From 1a493b0fa556a07c728862c3c3f70bfd8683bef0 Mon Sep 17 00:00:00 2001
From: Nemanja Ivanovic <nemanjai@ca.ibm.com>
Date: Fri, 22 May 2020 07:48:09 -0500
Subject: [PATCH] [PowerPC] Add missing handling for half precision

The fix for PR39865 took care of some of the handling for half precision
but it missed a number of issues that still exist. This patch fixes the
remaining issues that cause crashes in the PPC back end.

Fixes: https://bugs.llvm.org/show_bug.cgi?id=45776

Differential revision: https://reviews.llvm.org/D79283
---
 llvm/lib/Target/PowerPC/PPCISelLowering.cpp   |    6 +
 llvm/lib/Target/PowerPC/PPCISelLowering.h     |    2 +-
 llvm/lib/Target/PowerPC/PPCInstrVSX.td        |    7 +
 .../PowerPC/handle-f16-storage-type.ll        | 1081 +++++++++++++++++
 .../CodeGen/PowerPC/scalar_vector_test_2.ll   |   88 +-
 .../vector-constrained-fp-intrinsics.ll       |  146 +--
 6 files changed, 1151 insertions(+), 179 deletions(-)

diff --git a/llvm/lib/Target/PowerPC/PPCISelLowering.cpp b/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
index 7f6fc73dee75..8b1ebba596a0 100644
--- a/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
+++ b/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
@@ -10755,6 +10755,7 @@ SDValue PPCTargetLowering::LowerFP_EXTEND(SDValue Op, SelectionDAG &DAG) const {
   assert(Op.getOpcode() == ISD::FP_EXTEND &&
          "Should only be called for ISD::FP_EXTEND");
 
+  // FIXME: handle extends from half precision float vectors on P9.
   // We only want to custom lower an extend from v2f32 to v2f64.
   if (Op.getValueType() != MVT::v2f64 ||
       Op.getOperand(0).getValueType() != MVT::v2f32)
@@ -10968,6 +10969,11 @@ void PPCTargetLowering::ReplaceNodeResults(SDNode *N,
   case ISD::BITCAST:
     // Don't handle bitcast here.
     return;
+  case ISD::FP_EXTEND:
+    SDValue Lowered = LowerFP_EXTEND(SDValue(N, 0), DAG);
+    if (Lowered)
+      Results.push_back(Lowered);
+    return;
   }
 }
 
diff --git a/llvm/lib/Target/PowerPC/PPCISelLowering.h b/llvm/lib/Target/PowerPC/PPCISelLowering.h
index 41898eef3713..c34fd6aa78be 100644
--- a/llvm/lib/Target/PowerPC/PPCISelLowering.h
+++ b/llvm/lib/Target/PowerPC/PPCISelLowering.h
@@ -640,7 +640,7 @@ namespace llvm {
     /// then the VPERM for the shuffle. All in all a very slow sequence.
     TargetLoweringBase::LegalizeTypeAction getPreferredVectorAction(MVT VT)
       const override {
-      if (VT.getScalarSizeInBits() % 8 == 0)
+      if (VT.getVectorNumElements() != 1 && VT.getScalarSizeInBits() % 8 == 0)
         return TypeWidenVector;
       return TargetLoweringBase::getPreferredVectorAction(VT);
     }
diff --git a/llvm/lib/Target/PowerPC/PPCInstrVSX.td b/llvm/lib/Target/PowerPC/PPCInstrVSX.td
index 93fbdd69ea3f..acb8e3563ffb 100644
--- a/llvm/lib/Target/PowerPC/PPCInstrVSX.td
+++ b/llvm/lib/Target/PowerPC/PPCInstrVSX.td
@@ -3658,6 +3658,13 @@ def : Pat<(f32 (extloadf16 xoaddr:$src)),
           (f32 (COPY_TO_REGCLASS (XSCVHPDP (LXSIHZX xoaddr:$src)), VSSRC))>;
 def : Pat<(truncstoref16 f32:$src, xoaddr:$dst),
           (STXSIHX (XSCVDPHP (COPY_TO_REGCLASS $src, VSFRC)), xoaddr:$dst)>;
+def : Pat<(f64 (f16_to_fp i32:$A)),
+          (f64 (XSCVHPDP (MTVSRWZ $A)))>;
+def : Pat<(f32 (f16_to_fp i32:$A)),
+          (f32 (COPY_TO_REGCLASS (XSCVHPDP (MTVSRWZ $A)), VSSRC))>;
+def : Pat<(i32 (fp_to_f16 f32:$A)),
+          (i32 (MFVSRWZ (XSCVDPHP (COPY_TO_REGCLASS $A, VSFRC))))>;
+def : Pat<(i32 (fp_to_f16 f64:$A)), (i32 (MFVSRWZ (XSCVDPHP $A)))>;
 
 // Vector sign extensions
 def : Pat<(f64 (PPCVexts f64:$A, 1)),
diff --git a/llvm/test/CodeGen/PowerPC/handle-f16-storage-type.ll b/llvm/test/CodeGen/PowerPC/handle-f16-storage-type.ll
index f7ce464c5b61..c5968758cc4a 100644
--- a/llvm/test/CodeGen/PowerPC/handle-f16-storage-type.ll
+++ b/llvm/test/CodeGen/PowerPC/handle-f16-storage-type.ll
@@ -8,6 +8,8 @@
 ; RUN:   -verify-machineinstrs -ppc-asm-full-reg-names < %s | FileCheck %s \
 ; RUN:   --check-prefix=SOFT
 
+; Tests for various operations on half precison float. Much of the test is
+; copied from test/CodeGen/X86/half.ll.
 define dso_local double @loadd(i16* nocapture readonly %a) local_unnamed_addr #0 {
 ; P8-LABEL: loadd:
 ; P8:       # %bb.0: # %entry
@@ -197,4 +199,1083 @@ entry:
 }
 
 declare i16 @llvm.convert.to.fp16.f32(float)
+define void @test_load_store(half* %in, half* %out) #0 {
+; P8-LABEL: test_load_store:
+; P8:       # %bb.0:
+; P8-NEXT:    lhz r3, 0(r3)
+; P8-NEXT:    sth r3, 0(r4)
+; P8-NEXT:    blr
+;
+; CHECK-LABEL: test_load_store:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lhz r3, 0(r3)
+; CHECK-NEXT:    sth r3, 0(r4)
+; CHECK-NEXT:    blr
+;
+; SOFT-LABEL: test_load_store:
+; SOFT:       # %bb.0:
+; SOFT-NEXT:    mflr r0
+; SOFT-NEXT:    std r30, -16(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r0, 16(r1)
+; SOFT-NEXT:    stdu r1, -48(r1)
+; SOFT-NEXT:    lhz r3, 0(r3)
+; SOFT-NEXT:    mr r30, r4
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    sth r3, 0(r30)
+; SOFT-NEXT:    addi r1, r1, 48
+; SOFT-NEXT:    ld r0, 16(r1)
+; SOFT-NEXT:    ld r30, -16(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    mtlr r0
+; SOFT-NEXT:    blr
+  %val = load half, half* %in
+  store half %val, half* %out
+  ret void
+}
+define i16 @test_bitcast_from_half(half* %addr) #0 {
+; P8-LABEL: test_bitcast_from_half:
+; P8:       # %bb.0:
+; P8-NEXT:    lhz r3, 0(r3)
+; P8-NEXT:    blr
+;
+; CHECK-LABEL: test_bitcast_from_half:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lhz r3, 0(r3)
+; CHECK-NEXT:    blr
+;
+; SOFT-LABEL: test_bitcast_from_half:
+; SOFT:       # %bb.0:
+; SOFT-NEXT:    lhz r3, 0(r3)
+; SOFT-NEXT:    blr
+  %val = load half, half* %addr
+  %val_int = bitcast half %val to i16
+  ret i16 %val_int
+}
+define void @test_bitcast_to_half(half* %addr, i16 %in) #0 {
+; P8-LABEL: test_bitcast_to_half:
+; P8:       # %bb.0:
+; P8-NEXT:    sth r4, 0(r3)
+; P8-NEXT:    blr
+;
+; CHECK-LABEL: test_bitcast_to_half:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    sth r4, 0(r3)
+; CHECK-NEXT:    blr
+;
+; SOFT-LABEL: test_bitcast_to_half:
+; SOFT:       # %bb.0:
+; SOFT-NEXT:    sth r4, 0(r3)
+; SOFT-NEXT:    blr
+  %val_fp = bitcast i16 %in to half
+  store half %val_fp, half* %addr
+  ret void
+}
+define float @test_extend32(half* %addr) #0 {
+; P8-LABEL: test_extend32:
+; P8:       # %bb.0:
+; P8-NEXT:    mflr r0
+; P8-NEXT:    std r0, 16(r1)
+; P8-NEXT:    stdu r1, -32(r1)
+; P8-NEXT:    lhz r3, 0(r3)
+; P8-NEXT:    bl __gnu_h2f_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    addi r1, r1, 32
+; P8-NEXT:    ld r0, 16(r1)
+; P8-NEXT:    mtlr r0
+; P8-NEXT:    blr
+;
+; CHECK-LABEL: test_extend32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lxsihzx f0, 0, r3
+; CHECK-NEXT:    xscvhpdp f1, f0
+; CHECK-NEXT:    blr
+;
+; SOFT-LABEL: test_extend32:
+; SOFT:       # %bb.0:
+; SOFT-NEXT:    mflr r0
+; SOFT-NEXT:    std r0, 16(r1)
+; SOFT-NEXT:    stdu r1, -32(r1)
+; SOFT-NEXT:    lhz r3, 0(r3)
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    addi r1, r1, 32
+; SOFT-NEXT:    ld r0, 16(r1)
+; SOFT-NEXT:    mtlr r0
+; SOFT-NEXT:    blr
+  %val16 = load half, half* %addr
+  %val32 = fpext half %val16 to float
+  ret float %val32
+}
+define double @test_extend64(half* %addr) #0 {
+; P8-LABEL: test_extend64:
+; P8:       # %bb.0:
+; P8-NEXT:    mflr r0
+; P8-NEXT:    std r0, 16(r1)
+; P8-NEXT:    stdu r1, -32(r1)
+; P8-NEXT:    lhz r3, 0(r3)
+; P8-NEXT:    bl __gnu_h2f_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    addi r1, r1, 32
+; P8-NEXT:    ld r0, 16(r1)
+; P8-NEXT:    mtlr r0
+; P8-NEXT:    blr
+;
+; CHECK-LABEL: test_extend64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lxsihzx f0, 0, r3
+; CHECK-NEXT:    xscvhpdp f1, f0
