[RISCV] Custom lower FP_TO_FP16 and FP16_TO_FP to correct ABI of of libcall

[asb]

Summary

As introduced in D99148, RISC-V uses the softPromoteHalf legalisation for fp16 values without zfh, with logic ensuring that f16 values are passed in lower bits of FPRs (see D98670) when F or D support is present. This legalisation produces ISD::FP_TO_FP16 and ISD::FP16_TO_FP nodes which (as described in ISDOpcodes.h) provide a "semi-softened interface for dealing with f16 (as an i16)". i.e. the return type of the FP_TO_FP16 is an integer rather than a float (and the arg of FP16_TO_FP is an integer). The remainder of the description focuses primarily on FP_TO_FP16 for ease of explanation.

FP_TO_FP16 is lowered to a libcall to __truncsfhf2 (float) or __truncdfhf2 (double). As of D92241, _Float16 is used as the return type of these libcalls if the host compiler accepts _Float16 in a test input (i.e. dst_t is set to _Float16). _Float16 is enabled for the RISC-V target as of D105001 and so the return value should be passed in an FPR on hard float ABIs.

This patch fixes the ABI issue in what appears to be a minimally invasive way - leaving the softPromoteHalf logic undisturbed, and lowering FP_TO_FP16 to an f32-returning libcall, converting its result to an XLen integer value. Alternate strategies might be possible involving custom legalisation of FP_ROUND with an f16 target.

As can be seen in the test changes, the custom lowering for FP16_TO_FP means the libcall is no longer tail-callable.

Outstanding issues:

• Redundant fmv.x.w and fmv.w.x pairs produced in FP16_TO_FP lowering.
• Missing STRICT variants.

[craig.topper]

This seems like a good way to fix this to me. Custom lowering FP_ROUND wouldn't be enough. Soft promoting FADD and other arithmetic also creates FP16_TO_FP and FP_TO_FP16.

[craig.topper]

LGTM

[kito-cheng]

LGTM from the ABI view

[craig.topper]

llvm/lib/Target/RISCV/RISCVISelLowering.cpp

4927

I think we usually use the British spelling "legalisation"?

[asb]

Rebase, address review comment. I haven't had a chance yet to look at the strict opcodes or test related refactorings.

[craig.topper]

Ping. I would like to see this get commited. We started noticing this ABI mismatch a few weeks ago so I've cherry-picked this to our downstream.

[zixuan-wu]

Hi, I have a question that why other targets don't have such custom expand code instead of default expand behavior?

[asb]

@zixuan-wu I would assume due to different ABI rules.

@craig.topper I've gone ahead and fixed what I think was the only real blocker to merging this (ensuring coverage in the existing half-convert.ll test rather than introducing new files) and committed.

[zixuan-wu]

llvm/test/CodeGen/RISCV/half-convert.ll

149 ↗

(On Diff #536277)

why is there redundant fmv.x.w/fmv.w.x?

[asb]

llvm/test/CodeGen/RISCV/half-convert.ll

149 ↗

(On Diff #536277)

We're presumably missing a combine or peephole to get rid of the redundant pair. This is a limitation noted in the patch description. As this patch is addressing a correctness issue, and the suboptimal code quality isn't disastrous, it focuses solely on fixing the incorrect codegen. A follow-on patch can address this code quality issue.

[zixuan-wu]

llvm/test/CodeGen/RISCV/half-convert.ll

149 ↗

(On Diff #536277)

OK, sorry that I missed the description.

[joshua-arch1]

This comment was removed by joshua-arch1.

llvm/lib/Target/RISCV/RISCVISelLowering.cpp

RISCVTargetLowering::RISCVTargetLowering(const TargetMachine &TM,
 
   static const ISD::CondCode FPCCToExpand[] = {
       ISD::SETOGT, ISD::SETOGE, ISD::SETONE, ISD::SETUEQ, ISD::SETUGT,
       ISD::SETUGE, ISD::SETULT, ISD::SETULE, ISD::SETUNE, ISD::SETGT,
       ISD::SETGE,  ISD::SETNE,  ISD::SETO,   ISD::SETUO};
 
   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,
       ISD::FROUNDEVEN};
 
   if (Subtarget.hasStdExtZfhOrZfhminOrZhinxOrZhinxmin())
     setOperationAction(ISD::BITCAST, MVT::i16, Custom);
 
if (Subtarget.hasStdExtFOrZfinx()) {
     setCondCodeAction(FPCCToExpand, MVT::f32, Expand);
     setOperationAction(ISD::SELECT_CC, MVT::f32, Expand);
     setOperationAction(ISD::SELECT, MVT::f32, Custom);
     setOperationAction(ISD::BR_CC, MVT::f32, Expand);
     setOperationAction(FPOpToExpand, MVT::f32, Expand);
     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);
   }
 
