Index: llvm/lib/Target/RISCV/RISCVISelLowering.cpp
===================================================================
--- llvm/lib/Target/RISCV/RISCVISelLowering.cpp
+++ llvm/lib/Target/RISCV/RISCVISelLowering.cpp
@@ -347,7 +347,7 @@
 
   static const unsigned FPOpToExpand[] = {
       ISD::FSIN, ISD::FCOS,       ISD::FSINCOS,   ISD::FPOW,
-      ISD::FREM, ISD::FP16_TO_FP, ISD::FP_TO_FP16};
+      ISD::FREM};
 
   static const unsigned FPRndMode[] = {
       ISD::FCEIL, ISD::FFLOOR, ISD::FTRUNC, ISD::FRINT, ISD::FROUND,
@@ -422,6 +422,8 @@
     setLoadExtAction(ISD::EXTLOAD, MVT::f32, MVT::f16, Expand);
     setTruncStoreAction(MVT::f32, MVT::f16, Expand);
     setOperationAction(ISD::IS_FPCLASS, MVT::f32, Custom);
+    setOperationAction(ISD::FP_TO_FP16, MVT::f32, Custom);
+    setOperationAction(ISD::FP16_TO_FP, MVT::f32, Custom);
 
     if (Subtarget.hasStdExtZfa())
       setOperationAction(ISD::FNEARBYINT, MVT::f32, Legal);
@@ -456,6 +458,8 @@
     setLoadExtAction(ISD::EXTLOAD, MVT::f64, MVT::f16, Expand);
     setTruncStoreAction(MVT::f64, MVT::f16, Expand);
     setOperationAction(ISD::IS_FPCLASS, MVT::f64, Custom);
+    setOperationAction(ISD::FP_TO_FP16, MVT::f64, Custom);
+    setOperationAction(ISD::FP16_TO_FP, MVT::f64, Expand);
   }
 
   if (Subtarget.is64Bit()) {
@@ -4745,6 +4749,35 @@
   case ISD::FP_TO_SINT_SAT:
   case ISD::FP_TO_UINT_SAT:
     return lowerFP_TO_INT_SAT(Op, DAG, Subtarget);
+  case ISD::FP_TO_FP16: {
+    // Custom lower to ensure the libcall return is passed in an FPR on hard
+    // float ABIs.
+    assert(Subtarget.hasStdExtFOrZfinx() && "Unexpected custom legalization");
+    SDLoc DL(Op);
+    MakeLibCallOptions CallOptions;
+    RTLIB::Libcall LC =
+        RTLIB::getFPROUND(Op.getOperand(0).getValueType(), MVT::f16);
+    SDValue Res =
+        makeLibCall(DAG, LC, MVT::f32, Op.getOperand(0), CallOptions, DL).first;
+    if (Subtarget.is64Bit())
+      return DAG.getNode(RISCVISD::FMV_X_ANYEXTW_RV64, DL, MVT::i64, Res);
+    return DAG.getBitcast(MVT::i32, Res);
+  }
+  case ISD::FP16_TO_FP: {
+    // Custom lower to ensure the libcall argument is passed in an FPR on hard
+    // float ABIs.
+    assert(Subtarget.hasStdExtFOrZfinx() && "Unexpected custom legalization");
+    SDLoc DL(Op);
+    MakeLibCallOptions CallOptions;
+    SDValue Arg = Subtarget.is64Bit()
+                      ? DAG.getNode(RISCVISD::FMV_W_X_RV64, DL, MVT::f32,
+                                    Op.getOperand(0))
+                      : DAG.getBitcast(MVT::f32, Op.getOperand(0));
+    SDValue Res =
+        makeLibCall(DAG, RTLIB::FPEXT_F16_F32, MVT::f32, Arg, CallOptions, DL)
+            .first;
+    return Res;
+  }
   case ISD::FTRUNC:
   case ISD::FCEIL:
   case ISD::FFLOOR:
Index: llvm/test/CodeGen/RISCV/calling-conv-half.ll
===================================================================
--- llvm/test/CodeGen/RISCV/calling-conv-half.ll
+++ llvm/test/CodeGen/RISCV/calling-conv-half.ll
@@ -83,7 +83,6 @@
 ; RV32-ILP32F-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-ILP32F-NEXT:    sw s0, 8(sp) # 4-byte Folded Spill
 ; RV32-ILP32F-NEXT:    mv s0, a0
-; RV32-ILP32F-NEXT:    fmv.x.w a0, fa0
 ; RV32-ILP32F-NEXT:    call __extendhfsf2@plt
 ; RV32-ILP32F-NEXT:    fcvt.w.s a0, fa0, rtz
 ; RV32-ILP32F-NEXT:    add a0, s0, a0
@@ -99,6 +98,7 @@
 ; RV64-LP64F-NEXT:    sd s0, 0(sp) # 8-byte Folded Spill
 ; RV64-LP64F-NEXT:    mv s0, a0
 ; RV64-LP64F-NEXT:    fmv.x.w a0, fa0
+; RV64-LP64F-NEXT:    fmv.w.x fa0, a0
 ; RV64-LP64F-NEXT:    call __extendhfsf2@plt
 ; RV64-LP64F-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-LP64F-NEXT:    addw a0, s0, a0
@@ -292,7 +292,6 @@
 ; RV32-ILP32F-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-ILP32F-NEXT:    sw s0, 8(sp) # 4-byte Folded Spill
 ; RV32-ILP32F-NEXT:    mv s0, a7
-; RV32-ILP32F-NEXT:    fmv.x.w a0, fa0
 ; RV32-ILP32F-NEXT:    call __extendhfsf2@plt
 ; RV32-ILP32F-NEXT:    fcvt.w.s a0, fa0, rtz
 ; RV32-ILP32F-NEXT:    add a0, s0, a0
@@ -308,6 +307,7 @@
 ; RV64-LP64F-NEXT:    sd s0, 0(sp) # 8-byte Folded Spill
 ; RV64-LP64F-NEXT:    mv s0, a7
 ; RV64-LP64F-NEXT:    fmv.x.w a0, fa0
+; RV64-LP64F-NEXT:    fmv.w.x fa0, a0
 ; RV64-LP64F-NEXT:    call __extendhfsf2@plt
 ; RV64-LP64F-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-LP64F-NEXT:    addw a0, s0, a0
@@ -602,7 +602,6 @@
 ; RV32-ILP32F-NEXT:    addi sp, sp, -16
 ; RV32-ILP32F-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-ILP32F-NEXT:    call callee_half_ret@plt
-; RV32-ILP32F-NEXT:    fmv.x.w a0, fa0
 ; RV32-ILP32F-NEXT:    call __extendhfsf2@plt
 ; RV32-ILP32F-NEXT:    fcvt.w.s a0, fa0, rtz
 ; RV32-ILP32F-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
@@ -615,6 +614,7 @@
 ; RV64-LP64F-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-LP64F-NEXT:    call callee_half_ret@plt
 ; RV64-LP64F-NEXT:    fmv.x.w a0, fa0
+; RV64-LP64F-NEXT:    fmv.w.x fa0, a0
 ; RV64-LP64F-NEXT:    call __extendhfsf2@plt
 ; RV64-LP64F-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-LP64F-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
Index: llvm/test/CodeGen/RISCV/copysign-casts.ll
===================================================================
--- llvm/test/CodeGen/RISCV/copysign-casts.ll
+++ llvm/test/CodeGen/RISCV/copysign-casts.ll
@@ -163,7 +163,7 @@
 ; RV32IFD-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32IFD-NEXT:    fsd fs0, 0(sp) # 8-byte Folded Spill
 ; RV32IFD-NEXT:    fmv.d fs0, fa0
-; RV32IFD-NEXT:    fmv.x.w a0, fa1
+; RV32IFD-NEXT:    fmv.s fa0, fa1
 ; RV32IFD-NEXT:    call __extendhfsf2@plt
 ; RV32IFD-NEXT:    fcvt.d.s fa5, fa0
 ; RV32IFD-NEXT:    fsgnj.d fa0, fs0, fa5
@@ -179,6 +179,7 @@
 ; RV64IFD-NEXT:    fsd fs0, 0(sp) # 8-byte Folded Spill
 ; RV64IFD-NEXT:    fmv.d fs0, fa0
 ; RV64IFD-NEXT:    fmv.x.w a0, fa1
+; RV64IFD-NEXT:    fmv.w.x fa0, a0
 ; RV64IFD-NEXT:    call __extendhfsf2@plt
 ; RV64IFD-NEXT:    fcvt.d.s fa5, fa0
 ; RV64IFD-NEXT:    fsgnj.d fa0, fs0, fa5
@@ -264,7 +265,7 @@
 ; RV32IF-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32IF-NEXT:    fsw fs0, 8(sp) # 4-byte Folded Spill
 ; RV32IF-NEXT:    fmv.s fs0, fa0
-; RV32IF-NEXT:    fmv.x.w a0, fa1
+; RV32IF-NEXT:    fmv.s fa0, fa1
 ; RV32IF-NEXT:    call __extendhfsf2@plt
 ; RV32IF-NEXT:    fsgnj.s fa0, fs0, fa0
 ; RV32IF-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
@@ -278,7 +279,7 @@
 ; RV32IFD-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32IFD-NEXT:    fsd fs0, 0(sp) # 8-byte Folded Spill
 ; RV32IFD-NEXT:    fmv.s fs0, fa0
-; RV32IFD-NEXT:    fmv.x.w a0, fa1
+; RV32IFD-NEXT:    fmv.s fa0, fa1
 ; RV32IFD-NEXT:    call __extendhfsf2@plt
 ; RV32IFD-NEXT:    fsgnj.s fa0, fs0, fa0
 ; RV32IFD-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
@@ -293,6 +294,7 @@
 ; RV64IFD-NEXT:    fsd fs0, 0(sp) # 8-byte Folded Spill
 ; RV64IFD-NEXT:    fmv.s fs0, fa0
 ; RV64IFD-NEXT:    fmv.x.w a0, fa1
+; RV64IFD-NEXT:    fmv.w.x fa0, a0
 ; RV64IFD-NEXT:    call __extendhfsf2@plt
 ; RV64IFD-NEXT:    fsgnj.s fa0, fs0, fa0
 ; RV64IFD-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
Index: llvm/test/CodeGen/RISCV/fp16-promote.ll
===================================================================
--- llvm/test/CodeGen/RISCV/fp16-promote.ll
+++ llvm/test/CodeGen/RISCV/fp16-promote.ll
@@ -15,8 +15,14 @@
 define float @test_fpextend_float(ptr %p) nounwind {
 ; CHECK-LABEL: test_fpextend_float:
 ; CHECK:       # %bb.0:
+; CHECK-NEXT:    addi sp, sp, -16
+; CHECK-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; CHECK-NEXT:    lhu a0, 0(a0)
-; CHECK-NEXT:    tail __extendhfsf2@plt