+; CHECK-NEXT:    blr
+;
+; SOFT-LABEL: test_extend64:
+; SOFT:       # %bb.0:
+; SOFT-NEXT:    mflr r0
+; SOFT-NEXT:    std r0, 16(r1)
+; SOFT-NEXT:    stdu r1, -32(r1)
+; SOFT-NEXT:    lhz r3, 0(r3)
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    bl __extendsfdf2
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    addi r1, r1, 32
+; SOFT-NEXT:    ld r0, 16(r1)
+; SOFT-NEXT:    mtlr r0
+; SOFT-NEXT:    blr
+  %val16 = load half, half* %addr
+  %val32 = fpext half %val16 to double
+  ret double %val32
+}
+define void @test_trunc32(float %in, half* %addr) #0 {
+; P8-LABEL: test_trunc32:
+; P8:       # %bb.0:
+; P8-NEXT:    mflr r0
+; P8-NEXT:    std r30, -16(r1) # 8-byte Folded Spill
+; P8-NEXT:    std r0, 16(r1)
+; P8-NEXT:    stdu r1, -48(r1)
+; P8-NEXT:    mr r30, r4
+; P8-NEXT:    bl __gnu_f2h_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    sth r3, 0(r30)
+; P8-NEXT:    addi r1, r1, 48
+; P8-NEXT:    ld r0, 16(r1)
+; P8-NEXT:    ld r30, -16(r1) # 8-byte Folded Reload
+; P8-NEXT:    mtlr r0
+; P8-NEXT:    blr
+;
+; CHECK-LABEL: test_trunc32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    xscvdphp f0, f1
+; CHECK-NEXT:    stxsihx f0, 0, r4
+; CHECK-NEXT:    blr
+;
+; SOFT-LABEL: test_trunc32:
+; SOFT:       # %bb.0:
+; SOFT-NEXT:    mflr r0
+; SOFT-NEXT:    std r30, -16(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r0, 16(r1)
+; SOFT-NEXT:    stdu r1, -48(r1)
+; SOFT-NEXT:    clrldi r3, r3, 32
+; SOFT-NEXT:    mr r30, r4
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    clrldi r3, r3, 48
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    sth r3, 0(r30)
+; SOFT-NEXT:    addi r1, r1, 48
+; SOFT-NEXT:    ld r0, 16(r1)
+; SOFT-NEXT:    ld r30, -16(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    mtlr r0
+; SOFT-NEXT:    blr
+  %val16 = fptrunc float %in to half
+  store half %val16, half* %addr
+  ret void
+}
+define void @test_trunc64(double %in, half* %addr) #0 {
+; P8-LABEL: test_trunc64:
+; P8:       # %bb.0:
+; P8-NEXT:    mflr r0
+; P8-NEXT:    std r30, -16(r1) # 8-byte Folded Spill
+; P8-NEXT:    std r0, 16(r1)
+; P8-NEXT:    stdu r1, -48(r1)
+; P8-NEXT:    mr r30, r4
+; P8-NEXT:    bl __truncdfhf2
+; P8-NEXT:    nop
+; P8-NEXT:    sth r3, 0(r30)
+; P8-NEXT:    addi r1, r1, 48
+; P8-NEXT:    ld r0, 16(r1)
+; P8-NEXT:    ld r30, -16(r1) # 8-byte Folded Reload
+; P8-NEXT:    mtlr r0
+; P8-NEXT:    blr
+;
+; CHECK-LABEL: test_trunc64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    xscvdphp f0, f1
+; CHECK-NEXT:    stxsihx f0, 0, r4
+; CHECK-NEXT:    blr
+;
+; SOFT-LABEL: test_trunc64:
+; SOFT:       # %bb.0:
+; SOFT-NEXT:    mflr r0
+; SOFT-NEXT:    std r30, -16(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r0, 16(r1)
+; SOFT-NEXT:    stdu r1, -48(r1)
+; SOFT-NEXT:    mr r30, r4
+; SOFT-NEXT:    bl __truncdfhf2
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    clrldi r3, r3, 48
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    sth r3, 0(r30)
+; SOFT-NEXT:    addi r1, r1, 48
+; SOFT-NEXT:    ld r0, 16(r1)
+; SOFT-NEXT:    ld r30, -16(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    mtlr r0
+; SOFT-NEXT:    blr
+  %val16 = fptrunc double %in to half
+  store half %val16, half* %addr
+  ret void
+}
+define i64 @test_fptosi_i64(half* %p) #0 {
+; P8-LABEL: test_fptosi_i64:
+; P8:       # %bb.0:
+; P8-NEXT:    mflr r0
+; P8-NEXT:    std r0, 16(r1)
+; P8-NEXT:    stdu r1, -32(r1)
+; P8-NEXT:    lhz r3, 0(r3)
+; P8-NEXT:    bl __gnu_h2f_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    xscvdpsxds f0, f1
+; P8-NEXT:    mffprd r3, f0
+; P8-NEXT:    addi r1, r1, 32
+; P8-NEXT:    ld r0, 16(r1)
+; P8-NEXT:    mtlr r0
+; P8-NEXT:    blr
+;
+; CHECK-LABEL: test_fptosi_i64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lhz r3, 0(r3)
+; CHECK-NEXT:    mtfprwz f0, r3
+; CHECK-NEXT:    xscvhpdp f0, f0
+; CHECK-NEXT:    xscvdpsxds f0, f0
+; CHECK-NEXT:    mffprd r3, f0
+; CHECK-NEXT:    blr
+;
+; SOFT-LABEL: test_fptosi_i64:
+; SOFT:       # %bb.0:
+; SOFT-NEXT:    mflr r0
+; SOFT-NEXT:    std r0, 16(r1)
+; SOFT-NEXT:    stdu r1, -32(r1)
+; SOFT-NEXT:    lhz r3, 0(r3)
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    bl __fixsfdi
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    addi r1, r1, 32
+; SOFT-NEXT:    ld r0, 16(r1)
+; SOFT-NEXT:    mtlr r0
+; SOFT-NEXT:    blr
+  %a = load half, half* %p, align 2
+  %r = fptosi half %a to i64
+  ret i64 %r
+}
+define void @test_sitofp_i64(i64 %a, half* %p) #0 {
+; P8-LABEL: test_sitofp_i64:
+; P8:       # %bb.0:
+; P8-NEXT:    mflr r0
+; P8-NEXT:    std r30, -16(r1) # 8-byte Folded Spill
+; P8-NEXT:    std r0, 16(r1)
+; P8-NEXT:    stdu r1, -48(r1)
+; P8-NEXT:    mtfprd f0, r3
+; P8-NEXT:    mr r30, r4
+; P8-NEXT:    xscvsxdsp f1, f0
+; P8-NEXT:    bl __gnu_f2h_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    sth r3, 0(r30)
+; P8-NEXT:    addi r1, r1, 48
+; P8-NEXT:    ld r0, 16(r1)
+; P8-NEXT:    ld r30, -16(r1) # 8-byte Folded Reload
+; P8-NEXT:    mtlr r0
+; P8-NEXT:    blr
+;
+; CHECK-LABEL: test_sitofp_i64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    mtfprd f0, r3
+; CHECK-NEXT:    xscvsxdsp f0, f0
+; CHECK-NEXT:    xscvdphp f0, f0
+; CHECK-NEXT:    mffprwz r3, f0
+; CHECK-NEXT:    sth r3, 0(r4)
+; CHECK-NEXT:    blr
+;
+; SOFT-LABEL: test_sitofp_i64:
+; SOFT:       # %bb.0:
+; SOFT-NEXT:    mflr r0
+; SOFT-NEXT:    std r30, -16(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r0, 16(r1)
+; SOFT-NEXT:    stdu r1, -48(r1)
+; SOFT-NEXT:    mr r30, r4
+; SOFT-NEXT:    bl __floatdisf
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    clrldi r3, r3, 32
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    clrldi r3, r3, 48
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    sth r3, 0(r30)
+; SOFT-NEXT:    addi r1, r1, 48
+; SOFT-NEXT:    ld r0, 16(r1)
+; SOFT-NEXT:    ld r30, -16(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    mtlr r0
+; SOFT-NEXT:    blr
+  %r = sitofp i64 %a to half
+  store half %r, half* %p
+  ret void
+}
+define i64 @test_fptoui_i64(half* %p) #0 {
+; P8-LABEL: test_fptoui_i64:
+; P8:       # %bb.0:
+; P8-NEXT:    mflr r0
+; P8-NEXT:    std r0, 16(r1)
+; P8-NEXT:    stdu r1, -32(r1)
+; P8-NEXT:    lhz r3, 0(r3)
+; P8-NEXT:    bl __gnu_h2f_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    xscvdpuxds f0, f1
+; P8-NEXT:    mffprd r3, f0
+; P8-NEXT:    addi r1, r1, 32
+; P8-NEXT:    ld r0, 16(r1)
+; P8-NEXT:    mtlr r0
+; P8-NEXT:    blr
+;
+; CHECK-LABEL: test_fptoui_i64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lhz r3, 0(r3)
+; CHECK-NEXT:    mtfprwz f0, r3
+; CHECK-NEXT:    xscvhpdp f0, f0
+; CHECK-NEXT:    xscvdpuxds f0, f0
+; CHECK-NEXT:    mffprd r3, f0
+; CHECK-NEXT:    blr
+;
+; SOFT-LABEL: test_fptoui_i64:
+; SOFT:       # %bb.0:
+; SOFT-NEXT:    mflr r0
+; SOFT-NEXT:    std r0, 16(r1)
+; SOFT-NEXT:    stdu r1, -32(r1)
+; SOFT-NEXT:    lhz r3, 0(r3)
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    bl __fixunssfdi
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    addi r1, r1, 32
+; SOFT-NEXT:    ld r0, 16(r1)
+; SOFT-NEXT:    mtlr r0
+; SOFT-NEXT:    blr
+  %a = load half, half* %p, align 2
+  %r = fptoui half %a to i64
+  ret i64 %r
+}
+define void @test_uitofp_i64(i64 %a, half* %p) #0 {
+; P8-LABEL: test_uitofp_i64:
+; P8:       # %bb.0:
+; P8-NEXT:    mflr r0
+; P8-NEXT:    std r30, -16(r1) # 8-byte Folded Spill
+; P8-NEXT:    std r0, 16(r1)
+; P8-NEXT:    stdu r1, -48(r1)
+; P8-NEXT:    mtfprd f0, r3
+; P8-NEXT:    mr r30, r4
+; P8-NEXT:    xscvuxdsp f1, f0
+; P8-NEXT:    bl __gnu_f2h_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    sth r3, 0(r30)
+; P8-NEXT:    addi r1, r1, 48
+; P8-NEXT:    ld r0, 16(r1)
+; P8-NEXT:    ld r30, -16(r1) # 8-byte Folded Reload