   if (Subtarget.hasStdExtFOrZfinx() && Subtarget.is64Bit())
     setOperationAction(ISD::BITCAST, MVT::i32, Custom);
 
if (Subtarget.hasStdExtDOrZdinx()) {
     setOperationAction(ISD::SELECT, MVT::f64, Custom);
     setOperationAction(ISD::BR_CC, MVT::f64, Expand);
     setLoadExtAction(ISD::EXTLOAD, MVT::f64, MVT::f32, Expand);
     setTruncStoreAction(MVT::f64, MVT::f32, Expand);
     setOperationAction(FPOpToExpand, MVT::f64, Expand);
     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()) {
     setOperationAction({ISD::FP_TO_UINT, ISD::FP_TO_SINT,
                         ISD::STRICT_FP_TO_UINT, ISD::STRICT_FP_TO_SINT},
                        MVT::i32, Custom);
     setOperationAction(ISD::LROUND, MVT::i32, Custom);
   }
if (IsStrict) {
       return DAG.getMergeValues({SubVec, Src.getValue(1)}, DL);
     }
     Src = DAG.getNode(RVVOpc, DL, ContainerVT, Src, Mask, VL);
     return convertFromScalableVector(VT, Src, DAG, Subtarget);
   }
   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 legalisation");
+    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 legalisation");
+    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:
   case ISD::FNEARBYINT:
   case ISD::FRINT:
   case ISD::FROUND:
   case ISD::FROUNDEVEN:
     return lowerFTRUNC_FCEIL_FFLOOR_FROUND(Op, DAG, Subtarget);

llvm/test/CodeGen/RISCV/calling-conv-half.ll

 ; RV64IF-NEXT:    ret
 ;
 ; RV32-ILP32F-LABEL: callee_half_in_regs:
 ; RV32-ILP32F:       # %bb.0:
 ; RV32-ILP32F-NEXT:    addi sp, sp, -16
 ; 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
 ; RV32-ILP32F-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32-ILP32F-NEXT:    lw s0, 8(sp) # 4-byte Folded Reload
 ; RV32-ILP32F-NEXT:    addi sp, sp, 16
 ; RV32-ILP32F-NEXT:    ret
 ;
 ; RV64-LP64F-LABEL: callee_half_in_regs:
 ; RV64-LP64F:       # %bb.0:
 ; RV64-LP64F-NEXT:    addi sp, sp, -16
 ; RV64-LP64F-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; 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
 ; RV64-LP64F-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64-LP64F-NEXT:    ld s0, 0(sp) # 8-byte Folded Reload
 ; RV64-LP64F-NEXT:    addi sp, sp, 16
 ; RV64-LP64F-NEXT:    ret
 ;

 ; RV64IF-NEXT:    ret
 ;
 ; RV32-ILP32F-LABEL: callee_half_on_stack:
 ; RV32-ILP32F:       # %bb.0:
 ; RV32-ILP32F-NEXT:    addi sp, sp, -16
 ; 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
 ; RV32-ILP32F-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32-ILP32F-NEXT:    lw s0, 8(sp) # 4-byte Folded Reload
 ; RV32-ILP32F-NEXT:    addi sp, sp, 16
 ; RV32-ILP32F-NEXT:    ret
 ;
 ; RV64-LP64F-LABEL: callee_half_on_stack:
 ; RV64-LP64F:       # %bb.0:
 ; RV64-LP64F-NEXT:    addi sp, sp, -16
 ; RV64-LP64F-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; 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
 ; RV64-LP64F-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64-LP64F-NEXT:    ld s0, 0(sp) # 8-byte Folded Reload
 ; RV64-LP64F-NEXT:    addi sp, sp, 16
 ; RV64-LP64F-NEXT:    ret
 ;

 ; RV64IF-NEXT:    addi sp, sp, 16
 ; RV64IF-NEXT:    ret
 ;
 ; RV32-ILP32F-LABEL: caller_half_ret:
 ; RV32-ILP32F:       # %bb.0:
 ; 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
 ; RV32-ILP32F-NEXT:    addi sp, sp, 16
 ; RV32-ILP32F-NEXT:    ret
 ;
 ; RV64-LP64F-LABEL: caller_half_ret:
 ; RV64-LP64F:       # %bb.0:
 ; RV64-LP64F-NEXT:    addi sp, sp, -16
 ; 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
 ; RV64-LP64F-NEXT:    addi sp, sp, 16
 ; RV64-LP64F-NEXT:    ret
 ;
 ; RV32-ILP32ZFHMIN-LABEL: caller_half_ret:
 ; RV32-ILP32ZFHMIN:       # %bb.0:

llvm/test/CodeGen/RISCV/copysign-casts.ll

 ; RV32IF-NEXT:    ret
 ;
 ; RV32IFD-LABEL: fold_promote_d_h:
 ; RV32IFD:       # %bb.0:
 ; RV32IFD-NEXT:    addi sp, sp, -16
 ; 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
 ; RV32IFD-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32IFD-NEXT:    fld fs0, 0(sp) # 8-byte Folded Reload
 ; RV32IFD-NEXT:    addi sp, sp, 16
 ; RV32IFD-NEXT:    ret
 ;
 ; RV64IFD-LABEL: fold_promote_d_h:
 ; RV64IFD:       # %bb.0:
 ; RV64IFD-NEXT:    addi sp, sp, -16
 ; RV64IFD-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; 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
 ; RV64IFD-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64IFD-NEXT:    fld fs0, 0(sp) # 8-byte Folded Reload
 ; RV64IFD-NEXT:    addi sp, sp, 16
 ; RV64IFD-NEXT:    ret
 ;