+; CHECK-NEXT:    fmv.w.x fa0, a0
+; CHECK-NEXT:    call __extendhfsf2@plt
+; CHECK-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
+; CHECK-NEXT:    addi sp, sp, 16
+; CHECK-NEXT:    ret
   %a = load half, ptr %p
   %r = fpext half %a to float
   ret float %r
@@ -28,6 +34,7 @@
 ; CHECK-NEXT:    addi sp, sp, -16
 ; CHECK-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; CHECK-NEXT:    lhu a0, 0(a0)
+; CHECK-NEXT:    fmv.w.x fa0, a0
 ; CHECK-NEXT:    call __extendhfsf2@plt
 ; CHECK-NEXT:    fcvt.d.s fa0, fa0
 ; CHECK-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
@@ -46,6 +53,7 @@
 ; CHECK-NEXT:    sw s0, 8(sp) # 4-byte Folded Spill
 ; CHECK-NEXT:    mv s0, a0
 ; CHECK-NEXT:    call __truncsfhf2@plt
+; CHECK-NEXT:    fmv.x.w a0, fa0
 ; CHECK-NEXT:    sh a0, 0(s0)
 ; CHECK-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; CHECK-NEXT:    lw s0, 8(sp) # 4-byte Folded Reload
@@ -64,6 +72,7 @@
 ; CHECK-NEXT:    sw s0, 8(sp) # 4-byte Folded Spill
 ; CHECK-NEXT:    mv s0, a0
 ; CHECK-NEXT:    call __truncdfhf2@plt
+; CHECK-NEXT:    fmv.x.w a0, fa0
 ; CHECK-NEXT:    sh a0, 0(s0)
 ; CHECK-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; CHECK-NEXT:    lw s0, 8(sp) # 4-byte Folded Reload
@@ -80,22 +89,25 @@
 ; CHECK-NEXT:    addi sp, sp, -32
 ; CHECK-NEXT:    sw ra, 28(sp) # 4-byte Folded Spill
 ; CHECK-NEXT:    sw s0, 24(sp) # 4-byte Folded Spill
-; CHECK-NEXT:    sw s1, 20(sp) # 4-byte Folded Spill
-; CHECK-NEXT:    fsd fs0, 8(sp) # 8-byte Folded Spill
+; CHECK-NEXT:    fsd fs0, 16(sp) # 8-byte Folded Spill
+; CHECK-NEXT:    fsd fs1, 8(sp) # 8-byte Folded Spill
 ; CHECK-NEXT:    mv s0, a0
-; CHECK-NEXT:    lhu s1, 0(a0)
-; CHECK-NEXT:    lhu a0, 0(a1)
+; CHECK-NEXT:    lhu a0, 0(a0)
+; CHECK-NEXT:    lhu a1, 0(a1)
+; CHECK-NEXT:    fmv.w.x fs0, a0
+; CHECK-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NEXT:    call __extendhfsf2@plt
-; CHECK-NEXT:    fmv.s fs0, fa0
-; CHECK-NEXT:    mv a0, s1
+; CHECK-NEXT:    fmv.s fs1, fa0
+; CHECK-NEXT:    fmv.s fa0, fs0
 ; CHECK-NEXT:    call __extendhfsf2@plt
-; CHECK-NEXT:    fadd.s fa0, fa0, fs0
+; CHECK-NEXT:    fadd.s fa0, fa0, fs1
 ; CHECK-NEXT:    call __truncsfhf2@plt
+; CHECK-NEXT:    fmv.x.w a0, fa0
 ; CHECK-NEXT:    sh a0, 0(s0)
 ; CHECK-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; CHECK-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
-; CHECK-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
-; CHECK-NEXT:    fld fs0, 8(sp) # 8-byte Folded Reload
+; CHECK-NEXT:    fld fs0, 16(sp) # 8-byte Folded Reload
+; CHECK-NEXT:    fld fs1, 8(sp) # 8-byte Folded Reload
 ; CHECK-NEXT:    addi sp, sp, 32
 ; CHECK-NEXT:    ret
   %a = load half, ptr %p
@@ -111,22 +123,25 @@
 ; CHECK-NEXT:    addi sp, sp, -32
 ; CHECK-NEXT:    sw ra, 28(sp) # 4-byte Folded Spill
 ; CHECK-NEXT:    sw s0, 24(sp) # 4-byte Folded Spill
-; CHECK-NEXT:    sw s1, 20(sp) # 4-byte Folded Spill
-; CHECK-NEXT:    fsd fs0, 8(sp) # 8-byte Folded Spill
+; CHECK-NEXT:    fsd fs0, 16(sp) # 8-byte Folded Spill
+; CHECK-NEXT:    fsd fs1, 8(sp) # 8-byte Folded Spill
 ; CHECK-NEXT:    mv s0, a0
-; CHECK-NEXT:    lhu s1, 0(a0)
-; CHECK-NEXT:    lhu a0, 0(a1)
+; CHECK-NEXT:    lhu a0, 0(a0)
+; CHECK-NEXT:    lhu a1, 0(a1)
+; CHECK-NEXT:    fmv.w.x fs0, a0
+; CHECK-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NEXT:    call __extendhfsf2@plt
-; CHECK-NEXT:    fmv.s fs0, fa0
-; CHECK-NEXT:    mv a0, s1
+; CHECK-NEXT:    fmv.s fs1, fa0
+; CHECK-NEXT:    fmv.s fa0, fs0
 ; CHECK-NEXT:    call __extendhfsf2@plt
-; CHECK-NEXT:    fmul.s fa0, fa0, fs0
+; CHECK-NEXT:    fmul.s fa0, fa0, fs1
 ; CHECK-NEXT:    call __truncsfhf2@plt
+; CHECK-NEXT:    fmv.x.w a0, fa0
 ; CHECK-NEXT:    sh a0, 0(s0)
 ; CHECK-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; CHECK-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
-; CHECK-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
-; CHECK-NEXT:    fld fs0, 8(sp) # 8-byte Folded Reload
+; CHECK-NEXT:    fld fs0, 16(sp) # 8-byte Folded Reload
+; CHECK-NEXT:    fld fs1, 8(sp) # 8-byte Folded Reload
 ; CHECK-NEXT:    addi sp, sp, 32
 ; CHECK-NEXT:    ret
   %a = load half, ptr %p
Index: llvm/test/CodeGen/RISCV/fpclamptosat.ll
===================================================================
--- llvm/test/CodeGen/RISCV/fpclamptosat.ll
+++ llvm/test/CodeGen/RISCV/fpclamptosat.ll
@@ -355,7 +355,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    call __fixsfdi@plt
 ; RV32-NEXT:    lui a2, 524288
@@ -397,6 +396,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-NEXT:    lui a1, 524288
@@ -429,7 +429,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    call __fixunssfdi@plt
 ; RV32-NEXT:    sltiu a2, a0, -1
@@ -448,6 +447,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; RV64-NEXT:    li a1, -1
@@ -474,7 +474,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    call __fixsfdi@plt
 ; RV32-NEXT:    beqz a1, .LBB8_2
@@ -508,6 +507,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-NEXT:    li a1, -1
@@ -891,7 +891,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    fcvt.w.s a0, fa0, rtz
 ; RV32-NEXT:    lui a1, 8
@@ -916,6 +915,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-NEXT:    lui a1, 8
@@ -949,7 +949,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    fcvt.wu.s a0, fa0, rtz
 ; RV32-NEXT:    lui a1, 16
@@ -969,6 +968,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; RV64-NEXT:    lui a1, 16
@@ -995,7 +995,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    fcvt.w.s a0, fa0, rtz
 ; RV32-NEXT:    lui a1, 16
@@ -1018,6 +1017,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-NEXT:    lui a1, 16
@@ -1666,7 +1666,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 32
 ; RV32-NEXT:    sw ra, 28(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    addi a0, sp, 8
 ; RV32-NEXT:    call __fixsfti@plt
@@ -1729,6 +1728,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    call __fixsfti@plt
 ; RV64-NEXT:    li a2, -1
@@ -1779,7 +1779,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 32
 ; RV32-NEXT:    sw ra, 28(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    addi a0, sp, 8
 ; RV32-NEXT:    call __fixunssfti@plt
@@ -1808,6 +1807,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    call __fixunssfti@plt
 ; RV64-NEXT:    snez a1, a1
@@ -1831,7 +1831,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 32
 ; RV32-NEXT:    sw ra, 28(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    addi a0, sp, 8
 ; RV32-NEXT:    call __fixsfti@plt
@@ -1886,6 +1885,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    call __fixsfti@plt
 ; RV64-NEXT:    slti a2, a1, 1
@@ -2259,7 +2259,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    call __fixsfdi@plt
 ; RV32-NEXT:    lui a2, 524288
@@ -2311,6 +2310,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-NEXT:    lui a1, 524288
@@ -2341,7 +2341,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    call __fixunssfdi@plt
 ; RV32-NEXT:    seqz a1, a1
@@ -2358,6 +2357,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; RV64-NEXT:    li a1, -1
@@ -2383,7 +2383,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    call __fixsfdi@plt
 ; RV32-NEXT:    slti a2, a1, 0
@@ -2411,6 +2410,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-NEXT:    li a1, -1
@@ -2782,7 +2782,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    fcvt.w.s a0, fa0, rtz
 ; RV32-NEXT:    lui a1, 8
@@ -2807,6 +2806,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-NEXT:    lui a1, 8
@@ -2838,7 +2838,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    fcvt.wu.s a0, fa0, rtz
