i64> %tmp1, i32 32)
@@ -50,7 +50,7 @@ define void @vst2i64(i64* %A, <1 x i64>* %B) nounwind {
 define void @vst2Qi8(i8* %A, <16 x i8>* %B) nounwind {
 ;CHECK: vst2Qi8:
 ;Check the alignment value.  Max for this instruction is 256 bits:
-;CHECK: vst2.8 {d0, d1, d2, d3}, [r0, :64]
+;CHECK: vst2.8 {d16, d17, d18, d19}, [r0, :64]
 	%tmp1 = load <16 x i8>* %B
 	call void @llvm.arm.neon.vst2.v16i8(i8* %A, <16 x i8> %tmp1, <16 x i8> %tmp1, i32 8)
 	ret void
@@ -59,7 +59,7 @@ define void @vst2Qi8(i8* %A, <16 x i8>* %B) nounwind {
 define void @vst2Qi16(i16* %A, <8 x i16>* %B) nounwind {
 ;CHECK: vst2Qi16:
 ;Check the alignment value.  Max for this instruction is 256 bits:
-;CHECK: vst2.16 {d0, d1, d2, d3}, [r0, :128]
+;CHECK: vst2.16 {d16, d17, d18, d19}, [r0, :128]
 	%tmp0 = bitcast i16* %A to i8*
 	%tmp1 = load <8 x i16>* %B
 	call void @llvm.arm.neon.vst2.v8i16(i8* %tmp0, <8 x i16> %tmp1, <8 x i16> %tmp1, i32 16)
@@ -69,7 +69,7 @@ define void @vst2Qi16(i16* %A, <8 x i16>* %B) nounwind {
 define void @vst2Qi32(i32* %A, <4 x i32>* %B) nounwind {
 ;CHECK: vst2Qi32:
 ;Check the alignment value.  Max for this instruction is 256 bits:
-;CHECK: vst2.32 {d0, d1, d2, d3}, [r0, :256]
+;CHECK: vst2.32 {d16, d17, d18, d19}, [r0, :256]
 	%tmp0 = bitcast i32* %A to i8*
 	%tmp1 = load <4 x i32>* %B
 	call void @llvm.arm.neon.vst2.v4i32(i8* %tmp0, <4 x i32> %tmp1, <4 x i32> %tmp1, i32 64)
diff --git a/test/CodeGen/ARM/vst4.ll b/test/CodeGen/ARM/vst4.ll
index a339192..6a7e91d 100644
--- a/test/CodeGen/ARM/vst4.ll
+++ b/test/CodeGen/ARM/vst4.ll
@@ -3,7 +3,7 @@
 define void @vst4i8(i8* %A, <8 x i8>* %B) nounwind {
 ;CHECK: vst4i8:
 ;Check the alignment value.  Max for this instruction is 256 bits:
-;CHECK: vst4.8 {d0, d1, d2, d3}, [r0, :64]
+;CHECK: vst4.8 {d16, d17, d18, d19}, [r0, :64]
 	%tmp1 = load <8 x i8>* %B
 	call void @llvm.arm.neon.vst4.v8i8(i8* %A, <8 x i8> %tmp1, <8 x i8> %tmp1, <8 x i8> %tmp1, <8 x i8> %tmp1, i32 8)
 	ret void
@@ -12,7 +12,7 @@ define void @vst4i8(i8* %A, <8 x i8>* %B) nounwind {
 define void @vst4i16(i16* %A, <4 x i16>* %B) nounwind {
 ;CHECK: vst4i16:
 ;Check the alignment value.  Max for this instruction is 256 bits:
-;CHECK: vst4.16 {d0, d1, d2, d3}, [r0, :128]
+;CHECK: vst4.16 {d16, d17, d18, d19}, [r0, :128]
 	%tmp0 = bitcast i16* %A to i8*
 	%tmp1 = load <4 x i16>* %B
 	call void @llvm.arm.neon.vst4.v4i16(i8* %tmp0, <4 x i16> %tmp1, <4 x i16> %tmp1, <4 x i16> %tmp1, <4 x i16> %tmp1, i32 16)
@@ -22,7 +22,7 @@ define void @vst4i16(i16* %A, <4 x i16>* %B) nounwind {