 ; RV64I-NEXT:    ret
 ;
 ; RV32IF-LABEL: fold_promote_f_h:
 ; RV32IF:       # %bb.0:
 ; RV32IF-NEXT:    addi sp, sp, -16
 ; 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
 ; RV32IF-NEXT:    flw fs0, 8(sp) # 4-byte Folded Reload
 ; RV32IF-NEXT:    addi sp, sp, 16
 ; RV32IF-NEXT:    ret
 ;
 ; RV32IFD-LABEL: fold_promote_f_h:
 ; RV32IFD:       # %bb.0:
 ; RV32IFD-NEXT:    addi sp, sp, -16
 ; 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
 ; RV32IFD-NEXT:    fld fs0, 0(sp) # 8-byte Folded Reload
 ; RV32IFD-NEXT:    addi sp, sp, 16
 ; RV32IFD-NEXT:    ret
 ;
 ; RV64IFD-LABEL: fold_promote_f_h:
 ; RV64IFD:       # %bb.0:
 ; RV64IFD-NEXT:    addi sp, sp, -16
 ; RV64IFD-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; 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
 ; RV64IFD-NEXT:    fld fs0, 0(sp) # 8-byte Folded Reload
 ; RV64IFD-NEXT:    addi sp, sp, 16
 ; RV64IFD-NEXT:    ret
 ;
 ; RV32IFZFH-LABEL: fold_promote_f_h:

llvm/test/CodeGen/RISCV/fp16-promote.ll

; CHECK-NEXT:    ret
   %a = load half, ptr %p
   store half %a, ptr %q
   ret void
 }
 
 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
 }
 
 define double @test_fpextend_double(ptr %p) nounwind {
 ; CHECK-LABEL: test_fpextend_double:
 ; 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:    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
 ; CHECK-NEXT:    addi sp, sp, 16
 ; CHECK-NEXT:    ret
   %a = load half, ptr %p
   %r = fpext half %a to double
   ret double %r
 }
 
 define void @test_fptrunc_float(float %f, ptr %p) nounwind {
 ; CHECK-LABEL: test_fptrunc_float:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    addi sp, sp, -16
 ; CHECK-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; 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
 ; CHECK-NEXT:    addi sp, sp, 16
 ; CHECK-NEXT:    ret
   %a = fptrunc float %f to half
   store half %a, ptr %p
   ret void
 }
 
 define void @test_fptrunc_double(double %d, ptr %p) nounwind {
 ; CHECK-LABEL: test_fptrunc_double:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    addi sp, sp, -16
 ; CHECK-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; 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
 ; CHECK-NEXT:    addi sp, sp, 16
 ; CHECK-NEXT:    ret
   %a = fptrunc double %d to half
   store half %a, ptr %p
   ret void
 }
 
 define void @test_fadd(ptr %p, ptr %q) nounwind {
 ; CHECK-LABEL: test_fadd:
 ; CHECK:       # %bb.0:
 ; 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
   %b = load half, ptr %q
   %r = fadd half %a, %b
   store half %r, ptr %p
   ret void
 }
 
 define void @test_fmul(ptr %p, ptr %q) nounwind {
 ; CHECK-LABEL: test_fmul:
 ; CHECK:       # %bb.0:
 ; 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
   %b = load half, ptr %q
   %r = fmul half %a, %b
   store half %r, ptr %p
   ret void
 }

llvm/test/CodeGen/RISCV/fpclamptosat.ll

 
 define i32 @stest_f16i32(half %x) {
 ; RV32-LABEL: stest_f16i32:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -16
 ; 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
 ; RV32-NEXT:    addi a3, a2, -1
 ; RV32-NEXT:    beqz a1, .LBB6_2
 ; RV32-NEXT:  # %bb.1: # %entry
 ; RV32-NEXT:    slti a4, a1, 0
 ; RV32-NEXT:    j .LBB6_3