 ; RV32-NEXT:    lui a1, 16
@@ -2858,6 +2857,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; RV64-NEXT:    sext.w a0, a0
@@ -2884,7 +2884,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    fcvt.w.s a0, fa0, rtz
 ; RV32-NEXT:    lui a1, 16
@@ -2907,6 +2906,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-NEXT:    lui a1, 16
@@ -3708,7 +3708,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 32
 ; RV32-NEXT:    sw ra, 28(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    addi a0, sp, 8
 ; RV32-NEXT:    call __fixsfti@plt
@@ -3801,6 +3800,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    call __fixsfti@plt
 ; RV64-NEXT:    li a2, -1
@@ -3858,7 +3858,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 32
 ; RV32-NEXT:    sw ra, 28(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    addi a0, sp, 8
 ; RV32-NEXT:    call __fixunssfti@plt
@@ -3888,6 +3887,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    call __fixunssfti@plt
 ; RV64-NEXT:    snez a2, a1
@@ -3914,7 +3914,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 32
 ; RV32-NEXT:    sw ra, 28(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    addi a0, sp, 8
 ; RV32-NEXT:    call __fixsfti@plt
@@ -3986,6 +3985,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    call __fixsfti@plt
 ; RV64-NEXT:    mv a2, a1
Index: llvm/test/CodeGen/RISCV/half-fpext.ll
===================================================================
--- llvm/test/CodeGen/RISCV/half-fpext.ll
+++ llvm/test/CodeGen/RISCV/half-fpext.ll
@@ -31,7 +31,12 @@
 ;
 ; RV32ID-ILP32-LABEL: half_to_float:
 ; RV32ID-ILP32:       # %bb.0:
-; RV32ID-ILP32-NEXT:    tail __extendhfsf2@plt
+; RV32ID-ILP32-NEXT:    addi sp, sp, -16
+; RV32ID-ILP32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
+; RV32ID-ILP32-NEXT:    call __extendhfsf2@plt
+; RV32ID-ILP32-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
+; RV32ID-ILP32-NEXT:    addi sp, sp, 16
+; RV32ID-ILP32-NEXT:    ret
 ;
 ; RV64ID-LP64-LABEL: half_to_float:
 ; RV64ID-LP64:       # %bb.0:
@@ -44,13 +49,23 @@
 ;
 ; RV32ID-ILP32D-LABEL: half_to_float:
 ; RV32ID-ILP32D:       # %bb.0:
-; RV32ID-ILP32D-NEXT:    fmv.x.w a0, fa0
-; RV32ID-ILP32D-NEXT:    tail __extendhfsf2@plt
+; RV32ID-ILP32D-NEXT:    addi sp, sp, -16
+; RV32ID-ILP32D-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
+; RV32ID-ILP32D-NEXT:    call __extendhfsf2@plt
+; RV32ID-ILP32D-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
+; RV32ID-ILP32D-NEXT:    addi sp, sp, 16
+; RV32ID-ILP32D-NEXT:    ret
 ;
 ; RV64ID-LP64D-LABEL: half_to_float:
 ; RV64ID-LP64D:       # %bb.0:
+; RV64ID-LP64D-NEXT:    addi sp, sp, -16
+; RV64ID-LP64D-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64ID-LP64D-NEXT:    fmv.x.w a0, fa0
-; RV64ID-LP64D-NEXT:    tail __extendhfsf2@plt
+; RV64ID-LP64D-NEXT:    fmv.w.x fa0, a0
+; RV64ID-LP64D-NEXT:    call __extendhfsf2@plt
+; RV64ID-LP64D-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
+; RV64ID-LP64D-NEXT:    addi sp, sp, 16
+; RV64ID-LP64D-NEXT:    ret
   %1 = fpext half %a to float
   ret float %1
 }
@@ -112,7 +127,6 @@
 ; RV32ID-ILP32D:       # %bb.0:
 ; RV32ID-ILP32D-NEXT:    addi sp, sp, -16
 ; RV32ID-ILP32D-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
-; RV32ID-ILP32D-NEXT:    fmv.x.w a0, fa0
 ; RV32ID-ILP32D-NEXT:    call __extendhfsf2@plt
 ; RV32ID-ILP32D-NEXT:    fcvt.d.s fa0, fa0
 ; RV32ID-ILP32D-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
@@ -124,6 +138,7 @@
 ; RV64ID-LP64D-NEXT:    addi sp, sp, -16
 ; RV64ID-LP64D-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64ID-LP64D-NEXT:    fmv.x.w a0, fa0
+; RV64ID-LP64D-NEXT:    fmv.w.x fa0, a0
 ; RV64ID-LP64D-NEXT:    call __extendhfsf2@plt
 ; RV64ID-LP64D-NEXT:    fcvt.d.s fa0, fa0
 ; RV64ID-LP64D-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
Index: llvm/test/CodeGen/RISCV/half-fptrunc.ll
===================================================================
--- llvm/test/CodeGen/RISCV/half-fptrunc.ll
+++ llvm/test/CodeGen/RISCV/half-fptrunc.ll
@@ -52,6 +52,7 @@
 ; RV32ID-ILP32D-NEXT:    addi sp, sp, -16
 ; RV32ID-ILP32D-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32ID-ILP32D-NEXT:    call __truncsfhf2@plt
+; RV32ID-ILP32D-NEXT:    fmv.x.w a0, fa0
 ; RV32ID-ILP32D-NEXT:    lui a1, 1048560
 ; RV32ID-ILP32D-NEXT:    or a0, a0, a1
 ; RV32ID-ILP32D-NEXT:    fmv.w.x fa0, a0
@@ -64,6 +65,7 @@
 ; RV64ID-LP64D-NEXT:    addi sp, sp, -16
 ; RV64ID-LP64D-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64ID-LP64D-NEXT:    call __truncsfhf2@plt
+; RV64ID-LP64D-NEXT:    fmv.x.w a0, fa0
 ; RV64ID-LP64D-NEXT:    lui a1, 1048560
 ; RV64ID-LP64D-NEXT:    or a0, a0, a1
 ; RV64ID-LP64D-NEXT:    fmv.w.x fa0, a0
@@ -120,6 +122,7 @@
 ; RV32ID-ILP32D-NEXT:    addi sp, sp, -16
 ; RV32ID-ILP32D-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32ID-ILP32D-NEXT:    call __truncdfhf2@plt
+; RV32ID-ILP32D-NEXT:    fmv.x.w a0, fa0
 ; RV32ID-ILP32D-NEXT:    lui a1, 1048560
 ; RV32ID-ILP32D-NEXT:    or a0, a0, a1
 ; RV32ID-ILP32D-NEXT:    fmv.w.x fa0, a0
@@ -132,6 +135,7 @@
 ; RV64ID-LP64D-NEXT:    addi sp, sp, -16
 ; RV64ID-LP64D-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64ID-LP64D-NEXT:    call __truncdfhf2@plt
+; RV64ID-LP64D-NEXT:    fmv.x.w a0, fa0
 ; RV64ID-LP64D-NEXT:    lui a1, 1048560
 ; RV64ID-LP64D-NEXT:    or a0, a0, a1
 ; RV64ID-LP64D-NEXT:    fmv.w.x fa0, a0
Index: llvm/test/CodeGen/RISCV/libcall-tail-calls.ll
===================================================================
--- llvm/test/CodeGen/RISCV/libcall-tail-calls.ll
+++ llvm/test/CodeGen/RISCV/libcall-tail-calls.ll
@@ -120,10 +120,10 @@
 ; RV32IFD-ILP32D:       # %bb.0:
 ; RV32IFD-ILP32D-NEXT:    addi sp, sp, -16
 ; RV32IFD-ILP32D-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
-; RV32IFD-ILP32D-NEXT:    fmv.x.w a0, fa0
 ; RV32IFD-ILP32D-NEXT:    call __extendhfsf2@plt
 ; RV32IFD-ILP32D-NEXT:    call sinf@plt
 ; RV32IFD-ILP32D-NEXT:    call __truncsfhf2@plt
+; RV32IFD-ILP32D-NEXT:    fmv.x.w a0, fa0
 ; RV32IFD-ILP32D-NEXT:    lui a1, 1048560
 ; RV32IFD-ILP32D-NEXT:    or a0, a0, a1
 ; RV32IFD-ILP32D-NEXT:    fmv.w.x fa0, a0
@@ -135,10 +135,10 @@
 ; RV32IF-ILP32F:       # %bb.0:
 ; RV32IF-ILP32F-NEXT:    addi sp, sp, -16
 ; RV32IF-ILP32F-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
-; RV32IF-ILP32F-NEXT:    fmv.x.w a0, fa0
 ; RV32IF-ILP32F-NEXT:    call __extendhfsf2@plt
 ; RV32IF-ILP32F-NEXT:    call sinf@plt
 ; RV32IF-ILP32F-NEXT:    call __truncsfhf2@plt
+; RV32IF-ILP32F-NEXT:    fmv.x.w a0, fa0
 ; RV32IF-ILP32F-NEXT:    lui a1, 1048560
 ; RV32IF-ILP32F-NEXT:    or a0, a0, a1
 ; RV32IF-ILP32F-NEXT:    fmv.w.x fa0, a0
@@ -177,9 +177,11 @@
 ; RV64IFD-LP64D-NEXT:    addi sp, sp, -16
 ; RV64IFD-LP64D-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64IFD-LP64D-NEXT:    fmv.x.w a0, fa0
+; RV64IFD-LP64D-NEXT:    fmv.w.x fa0, a0
 ; RV64IFD-LP64D-NEXT:    call __extendhfsf2@plt
 ; RV64IFD-LP64D-NEXT:    call sinf@plt
 ; RV64IFD-LP64D-NEXT:    call __truncsfhf2@plt
+; RV64IFD-LP64D-NEXT:    fmv.x.w a0, fa0
 ; RV64IFD-LP64D-NEXT:    lui a1, 1048560
 ; RV64IFD-LP64D-NEXT:    or a0, a0, a1
 ; RV64IFD-LP64D-NEXT:    fmv.w.x fa0, a0
@@ -192,9 +194,11 @@
 ; RV64IF-LP64F-NEXT:    addi sp, sp, -16
 ; RV64IF-LP64F-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64IF-LP64F-NEXT:    fmv.x.w a0, fa0
+; RV64IF-LP64F-NEXT:    fmv.w.x fa0, a0
 ; RV64IF-LP64F-NEXT:    call __extendhfsf2@plt
 ; RV64IF-LP64F-NEXT:    call sinf@plt
 ; RV64IF-LP64F-NEXT:    call __truncsfhf2@plt
+; RV64IF-LP64F-NEXT:    fmv.x.w a0, fa0
 ; RV64IF-LP64F-NEXT:    lui a1, 1048560