+; P8-NEXT:    mtlr r0
+; P8-NEXT:    blr
+;
+; CHECK-LABEL: test_uitofp_i64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    mtfprd f0, r3
+; CHECK-NEXT:    xscvuxdsp f0, f0
+; CHECK-NEXT:    xscvdphp f0, f0
+; CHECK-NEXT:    mffprwz r3, f0
+; CHECK-NEXT:    sth r3, 0(r4)
+; CHECK-NEXT:    blr
+;
+; SOFT-LABEL: test_uitofp_i64:
+; SOFT:       # %bb.0:
+; SOFT-NEXT:    mflr r0
+; SOFT-NEXT:    std r30, -16(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r0, 16(r1)
+; SOFT-NEXT:    stdu r1, -48(r1)
+; SOFT-NEXT:    mr r30, r4
+; SOFT-NEXT:    bl __floatundisf
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    clrldi r3, r3, 48
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    sth r3, 0(r30)
+; SOFT-NEXT:    addi r1, r1, 48
+; SOFT-NEXT:    ld r0, 16(r1)
+; SOFT-NEXT:    ld r30, -16(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    mtlr r0
+; SOFT-NEXT:    blr
+  %r = uitofp i64 %a to half
+  store half %r, half* %p
+  ret void
+}
+define <4 x float> @test_extend32_vec4(<4 x half>* %p) #0 {
+; P8-LABEL: test_extend32_vec4:
+; P8:       # %bb.0:
+; P8-NEXT:    mflr r0
+; P8-NEXT:    std r30, -16(r1) # 8-byte Folded Spill
+; P8-NEXT:    std r0, 16(r1)
+; P8-NEXT:    stdu r1, -112(r1)
+; P8-NEXT:    mr r30, r3
+; P8-NEXT:    lhz r3, 6(r3)
+; P8-NEXT:    bl __gnu_h2f_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    li r3, 80
+; P8-NEXT:    # kill: def $f1 killed $f1 def $vsl1
+; P8-NEXT:    stxvd2x vs1, r1, r3 # 16-byte Folded Spill
+; P8-NEXT:    lhz r3, 2(r30)
+; P8-NEXT:    bl __gnu_h2f_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    li r3, 64
+; P8-NEXT:    # kill: def $f1 killed $f1 def $vsl1
+; P8-NEXT:    stxvd2x vs1, r1, r3 # 16-byte Folded Spill
+; P8-NEXT:    lhz r3, 4(r30)
+; P8-NEXT:    bl __gnu_h2f_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    li r3, 48
+; P8-NEXT:    # kill: def $f1 killed $f1 def $vsl1
+; P8-NEXT:    stxvd2x vs1, r1, r3 # 16-byte Folded Spill
+; P8-NEXT:    lhz r3, 0(r30)
+; P8-NEXT:    bl __gnu_h2f_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    li r3, 80
+; P8-NEXT:    # kill: def $f1 killed $f1 def $vsl1
+; P8-NEXT:    lxvd2x vs0, r1, r3 # 16-byte Folded Reload
+; P8-NEXT:    li r3, 64
+; P8-NEXT:    lxvd2x vs2, r1, r3 # 16-byte Folded Reload
+; P8-NEXT:    li r3, 48
+; P8-NEXT:    xxmrghd vs0, vs0, vs2
+; P8-NEXT:    lxvd2x vs2, r1, r3 # 16-byte Folded Reload
+; P8-NEXT:    xxmrghd vs1, vs2, vs1
+; P8-NEXT:    xvcvdpsp vs34, vs0
+; P8-NEXT:    xvcvdpsp vs35, vs1
+; P8-NEXT:    vmrgew v2, v2, v3
+; P8-NEXT:    addi r1, r1, 112
+; P8-NEXT:    ld r0, 16(r1)
+; P8-NEXT:    ld r30, -16(r1) # 8-byte Folded Reload
+; P8-NEXT:    mtlr r0
+; P8-NEXT:    blr
+;
+; CHECK-LABEL: test_extend32_vec4:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lhz r4, 6(r3)
+; CHECK-NEXT:    mtfprwz f0, r4
+; CHECK-NEXT:    xscvhpdp f0, f0
+; CHECK-NEXT:    lhz r4, 2(r3)
+; CHECK-NEXT:    mtfprwz f1, r4
+; CHECK-NEXT:    xscvhpdp f1, f1
+; CHECK-NEXT:    lhz r4, 4(r3)
+; CHECK-NEXT:    mtfprwz f2, r4
+; CHECK-NEXT:    xscvhpdp f2, f2
+; CHECK-NEXT:    lhz r3, 0(r3)
+; CHECK-NEXT:    xxmrghd vs0, vs0, vs1
+; CHECK-NEXT:    mtfprwz f3, r3
+; CHECK-NEXT:    xscvhpdp f3, f3
+; CHECK-NEXT:    xxmrghd vs2, vs2, vs3
+; CHECK-NEXT:    xvcvdpsp vs34, vs2
+; CHECK-NEXT:    xvcvdpsp vs35, vs0
+; CHECK-NEXT:    vmrgew v2, v3, v2
+; CHECK-NEXT:    blr
+;
+; SOFT-LABEL: test_extend32_vec4:
+; SOFT:       # %bb.0:
+; SOFT-NEXT:    mflr r0
+; SOFT-NEXT:    std r27, -40(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r28, -32(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r29, -24(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r30, -16(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r0, 16(r1)
+; SOFT-NEXT:    stdu r1, -80(r1)
+; SOFT-NEXT:    mr r30, r3
+; SOFT-NEXT:    lhz r3, 0(r3)
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r29, r3
+; SOFT-NEXT:    lhz r3, 2(r30)
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r28, r3
+; SOFT-NEXT:    lhz r3, 4(r30)
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r27, r3
+; SOFT-NEXT:    lhz r3, 6(r30)
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r6, r3
+; SOFT-NEXT:    mr r3, r29
+; SOFT-NEXT:    mr r4, r28
+; SOFT-NEXT:    mr r5, r27
+; SOFT-NEXT:    addi r1, r1, 80
+; SOFT-NEXT:    ld r0, 16(r1)
+; SOFT-NEXT:    ld r30, -16(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    ld r29, -24(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    ld r28, -32(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    mtlr r0
+; SOFT-NEXT:    ld r27, -40(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    blr
+  %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
+}
+define <4 x double> @test_extend64_vec4(<4 x half>* %p) #0 {
+; P8-LABEL: test_extend64_vec4:
+; P8:       # %bb.0:
+; P8-NEXT:    mflr r0
+; P8-NEXT:    std r30, -16(r1) # 8-byte Folded Spill
+; P8-NEXT:    std r0, 16(r1)
+; P8-NEXT:    stdu r1, -112(r1)
+; P8-NEXT:    mr r30, r3
+; P8-NEXT:    lhz r3, 6(r3)
+; P8-NEXT:    bl __gnu_h2f_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    li r3, 80
+; P8-NEXT:    # kill: def $f1 killed $f1 def $vsl1
+; P8-NEXT:    stxvd2x vs1, r1, r3 # 16-byte Folded Spill
+; P8-NEXT:    lhz r3, 4(r30)
+; P8-NEXT:    bl __gnu_h2f_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    li r3, 64
+; P8-NEXT:    # kill: def $f1 killed $f1 def $vsl1
+; P8-NEXT:    stxvd2x vs1, r1, r3 # 16-byte Folded Spill
+; P8-NEXT:    lhz r3, 2(r30)
+; P8-NEXT:    bl __gnu_h2f_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    li r3, 48
+; P8-NEXT:    # kill: def $f1 killed $f1 def $vsl1
+; P8-NEXT:    stxvd2x vs1, r1, r3 # 16-byte Folded Spill
+; P8-NEXT:    lhz r3, 0(r30)
+; P8-NEXT:    bl __gnu_h2f_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    li r3, 80
+; P8-NEXT:    # kill: def $f1 killed $f1 def $vsl1
+; P8-NEXT:    lxvd2x vs0, r1, r3 # 16-byte Folded Reload
+; P8-NEXT:    li r3, 64
+; P8-NEXT:    lxvd2x vs2, r1, r3 # 16-byte Folded Reload
+; P8-NEXT:    li r3, 48
+; P8-NEXT:    xxmrghd vs35, vs0, vs2
+; P8-NEXT:    lxvd2x vs0, r1, r3 # 16-byte Folded Reload
+; P8-NEXT:    xxmrghd vs34, vs0, vs1
+; P8-NEXT:    addi r1, r1, 112
+; P8-NEXT:    ld r0, 16(r1)
+; P8-NEXT:    ld r30, -16(r1) # 8-byte Folded Reload
+; P8-NEXT:    mtlr r0
+; P8-NEXT:    blr
+;
+; CHECK-LABEL: test_extend64_vec4:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    lhz r4, 6(r3)
+; CHECK-NEXT:    lhz r5, 4(r3)
+; CHECK-NEXT:    lhz r6, 2(r3)
+; CHECK-NEXT:    lhz r3, 0(r3)
+; CHECK-NEXT:    mtfprwz f0, r3
+; CHECK-NEXT:    mtfprwz f1, r6
+; CHECK-NEXT:    xscvhpdp f0, f0
+; CHECK-NEXT:    xscvhpdp f1, f1
+; CHECK-NEXT:    xxmrghd vs34, vs1, vs0
+; CHECK-NEXT:    mtfprwz f0, r5
+; CHECK-NEXT:    mtfprwz f1, r4
+; CHECK-NEXT:    xscvhpdp f0, f0
+; CHECK-NEXT:    xscvhpdp f1, f1
+; CHECK-NEXT:    xxmrghd vs35, vs1, vs0
+; CHECK-NEXT:    blr
+;
+; SOFT-LABEL: test_extend64_vec4:
+; SOFT:       # %bb.0:
+; SOFT-NEXT:    mflr r0
+; SOFT-NEXT:    std r27, -40(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r28, -32(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r29, -24(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r30, -16(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r0, 16(r1)
+; SOFT-NEXT:    stdu r1, -80(r1)
+; SOFT-NEXT:    mr r30, r3
+; SOFT-NEXT:    lhz r3, 0(r3)
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    bl __extendsfdf2
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r29, r3
+; SOFT-NEXT:    lhz r3, 2(r30)
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    bl __extendsfdf2
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r28, r3
+; SOFT-NEXT:    lhz r3, 4(r30)
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    bl __extendsfdf2