 ;
 ; RV64-LABEL: stest_f16i32:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    addiw a2, a1, -1
 ; RV64-NEXT:    blt a0, a2, .LBB6_2
 ; RV64-NEXT:  # %bb.1: # %entry
 ; RV64-NEXT:    mv a0, a2
 ; RV64-NEXT:  .LBB6_2: # %entry

 
 define i32 @utesth_f16i32(half %x) {
 ; RV32-LABEL: utesth_f16i32:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -16
 ; 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
 ; RV32-NEXT:    seqz a1, a1
 ; RV32-NEXT:    and a1, a1, a2
 ; RV32-NEXT:    addi a1, a1, -1
 ; RV32-NEXT:    or a0, a1, a0
 ; RV32-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32-NEXT:    addi sp, sp, 16
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: utesth_f16i32:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    srli a1, a1, 32
 ; RV64-NEXT:    bltu a0, a1, .LBB7_2
 ; RV64-NEXT:  # %bb.1: # %entry
 ; RV64-NEXT:    mv a0, a1
 ; RV64-NEXT:  .LBB7_2: # %entry

 
 define i32 @ustest_f16i32(half %x) {
 ; RV32-LABEL: ustest_f16i32:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -16
 ; 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
 ; RV32-NEXT:  # %bb.1: # %entry
 ; RV32-NEXT:    slti a2, a1, 0
 ; RV32-NEXT:    j .LBB8_3
 ; RV32-NEXT:  .LBB8_2:
 ; RV32-NEXT:    sltiu a2, a0, -1

 ;
 ; RV64-LABEL: ustest_f16i32:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    srli a1, a1, 32
 ; RV64-NEXT:    blt a0, a1, .LBB8_2
 ; RV64-NEXT:  # %bb.1: # %entry
 ; RV64-NEXT:    mv a0, a1
 ; RV64-NEXT:  .LBB8_2: # %entry

 
 define i16 @stest_f16i16(half %x) {
 ; RV32-LABEL: stest_f16i16:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -16
 ; 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
 ; RV32-NEXT:    addi a1, a1, -1
 ; RV32-NEXT:    blt a0, a1, .LBB15_2
 ; RV32-NEXT:  # %bb.1: # %entry
 ; RV32-NEXT:    mv a0, a1
 ; RV32-NEXT:  .LBB15_2: # %entry
 ; RV32-NEXT:    lui a1, 1048568
 ; RV32-NEXT:    blt a1, a0, .LBB15_4
 ; RV32-NEXT:  # %bb.3: # %entry
 ; RV32-NEXT:    lui a0, 1048568
 ; RV32-NEXT:  .LBB15_4: # %entry
 ; RV32-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32-NEXT:    addi sp, sp, 16
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: stest_f16i16:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    addiw a1, a1, -1
 ; RV64-NEXT:    blt a0, a1, .LBB15_2
 ; RV64-NEXT:  # %bb.1: # %entry
 ; RV64-NEXT:    mv a0, a1
 ; RV64-NEXT:  .LBB15_2: # %entry

 
 define i16 @utesth_f16i16(half %x) {
 ; RV32-LABEL: utesth_f16i16:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -16
 ; 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
 ; RV32-NEXT:    addi a1, a1, -1
 ; RV32-NEXT:    bltu a0, a1, .LBB16_2
 ; RV32-NEXT:  # %bb.1: # %entry
 ; RV32-NEXT:    mv a0, a1
 ; RV32-NEXT:  .LBB16_2: # %entry
 ; RV32-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32-NEXT:    addi sp, sp, 16
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: utesth_f16i16:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    addiw a1, a1, -1
 ; RV64-NEXT:    bltu a0, a1, .LBB16_2
 ; RV64-NEXT:  # %bb.1: # %entry
 ; RV64-NEXT:    mv a0, a1
 ; RV64-NEXT:  .LBB16_2: # %entry

 
 define i16 @ustest_f16i16(half %x) {
 ; RV32-LABEL: ustest_f16i16:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -16
 ; 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
 ; RV32-NEXT:    addi a1, a1, -1
 ; RV32-NEXT:    blt a0, a1, .LBB17_2
 ; RV32-NEXT:  # %bb.1: # %entry
 ; RV32-NEXT:    mv a0, a1
 ; RV32-NEXT:  .LBB17_2: # %entry
 ; RV32-NEXT:    sgtz a1, a0
 ; RV32-NEXT:    neg a1, a1
 ; RV32-NEXT:    and a0, a1, a0
 ; RV32-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32-NEXT:    addi sp, sp, 16
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: ustest_f16i16:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    addiw a1, a1, -1
 ; RV64-NEXT:    blt a0, a1, .LBB17_2
 ; RV64-NEXT:  # %bb.1: # %entry
 ; RV64-NEXT:    mv a0, a1
 ; RV64-NEXT:  .LBB17_2: # %entry

 
 define i64 @stest_f16i64(half %x) {
 ; RV32-LABEL: stest_f16i64:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -32
 ; 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
 ; RV32-NEXT:    lw a0, 20(sp)
 ; RV32-NEXT:    lw a2, 16(sp)
 ; RV32-NEXT:    lw a1, 12(sp)
 ; RV32-NEXT:    lw a4, 8(sp)
 ; RV32-NEXT:    lui a3, 524288