 ; RV64IF-LP64F-NEXT:    or a0, a0, a1
 ; RV64IF-LP64F-NEXT:    fmv.w.x fa0, a0
Index: llvm/test/CodeGen/RISCV/rvv/fpclamptosat_vec.ll
===================================================================
--- llvm/test/CodeGen/RISCV/rvv/fpclamptosat_vec.ll
+++ llvm/test/CodeGen/RISCV/rvv/fpclamptosat_vec.ll
@@ -369,17 +369,17 @@
 ; CHECK-NOV-NEXT:    lhu s3, 8(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 16(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.l.s s2, fs2, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.l.s s1, fs2, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    lui a1, 524288
@@ -387,7 +387,7 @@
 ; CHECK-NOV-NEXT:    bge a0, a4, .LBB6_10
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs1, rtz
-; CHECK-NOV-NEXT:    bge s2, a4, .LBB6_11
+; CHECK-NOV-NEXT:    bge s1, a4, .LBB6_11
 ; CHECK-NOV-NEXT:  .LBB6_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    bge a2, a4, .LBB6_12
@@ -398,14 +398,14 @@
 ; CHECK-NOV-NEXT:  .LBB6_5: # %entry
 ; CHECK-NOV-NEXT:    bge a1, a2, .LBB6_15
 ; CHECK-NOV-NEXT:  .LBB6_6: # %entry
-; CHECK-NOV-NEXT:    bge a1, s2, .LBB6_16
+; CHECK-NOV-NEXT:    bge a1, s1, .LBB6_16
 ; CHECK-NOV-NEXT:  .LBB6_7: # %entry
 ; CHECK-NOV-NEXT:    blt a1, a0, .LBB6_9
 ; CHECK-NOV-NEXT:  .LBB6_8: # %entry
 ; CHECK-NOV-NEXT:    lui a0, 524288
 ; CHECK-NOV-NEXT:  .LBB6_9: # %entry
 ; CHECK-NOV-NEXT:    sw a0, 12(s0)
-; CHECK-NOV-NEXT:    sw s2, 8(s0)
+; CHECK-NOV-NEXT:    sw s1, 8(s0)
 ; CHECK-NOV-NEXT:    sw a2, 4(s0)
 ; CHECK-NOV-NEXT:    sw a3, 0(s0)
 ; CHECK-NOV-NEXT:    ld ra, 56(sp) # 8-byte Folded Reload
@@ -421,9 +421,9 @@
 ; CHECK-NOV-NEXT:  .LBB6_10: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a4
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs1, rtz
-; CHECK-NOV-NEXT:    blt s2, a4, .LBB6_2
+; CHECK-NOV-NEXT:    blt s1, a4, .LBB6_2
 ; CHECK-NOV-NEXT:  .LBB6_11: # %entry
-; CHECK-NOV-NEXT:    mv s2, a4
+; CHECK-NOV-NEXT:    mv s1, a4
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    blt a2, a4, .LBB6_3
 ; CHECK-NOV-NEXT:  .LBB6_12: # %entry
@@ -437,9 +437,9 @@
 ; CHECK-NOV-NEXT:    blt a1, a2, .LBB6_6
 ; CHECK-NOV-NEXT:  .LBB6_15: # %entry
 ; CHECK-NOV-NEXT:    lui a2, 524288
-; CHECK-NOV-NEXT:    blt a1, s2, .LBB6_7
+; CHECK-NOV-NEXT:    blt a1, s1, .LBB6_7
 ; CHECK-NOV-NEXT:  .LBB6_16: # %entry
-; CHECK-NOV-NEXT:    lui s2, 524288
+; CHECK-NOV-NEXT:    lui s1, 524288
 ; CHECK-NOV-NEXT:    bge a1, a0, .LBB6_8
 ; CHECK-NOV-NEXT:    j .LBB6_9
 ;
@@ -463,13 +463,14 @@
 ; CHECK-V-NEXT:    lhu s1, 16(a0)
 ; CHECK-V-NEXT:    lhu s2, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
@@ -479,7 +480,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
@@ -489,7 +490,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
@@ -548,17 +549,17 @@
 ; CHECK-NOV-NEXT:    lhu s3, 16(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 8(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.lu.s s2, fs2, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.lu.s s1, fs2, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    li a1, -1
@@ -566,7 +567,7 @@
 ; CHECK-NOV-NEXT:    bgeu a0, a1, .LBB7_6
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a2, fs1, rtz
-; CHECK-NOV-NEXT:    bgeu s2, a1, .LBB7_7
+; CHECK-NOV-NEXT:    bgeu s1, a1, .LBB7_7
 ; CHECK-NOV-NEXT:  .LBB7_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    bgeu a2, a1, .LBB7_8
@@ -577,7 +578,7 @@
 ; CHECK-NOV-NEXT:  .LBB7_5: # %entry
 ; CHECK-NOV-NEXT:    sw a3, 12(s0)
 ; CHECK-NOV-NEXT:    sw a2, 8(s0)
-; CHECK-NOV-NEXT:    sw s2, 4(s0)
+; CHECK-NOV-NEXT:    sw s1, 4(s0)
 ; CHECK-NOV-NEXT:    sw a0, 0(s0)
 ; CHECK-NOV-NEXT:    ld ra, 56(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s0, 48(sp) # 8-byte Folded Reload
@@ -592,9 +593,9 @@
 ; CHECK-NOV-NEXT:  .LBB7_6: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a1
 ; CHECK-NOV-NEXT:    fcvt.lu.s a2, fs1, rtz
-; CHECK-NOV-NEXT:    bltu s2, a1, .LBB7_2
+; CHECK-NOV-NEXT:    bltu s1, a1, .LBB7_2
 ; CHECK-NOV-NEXT:  .LBB7_7: # %entry
-; CHECK-NOV-NEXT:    mv s2, a1
+; CHECK-NOV-NEXT:    mv s1, a1
 ; CHECK-NOV-NEXT:    fcvt.lu.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    bltu a2, a1, .LBB7_3
 ; CHECK-NOV-NEXT:  .LBB7_8: # %entry
@@ -622,13 +623,14 @@
 ; CHECK-V-NEXT:    lhu s1, 16(a0)
 ; CHECK-V-NEXT:    lhu s2, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
@@ -638,7 +640,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
@@ -648,7 +650,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
@@ -704,17 +706,17 @@
 ; CHECK-NOV-NEXT:    lhu s3, 16(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 8(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.l.s s2, fs2, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.l.s s1, fs2, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    li a2, -1
@@ -722,7 +724,7 @@
 ; CHECK-NOV-NEXT:    bge a0, a2, .LBB8_6
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs1, rtz
-; CHECK-NOV-NEXT:    bge s2, a2, .LBB8_7
+; CHECK-NOV-NEXT:    bge s1, a2, .LBB8_7
 ; CHECK-NOV-NEXT:  .LBB8_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    bge a1, a2, .LBB8_8
@@ -733,12 +735,12 @@
 ; CHECK-NOV-NEXT:  .LBB8_5: # %entry
 ; CHECK-NOV-NEXT:    sgtz a2, a3
 ; CHECK-NOV-NEXT:    sgtz a4, a1
-; CHECK-NOV-NEXT:    sgtz a5, s2
+; CHECK-NOV-NEXT:    sgtz a5, s1
 ; CHECK-NOV-NEXT:    sgtz a6, a0
 ; CHECK-NOV-NEXT:    negw a6, a6
 ; CHECK-NOV-NEXT:    and a0, a6, a0
 ; CHECK-NOV-NEXT:    negw a5, a5
-; CHECK-NOV-NEXT:    and a5, a5, s2
+; CHECK-NOV-NEXT:    and a5, a5, s1
 ; CHECK-NOV-NEXT:    negw a4, a4
 ; CHECK-NOV-NEXT:    and a1, a4, a1
 ; CHECK-NOV-NEXT:    negw a2, a2
@@ -760,9 +762,9 @@
 ; CHECK-NOV-NEXT:  .LBB8_6: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a2
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs1, rtz
-; CHECK-NOV-NEXT:    blt s2, a2, .LBB8_2
+; CHECK-NOV-NEXT:    blt s1, a2, .LBB8_2
 ; CHECK-NOV-NEXT:  .LBB8_7: # %entry
-; CHECK-NOV-NEXT:    mv s2, a2
+; CHECK-NOV-NEXT:    mv s1, a2
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    blt a1, a2, .LBB8_3
 ; CHECK-NOV-NEXT:  .LBB8_8: # %entry
@@ -790,13 +792,14 @@
 ; CHECK-V-NEXT:    lhu s1, 16(a0)
 ; CHECK-V-NEXT:    lhu s2, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
@@ -806,7 +809,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
@@ -816,7 +819,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
@@ -1237,29 +1240,29 @@
 ; CHECK-NOV-NEXT:    lhu s7, 40(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 48(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs6, fa0
-; CHECK-NOV-NEXT:    mv a0, s7
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s7
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs5, fa0
-; CHECK-NOV-NEXT:    mv a0, s6
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s6
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs4, fa0
-; CHECK-NOV-NEXT:    mv a0, s5
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s5
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs3, fa0
-; CHECK-NOV-NEXT:    mv a0, s4
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s4
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.l.s s2, fs6, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.l.s s1, fs6, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    lui a7, 8
@@ -1267,7 +1270,7 @@
 ; CHECK-NOV-NEXT:    bge a0, a7, .LBB15_18
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs5, rtz
-; CHECK-NOV-NEXT:    bge s2, a7, .LBB15_19
+; CHECK-NOV-NEXT:    bge s1, a7, .LBB15_19
 ; CHECK-NOV-NEXT:  .LBB15_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs4, rtz