 define void @vst4i32(i32* %A, <2 x i32>* %B) nounwind {
 ;CHECK: vst4i32:
 ;Check the alignment value.  Max for this instruction is 256 bits:
-;CHECK: vst4.32 {d0, d1, d2, d3}, [r0, :256]
+;CHECK: vst4.32 {d16, d17, d18, d19}, [r0, :256]
 	%tmp0 = bitcast i32* %A to i8*
 	%tmp1 = load <2 x i32>* %B
 	call void @llvm.arm.neon.vst4.v2i32(i8* %tmp0, <2 x i32> %tmp1, <2 x i32> %tmp1, <2 x i32> %tmp1, <2 x i32> %tmp1, i32 32)
@@ -41,7 +41,7 @@ define void @vst4f(float* %A, <2 x float>* %B) nounwind {
 define void @vst4i64(i64* %A, <1 x i64>* %B) nounwind {
 ;CHECK: vst4i64:
 ;Check the alignment value.  Max for this instruction is 256 bits:
-;CHECK: vst1.64 {d0, d1, d2, d3}, [r0, :256]
+;CHECK: vst1.64 {d16, d17, d18, d19}, [r0, :256]
 	%tmp0 = bitcast i64* %A to i8*
 	%tmp1 = load <1 x i64>* %B
 	call void @llvm.arm.neon.vst4.v1i64(i8* %tmp0, <1 x i64> %tmp1, <1 x i64> %tmp1, <1 x i64> %tmp1, <1 x i64> %tmp1, i32 64)
@@ -51,8 +51,8 @@ define void @vst4i64(i64* %A, <1 x i64>* %B) nounwind {
 define void @vst4Qi8(i8* %A, <16 x i8>* %B) nounwind {
 ;CHECK: vst4Qi8:
 ;Check the alignment value.  Max for this instruction is 256 bits:
-;CHECK: vst4.8 {d0, d2, d4, d6}, [r0, :256]!
-;CHECK: vst4.8 {d1, d3, d5, d7}, [r0, :256]
+;CHECK: vst4.8 {d16, d18, d20, d22}, [r0, :256]!
+;CHECK: vst4.8 {d17, d19, d21, d23}, [r0, :256]
 	%tmp1 = load <16 x i8>* %B
 	call void @llvm.arm.neon.vst4.v16i8(i8* %A, <16 x i8> %tmp1, <16 x i8> %tmp1, <16 x i8> %tmp1, <16 x i8> %tmp1, i32 64)
 	ret void
@@ -61,8 +61,8 @@ define void @vst4Qi8(i8* %A, <16 x i8>* %B) nounwind {
 define void @vst4Qi16(i16* %A, <8 x i16>* %B) nounwind {
 ;CHECK: vst4Qi16:
 ;Check for no alignment specifier.
-;CHECK: vst4.16 {d0, d2, d4, d6}, [r0]!
-;CHECK: vst4.16 {d1, d3, d5, d7}, [r0]
+;CHECK: vst4.16 {d16, d18, d20, d22}, [r0]!
+;CHECK: vst4.16 {d17, d19, d21, d23}, [r0]
 	%tmp0 = bitcast i16* %A to i8*
 	%tmp1 = load <8 x i16>* %B
 	call void @llvm.arm.neon.vst4.v8i16(i8* %tmp0, <8 x i16> %tmp1, <8 x i16> %tmp1, <8 x i16> %tmp1, <8 x i16> %tmp1, i32 1)
diff --git a/test/CodeGen/Thumb2/2010-06-14-NEONCoalescer.ll b/test/CodeGen/Thumb2/2010-06-14-NEONCoalescer.ll
index 2675006..0b6c92b 100644
--- a/test/CodeGen/Thumb2/2010-06-14-NEONCoalescer.ll
+++ b/test/CodeGen/Thumb2/2010-06-14-NEONCoalescer.ll
@@ -23,8 +23,8 @@ entry:
   %4 = insertelement <2 x double> %2, double %V.0.ph, i32 1 ; <<2 x double>> [#uses=2]
 ; Constant pool load followed by add.
 ; Then clobber the loaded register, not the sum.
-; CHECK: vldr.64 [[LDR:d.]]