 ;
 ; RV64-LABEL: stest_f16i64:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    srli a3, a2, 1
 ; RV64-NEXT:    beqz a1, .LBB24_2
 ; RV64-NEXT:  # %bb.1: # %entry
 ; RV64-NEXT:    slti a4, a1, 0
 ; RV64-NEXT:    j .LBB24_3

 
 define i64 @utesth_f16i64(half %x) {
 ; RV32-LABEL: utesth_f16i64:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -32
 ; 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
 ; RV32-NEXT:    lw a0, 16(sp)
 ; RV32-NEXT:    lw a1, 20(sp)
 ; RV32-NEXT:    lw a2, 12(sp)
 ; RV32-NEXT:    lw a3, 8(sp)
 ; RV32-NEXT:    or a4, a1, a0

 ;
 ; RV64-LABEL: utesth_f16i64:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    addi a1, a1, -1
 ; RV64-NEXT:    and a0, a1, a0
 ; RV64-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64-NEXT:    addi sp, sp, 16
 ; RV64-NEXT:    ret
 entry:
   %conv = fptoui half %x to i128
   %0 = icmp ult i128 %conv, 18446744073709551616
   %spec.store.select = select i1 %0, i128 %conv, i128 18446744073709551616
   %conv6 = trunc i128 %spec.store.select to i64
   ret i64 %conv6
 }
 
 define i64 @ustest_f16i64(half %x) {
 ; RV32-LABEL: ustest_f16i64:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -32
 ; 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
 ; RV32-NEXT:    lw a1, 20(sp)
 ; RV32-NEXT:    lw a0, 16(sp)
 ; RV32-NEXT:    beqz a1, .LBB26_2
 ; RV32-NEXT:  # %bb.1: # %entry
 ; RV32-NEXT:    slti a2, a1, 0

 ;
 ; RV64-LABEL: ustest_f16i64:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    blez a1, .LBB26_2
 ; RV64-NEXT:  # %bb.1: # %entry
 ; RV64-NEXT:    li a1, 1
 ; RV64-NEXT:  .LBB26_2: # %entry
 ; RV64-NEXT:    neg a2, a2

 
 define i32 @stest_f16i32_mm(half %x) {
 ; RV32-LABEL: stest_f16i32_mm:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -16
 ; 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
 ; RV32-NEXT:    addi a4, a2, -1
 ; RV32-NEXT:    mv a3, a0
 ; RV32-NEXT:    bgez a1, .LBB33_9
 ; RV32-NEXT:  # %bb.1: # %entry
 ; RV32-NEXT:    bgeu a0, a4, .LBB33_10

 ;
 ; RV64-LABEL: stest_f16i32_mm:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    addiw a2, a1, -1
 ; RV64-NEXT:    blt a0, a2, .LBB33_2
 ; RV64-NEXT:  # %bb.1: # %entry
 ; RV64-NEXT:    mv a0, a2
 ; RV64-NEXT:  .LBB33_2: # %entry

 
 define i32 @utesth_f16i32_mm(half %x) {
 ; RV32-LABEL: utesth_f16i32_mm:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -16
 ; 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
 ; RV32-NEXT:    addi a1, a1, -1
 ; RV32-NEXT:    or a0, a1, a0
 ; RV32-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32-NEXT:    addi sp, sp, 16
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: utesth_f16i32_mm:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    srli a1, a1, 32
 ; RV64-NEXT:    bltu a0, a1, .LBB34_2
 ; RV64-NEXT:  # %bb.1: # %entry
 ; RV64-NEXT:    mv a0, a1
 ; RV64-NEXT:  .LBB34_2: # %entry

 
 define i32 @ustest_f16i32_mm(half %x) {
 ; RV32-LABEL: ustest_f16i32_mm:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -16
 ; 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
 ; RV32-NEXT:    beqz a1, .LBB35_2
 ; RV32-NEXT:  # %bb.1: # %entry
 ; RV32-NEXT:    addi a3, a2, -1
 ; RV32-NEXT:    or a0, a3, a0
 ; RV32-NEXT:  .LBB35_2: # %entry
 ; RV32-NEXT:    neg a2, a2
 ; RV32-NEXT:    and a1, a2, a1
 ; RV32-NEXT:    slti a1, a1, 0
 ; RV32-NEXT:    addi a1, a1, -1
 ; RV32-NEXT:    and a0, a1, a0
 ; RV32-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32-NEXT:    addi sp, sp, 16
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: ustest_f16i32_mm:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    srli a1, a1, 32
 ; RV64-NEXT:    blt a0, a1, .LBB35_2
 ; RV64-NEXT:  # %bb.1: # %entry
 ; RV64-NEXT:    mv a0, a1
 ; RV64-NEXT:  .LBB35_2: # %entry