 ; CHECK-NOV-NEXT:    bge a1, a7, .LBB15_20
@@ -1299,14 +1302,14 @@
 ; CHECK-NOV-NEXT:  .LBB15_13: # %entry
 ; CHECK-NOV-NEXT:    bge a7, a1, .LBB15_31
 ; CHECK-NOV-NEXT:  .LBB15_14: # %entry
-; CHECK-NOV-NEXT:    bge a7, s2, .LBB15_32
+; CHECK-NOV-NEXT:    bge a7, s1, .LBB15_32
 ; CHECK-NOV-NEXT:  .LBB15_15: # %entry
 ; CHECK-NOV-NEXT:    blt a7, a0, .LBB15_17
 ; CHECK-NOV-NEXT:  .LBB15_16: # %entry
 ; CHECK-NOV-NEXT:    lui a0, 1048568
 ; CHECK-NOV-NEXT:  .LBB15_17: # %entry
 ; CHECK-NOV-NEXT:    sh a0, 14(s0)
-; CHECK-NOV-NEXT:    sh s2, 12(s0)
+; CHECK-NOV-NEXT:    sh s1, 12(s0)
 ; CHECK-NOV-NEXT:    sh a1, 10(s0)
 ; CHECK-NOV-NEXT:    sh a2, 8(s0)
 ; CHECK-NOV-NEXT:    sh a3, 6(s0)
@@ -1334,9 +1337,9 @@
 ; CHECK-NOV-NEXT:  .LBB15_18: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a7
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs5, rtz
-; CHECK-NOV-NEXT:    blt s2, a7, .LBB15_2
+; CHECK-NOV-NEXT:    blt s1, a7, .LBB15_2
 ; CHECK-NOV-NEXT:  .LBB15_19: # %entry
-; CHECK-NOV-NEXT:    mv s2, a7
+; CHECK-NOV-NEXT:    mv s1, a7
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs4, rtz
 ; CHECK-NOV-NEXT:    blt a1, a7, .LBB15_3
 ; CHECK-NOV-NEXT:  .LBB15_20: # %entry
@@ -1379,9 +1382,9 @@
 ; CHECK-NOV-NEXT:    blt a7, a1, .LBB15_14
 ; CHECK-NOV-NEXT:  .LBB15_31: # %entry
 ; CHECK-NOV-NEXT:    lui a1, 1048568
-; CHECK-NOV-NEXT:    blt a7, s2, .LBB15_15
+; CHECK-NOV-NEXT:    blt a7, s1, .LBB15_15
 ; CHECK-NOV-NEXT:  .LBB15_32: # %entry
-; CHECK-NOV-NEXT:    lui s2, 1048568
+; CHECK-NOV-NEXT:    lui s1, 1048568
 ; CHECK-NOV-NEXT:    bge a7, a0, .LBB15_16
 ; CHECK-NOV-NEXT:    j .LBB15_17
 ;
@@ -1417,13 +1420,14 @@
 ; CHECK-V-NEXT:    lhu s5, 16(a0)
 ; CHECK-V-NEXT:    lhu s6, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s6
+; CHECK-V-NEXT:    fmv.w.x fa0, s6
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -1433,7 +1437,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s5
+; CHECK-V-NEXT:    fmv.w.x fa0, s5
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -1443,7 +1447,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s4
+; CHECK-V-NEXT:    fmv.w.x fa0, s4
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -1453,7 +1457,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s3
+; CHECK-V-NEXT:    fmv.w.x fa0, s3
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -1463,7 +1467,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 4
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -1473,7 +1477,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 5
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -1483,7 +1487,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 6
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
@@ -1567,29 +1571,29 @@
 ; CHECK-NOV-NEXT:    lhu s7, 16(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 8(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs6, fa0
-; CHECK-NOV-NEXT:    mv a0, s7
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s7
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs5, fa0
-; CHECK-NOV-NEXT:    mv a0, s6
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s6
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs4, fa0
-; CHECK-NOV-NEXT:    mv a0, s5
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s5
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs3, fa0
-; CHECK-NOV-NEXT:    mv a0, s4
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s4
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.lu.s s2, fs6, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.lu.s s1, fs6, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    lui a1, 16
@@ -1597,7 +1601,7 @@
 ; CHECK-NOV-NEXT:    bgeu a0, a1, .LBB16_10
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a2, fs5, rtz
-; CHECK-NOV-NEXT:    bgeu s2, a1, .LBB16_11
+; CHECK-NOV-NEXT:    bgeu s1, a1, .LBB16_11
 ; CHECK-NOV-NEXT:  .LBB16_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a3, fs4, rtz
 ; CHECK-NOV-NEXT:    bgeu a2, a1, .LBB16_12
@@ -1624,7 +1628,7 @@
 ; CHECK-NOV-NEXT:    sh a4, 8(s0)
 ; CHECK-NOV-NEXT:    sh a3, 6(s0)
 ; CHECK-NOV-NEXT:    sh a2, 4(s0)
-; CHECK-NOV-NEXT:    sh s2, 2(s0)
+; CHECK-NOV-NEXT:    sh s1, 2(s0)
 ; CHECK-NOV-NEXT:    sh a0, 0(s0)
 ; CHECK-NOV-NEXT:    ld ra, 120(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s0, 112(sp) # 8-byte Folded Reload
@@ -1647,9 +1651,9 @@
 ; CHECK-NOV-NEXT:  .LBB16_10: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a1
 ; CHECK-NOV-NEXT:    fcvt.lu.s a2, fs5, rtz
-; CHECK-NOV-NEXT:    bltu s2, a1, .LBB16_2
+; CHECK-NOV-NEXT:    bltu s1, a1, .LBB16_2
 ; CHECK-NOV-NEXT:  .LBB16_11: # %entry
-; CHECK-NOV-NEXT:    mv s2, a1
+; CHECK-NOV-NEXT:    mv s1, a1
 ; CHECK-NOV-NEXT:    fcvt.lu.s a3, fs4, rtz
 ; CHECK-NOV-NEXT:    bltu a2, a1, .LBB16_3
 ; CHECK-NOV-NEXT:  .LBB16_12: # %entry
@@ -1705,13 +1709,14 @@
 ; CHECK-V-NEXT:    lhu s5, 16(a0)
 ; CHECK-V-NEXT:    lhu s6, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s6
+; CHECK-V-NEXT:    fmv.w.x fa0, s6
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -1721,7 +1726,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s5
+; CHECK-V-NEXT:    fmv.w.x fa0, s5
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -1731,7 +1736,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s4
+; CHECK-V-NEXT:    fmv.w.x fa0, s4
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -1741,7 +1746,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s3
+; CHECK-V-NEXT:    fmv.w.x fa0, s3
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -1751,7 +1756,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 4
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -1761,7 +1766,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 5
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -1771,7 +1776,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 6
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
@@ -1851,29 +1856,29 @@
 ; CHECK-NOV-NEXT:    lhu s7, 16(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 8(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs6, fa0
-; CHECK-NOV-NEXT:    mv a0, s7
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s7
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs5, fa0
-; CHECK-NOV-NEXT:    mv a0, s6
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s6
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs4, fa0
-; CHECK-NOV-NEXT:    mv a0, s5
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s5
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs3, fa0
-; CHECK-NOV-NEXT:    mv a0, s4
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s4
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.l.s s2, fs6, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.l.s s1, fs6, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    lui a3, 16
@@ -1881,7 +1886,7 @@
 ; CHECK-NOV-NEXT:    bge a0, a3, .LBB17_10
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs5, rtz
-; CHECK-NOV-NEXT:    bge s2, a3, .LBB17_11
+; CHECK-NOV-NEXT:    bge s1, a3, .LBB17_11
 ; CHECK-NOV-NEXT:  .LBB17_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs4, rtz
 ; CHECK-NOV-NEXT:    bge a1, a3, .LBB17_12
@@ -1908,12 +1913,12 @@
 ; CHECK-NOV-NEXT:    sgtz t2, a4
 ; CHECK-NOV-NEXT:    sgtz t3, a2
 ; CHECK-NOV-NEXT:    sgtz t4, a1
-; CHECK-NOV-NEXT:    sgtz t5, s2
+; CHECK-NOV-NEXT:    sgtz t5, s1
 ; CHECK-NOV-NEXT:    sgtz t6, a0
 ; CHECK-NOV-NEXT:    negw t6, t6
 ; CHECK-NOV-NEXT:    and a0, t6, a0
 ; CHECK-NOV-NEXT:    negw t5, t5
-; CHECK-NOV-NEXT:    and t5, t5, s2
+; CHECK-NOV-NEXT:    and t5, t5, s1
 ; CHECK-NOV-NEXT:    negw t4, t4