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r27, r3
+; SOFT-NEXT:    lhz r3, 6(r30)
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    bl __extendsfdf2
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r6, r3
+; SOFT-NEXT:    mr r3, r29
+; SOFT-NEXT:    mr r4, r28
+; SOFT-NEXT:    mr r5, r27
+; SOFT-NEXT:    addi r1, r1, 80
+; SOFT-NEXT:    ld r0, 16(r1)
+; SOFT-NEXT:    ld r30, -16(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    ld r29, -24(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    ld r28, -32(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    mtlr r0
+; SOFT-NEXT:    ld r27, -40(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    blr
+  %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
+}
+define void @test_trunc32_vec4(<4 x float> %a, <4 x half>* %p) #0 {
+; P8-LABEL: test_trunc32_vec4:
+; P8:       # %bb.0:
+; P8-NEXT:    mflr r0
+; P8-NEXT:    std r0, 16(r1)
+; P8-NEXT:    stdu r1, -112(r1)
+; P8-NEXT:    xxsldwi vs0, vs34, vs34, 3
+; P8-NEXT:    li r3, 48
+; P8-NEXT:    std r27, 72(r1) # 8-byte Folded Spill
+; P8-NEXT:    std r28, 80(r1) # 8-byte Folded Spill
+; P8-NEXT:    std r29, 88(r1) # 8-byte Folded Spill
+; P8-NEXT:    std r30, 96(r1) # 8-byte Folded Spill
+; P8-NEXT:    mr r30, r5
+; P8-NEXT:    xscvspdpn f1, vs0
+; P8-NEXT:    stxvd2x vs63, r1, r3 # 16-byte Folded Spill
+; P8-NEXT:    vmr v31, v2
+; P8-NEXT:    bl __gnu_f2h_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    xxswapd vs0, vs63
+; P8-NEXT:    mr r29, r3
+; P8-NEXT:    xscvspdpn f1, vs0
+; P8-NEXT:    bl __gnu_f2h_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    xxsldwi vs0, vs63, vs63, 1
+; P8-NEXT:    mr r28, r3
+; P8-NEXT:    xscvspdpn f1, vs0
+; P8-NEXT:    bl __gnu_f2h_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    xscvspdpn f1, vs63
+; P8-NEXT:    mr r27, r3
+; P8-NEXT:    bl __gnu_f2h_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    sth r3, 6(r30)
+; P8-NEXT:    li r3, 48
+; P8-NEXT:    sth r27, 4(r30)
+; P8-NEXT:    ld r27, 72(r1) # 8-byte Folded Reload
+; P8-NEXT:    sth r28, 2(r30)
+; P8-NEXT:    sth r29, 0(r30)
+; P8-NEXT:    ld r30, 96(r1) # 8-byte Folded Reload
+; P8-NEXT:    ld r29, 88(r1) # 8-byte Folded Reload
+; P8-NEXT:    lxvd2x vs63, r1, r3 # 16-byte Folded Reload
+; P8-NEXT:    ld r28, 80(r1) # 8-byte Folded Reload
+; P8-NEXT:    addi r1, r1, 112
+; P8-NEXT:    ld r0, 16(r1)
+; P8-NEXT:    mtlr r0
+; P8-NEXT:    blr
+;
+; CHECK-LABEL: test_trunc32_vec4:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    xxsldwi vs0, vs34, vs34, 3
+; CHECK-NEXT:    xscvspdpn f0, vs0
+; CHECK-NEXT:    xscvdphp f0, f0
+; CHECK-NEXT:    mffprwz r3, f0
+; CHECK-NEXT:    xxswapd vs0, vs34
+; CHECK-NEXT:    xxsldwi vs1, vs34, vs34, 1
+; CHECK-NEXT:    xscvspdpn f1, vs1
+; CHECK-NEXT:    xscvspdpn f0, vs0
+; CHECK-NEXT:    xscvdphp f0, f0
+; CHECK-NEXT:    xscvdphp f1, f1
+; CHECK-NEXT:    mffprwz r4, f1
+; CHECK-NEXT:    xscvspdpn f1, vs34
+; CHECK-NEXT:    xscvdphp f1, f1
+; CHECK-NEXT:    sth r4, 4(r5)
+; CHECK-NEXT:    mffprwz r4, f0
+; CHECK-NEXT:    sth r4, 2(r5)
+; CHECK-NEXT:    sth r3, 0(r5)
+; CHECK-NEXT:    mffprwz r6, f1
+; CHECK-NEXT:    sth r6, 6(r5)
+; CHECK-NEXT:    blr
+;
+; SOFT-LABEL: test_trunc32_vec4:
+; SOFT:       # %bb.0:
+; SOFT-NEXT:    mflr r0
+; SOFT-NEXT:    std r26, -48(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r27, -40(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r28, -32(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r29, -24(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r30, -16(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r0, 16(r1)
+; SOFT-NEXT:    stdu r1, -80(r1)
+; SOFT-NEXT:    mr r27, r3
+; SOFT-NEXT:    clrldi r3, r6, 32
+; SOFT-NEXT:    mr r30, r7
+; SOFT-NEXT:    mr r29, r5
+; SOFT-NEXT:    mr r28, r4
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r26, r3
+; SOFT-NEXT:    clrldi r3, r29, 32
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r29, r3
+; SOFT-NEXT:    clrldi r3, r28, 32
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r28, r3
+; SOFT-NEXT:    clrldi r3, r27, 32
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    clrldi r3, r3, 48
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r27, r3
+; SOFT-NEXT:    clrldi r3, r28, 48
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r28, r3
+; SOFT-NEXT:    clrldi r3, r29, 48
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r29, r3
+; SOFT-NEXT:    clrldi r3, r26, 48
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    sth r3, 6(r30)
+; SOFT-NEXT:    mr r3, r29
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    sth r3, 4(r30)
+; SOFT-NEXT:    mr r3, r28
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    sth r3, 2(r30)
+; SOFT-NEXT:    mr r3, r27
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    sth r3, 0(r30)
+; SOFT-NEXT:    addi r1, r1, 80
+; SOFT-NEXT:    ld r0, 16(r1)
+; SOFT-NEXT:    ld r30, -16(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    ld r29, -24(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    ld r28, -32(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    mtlr r0
+; SOFT-NEXT:    ld r27, -40(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    ld r26, -48(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    blr
+  %v = fptrunc <4 x float> %a to <4 x half>
+  store <4 x half> %v, <4 x half>* %p
+  ret void
+}
+define void @test_trunc64_vec4(<4 x double> %a, <4 x half>* %p) #0 {
+; P8-LABEL: test_trunc64_vec4:
+; P8:       # %bb.0:
+; P8-NEXT:    mflr r0
+; P8-NEXT:    std r0, 16(r1)
+; P8-NEXT:    stdu r1, -128(r1)
+; P8-NEXT:    li r3, 48
+; P8-NEXT:    xxswapd vs1, vs34
+; P8-NEXT:    std r27, 88(r1) # 8-byte Folded Spill
+; P8-NEXT:    std r28, 96(r1) # 8-byte Folded Spill
+; P8-NEXT:    std r29, 104(r1) # 8-byte Folded Spill
+; P8-NEXT:    std r30, 112(r1) # 8-byte Folded Spill
+; P8-NEXT:    mr r30, r7
+; P8-NEXT:    # kill: def $f1 killed $f1 killed $vsl1
+; P8-NEXT:    stxvd2x vs62, r1, r3 # 16-byte Folded Spill
+; P8-NEXT:    li r3, 64
+; P8-NEXT:    vmr v30, v2
+; P8-NEXT:    stxvd2x vs63, r1, r3 # 16-byte Folded Spill
+; P8-NEXT:    vmr v31, v3
+; P8-NEXT:    bl __truncdfhf2
+; P8-NEXT:    nop
+; P8-NEXT:    xxswapd vs1, vs63
+; P8-NEXT:    mr r29, r3
+; P8-NEXT:    # kill: def $f1 killed $f1 killed $vsl1
+; P8-NEXT:    bl __truncdfhf2
+; P8-NEXT:    nop
+; P8-NEXT:    xxlor f1, vs62, vs62
+; P8-NEXT:    mr r28, r3
+; P8-NEXT:    bl __truncdfhf2
+; P8-NEXT:    nop
+; P8-NEXT:    xxlor f1, vs63, vs63
+; P8-NEXT:    mr r27, r3
+; P8-NEXT:    bl __truncdfhf2
+; P8-NEXT:    nop
+; P8-NEXT:    sth r3, 6(r30)
+; P8-NEXT:    li r3, 64
+; P8-NEXT:    sth r27, 2(r30)
+; P8-NEXT:    ld r27, 88(r1) # 8-byte Folded Reload
+; P8-NEXT:    sth r28, 4(r30)
+; P8-NEXT:    sth r29, 0(r30)
+; P8-NEXT:    ld r30, 112(r1) # 8-byte Folded Reload
+; P8-NEXT:    ld r29, 104(r1) # 8-byte Folded Reload
+; P8-NEXT:    lxvd2x vs63, r1, r3 # 16-byte Folded Reload
+; P8-NEXT:    li r3, 48
+; P8-NEXT:    ld r28, 96(r1) # 8-byte Folded Reload
+; P8-NEXT:    lxvd2x vs62, r1, r3 # 16-byte Folded Reload
+; P8-NEXT:    addi r1, r1, 128
+; P8-NEXT:    ld r0, 16(r1)
+; P8-NEXT:    mtlr r0
+; P8-NEXT:    blr
+;
+; CHECK-LABEL: test_trunc64_vec4:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    xxswapd vs0, vs34
+; CHECK-NEXT:    xscvdphp f0, f0
+; CHECK-NEXT:    mffprwz r3, f0
+; CHECK-NEXT:    xxswapd vs0, vs35
+; CHECK-NEXT:    xscvdphp f0, f0
+; CHECK-NEXT:    xscvdphp f1, vs34
+; CHECK-NEXT:    mffprwz r4, f1
+; CHECK-NEXT:    xscvdphp f1, vs35
+; CHECK-NEXT:    sth r4, 2(r7)
+; CHECK-NEXT:    mffprwz r4, f0
+; CHECK-NEXT:    sth r4, 4(r7)
+; CHECK-NEXT:    sth r3, 0(r7)
+; CHECK-NEXT:    mffprwz r5, f1