 
 define i16 @stest_f16i16_mm(half %x) {
 ; RV32-LABEL: stest_f16i16_mm:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -16
 ; 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
 ; RV32-NEXT:    addi a1, a1, -1
 ; RV32-NEXT:    blt a0, a1, .LBB42_2
 ; RV32-NEXT:  # %bb.1: # %entry
 ; RV32-NEXT:    mv a0, a1
 ; RV32-NEXT:  .LBB42_2: # %entry
 ; RV32-NEXT:    lui a1, 1048568
 ; RV32-NEXT:    blt a1, a0, .LBB42_4
 ; RV32-NEXT:  # %bb.3: # %entry
 ; RV32-NEXT:    lui a0, 1048568
 ; RV32-NEXT:  .LBB42_4: # %entry
 ; RV32-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32-NEXT:    addi sp, sp, 16
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: stest_f16i16_mm:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    addiw a1, a1, -1
 ; RV64-NEXT:    blt a0, a1, .LBB42_2
 ; RV64-NEXT:  # %bb.1: # %entry
 ; RV64-NEXT:    mv a0, a1
 ; RV64-NEXT:  .LBB42_2: # %entry

 
 define i16 @utesth_f16i16_mm(half %x) {
 ; RV32-LABEL: utesth_f16i16_mm:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -16
 ; 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
 ; RV32-NEXT:    addi a1, a1, -1
 ; RV32-NEXT:    bltu a0, a1, .LBB43_2
 ; RV32-NEXT:  # %bb.1: # %entry
 ; RV32-NEXT:    mv a0, a1
 ; RV32-NEXT:  .LBB43_2: # %entry
 ; RV32-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32-NEXT:    addi sp, sp, 16
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: utesth_f16i16_mm:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    lui a1, 16
 ; RV64-NEXT:    addiw a1, a1, -1
 ; RV64-NEXT:    bltu a0, a1, .LBB43_2
 ; RV64-NEXT:  # %bb.1: # %entry
 ; RV64-NEXT:    mv a0, a1

 
 define i16 @ustest_f16i16_mm(half %x) {
 ; RV32-LABEL: ustest_f16i16_mm:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -16
 ; 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
 ; RV32-NEXT:    addi a1, a1, -1
 ; RV32-NEXT:    blt a0, a1, .LBB44_2
 ; RV32-NEXT:  # %bb.1: # %entry
 ; RV32-NEXT:    mv a0, a1
 ; RV32-NEXT:  .LBB44_2: # %entry
 ; RV32-NEXT:    sgtz a1, a0
 ; RV32-NEXT:    neg a1, a1
 ; RV32-NEXT:    and a0, a1, a0
 ; RV32-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32-NEXT:    addi sp, sp, 16
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: ustest_f16i16_mm:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    addiw a1, a1, -1
 ; RV64-NEXT:    blt a0, a1, .LBB44_2
 ; RV64-NEXT:  # %bb.1: # %entry
 ; RV64-NEXT:    mv a0, a1
 ; RV64-NEXT:  .LBB44_2: # %entry

 
 define i64 @stest_f16i64_mm(half %x) {
 ; RV32-LABEL: stest_f16i64_mm:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -32
 ; 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
 ; RV32-NEXT:    lw a0, 20(sp)
 ; RV32-NEXT:    lw t0, 8(sp)
 ; RV32-NEXT:    lw a4, 12(sp)
 ; RV32-NEXT:    lw a1, 16(sp)
 ; RV32-NEXT:    lui a3, 524288

 ;
 ; RV64-LABEL: stest_f16i64_mm:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    srli a4, a2, 1
 ; RV64-NEXT:    mv a3, a0
 ; RV64-NEXT:    bgez a1, .LBB51_9
 ; RV64-NEXT:  # %bb.1: # %entry
 ; RV64-NEXT:    bgeu a0, a4, .LBB51_10

 
 define i64 @utesth_f16i64_mm(half %x) {
 ; RV32-LABEL: utesth_f16i64_mm:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -32
 ; 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
 ; RV32-NEXT:    lw a0, 16(sp)
 ; RV32-NEXT:    lw a1, 20(sp)
 ; RV32-NEXT:    lw a2, 12(sp)
 ; RV32-NEXT:    lw a3, 8(sp)
 ; RV32-NEXT:    or a4, a1, a0

 ;
 ; RV64-LABEL: utesth_f16i64_mm:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    addi a2, a2, -1
 ; RV64-NEXT:    and a0, a2, a0
 ; RV64-NEXT:    addi a1, a1, -1
 ; RV64-NEXT:    seqz a1, a1
 ; RV64-NEXT:    addi a1, a1, -1

 
 define i64 @ustest_f16i64_mm(half %x) {
 ; RV32-LABEL: ustest_f16i64_mm:
 ; RV32:       # %bb.0: # %entry
 ; RV32-NEXT:    addi sp, sp, -32
 ; 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
 ; RV32-NEXT:    lw a0, 8(sp)
 ; RV32-NEXT:    lw a3, 12(sp)
 ; RV32-NEXT:    lw a1, 20(sp)
 ; RV32-NEXT:    lw a4, 16(sp)
 ; RV32-NEXT:    slti a2, a1, 0