 ; CHECK-NOV-NEXT:    and a1, t4, a1
 ; CHECK-NOV-NEXT:    negw t3, t3
@@ -1955,9 +1960,9 @@
 ; CHECK-NOV-NEXT:  .LBB17_10: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a3
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs5, rtz
-; CHECK-NOV-NEXT:    blt s2, a3, .LBB17_2
+; CHECK-NOV-NEXT:    blt s1, a3, .LBB17_2
 ; CHECK-NOV-NEXT:  .LBB17_11: # %entry
-; CHECK-NOV-NEXT:    mv s2, a3
+; CHECK-NOV-NEXT:    mv s1, a3
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs4, rtz
 ; CHECK-NOV-NEXT:    blt a1, a3, .LBB17_3
 ; CHECK-NOV-NEXT:  .LBB17_12: # %entry
@@ -2013,13 +2018,14 @@
 ; CHECK-V-NEXT:    lhu s5, 16(a0)
 ; CHECK-V-NEXT:    lhu s6, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s6
+; CHECK-V-NEXT:    fmv.w.x fa0, s6
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -2029,7 +2035,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s5
+; CHECK-V-NEXT:    fmv.w.x fa0, s5
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -2039,7 +2045,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s4
+; CHECK-V-NEXT:    fmv.w.x fa0, s4
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -2049,7 +2055,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s3
+; CHECK-V-NEXT:    fmv.w.x fa0, s3
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -2059,7 +2065,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 4
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -2069,7 +2075,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 5
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -2079,7 +2085,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 6
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
@@ -2948,11 +2954,12 @@
 ; CHECK-NOV-NEXT:    .cfi_offset s1, -24
 ; CHECK-NOV-NEXT:    .cfi_offset s2, -32
 ; CHECK-NOV-NEXT:    mv s2, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a0
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixsfti@plt
 ; CHECK-NOV-NEXT:    mv s0, a0
 ; CHECK-NOV-NEXT:    mv s1, a1
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixsfti@plt
 ; CHECK-NOV-NEXT:    mv a2, a0
@@ -3029,11 +3036,12 @@
 ; CHECK-V-NEXT:    .cfi_offset s1, -24
 ; CHECK-V-NEXT:    .cfi_offset s2, -32
 ; CHECK-V-NEXT:    mv s2, a1
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixsfti@plt
 ; CHECK-V-NEXT:    mv s0, a0
 ; CHECK-V-NEXT:    mv s1, a1
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixsfti@plt
 ; CHECK-V-NEXT:    li a2, -1
@@ -3121,11 +3129,12 @@
 ; CHECK-NOV-NEXT:    .cfi_offset s1, -24
 ; CHECK-NOV-NEXT:    .cfi_offset s2, -32
 ; CHECK-NOV-NEXT:    mv s0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a0
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixunssfti@plt
 ; CHECK-NOV-NEXT:    mv s1, a0
 ; CHECK-NOV-NEXT:    mv s2, a1
-; CHECK-NOV-NEXT:    mv a0, s0
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s0
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixunssfti@plt
 ; CHECK-NOV-NEXT:    snez a1, a1
@@ -3155,11 +3164,12 @@
 ; CHECK-V-NEXT:    .cfi_offset s1, -24
 ; CHECK-V-NEXT:    .cfi_offset s2, -32
 ; CHECK-V-NEXT:    mv s0, a1
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixunssfti@plt
 ; CHECK-V-NEXT:    mv s1, a0
 ; CHECK-V-NEXT:    mv s2, a1
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixunssfti@plt
 ; CHECK-V-NEXT:    snez a1, a1
@@ -3200,11 +3210,12 @@
 ; CHECK-NOV-NEXT:    .cfi_offset s1, -24
 ; CHECK-NOV-NEXT:    .cfi_offset s2, -32
 ; CHECK-NOV-NEXT:    mv s2, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a0
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixsfti@plt
 ; CHECK-NOV-NEXT:    mv s0, a0
 ; CHECK-NOV-NEXT:    mv s1, a1
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixsfti@plt
 ; CHECK-NOV-NEXT:    mv a2, s1
@@ -3260,11 +3271,12 @@
 ; CHECK-V-NEXT:    .cfi_offset s1, -24
 ; CHECK-V-NEXT:    .cfi_offset s2, -32
 ; CHECK-V-NEXT:    mv s2, a1
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixsfti@plt
 ; CHECK-V-NEXT:    mv s0, a0
 ; CHECK-V-NEXT:    mv s1, a1
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixsfti@plt
 ; CHECK-V-NEXT:    mv a2, s1
@@ -3677,17 +3689,17 @@
 ; CHECK-NOV-NEXT:    lhu s3, 8(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 16(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.l.s s2, fs2, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.l.s s1, fs2, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    lui a1, 524288
@@ -3695,7 +3707,7 @@
 ; CHECK-NOV-NEXT:    bge a0, a4, .LBB33_10
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs1, rtz
-; CHECK-NOV-NEXT:    bge s2, a4, .LBB33_11
+; CHECK-NOV-NEXT:    bge s1, a4, .LBB33_11
 ; CHECK-NOV-NEXT:  .LBB33_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    bge a2, a4, .LBB33_12
@@ -3706,14 +3718,14 @@
 ; CHECK-NOV-NEXT:  .LBB33_5: # %entry
 ; CHECK-NOV-NEXT:    bge a1, a2, .LBB33_15
 ; CHECK-NOV-NEXT:  .LBB33_6: # %entry
-; CHECK-NOV-NEXT:    bge a1, s2, .LBB33_16
+; CHECK-NOV-NEXT:    bge a1, s1, .LBB33_16
 ; CHECK-NOV-NEXT:  .LBB33_7: # %entry
 ; CHECK-NOV-NEXT:    blt a1, a0, .LBB33_9
 ; CHECK-NOV-NEXT:  .LBB33_8: # %entry
 ; CHECK-NOV-NEXT:    lui a0, 524288
 ; CHECK-NOV-NEXT:  .LBB33_9: # %entry
 ; CHECK-NOV-NEXT:    sw a0, 12(s0)
-; CHECK-NOV-NEXT:    sw s2, 8(s0)
+; CHECK-NOV-NEXT:    sw s1, 8(s0)
 ; CHECK-NOV-NEXT:    sw a2, 4(s0)
 ; CHECK-NOV-NEXT:    sw a3, 0(s0)
 ; CHECK-NOV-NEXT:    ld ra, 56(sp) # 8-byte Folded Reload
@@ -3729,9 +3741,9 @@
 ; CHECK-NOV-NEXT:  .LBB33_10: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a4
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs1, rtz
-; CHECK-NOV-NEXT:    blt s2, a4, .LBB33_2
+; CHECK-NOV-NEXT:    blt s1, a4, .LBB33_2
 ; CHECK-NOV-NEXT:  .LBB33_11: # %entry
-; CHECK-NOV-NEXT:    mv s2, a4
+; CHECK-NOV-NEXT:    mv s1, a4
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    blt a2, a4, .LBB33_3
 ; CHECK-NOV-NEXT:  .LBB33_12: # %entry
@@ -3745,9 +3757,9 @@
 ; CHECK-NOV-NEXT:    blt a1, a2, .LBB33_6
 ; CHECK-NOV-NEXT:  .LBB33_15: # %entry
 ; CHECK-NOV-NEXT:    lui a2, 524288
-; CHECK-NOV-NEXT:    blt a1, s2, .LBB33_7
+; CHECK-NOV-NEXT:    blt a1, s1, .LBB33_7
 ; CHECK-NOV-NEXT:  .LBB33_16: # %entry
-; CHECK-NOV-NEXT:    lui s2, 524288
+; CHECK-NOV-NEXT:    lui s1, 524288
 ; CHECK-NOV-NEXT:    bge a1, a0, .LBB33_8
 ; CHECK-NOV-NEXT:    j .LBB33_9
 ;
@@ -3771,13 +3783,14 @@
 ; CHECK-V-NEXT:    lhu s1, 16(a0)
 ; CHECK-V-NEXT:    lhu s2, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
@@ -3787,7 +3800,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
@@ -3797,7 +3810,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
@@ -3854,17 +3867,17 @@
 ; CHECK-NOV-NEXT:    lhu s3, 16(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 8(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.lu.s s2, fs2, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.lu.s s1, fs2, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    li a1, -1
@@ -3872,7 +3885,7 @@
 ; CHECK-NOV-NEXT:    bgeu a0, a1, .LBB34_6
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a2, fs1, rtz
-; CHECK-NOV-NEXT:    bgeu s2, a1, .LBB34_7
+; CHECK-NOV-NEXT:    bgeu s1, a1, .LBB34_7
 ; CHECK-NOV-NEXT:  .LBB34_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    bgeu a2, a1, .LBB34_8
@@ -3883,7 +3896,7 @@
 ; CHECK-NOV-NEXT:  .LBB34_5: # %entry
 ; CHECK-NOV-NEXT:    sw a3, 12(s0)
 ; CHECK-NOV-NEXT:    sw a2, 8(s0)
-; CHECK-NOV-NEXT:    sw s2, 4(s0)
+; CHECK-NOV-NEXT:    sw s1, 4(s0)
 ; CHECK-NOV-NEXT:    sw a0, 0(s0)
 ; CHECK-NOV-NEXT:    ld ra, 56(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s0, 48(sp) # 8-byte Folded Reload
@@ -3898,9 +3911,9 @@