+; CHECK-NEXT:    sth r5, 6(r7)
+; CHECK-NEXT:    blr
+;
+; SOFT-LABEL: test_trunc64_vec4:
+; SOFT:       # %bb.0:
+; SOFT-NEXT:    mflr r0
+; SOFT-NEXT:    std r26, -48(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r27, -40(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r28, -32(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r29, -24(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r30, -16(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r0, 16(r1)
+; SOFT-NEXT:    stdu r1, -80(r1)
+; SOFT-NEXT:    mr r27, r3
+; SOFT-NEXT:    mr r3, r6
+; SOFT-NEXT:    mr r30, r7
+; SOFT-NEXT:    mr r29, r5
+; SOFT-NEXT:    mr r28, r4
+; SOFT-NEXT:    bl __truncdfhf2
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r26, r3
+; SOFT-NEXT:    mr r3, r29
+; SOFT-NEXT:    bl __truncdfhf2
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r29, r3
+; SOFT-NEXT:    mr r3, r28
+; SOFT-NEXT:    bl __truncdfhf2
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r28, r3
+; SOFT-NEXT:    mr r3, r27
+; SOFT-NEXT:    bl __truncdfhf2
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    clrldi r3, r3, 48
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r27, r3
+; SOFT-NEXT:    clrldi r3, r28, 48
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r28, r3
+; SOFT-NEXT:    clrldi r3, r29, 48
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r29, r3
+; SOFT-NEXT:    clrldi r3, r26, 48
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    sth r3, 6(r30)
+; SOFT-NEXT:    mr r3, r29
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    sth r3, 4(r30)
+; SOFT-NEXT:    mr r3, r28
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    sth r3, 2(r30)
+; SOFT-NEXT:    mr r3, r27
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    sth r3, 0(r30)
+; SOFT-NEXT:    addi r1, r1, 80
+; SOFT-NEXT:    ld r0, 16(r1)
+; SOFT-NEXT:    ld r30, -16(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    ld r29, -24(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    ld r28, -32(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    mtlr r0
+; SOFT-NEXT:    ld r27, -40(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    ld r26, -48(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    blr
+  %v = fptrunc <4 x double> %a to <4 x half>
+  store <4 x half> %v, <4 x half>* %p
+  ret void
+}
+define float @test_sitofp_fadd_i32(i32 %a, half* %b) #0 {
+; P8-LABEL: test_sitofp_fadd_i32:
+; P8:       # %bb.0:
+; P8-NEXT:    mflr r0
+; P8-NEXT:    std r30, -24(r1) # 8-byte Folded Spill
+; P8-NEXT:    stfd f31, -8(r1) # 8-byte Folded Spill
+; P8-NEXT:    std r0, 16(r1)
+; P8-NEXT:    stdu r1, -64(r1)
+; P8-NEXT:    mr r30, r3
+; P8-NEXT:    lhz r3, 0(r4)
+; P8-NEXT:    bl __gnu_h2f_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    mtfprwa f0, r30
+; P8-NEXT:    fmr f31, f1
+; P8-NEXT:    xscvsxdsp f1, f0
+; P8-NEXT:    bl __gnu_f2h_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    bl __gnu_h2f_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    xsaddsp f1, f31, f1
+; P8-NEXT:    addi r1, r1, 64
+; P8-NEXT:    ld r0, 16(r1)
+; P8-NEXT:    lfd f31, -8(r1) # 8-byte Folded Reload
+; P8-NEXT:    ld r30, -24(r1) # 8-byte Folded Reload
+; P8-NEXT:    mtlr r0
+; P8-NEXT:    blr
+;
+; CHECK-LABEL: test_sitofp_fadd_i32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    mtfprwa f1, r3
+; CHECK-NEXT:    xscvsxdsp f1, f1
+; CHECK-NEXT:    lhz r4, 0(r4)
+; CHECK-NEXT:    mtfprwz f0, r4
+; CHECK-NEXT:    xscvhpdp f0, f0
+; CHECK-NEXT:    xscvdphp f1, f1
+; CHECK-NEXT:    mffprwz r3, f1
+; CHECK-NEXT:    mtfprwz f1, r3
+; CHECK-NEXT:    xscvhpdp f1, f1
+; CHECK-NEXT:    xsaddsp f1, f0, f1
+; CHECK-NEXT:    blr
+;
+; SOFT-LABEL: test_sitofp_fadd_i32:
+; SOFT:       # %bb.0:
+; SOFT-NEXT:    mflr r0
+; SOFT-NEXT:    std r29, -24(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r30, -16(r1) # 8-byte Folded Spill
+; SOFT-NEXT:    std r0, 16(r1)
+; SOFT-NEXT:    stdu r1, -64(r1)
+; SOFT-NEXT:    mr r30, r3
+; SOFT-NEXT:    lhz r3, 0(r4)
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r29, r3
+; SOFT-NEXT:    extsw r3, r30
+; SOFT-NEXT:    bl __floatsisf
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    clrldi r3, r3, 32
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    clrldi r3, r3, 48
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    mr r4, r3
+; SOFT-NEXT:    mr r3, r29
+; SOFT-NEXT:    bl __addsf3
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    addi r1, r1, 64
+; SOFT-NEXT:    ld r0, 16(r1)
+; SOFT-NEXT:    ld r30, -16(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    ld r29, -24(r1) # 8-byte Folded Reload
+; SOFT-NEXT:    mtlr r0
+; SOFT-NEXT:    blr
+  %tmp0 = load half, half* %b
+  %tmp1 = sitofp i32 %a to half
+  %tmp2 = fadd half %tmp0, %tmp1
+  %tmp3 = fpext half %tmp2 to float
+  ret float %tmp3
+}
+define half @PR40273(half) #0 {
+; P8-LABEL: PR40273:
+; P8:       # %bb.0:
+; P8-NEXT:    mflr r0
+; P8-NEXT:    std r0, 16(r1)
+; P8-NEXT:    stdu r1, -32(r1)
+; P8-NEXT:    bl __gnu_f2h_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    bl __gnu_h2f_ieee
+; P8-NEXT:    nop
+; P8-NEXT:    xxlxor f0, f0, f0
+; P8-NEXT:    fcmpu cr0, f1, f0
+; P8-NEXT:    beq cr0, .LBB20_2
+; P8-NEXT:  # %bb.1:
+; P8-NEXT:    addis r3, r2, .LCPI20_0@toc@ha
+; P8-NEXT:    lfs f0, .LCPI20_0@toc@l(r3)
+; P8-NEXT:  .LBB20_2:
+; P8-NEXT:    fmr f1, f0
+; P8-NEXT:    addi r1, r1, 32
+; P8-NEXT:    ld r0, 16(r1)
+; P8-NEXT:    mtlr r0
+; P8-NEXT:    blr
+;
+; CHECK-LABEL: PR40273:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    xscvdphp f0, f1
+; CHECK-NEXT:    xxlxor f1, f1, f1
+; CHECK-NEXT:    mffprwz r3, f0
+; CHECK-NEXT:    mtfprwz f0, r3
+; CHECK-NEXT:    xscvhpdp f0, f0
+; CHECK-NEXT:    fcmpu cr0, f0, f1
+; CHECK-NEXT:    beqlr cr0
+; CHECK-NEXT:  # %bb.1:
+; CHECK-NEXT:    addis r3, r2, .LCPI20_0@toc@ha
+; CHECK-NEXT:    lfs f1, .LCPI20_0@toc@l(r3)
+; CHECK-NEXT:    blr
+;
+; SOFT-LABEL: PR40273:
+; SOFT:       # %bb.0:
+; SOFT-NEXT:    mflr r0
+; SOFT-NEXT:    std r0, 16(r1)
+; SOFT-NEXT:    stdu r1, -32(r1)
+; SOFT-NEXT:    clrldi r3, r3, 48
+; SOFT-NEXT:    bl __gnu_h2f_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    li r4, 0
+; SOFT-NEXT:    bl __nesf2
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    cmplwi r3, 0
+; SOFT-NEXT:    lis r3, 16256
+; SOFT-NEXT:    isel r3, 0, r3, eq
+; SOFT-NEXT:    bl __gnu_f2h_ieee
+; SOFT-NEXT:    nop
+; SOFT-NEXT:    addi r1, r1, 32
+; SOFT-NEXT:    ld r0, 16(r1)
+; SOFT-NEXT:    mtlr r0
+; SOFT-NEXT:    blr
+  %2 = fcmp une half %0, 0xH0000
+  %3 = uitofp i1 %2 to half
+  ret half %3
+}
 attributes #0 = { nounwind }
diff --git a/llvm/test/CodeGen/PowerPC/scalar_vector_test_2.ll b/llvm/test/CodeGen/PowerPC/scalar_vector_test_2.ll
index 6dcd14d06c62..783ea3a11cce 100644
--- a/llvm/test/CodeGen/PowerPC/scalar_vector_test_2.ll
+++ b/llvm/test/CodeGen/PowerPC/scalar_vector_test_2.ll
@@ -11,46 +11,34 @@
 define void @test_liwzx1(<1 x float>* %A, <1 x float>* %B, <1 x float>* %C) {
 ; P9LE-LABEL: test_liwzx1:
 ; P9LE:       # %bb.0:
-; P9LE-NEXT:    lfiwzx f0, 0, r3
-; P9LE-NEXT:    lfiwzx f1, 0, r4
-; P9LE-NEXT:    xxswapd vs0, f0
-; P9LE-NEXT:    xxswapd vs1, f1
-; P9LE-NEXT:    xvaddsp vs0, vs0, vs1
-; P9LE-NEXT:    xxsldwi vs0, vs0, vs0, 2
-; P9LE-NEXT:    stfiwx f0, 0, r5
+; P9LE-NEXT:    lfs f0, 0(r3)
+; P9LE-NEXT:    lfs f1, 0(r4)
+; P9LE-NEXT:    xsaddsp f0, f0, f1
+; P9LE-NEXT:    stfs f0, 0(r5)
 ; P9LE-NEXT:    blr
 ;
 ; P9BE-LABEL: test_liwzx1:
 ; P9BE:       # %bb.0:
-; P9BE-NEXT:    lfiwzx f0, 0, r3
-; P9BE-NEXT:    lfiwzx f1, 0, r4
-; P9BE-NEXT:    xxsldwi vs0, f0, f0, 1
-; P9BE-NEXT:    xxsldwi vs1, f1, f1, 1
-; P9BE-NEXT:    xvaddsp vs0, vs0, vs1
-; P9BE-NEXT:    xxsldwi vs0, vs0, vs0, 3