 ;
 ; RV64-LABEL: ustest_f16i64_mm:
 ; RV64:       # %bb.0: # %entry
 ; RV64-NEXT:    addi sp, sp, -16
 ; RV64-NEXT:    .cfi_def_cfa_offset 16
 ; 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
 ; RV64-NEXT:    blez a1, .LBB53_2
 ; RV64-NEXT:  # %bb.1: # %entry
 ; RV64-NEXT:    li a2, 1
 ; RV64-NEXT:  .LBB53_2: # %entry
 ; RV64-NEXT:    slti a3, a1, 1

llvm/test/CodeGen/RISCV/half-fpext.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 2
 ; RUN: llc -mtriple=riscv32 -verify-machineinstrs < %s | FileCheck %s -check-prefix=RV32I-ILP32
 ; RUN: llc -mtriple=riscv64 -verify-machineinstrs < %s | FileCheck %s -check-prefix=RV64I-LP64
 ; RUN: llc -mtriple=riscv32 -mattr=+d -target-abi=ilp32 -verify-machineinstrs < %s | FileCheck %s -check-prefix=RV32ID-ILP32
 ; RUN: llc -mtriple=riscv64 -mattr=+d -target-abi=lp64 -verify-machineinstrs < %s | FileCheck %s -check-prefix=RV64ID-LP64
 ; RUN: llc -mtriple=riscv32 -mattr=+d -target-abi=ilp32d -verify-machineinstrs < %s | FileCheck %s -check-prefix=RV32ID-ILP32D
 ; RUN: llc -mtriple=riscv64 -mattr=+d -target-abi=lp64d -verify-machineinstrs < %s  | FileCheck %s -check-prefix=RV64ID-LP64D
 
 define float @half_to_float(half %a) nounwind {
 ; RV32I-ILP32-LABEL: half_to_float:
 ; RV32I-ILP32:       # %bb.0:
 ; RV32I-ILP32-NEXT:    addi sp, sp, -16
 ; RV32I-ILP32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
-; RV32I-ILP32-NEXT:    slli a0, a0, 16
-; RV32I-ILP32-NEXT:    srli a0, a0, 16
 ; RV32I-ILP32-NEXT:    call __extendhfsf2@plt
 ; RV32I-ILP32-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-ILP32-NEXT:    addi sp, sp, 16
 ; RV32I-ILP32-NEXT:    ret
 ;
 ; RV64I-LP64-LABEL: half_to_float:
 ; RV64I-LP64:       # %bb.0:
 ; RV64I-LP64-NEXT:    addi sp, sp, -16
 ; RV64I-LP64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
-; RV64I-LP64-NEXT:    slli a0, a0, 48
-; RV64I-LP64-NEXT:    srli a0, a0, 48
 ; RV64I-LP64-NEXT:    call __extendhfsf2@plt
 ; RV64I-LP64-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-LP64-NEXT:    addi sp, sp, 16
 ; RV64I-LP64-NEXT:    ret
 ;
 ; 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:
 ; RV64ID-LP64-NEXT:    addi sp, sp, -16
 ; RV64ID-LP64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64ID-LP64-NEXT:    call __extendhfsf2@plt
 ; RV64ID-LP64-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64ID-LP64-NEXT:    addi sp, sp, 16
 ; RV64ID-LP64-NEXT:    ret
 ;
 ; 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
 }
 
 define double @half_to_double(half %a) nounwind {
 ; RV32I-ILP32-LABEL: half_to_double:
 ; RV32I-ILP32:       # %bb.0:
 ; RV32I-ILP32-NEXT:    addi sp, sp, -16
 ; RV32I-ILP32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
-; RV32I-ILP32-NEXT:    slli a0, a0, 16
-; RV32I-ILP32-NEXT:    srli a0, a0, 16
 ; RV32I-ILP32-NEXT:    call __extendhfsf2@plt
 ; RV32I-ILP32-NEXT:    call __extendsfdf2@plt
 ; RV32I-ILP32-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-ILP32-NEXT:    addi sp, sp, 16
 ; RV32I-ILP32-NEXT:    ret
 ;
 ; RV64I-LP64-LABEL: half_to_double:
 ; RV64I-LP64:       # %bb.0:
 ; RV64I-LP64-NEXT:    addi sp, sp, -16
 ; RV64I-LP64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
-; RV64I-LP64-NEXT:    slli a0, a0, 48
-; RV64I-LP64-NEXT:    srli a0, a0, 48
 ; RV64I-LP64-NEXT:    call __extendhfsf2@plt
-; RV64I-LP64-NEXT:    slli a0, a0, 32
-; RV64I-LP64-NEXT:    srli a0, a0, 32
 ; RV64I-LP64-NEXT:    call __extendsfdf2@plt
 ; RV64I-LP64-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-LP64-NEXT:    addi sp, sp, 16
 ; RV64I-LP64-NEXT:    ret
 ;
 ; RV32ID-ILP32-LABEL: half_to_double:
 ; RV32ID-ILP32:       # %bb.0:
 ; RV32ID-ILP32-NEXT:    addi sp, sp, -16