 ; CHECK-NOV-NEXT:  .LBB34_6: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a1
 ; CHECK-NOV-NEXT:    fcvt.lu.s a2, fs1, rtz
-; CHECK-NOV-NEXT:    bltu s2, a1, .LBB34_2
+; CHECK-NOV-NEXT:    bltu s1, a1, .LBB34_2
 ; CHECK-NOV-NEXT:  .LBB34_7: # %entry
-; CHECK-NOV-NEXT:    mv s2, a1
+; CHECK-NOV-NEXT:    mv s1, a1
 ; CHECK-NOV-NEXT:    fcvt.lu.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    bltu a2, a1, .LBB34_3
 ; CHECK-NOV-NEXT:  .LBB34_8: # %entry
@@ -3928,13 +3941,14 @@
 ; CHECK-V-NEXT:    lhu s1, 16(a0)
 ; CHECK-V-NEXT:    lhu s2, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
@@ -3944,7 +3958,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
@@ -3954,7 +3968,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
@@ -4009,17 +4023,17 @@
 ; CHECK-NOV-NEXT:    lhu s3, 8(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 16(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.l.s s2, fs2, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.l.s s1, fs2, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    li a2, -1
@@ -4027,7 +4041,7 @@
 ; CHECK-NOV-NEXT:    bge a0, a2, .LBB35_6
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs1, rtz
-; CHECK-NOV-NEXT:    bge s2, a2, .LBB35_7
+; CHECK-NOV-NEXT:    bge s1, a2, .LBB35_7
 ; CHECK-NOV-NEXT:  .LBB35_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    bge a1, a2, .LBB35_8
@@ -4042,9 +4056,9 @@
 ; CHECK-NOV-NEXT:    sgtz a3, a1
 ; CHECK-NOV-NEXT:    negw a3, a3
 ; CHECK-NOV-NEXT:    and a1, a3, a1
-; CHECK-NOV-NEXT:    sgtz a3, s2
+; CHECK-NOV-NEXT:    sgtz a3, s1
 ; CHECK-NOV-NEXT:    negw a3, a3
-; CHECK-NOV-NEXT:    and a3, a3, s2
+; CHECK-NOV-NEXT:    and a3, a3, s1
 ; CHECK-NOV-NEXT:    sgtz a4, a0
 ; CHECK-NOV-NEXT:    negw a4, a4
 ; CHECK-NOV-NEXT:    and a0, a4, a0
@@ -4065,9 +4079,9 @@
 ; CHECK-NOV-NEXT:  .LBB35_6: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a2
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs1, rtz
-; CHECK-NOV-NEXT:    blt s2, a2, .LBB35_2
+; CHECK-NOV-NEXT:    blt s1, a2, .LBB35_2
 ; CHECK-NOV-NEXT:  .LBB35_7: # %entry
-; CHECK-NOV-NEXT:    mv s2, a2
+; CHECK-NOV-NEXT:    mv s1, a2
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    blt a1, a2, .LBB35_3
 ; CHECK-NOV-NEXT:  .LBB35_8: # %entry
@@ -4095,13 +4109,14 @@
 ; CHECK-V-NEXT:    lhu s1, 16(a0)
 ; CHECK-V-NEXT:    lhu s2, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
@@ -4111,7 +4126,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
@@ -4121,7 +4136,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
@@ -4530,29 +4545,29 @@
 ; CHECK-NOV-NEXT:    lhu s7, 40(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 48(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs6, fa0
-; CHECK-NOV-NEXT:    mv a0, s7
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s7
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs5, fa0
-; CHECK-NOV-NEXT:    mv a0, s6
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s6
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs4, fa0
-; CHECK-NOV-NEXT:    mv a0, s5
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s5
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs3, fa0
-; CHECK-NOV-NEXT:    mv a0, s4
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s4
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.l.s s2, fs6, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.l.s s1, fs6, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    lui a7, 8
@@ -4560,7 +4575,7 @@
 ; CHECK-NOV-NEXT:    bge a0, a7, .LBB42_18
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs5, rtz
-; CHECK-NOV-NEXT:    bge s2, a7, .LBB42_19
+; CHECK-NOV-NEXT:    bge s1, a7, .LBB42_19
 ; CHECK-NOV-NEXT:  .LBB42_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs4, rtz
 ; CHECK-NOV-NEXT:    bge a1, a7, .LBB42_20
@@ -4592,14 +4607,14 @@
 ; CHECK-NOV-NEXT:  .LBB42_13: # %entry
 ; CHECK-NOV-NEXT:    bge a7, a1, .LBB42_31
 ; CHECK-NOV-NEXT:  .LBB42_14: # %entry
-; CHECK-NOV-NEXT:    bge a7, s2, .LBB42_32
+; CHECK-NOV-NEXT:    bge a7, s1, .LBB42_32
 ; CHECK-NOV-NEXT:  .LBB42_15: # %entry
 ; CHECK-NOV-NEXT:    blt a7, a0, .LBB42_17
 ; CHECK-NOV-NEXT:  .LBB42_16: # %entry
 ; CHECK-NOV-NEXT:    lui a0, 1048568
 ; CHECK-NOV-NEXT:  .LBB42_17: # %entry
 ; CHECK-NOV-NEXT:    sh a0, 14(s0)
-; CHECK-NOV-NEXT:    sh s2, 12(s0)
+; CHECK-NOV-NEXT:    sh s1, 12(s0)
 ; CHECK-NOV-NEXT:    sh a1, 10(s0)
 ; CHECK-NOV-NEXT:    sh a2, 8(s0)
 ; CHECK-NOV-NEXT:    sh a3, 6(s0)
@@ -4627,9 +4642,9 @@
 ; CHECK-NOV-NEXT:  .LBB42_18: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a7
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs5, rtz
-; CHECK-NOV-NEXT:    blt s2, a7, .LBB42_2
+; CHECK-NOV-NEXT:    blt s1, a7, .LBB42_2
 ; CHECK-NOV-NEXT:  .LBB42_19: # %entry
-; CHECK-NOV-NEXT:    mv s2, a7
+; CHECK-NOV-NEXT:    mv s1, a7
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs4, rtz
 ; CHECK-NOV-NEXT:    blt a1, a7, .LBB42_3
 ; CHECK-NOV-NEXT:  .LBB42_20: # %entry
@@ -4672,9 +4687,9 @@
 ; CHECK-NOV-NEXT:    blt a7, a1, .LBB42_14
 ; CHECK-NOV-NEXT:  .LBB42_31: # %entry
 ; CHECK-NOV-NEXT:    lui a1, 1048568
-; CHECK-NOV-NEXT:    blt a7, s2, .LBB42_15
+; CHECK-NOV-NEXT:    blt a7, s1, .LBB42_15
 ; CHECK-NOV-NEXT:  .LBB42_32: # %entry
-; CHECK-NOV-NEXT:    lui s2, 1048568
+; CHECK-NOV-NEXT:    lui s1, 1048568
 ; CHECK-NOV-NEXT:    bge a7, a0, .LBB42_16
 ; CHECK-NOV-NEXT:    j .LBB42_17
 ;
@@ -4710,13 +4725,14 @@
 ; CHECK-V-NEXT:    lhu s5, 16(a0)
 ; CHECK-V-NEXT:    lhu s6, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s6
+; CHECK-V-NEXT:    fmv.w.x fa0, s6
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -4726,7 +4742,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s5
+; CHECK-V-NEXT:    fmv.w.x fa0, s5
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -4736,7 +4752,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s4
+; CHECK-V-NEXT:    fmv.w.x fa0, s4
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -4746,7 +4762,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s3
+; CHECK-V-NEXT:    fmv.w.x fa0, s3
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -4756,7 +4772,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 4
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -4766,7 +4782,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 5
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -4776,7 +4792,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 6
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
@@ -4858,31 +4874,31 @@
 ; CHECK-NOV-NEXT:    lhu s7, 16(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 8(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs5, fa0
-; CHECK-NOV-NEXT:    mv a0, s7
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s7
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs6, fa0
-; CHECK-NOV-NEXT:    mv a0, s6
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s6
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs4, fa0
-; CHECK-NOV-NEXT:    mv a0, s5
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s5
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs3, fa0
-; CHECK-NOV-NEXT:    mv a0, s4
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s4
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.lu.s s3, fs6, rtz
-; CHECK-NOV-NEXT:    fcvt.lu.s s2, fs5, rtz
-; CHECK-NOV-NEXT:    sext.w s2, s2
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fcvt.lu.s s2, fs6, rtz
+; CHECK-NOV-NEXT:    fcvt.lu.s a0, fs5, rtz
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    sext.w s1, a0
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    sext.w a0, a0
@@ -4893,10 +4909,10 @@