-; P9BE-NEXT:    stfiwx f0, 0, r5
+; P9BE-NEXT:    lfs f0, 0(r3)
+; P9BE-NEXT:    lfs f1, 0(r4)
+; P9BE-NEXT:    xsaddsp f0, f0, f1
+; P9BE-NEXT:    stfs f0, 0(r5)
 ; P9BE-NEXT:    blr
 ;
 ; P8LE-LABEL: test_liwzx1:
 ; P8LE:       # %bb.0:
-; P8LE-NEXT:    lfiwzx f0, 0, r3
-; P8LE-NEXT:    lfiwzx f1, 0, r4
-; P8LE-NEXT:    xxswapd vs0, f0
-; P8LE-NEXT:    xxswapd vs1, f1
-; P8LE-NEXT:    xvaddsp vs0, vs0, vs1
-; P8LE-NEXT:    xxsldwi vs0, vs0, vs0, 2
-; P8LE-NEXT:    stfiwx f0, 0, r5
+; P8LE-NEXT:    lfsx f0, 0, r3
+; P8LE-NEXT:    lfsx f1, 0, r4
+; P8LE-NEXT:    xsaddsp f0, f0, f1
+; P8LE-NEXT:    stfsx f0, 0, r5
 ; P8LE-NEXT:    blr
 ;
 ; P8BE-LABEL: test_liwzx1:
 ; P8BE:       # %bb.0:
-; P8BE-NEXT:    lfiwzx f0, 0, r3
-; P8BE-NEXT:    lfiwzx f1, 0, r4
-; P8BE-NEXT:    xxsldwi vs0, f0, f0, 1
-; P8BE-NEXT:    xxsldwi vs1, f1, f1, 1
-; P8BE-NEXT:    xvaddsp vs0, vs0, vs1
-; P8BE-NEXT:    xxsldwi vs0, vs0, vs0, 3
-; P8BE-NEXT:    stfiwx f0, 0, r5
+; P8BE-NEXT:    lfsx f0, 0, r3
+; P8BE-NEXT:    lfsx f1, 0, r4
+; P8BE-NEXT:    xsaddsp f0, f0, f1
+; P8BE-NEXT:    stfsx f0, 0, r5
 ; P8BE-NEXT:    blr
 
 
@@ -65,50 +53,38 @@ define void @test_liwzx1(<1 x float>* %A, <1 x float>* %B, <1 x float>* %C) {
 define <1 x float>* @test_liwzx2(<1 x float>* %A, <1 x float>* %B, <1 x float>* %C) {
 ; P9LE-LABEL: test_liwzx2:
 ; P9LE:       # %bb.0:
-; P9LE-NEXT:    lfiwzx f0, 0, r3
-; P9LE-NEXT:    lfiwzx f1, 0, r4
-; P9LE-NEXT:    xxswapd vs0, f0
-; P9LE-NEXT:    xxswapd vs1, f1
-; P9LE-NEXT:    xvsubsp vs0, vs0, vs1
-; P9LE-NEXT:    xxsldwi vs0, vs0, vs0, 2
+; P9LE-NEXT:    lfs f0, 0(r3)
 ; P9LE-NEXT:    mr r3, r5
-; P9LE-NEXT:    stfiwx f0, 0, r5
+; P9LE-NEXT:    lfs f1, 0(r4)
+; P9LE-NEXT:    xssubsp f0, f0, f1
+; P9LE-NEXT:    stfs f0, 0(r5)
 ; P9LE-NEXT:    blr
 ;
 ; P9BE-LABEL: test_liwzx2:
 ; P9BE:       # %bb.0:
-; P9BE-NEXT:    lfiwzx f0, 0, r3
-; P9BE-NEXT:    lfiwzx f1, 0, r4
-; P9BE-NEXT:    xxsldwi vs0, f0, f0, 1
-; P9BE-NEXT:    xxsldwi vs1, f1, f1, 1
-; P9BE-NEXT:    xvsubsp vs0, vs0, vs1
-; P9BE-NEXT:    xxsldwi vs0, vs0, vs0, 3
+; P9BE-NEXT:    lfs f0, 0(r3)
 ; P9BE-NEXT:    mr r3, r5
-; P9BE-NEXT:    stfiwx f0, 0, r5
+; P9BE-NEXT:    lfs f1, 0(r4)
+; P9BE-NEXT:    xssubsp f0, f0, f1
+; P9BE-NEXT:    stfs f0, 0(r5)
 ; P9BE-NEXT:    blr
 ;
 ; P8LE-LABEL: test_liwzx2:
 ; P8LE:       # %bb.0:
-; P8LE-NEXT:    lfiwzx f0, 0, r3
-; P8LE-NEXT:    lfiwzx f1, 0, r4
+; P8LE-NEXT:    lfsx f0, 0, r3
+; P8LE-NEXT:    lfsx f1, 0, r4
 ; P8LE-NEXT:    mr r3, r5
-; P8LE-NEXT:    xxswapd vs0, f0
-; P8LE-NEXT:    xxswapd vs1, f1
-; P8LE-NEXT:    xvsubsp vs0, vs0, vs1
-; P8LE-NEXT:    xxsldwi vs0, vs0, vs0, 2
-; P8LE-NEXT:    stfiwx f0, 0, r5
+; P8LE-NEXT:    xssubsp f0, f0, f1
+; P8LE-NEXT:    stfsx f0, 0, r5
 ; P8LE-NEXT:    blr
 ;
 ; P8BE-LABEL: test_liwzx2:
 ; P8BE:       # %bb.0:
-; P8BE-NEXT:    lfiwzx f0, 0, r3
-; P8BE-NEXT:    lfiwzx f1, 0, r4
+; P8BE-NEXT:    lfsx f0, 0, r3
+; P8BE-NEXT:    lfsx f1, 0, r4
 ; P8BE-NEXT:    mr r3, r5
-; P8BE-NEXT:    xxsldwi vs0, f0, f0, 1
-; P8BE-NEXT:    xxsldwi vs1, f1, f1, 1
-; P8BE-NEXT:    xvsubsp vs0, vs0, vs1
-; P8BE-NEXT:    xxsldwi vs0, vs0, vs0, 3
-; P8BE-NEXT:    stfiwx f0, 0, r5
+; P8BE-NEXT:    xssubsp f0, f0, f1
+; P8BE-NEXT:    stfsx f0, 0, r5
 ; P8BE-NEXT:    blr
 
 
diff --git a/llvm/test/CodeGen/PowerPC/vector-constrained-fp-intrinsics.ll b/llvm/test/CodeGen/PowerPC/vector-constrained-fp-intrinsics.ll
index 4117f279eb16..c188cd6bf14a 100644
--- a/llvm/test/CodeGen/PowerPC/vector-constrained-fp-intrinsics.ll
+++ b/llvm/test/CodeGen/PowerPC/vector-constrained-fp-intrinsics.ll
@@ -9,9 +9,7 @@ define <1 x float> @constrained_vector_fdiv_v1f32() #0 {
 ; PC64LE-NEXT:    addis 4, 2, .LCPI0_1@toc@ha
 ; PC64LE-NEXT:    lfs 0, .LCPI0_0@toc@l(3)
 ; PC64LE-NEXT:    lfs 1, .LCPI0_1@toc@l(4)
-; PC64LE-NEXT:    xsdivsp 0, 1, 0
-; PC64LE-NEXT:    xscvdpspn 0, 0
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
+; PC64LE-NEXT:    xsdivsp 1, 1, 0
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_fdiv_v1f32:
@@ -20,9 +18,7 @@ define <1 x float> @constrained_vector_fdiv_v1f32() #0 {
 ; PC64LE9-NEXT:    lfs 0, .LCPI0_0@toc@l(3)
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI0_1@toc@ha
 ; PC64LE9-NEXT:    lfs 1, .LCPI0_1@toc@l(3)
-; PC64LE9-NEXT:    xsdivsp 0, 1, 0
-; PC64LE9-NEXT:    xscvdpspn 0, 0
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
+; PC64LE9-NEXT:    xsdivsp 1, 1, 0
 ; PC64LE9-NEXT:    blr
 entry:
   %div = call <1 x float> @llvm.experimental.constrained.fdiv.v1f32(
@@ -232,8 +228,6 @@ define <1 x float> @constrained_vector_frem_v1f32() #0 {
 ; PC64LE-NEXT:    lfs 2, .LCPI5_1@toc@l(4)
 ; PC64LE-NEXT:    bl fmodf
 ; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    xscvdpspn 0, 1
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE-NEXT:    addi 1, 1, 32
 ; PC64LE-NEXT:    ld 0, 16(1)
 ; PC64LE-NEXT:    mtlr 0
@@ -250,8 +244,6 @@ define <1 x float> @constrained_vector_frem_v1f32() #0 {
 ; PC64LE9-NEXT:    lfs 2, .LCPI5_1@toc@l(3)
 ; PC64LE9-NEXT:    bl fmodf
 ; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    xscvdpspn 0, 1
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE9-NEXT:    addi 1, 1, 32
 ; PC64LE9-NEXT:    ld 0, 16(1)
 ; PC64LE9-NEXT:    mtlr 0
@@ -645,9 +637,7 @@ define <1 x float> @constrained_vector_fmul_v1f32() #0 {
 ; PC64LE-NEXT:    addis 4, 2, .LCPI10_1@toc@ha
 ; PC64LE-NEXT:    lfs 0, .LCPI10_0@toc@l(3)
 ; PC64LE-NEXT:    lfs 1, .LCPI10_1@toc@l(4)
-; PC64LE-NEXT:    xsmulsp 0, 1, 0
-; PC64LE-NEXT:    xscvdpspn 0, 0
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
+; PC64LE-NEXT:    xsmulsp 1, 1, 0
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_fmul_v1f32:
@@ -656,9 +646,7 @@ define <1 x float> @constrained_vector_fmul_v1f32() #0 {
 ; PC64LE9-NEXT:    lfs 0, .LCPI10_0@toc@l(3)
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI10_1@toc@ha
 ; PC64LE9-NEXT:    lfs 1, .LCPI10_1@toc@l(3)
-; PC64LE9-NEXT:    xsmulsp 0, 1, 0
-; PC64LE9-NEXT:    xscvdpspn 0, 0
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
+; PC64LE9-NEXT:    xsmulsp 1, 1, 0
 ; PC64LE9-NEXT:    blr
 entry:
   %mul = call <1 x float> @llvm.experimental.constrained.fmul.v1f32(
@@ -865,9 +853,7 @@ define <1 x float> @constrained_vector_fadd_v1f32() #0 {
 ; PC64LE-NEXT:    addis 4, 2, .LCPI15_1@toc@ha
 ; PC64LE-NEXT:    lfs 0, .LCPI15_0@toc@l(3)
 ; PC64LE-NEXT:    lfs 1, .LCPI15_1@toc@l(4)
-; PC64LE-NEXT:    xsaddsp 0, 1, 0
-; PC64LE-NEXT:    xscvdpspn 0, 0
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
+; PC64LE-NEXT:    xsaddsp 1, 1, 0
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_fadd_v1f32:
@@ -876,9 +862,7 @@ define <1 x float> @constrained_vector_fadd_v1f32() #0 {
 ; PC64LE9-NEXT:    lfs 0, .LCPI15_0@toc@l(3)
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI15_1@toc@ha
 ; PC64LE9-NEXT:    lfs 1, .LCPI15_1@toc@l(3)
-; PC64LE9-NEXT:    xsaddsp 0, 1, 0
-; PC64LE9-NEXT:    xscvdpspn 0, 0
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
+; PC64LE9-NEXT:    xsaddsp 1, 1, 0
 ; PC64LE9-NEXT:    blr
 entry:
   %add = call <1 x float> @llvm.experimental.constrained.fadd.v1f32(
@@ -1081,9 +1065,7 @@ define <1 x float> @constrained_vector_fsub_v1f32() #0 {
 ; PC64LE-NEXT:    addis 4, 2, .LCPI20_1@toc@ha
 ; PC64LE-NEXT:    lfs 0, .LCPI20_0@toc@l(3)
 ; PC64LE-NEXT:    lfs 1, .LCPI20_1@toc@l(4)
-; PC64LE-NEXT:    xssubsp 0, 1, 0
-; PC64LE-NEXT:    xscvdpspn 0, 0
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
+; PC64LE-NEXT:    xssubsp 1, 1, 0
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_fsub_v1f32:
@@ -1092,9 +1074,7 @@ define <1 x float> @constrained_vector_fsub_v1f32() #0 {
 ; PC64LE9-NEXT:    lfs 0, .LCPI20_0@toc@l(3)
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI20_1@toc@ha