-; CHECK: vadd.f64 [[ADD:d.]], [[LDR]], [[LDR]]
+; CHECK: vldr.64 [[LDR:d.*]],
+; CHECK: vadd.f64 [[ADD:d.*]], [[LDR]], [[LDR]]
 ; CHECK: vmov.f64 [[LDR]]
   %5 = fadd <2 x double> %3, %3                   ; <<2 x double>> [#uses=2]
   %6 = fadd <2 x double> %4, %4                   ; <<2 x double>> [#uses=2]
diff --git a/test/CodeGen/Thumb2/cortex-fp.ll b/test/CodeGen/Thumb2/cortex-fp.ll
index f7ec5a3..d06f8a7 100644
--- a/test/CodeGen/Thumb2/cortex-fp.ll
+++ b/test/CodeGen/Thumb2/cortex-fp.ll
@@ -19,6 +19,6 @@ entry:
   %0 = fmul double %a, %b
 ; CORTEXM3: blx ___muldf3
 ; CORTEXM4: blx ___muldf3
-; CORTEXA8: vmul.f64  d0, d1, d0
+; CORTEXA8: vmul.f64  d16, d17, d16
   ret double %0
 }
diff --git a/test/CodeGen/Thumb2/machine-licm.ll b/test/CodeGen/Thumb2/machine-licm.ll
index b949b2f..3fac512 100644
--- a/test/CodeGen/Thumb2/machine-licm.ll
+++ b/test/CodeGen/Thumb2/machine-licm.ll
@@ -56,7 +56,7 @@ define void @t2(i8* %ptr1, i8* %ptr2) nounwind {
 entry:
 ; CHECK: t2:
 ; CHECK: adr r{{.}}, #LCPI1_0
-; CHECK: vldmia r3, {d0, d1}
+; CHECK: vldmia r3, {d16, d17}
   br i1 undef, label %bb1, label %bb2
 
 bb1:
diff --git a/test/CodeGen/Thumb2/thumb2-spill-q.ll b/test/CodeGen/Thumb2/thumb2-spill-q.ll
index 4f92c93..0d73fba 100644
--- a/test/CodeGen/Thumb2/thumb2-spill-q.ll
+++ b/test/CodeGen/Thumb2/thumb2-spill-q.ll
@@ -20,6 +20,26 @@ entry:
   %1 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind ; <<4 x float>> [#uses=1]
   store float 0.000000e+00, float* undef, align 4
   %2 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind ; <<4 x float>> [#uses=1]
+  %ld3 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld4 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld5 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld6 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld7 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld8 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld9 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld10 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld11 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
+  %ld12 = call <4 x float> @llvm.arm.neon.vld1.v4f32(i8* undef, i32 1) nounwind
+  store float 0.000000e+00, float* undef, align 4
   %val173 = load <4 x float>* undef               ; <<4 x float>> [#uses=1]
   br label %bb4
 
@@ -44,7 +64,16 @@ bb4:                                              ; preds = %bb193, %entry
   %18 = fmul <4 x float> %17, %val173             ; <<4 x float>> [#uses=1]
   %19 = shufflevector <4 x float> %18, <4 x float> undef, <2 x i32> <i32 2, i32 3> ; <<2 x float>> [#uses=1]
   %20 = shufflevector <2 x float> %19, <2 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
-  %21 = fadd <4 x float> zeroinitializer, %20     ; <<4 x float>> [#uses=2]
+  %tmp1 = fadd <4 x float> %20, %ld3
+  %tmp2 = fadd <4 x float> %tmp1, %ld4
+  %tmp3 = fadd <4 x float> %tmp2, %ld5
+  %tmp4 = fadd <4 x float> %tmp3, %ld6
+  %tmp5 = fadd <4 x float> %tmp4, %ld7
+  %tmp6 = fadd <4 x float> %tmp5, %ld8
+  %tmp7 = fadd <4 x float> %tmp6, %ld9
+  %tmp8 = fadd <4 x float> %tmp7, %ld10
+  %tmp9 = fadd <4 x float> %tmp8, %ld11
+  %21 = fadd <4 x float> %tmp9, %ld12
   %22 = fcmp ogt <4 x float> %besterror.0.2264, %21 ; <<4 x i1>> [#uses=0]
   %tmp = extractelement <4 x i1> %22, i32 0
   br i1 %tmp, label %bb193, label %bb186