 ; CHECK-NOV-NEXT:    mv a0, a1
 ; CHECK-NOV-NEXT:  .LBB43_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a3, fs4, rtz
-; CHECK-NOV-NEXT:    sext.w a2, s3
-; CHECK-NOV-NEXT:    bltu s2, a1, .LBB43_4
+; CHECK-NOV-NEXT:    sext.w a2, s2
+; CHECK-NOV-NEXT:    bltu s1, a1, .LBB43_4
 ; CHECK-NOV-NEXT:  # %bb.3: # %entry
-; CHECK-NOV-NEXT:    mv s2, a1
+; CHECK-NOV-NEXT:    mv s1, a1
 ; CHECK-NOV-NEXT:  .LBB43_4: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a4, fs3, rtz
 ; CHECK-NOV-NEXT:    sext.w a3, a3
@@ -4933,7 +4949,7 @@
 ; CHECK-NOV-NEXT:    sh a4, 8(s0)
 ; CHECK-NOV-NEXT:    sh a3, 6(s0)
 ; CHECK-NOV-NEXT:    sh a2, 4(s0)
-; CHECK-NOV-NEXT:    sh s2, 2(s0)
+; CHECK-NOV-NEXT:    sh s1, 2(s0)
 ; CHECK-NOV-NEXT:    sh a0, 0(s0)
 ; CHECK-NOV-NEXT:    ld ra, 120(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s0, 112(sp) # 8-byte Folded Reload
@@ -4994,13 +5010,14 @@
 ; CHECK-V-NEXT:    lhu s5, 16(a0)
 ; CHECK-V-NEXT:    lhu s6, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s6
+; CHECK-V-NEXT:    fmv.w.x fa0, s6
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -5010,7 +5027,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s5
+; CHECK-V-NEXT:    fmv.w.x fa0, s5
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -5020,7 +5037,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s4
+; CHECK-V-NEXT:    fmv.w.x fa0, s4
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -5030,7 +5047,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s3
+; CHECK-V-NEXT:    fmv.w.x fa0, s3
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -5040,7 +5057,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 4
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -5050,7 +5067,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 5
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -5060,7 +5077,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 6
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
@@ -5139,29 +5156,29 @@
 ; CHECK-NOV-NEXT:    lhu s7, 40(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 48(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs6, fa0
-; CHECK-NOV-NEXT:    mv a0, s7
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s7
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs5, fa0
-; CHECK-NOV-NEXT:    mv a0, s6
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s6
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs4, fa0
-; CHECK-NOV-NEXT:    mv a0, s5
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s5
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs3, fa0
-; CHECK-NOV-NEXT:    mv a0, s4
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s4
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.l.s s2, fs6, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.l.s s1, fs6, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    lui a3, 16
@@ -5169,7 +5186,7 @@
 ; CHECK-NOV-NEXT:    bge a0, a3, .LBB44_10
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs5, rtz
-; CHECK-NOV-NEXT:    bge s2, a3, .LBB44_11
+; CHECK-NOV-NEXT:    bge s1, a3, .LBB44_11
 ; CHECK-NOV-NEXT:  .LBB44_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs4, rtz
 ; CHECK-NOV-NEXT:    bge a1, a3, .LBB44_12
@@ -5208,9 +5225,9 @@
 ; CHECK-NOV-NEXT:    sgtz a7, a1
 ; CHECK-NOV-NEXT:    negw a7, a7
 ; CHECK-NOV-NEXT:    and a1, a7, a1
-; CHECK-NOV-NEXT:    sgtz a7, s2
+; CHECK-NOV-NEXT:    sgtz a7, s1
 ; CHECK-NOV-NEXT:    negw a7, a7
-; CHECK-NOV-NEXT:    and a7, a7, s2
+; CHECK-NOV-NEXT:    and a7, a7, s1
 ; CHECK-NOV-NEXT:    sgtz t0, a0
 ; CHECK-NOV-NEXT:    negw t0, t0
 ; CHECK-NOV-NEXT:    and a0, t0, a0
@@ -5243,9 +5260,9 @@
 ; CHECK-NOV-NEXT:  .LBB44_10: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a3
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs5, rtz
-; CHECK-NOV-NEXT:    blt s2, a3, .LBB44_2
+; CHECK-NOV-NEXT:    blt s1, a3, .LBB44_2
 ; CHECK-NOV-NEXT:  .LBB44_11: # %entry
-; CHECK-NOV-NEXT:    mv s2, a3
+; CHECK-NOV-NEXT:    mv s1, a3
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs4, rtz
 ; CHECK-NOV-NEXT:    blt a1, a3, .LBB44_3
 ; CHECK-NOV-NEXT:  .LBB44_12: # %entry
@@ -5301,13 +5318,14 @@
 ; CHECK-V-NEXT:    lhu s5, 16(a0)
 ; CHECK-V-NEXT:    lhu s6, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s6
+; CHECK-V-NEXT:    fmv.w.x fa0, s6
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -5317,7 +5335,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s5
+; CHECK-V-NEXT:    fmv.w.x fa0, s5
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -5327,7 +5345,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s4
+; CHECK-V-NEXT:    fmv.w.x fa0, s4
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -5337,7 +5355,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s3
+; CHECK-V-NEXT:    fmv.w.x fa0, s3
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -5347,7 +5365,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 4
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -5357,7 +5375,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 5
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
@@ -5367,7 +5385,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 6
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
@@ -6360,11 +6378,12 @@
 ; CHECK-NOV-NEXT:    .cfi_offset s1, -24
 ; CHECK-NOV-NEXT:    .cfi_offset s2, -32
 ; CHECK-NOV-NEXT:    mv s2, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a0
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixsfti@plt
 ; CHECK-NOV-NEXT:    mv s0, a0
 ; CHECK-NOV-NEXT:    mv s1, a1
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixsfti@plt
 ; CHECK-NOV-NEXT:    mv a2, a0
@@ -6465,11 +6484,12 @@
 ; CHECK-V-NEXT:    .cfi_offset s1, -24
 ; CHECK-V-NEXT:    .cfi_offset s2, -32
 ; CHECK-V-NEXT:    mv s2, a1
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixsfti@plt
 ; CHECK-V-NEXT:    mv s0, a0
 ; CHECK-V-NEXT:    mv s1, a1
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixsfti@plt
 ; CHECK-V-NEXT:    li a2, -1
@@ -6579,12 +6599,12 @@
 ; CHECK-NOV-NEXT:    .cfi_offset s1, -24
 ; CHECK-NOV-NEXT:    .cfi_offset s2, -32
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixunssfti@plt
 ; CHECK-NOV-NEXT:    mv s1, a0
 ; CHECK-NOV-NEXT:    mv s2, a1
-; CHECK-NOV-NEXT:    mv a0, s0
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s0
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixunssfti@plt
 ; CHECK-NOV-NEXT:    snez a2, a1
@@ -6621,12 +6641,12 @@
 ; CHECK-V-NEXT:    .cfi_offset s1, -24
 ; CHECK-V-NEXT:    .cfi_offset s2, -32
 ; CHECK-V-NEXT:    mv s0, a0
-; CHECK-V-NEXT:    mv a0, a1
+; CHECK-V-NEXT:    fmv.w.x fa0, a1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixunssfti@plt
 ; CHECK-V-NEXT:    mv s1, a0
 ; CHECK-V-NEXT:    mv s2, a1
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixunssfti@plt
 ; CHECK-V-NEXT:    snez a2, a1
@@ -6674,11 +6694,12 @@
 ; CHECK-NOV-NEXT:    .cfi_offset s1, -24
 ; CHECK-NOV-NEXT:    .cfi_offset s2, -32
 ; CHECK-NOV-NEXT:    mv s2, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a0
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixsfti@plt
 ; CHECK-NOV-NEXT:    mv s0, a0
 ; CHECK-NOV-NEXT:    mv s1, a1
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixsfti@plt
 ; CHECK-NOV-NEXT:    mv a2, a1
@@ -6737,11 +6758,12 @@
 ; CHECK-V-NEXT:    .cfi_offset s1, -24
 ; CHECK-V-NEXT:    .cfi_offset s2, -32
 ; CHECK-V-NEXT:    mv s2, a1
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixsfti@plt
 ; CHECK-V-NEXT:    mv s0, a0
 ; CHECK-V-NEXT:    mv s1, a1
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixsfti@plt
 ; CHECK-V-NEXT:    mv a2, a1