 ; PC64LE9-NEXT:    lfs 1, .LCPI20_1@toc@l(3)
-; PC64LE9-NEXT:    xssubsp 0, 1, 0
-; PC64LE9-NEXT:    xscvdpspn 0, 0
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
+; PC64LE9-NEXT:    xssubsp 1, 1, 0
 ; PC64LE9-NEXT:    blr
 entry:
   %sub = call <1 x float> @llvm.experimental.constrained.fsub.v1f32(
@@ -1295,18 +1275,14 @@ define <1 x float> @constrained_vector_sqrt_v1f32() #0 {
 ; PC64LE:       # %bb.0: # %entry
 ; PC64LE-NEXT:    addis 3, 2, .LCPI25_0@toc@ha
 ; PC64LE-NEXT:    lfs 0, .LCPI25_0@toc@l(3)
-; PC64LE-NEXT:    xssqrtsp 0, 0
-; PC64LE-NEXT:    xscvdpspn 0, 0
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
+; PC64LE-NEXT:    xssqrtsp 1, 0
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_sqrt_v1f32:
 ; PC64LE9:       # %bb.0: # %entry
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI25_0@toc@ha
 ; PC64LE9-NEXT:    lfs 0, .LCPI25_0@toc@l(3)
-; PC64LE9-NEXT:    xssqrtsp 0, 0
-; PC64LE9-NEXT:    xscvdpspn 0, 0
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
+; PC64LE9-NEXT:    xssqrtsp 1, 0
 ; PC64LE9-NEXT:    blr
 entry:
   %sqrt = call <1 x float> @llvm.experimental.constrained.sqrt.v1f32(
@@ -1481,8 +1457,6 @@ define <1 x float> @constrained_vector_pow_v1f32() #0 {
 ; PC64LE-NEXT:    lfs 2, .LCPI30_1@toc@l(4)
 ; PC64LE-NEXT:    bl powf
 ; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    xscvdpspn 0, 1
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE-NEXT:    addi 1, 1, 32
 ; PC64LE-NEXT:    ld 0, 16(1)
 ; PC64LE-NEXT:    mtlr 0
@@ -1499,8 +1473,6 @@ define <1 x float> @constrained_vector_pow_v1f32() #0 {
 ; PC64LE9-NEXT:    lfs 2, .LCPI30_1@toc@l(3)
 ; PC64LE9-NEXT:    bl powf
 ; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    xscvdpspn 0, 1
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE9-NEXT:    addi 1, 1, 32
 ; PC64LE9-NEXT:    ld 0, 16(1)
 ; PC64LE9-NEXT:    mtlr 0
@@ -1899,8 +1871,6 @@ define <1 x float> @constrained_vector_powi_v1f32() #0 {
 ; PC64LE-NEXT:    lfs 1, .LCPI35_0@toc@l(3)
 ; PC64LE-NEXT:    bl __powisf2
 ; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    xscvdpspn 0, 1
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE-NEXT:    addi 1, 1, 32
 ; PC64LE-NEXT:    ld 0, 16(1)
 ; PC64LE-NEXT:    mtlr 0
@@ -1916,8 +1886,6 @@ define <1 x float> @constrained_vector_powi_v1f32() #0 {
 ; PC64LE9-NEXT:    li 4, 3
 ; PC64LE9-NEXT:    bl __powisf2
 ; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    xscvdpspn 0, 1
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE9-NEXT:    addi 1, 1, 32
 ; PC64LE9-NEXT:    ld 0, 16(1)
 ; PC64LE9-NEXT:    mtlr 0
@@ -2284,8 +2252,6 @@ define <1 x float> @constrained_vector_sin_v1f32() #0 {
 ; PC64LE-NEXT:    lfs 1, .LCPI40_0@toc@l(3)
 ; PC64LE-NEXT:    bl sinf
 ; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    xscvdpspn 0, 1
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE-NEXT:    addi 1, 1, 32
 ; PC64LE-NEXT:    ld 0, 16(1)
 ; PC64LE-NEXT:    mtlr 0
@@ -2300,8 +2266,6 @@ define <1 x float> @constrained_vector_sin_v1f32() #0 {
 ; PC64LE9-NEXT:    lfs 1, .LCPI40_0@toc@l(3)
 ; PC64LE9-NEXT:    bl sinf
 ; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    xscvdpspn 0, 1
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE9-NEXT:    addi 1, 1, 32
 ; PC64LE9-NEXT:    ld 0, 16(1)
 ; PC64LE9-NEXT:    mtlr 0
@@ -2637,8 +2601,6 @@ define <1 x float> @constrained_vector_cos_v1f32() #0 {
 ; PC64LE-NEXT:    lfs 1, .LCPI45_0@toc@l(3)
 ; PC64LE-NEXT:    bl cosf
 ; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    xscvdpspn 0, 1
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE-NEXT:    addi 1, 1, 32
 ; PC64LE-NEXT:    ld 0, 16(1)
 ; PC64LE-NEXT:    mtlr 0
@@ -2653,8 +2615,6 @@ define <1 x float> @constrained_vector_cos_v1f32() #0 {
 ; PC64LE9-NEXT:    lfs 1, .LCPI45_0@toc@l(3)
 ; PC64LE9-NEXT:    bl cosf
 ; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    xscvdpspn 0, 1
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE9-NEXT:    addi 1, 1, 32
 ; PC64LE9-NEXT:    ld 0, 16(1)
 ; PC64LE9-NEXT:    mtlr 0
@@ -2990,8 +2950,6 @@ define <1 x float> @constrained_vector_exp_v1f32() #0 {
 ; PC64LE-NEXT:    lfs 1, .LCPI50_0@toc@l(3)
 ; PC64LE-NEXT:    bl expf
 ; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    xscvdpspn 0, 1
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE-NEXT:    addi 1, 1, 32
 ; PC64LE-NEXT:    ld 0, 16(1)
 ; PC64LE-NEXT:    mtlr 0
@@ -3006,8 +2964,6 @@ define <1 x float> @constrained_vector_exp_v1f32() #0 {
 ; PC64LE9-NEXT:    lfs 1, .LCPI50_0@toc@l(3)
 ; PC64LE9-NEXT:    bl expf
 ; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    xscvdpspn 0, 1
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE9-NEXT:    addi 1, 1, 32
 ; PC64LE9-NEXT:    ld 0, 16(1)
 ; PC64LE9-NEXT:    mtlr 0
@@ -3343,8 +3299,6 @@ define <1 x float> @constrained_vector_exp2_v1f32() #0 {
 ; PC64LE-NEXT:    lfs 1, .LCPI55_0@toc@l(3)
 ; PC64LE-NEXT:    bl exp2f
 ; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    xscvdpspn 0, 1
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE-NEXT:    addi 1, 1, 32
 ; PC64LE-NEXT:    ld 0, 16(1)
 ; PC64LE-NEXT:    mtlr 0
@@ -3359,8 +3313,6 @@ define <1 x float> @constrained_vector_exp2_v1f32() #0 {
 ; PC64LE9-NEXT:    lfs 1, .LCPI55_0@toc@l(3)
 ; PC64LE9-NEXT:    bl exp2f
 ; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    xscvdpspn 0, 1
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE9-NEXT:    addi 1, 1, 32
 ; PC64LE9-NEXT:    ld 0, 16(1)
 ; PC64LE9-NEXT:    mtlr 0
@@ -3696,8 +3648,6 @@ define <1 x float> @constrained_vector_log_v1f32() #0 {
 ; PC64LE-NEXT:    lfs 1, .LCPI60_0@toc@l(3)
 ; PC64LE-NEXT:    bl logf
 ; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    xscvdpspn 0, 1
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE-NEXT:    addi 1, 1, 32
 ; PC64LE-NEXT:    ld 0, 16(1)
 ; PC64LE-NEXT:    mtlr 0
@@ -3712,8 +3662,6 @@ define <1 x float> @constrained_vector_log_v1f32() #0 {
 ; PC64LE9-NEXT:    lfs 1, .LCPI60_0@toc@l(3)
 ; PC64LE9-NEXT:    bl logf
 ; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    xscvdpspn 0, 1
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE9-NEXT:    addi 1, 1, 32
 ; PC64LE9-NEXT:    ld 0, 16(1)
 ; PC64LE9-NEXT:    mtlr 0
@@ -4049,8 +3997,6 @@ define <1 x float> @constrained_vector_log10_v1f32() #0 {
 ; PC64LE-NEXT:    lfs 1, .LCPI65_0@toc@l(3)
 ; PC64LE-NEXT:    bl log10f
 ; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    xscvdpspn 0, 1
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE-NEXT:    addi 1, 1, 32
 ; PC64LE-NEXT:    ld 0, 16(1)
 ; PC64LE-NEXT:    mtlr 0
@@ -4065,8 +4011,6 @@ define <1 x float> @constrained_vector_log10_v1f32() #0 {
 ; PC64LE9-NEXT:    lfs 1, .LCPI65_0@toc@l(3)
 ; PC64LE9-NEXT:    bl log10f
 ; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    xscvdpspn 0, 1
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE9-NEXT:    addi 1, 1, 32
 ; PC64LE9-NEXT:    ld 0, 16(1)
 ; PC64LE9-NEXT:    mtlr 0
@@ -4402,8 +4346,6 @@ define <1 x float> @constrained_vector_log2_v1f32() #0 {
 ; PC64LE-NEXT:    lfs 1, .LCPI70_0@toc@l(3)
 ; PC64LE-NEXT:    bl log2f
 ; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    xscvdpspn 0, 1
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE-NEXT:    addi 1, 1, 32
 ; PC64LE-NEXT:    ld 0, 16(1)
 ; PC64LE-NEXT:    mtlr 0
@@ -4418,8 +4360,6 @@ define <1 x float> @constrained_vector_log2_v1f32() #0 {
 ; PC64LE9-NEXT:    lfs 1, .LCPI70_0@toc@l(3)
 ; PC64LE9-NEXT:    bl log2f
 ; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    xscvdpspn 0, 1
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE9-NEXT:    addi 1, 1, 32
 ; PC64LE9-NEXT:    ld 0, 16(1)
 ; PC64LE9-NEXT:    mtlr 0
@@ -4750,18 +4690,14 @@ define <1 x float> @constrained_vector_rint_v1f32() #0 {
 ; PC64LE:       # %bb.0: # %entry
 ; PC64LE-NEXT:    addis 3, 2, .LCPI75_0@toc@ha
 ; PC64LE-NEXT:    lfs 0, .LCPI75_0@toc@l(3)
-; PC64LE-NEXT:    xsrdpic 0, 0
-; PC64LE-NEXT:    xscvdpspn 0, 0
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
+; PC64LE-NEXT:    xsrdpic 1, 0
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_rint_v1f32:
 ; PC64LE9:       # %bb.0: # %entry
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI75_0@toc@ha
 ; PC64LE9-NEXT:    lfs 0, .LCPI75_0@toc@l(3)
-; PC64LE9-NEXT:    xsrdpic 0, 0
-; PC64LE9-NEXT:    xscvdpspn 0, 0
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
+; PC64LE9-NEXT:    xsrdpic 1, 0
 ; PC64LE9-NEXT:    blr
 entry:
   %rint = call <1 x float> @llvm.experimental.constrained.rint.v1f32(
@@ -4934,8 +4870,6 @@ define <1 x float> @constrained_vector_nearbyint_v1f32() #0 {
 ; PC64LE-NEXT:    lfs 1, .LCPI80_0@toc@l(3)
 ; PC64LE-NEXT:    bl nearbyintf
 ; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    xscvdpspn 0, 1
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE-NEXT:    addi 1, 1, 32
 ; PC64LE-NEXT:    ld 0, 16(1)
 ; PC64LE-NEXT:    mtlr 0
@@ -4950,8 +4884,6 @@ define <1 x float> @constrained_vector_nearbyint_v1f32() #0 {
 ; PC64LE9-NEXT:    lfs 1, .LCPI80_0@toc@l(3)
 ; PC64LE9-NEXT:    bl nearbyintf
 ; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    xscvdpspn 0, 1
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE9-NEXT:    addi 1, 1, 32
 ; PC64LE9-NEXT:    ld 0, 16(1)
 ; PC64LE9-NEXT:    mtlr 0
@@ -5289,8 +5221,6 @@ define <1 x float> @constrained_vector_maxnum_v1f32() #0 {
 ; PC64LE-NEXT:    lfs 2, .LCPI85_1@toc@l(4)
 ; PC64LE-NEXT:    bl fmaxf
 ; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    xscvdpspn 0, 1
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE-NEXT:    addi 1, 1, 32
 ; PC64LE-NEXT:    ld 0, 16(1)
 ; PC64LE-NEXT:    mtlr 0
@@ -5307,8 +5237,6 @@ define <1 x float> @constrained_vector_maxnum_v1f32() #0 {
 ; PC64LE9-NEXT:    lfs 2, .LCPI85_1@toc@l(3)
 ; PC64LE9-NEXT:    bl fmaxf
 ; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    xscvdpspn 0, 1
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE9-NEXT:    addi 1, 1, 32
 ; PC64LE9-NEXT:    ld 0, 16(1)
 ; PC64LE9-NEXT:    mtlr 0
@@ -5580,8 +5508,6 @@ define <1 x float> @constrained_vector_minnum_v1f32() #0 {
 ; PC64LE-NEXT:    lfs 2, .LCPI90_1@toc@l(4)
 ; PC64LE-NEXT:    bl fminf
 ; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    xscvdpspn 0, 1
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE-NEXT:    addi 1, 1, 32
 ; PC64LE-NEXT:    ld 0, 16(1)
 ; PC64LE-NEXT:    mtlr 0
@@ -5598,8 +5524,6 @@ define <1 x float> @constrained_vector_minnum_v1f32() #0 {
 ; PC64LE9-NEXT:    lfs 2, .LCPI90_1@toc@l(3)
 ; PC64LE9-NEXT:    bl fminf
 ; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    xscvdpspn 0, 1
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
 ; PC64LE9-NEXT:    addi 1, 1, 32
 ; PC64LE9-NEXT:    ld 0, 16(1)
 ; PC64LE9-NEXT:    mtlr 0
@@ -5864,18 +5788,14 @@ define <1 x float> @constrained_vector_fptrunc_v1f64() #0 {
 ; PC64LE:       # %bb.0: # %entry
 ; PC64LE-NEXT:    addis 3, 2, .LCPI95_0@toc@ha
 ; PC64LE-NEXT:    lfd 0, .LCPI95_0@toc@l(3)
-; PC64LE-NEXT:    frsp 0, 0
-; PC64LE-NEXT:    xscvdpspn 0, 0
-; PC64LE-NEXT:    xxsldwi 34, 0, 0, 1
+; PC64LE-NEXT:    frsp 1, 0
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_fptrunc_v1f64:
 ; PC64LE9:       # %bb.0: # %entry
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI95_0@toc@ha
 ; PC64LE9-NEXT:    lfd 0, .LCPI95_0@toc@l(3)
-; PC64LE9-NEXT:    frsp 0, 0
-; PC64LE9-NEXT:    xscvdpspn 0, 0
-; PC64LE9-NEXT:    xxsldwi 34, 0, 0, 1
+; PC64LE9-NEXT:    frsp 1, 0
 ; PC64LE9-NEXT:    blr
 entry:
   %result = call <1 x float> @llvm.experimental.constrained.fptrunc.v1f32.v1f64(
@@ -6027,15 +5947,13 @@ define <1 x double> @constrained_vector_fpext_v1f32() #0 {
 ; PC64LE-LABEL: constrained_vector_fpext_v1f32:
 ; PC64LE:       # %bb.0: # %entry
 ; PC64LE-NEXT:    addis 3, 2, .LCPI99_0@toc@ha
-; PC64LE-NEXT:    lfs 0, .LCPI99_0@toc@l(3)
-; PC64LE-NEXT:    xxspltd 34, 0, 0
+; PC64LE-NEXT:    lfs 1, .LCPI99_0@toc@l(3)
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_fpext_v1f32:
 ; PC64LE9:       # %bb.0: # %entry
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI99_0@toc@ha
-; PC64LE9-NEXT:    lfs 0, .LCPI99_0@toc@l(3)
-; PC64LE9-NEXT:    xxspltd 34, 0, 0
+; PC64LE9-NEXT:    lfs 1, .LCPI99_0@toc@l(3)
 ; PC64LE9-NEXT:    blr
 entry:
   %result = call <1 x double> @llvm.experimental.constrained.fpext.v1f64.v1f32(
@@ -6137,17 +6055,13 @@ define <1 x float> @constrained_vector_ceil_v1f32() #0 {
 ; PC64LE-LABEL: constrained_vector_ceil_v1f32:
 ; PC64LE:       # %bb.0: # %entry
 ; PC64LE-NEXT:    addis 3, 2, .LCPI103_0@toc@ha
-; PC64LE-NEXT:    addi 3, 3, .LCPI103_0@toc@l
-; PC64LE-NEXT:    lfiwzx 0, 0, 3
-; PC64LE-NEXT:    xxswapd 34, 0
+; PC64LE-NEXT:    lfs 1, .LCPI103_0@toc@l(3)
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_ceil_v1f32:
 ; PC64LE9:       # %bb.0: # %entry
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI103_0@toc@ha
-; PC64LE9-NEXT:    addi 3, 3, .LCPI103_0@toc@l
-; PC64LE9-NEXT:    lfiwzx 0, 0, 3
-; PC64LE9-NEXT:    xxswapd 34, 0
+; PC64LE9-NEXT:    lfs 1, .LCPI103_0@toc@l(3)
 ; PC64LE9-NEXT:    blr
 entry:
   %ceil = call <1 x float> @llvm.experimental.constrained.ceil.v1f32(
@@ -6226,17 +6140,13 @@ define <1 x float> @constrained_vector_floor_v1f32() #0 {
 ; PC64LE-LABEL: constrained_vector_floor_v1f32:
 ; PC64LE:       # %bb.0: # %entry
 ; PC64LE-NEXT:    addis 3, 2, .LCPI107_0@toc@ha
-; PC64LE-NEXT:    addi 3, 3, .LCPI107_0@toc@l
-; PC64LE-NEXT:    lfiwzx 0, 0, 3
-; PC64LE-NEXT:    xxswapd 34, 0
+; PC64LE-NEXT:    lfs 1, .LCPI107_0@toc@l(3)
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_floor_v1f32:
 ; PC64LE9:       # %bb.0: # %entry
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI107_0@toc@ha
-; PC64LE9-NEXT:    addi 3, 3, .LCPI107_0@toc@l
-; PC64LE9-NEXT:    lfiwzx 0, 0, 3
-; PC64LE9-NEXT:    xxswapd 34, 0
+; PC64LE9-NEXT:    lfs 1, .LCPI107_0@toc@l(3)
 ; PC64LE9-NEXT:    blr
 entry:
   %floor = call <1 x float> @llvm.experimental.constrained.floor.v1f32(
@@ -6316,17 +6226,13 @@ define <1 x float> @constrained_vector_round_v1f32() #0 {
 ; PC64LE-LABEL: constrained_vector_round_v1f32:
 ; PC64LE:       # %bb.0: # %entry
 ; PC64LE-NEXT:    addis 3, 2, .LCPI111_0@toc@ha
-; PC64LE-NEXT:    addi 3, 3, .LCPI111_0@toc@l
-; PC64LE-NEXT:    lfiwzx 0, 0, 3
-; PC64LE-NEXT:    xxswapd 34, 0
+; PC64LE-NEXT:    lfs 1, .LCPI111_0@toc@l(3)
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_round_v1f32:
 ; PC64LE9:       # %bb.0: # %entry
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI111_0@toc@ha
-; PC64LE9-NEXT:    addi 3, 3, .LCPI111_0@toc@l
-; PC64LE9-NEXT:    lfiwzx 0, 0, 3
-; PC64LE9-NEXT:    xxswapd 34, 0
+; PC64LE9-NEXT:    lfs 1, .LCPI111_0@toc@l(3)
 ; PC64LE9-NEXT:    blr
 entry:
   %round = call <1 x float> @llvm.experimental.constrained.round.v1f32(
@@ -6408,17 +6314,13 @@ define <1 x float> @constrained_vector_trunc_v1f32() #0 {
 ; PC64LE-LABEL: constrained_vector_trunc_v1f32:
 ; PC64LE:       # %bb.0: # %entry
 ; PC64LE-NEXT:    addis 3, 2, .LCPI115_0@toc@ha
-; PC64LE-NEXT:    addi 3, 3, .LCPI115_0@toc@l
-; PC64LE-NEXT:    lfiwzx 0, 0, 3
-; PC64LE-NEXT:    xxswapd 34, 0
+; PC64LE-NEXT:    lfs 1, .LCPI115_0@toc@l(3)
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_trunc_v1f32:
 ; PC64LE9:       # %bb.0: # %entry
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI115_0@toc@ha
-; PC64LE9-NEXT:    addi 3, 3, .LCPI115_0@toc@l
-; PC64LE9-NEXT:    lfiwzx 0, 0, 3
-; PC64LE9-NEXT:    xxswapd 34, 0
+; PC64LE9-NEXT:    lfs 1, .LCPI115_0@toc@l(3)
 ; PC64LE9-NEXT:    blr
 entry:
   %trunc = call <1 x float> @llvm.experimental.constrained.trunc.v1f32(