 ; RV64ID-LP64-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64ID-LP64-NEXT:    addi sp, sp, 16
 ; RV64ID-LP64-NEXT:    ret
 ;
 ; RV32ID-ILP32D-LABEL: half_to_double:
 ; 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
 ; RV32ID-ILP32D-NEXT:    addi sp, sp, 16
 ; RV32ID-ILP32D-NEXT:    ret
 ;
 ; RV64ID-LP64D-LABEL: half_to_double:
 ; 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:    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
 ; RV64ID-LP64D-NEXT:    addi sp, sp, 16
 ; RV64ID-LP64D-NEXT:    ret
   %1 = fpext half %a to double
   ret double %1
 }

llvm/test/CodeGen/RISCV/half-fptrunc.ll

 ; RV64ID-LP64-NEXT:    addi sp, sp, 16
 ; RV64ID-LP64-NEXT:    ret
 ;
 ; RV32ID-ILP32D-LABEL: float_to_half:
 ; 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:    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
 ; 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: float_to_half:
 ; 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:    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
 ; RV64ID-LP64D-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64ID-LP64D-NEXT:    addi sp, sp, 16
 ; RV64ID-LP64D-NEXT:    ret
   %1 = fptrunc float %a to half
   ret half %1

 ; RV64ID-LP64-NEXT:    addi sp, sp, 16
 ; RV64ID-LP64-NEXT:    ret
 ;
 ; RV32ID-ILP32D-LABEL: double_to_half:
 ; 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:    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
 ; 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: double_to_half:
 ; 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:    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
 ; RV64ID-LP64D-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64ID-LP64D-NEXT:    addi sp, sp, 16
 ; RV64ID-LP64D-NEXT:    ret
   %1 = fptrunc double %a to half
   ret half %1
 }

llvm/test/CodeGen/RISCV/libcall-tail-calls.ll

 
 declare half @llvm.sin.f16(half)
 
 define half @sin_f16(half %a) nounwind {
 ; RV32IFD-ILP32D-LABEL: sin_f16:
 ; 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
 ; RV32IFD-ILP32D-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32IFD-ILP32D-NEXT:    addi sp, sp, 16
 ; RV32IFD-ILP32D-NEXT:    ret
 ;
 ; RV32IF-ILP32F-LABEL: sin_f16:
 ; 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
 ; RV32IF-ILP32F-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32IF-ILP32F-NEXT:    addi sp, sp, 16
 ; RV32IF-ILP32F-NEXT:    ret
 ;
 ; RV32IFD-ILP32-LABEL: sin_f16:

 ; RV32I-ILP32-NEXT:    addi sp, sp, 16
 ; RV32I-ILP32-NEXT:    ret
 ;
 ; RV64IFD-LP64D-LABEL: sin_f16:
 ; RV64IFD-LP64D:       # %bb.0:
 ; 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
 ; RV64IFD-LP64D-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64IFD-LP64D-NEXT:    addi sp, sp, 16
 ; RV64IFD-LP64D-NEXT:    ret
 ;
 ; RV64IF-LP64F-LABEL: sin_f16:
 ; RV64IF-LP64F:       # %bb.0:
 ; 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
 ; RV64IF-LP64F-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64IF-LP64F-NEXT:    addi sp, sp, 16
 ; RV64IF-LP64F-NEXT:    ret
 ;
 ; RV64IFD-LP64-LABEL: sin_f16:

llvm/test/CodeGen/RISCV/rvv/fpclamptosat_vec.ll

 ; CHECK-NOV-NEXT:    .cfi_offset fs0, -48
 ; CHECK-NOV-NEXT:    .cfi_offset fs1, -56
 ; CHECK-NOV-NEXT:    .cfi_offset fs2, -64
 ; CHECK-NOV-NEXT:    lhu s1, 24(a1)
 ; CHECK-NOV-NEXT:    lhu s2, 0(a1)
 ; 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
 ; CHECK-NOV-NEXT:    addiw a4, a1, -1
 ; 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
 ; CHECK-NOV-NEXT:  .LBB6_3: # %entry
 ; CHECK-NOV-NEXT:    bge a3, a4, .LBB6_13
 ; CHECK-NOV-NEXT:  .LBB6_4: # %entry
 ; CHECK-NOV-NEXT:    bge a1, a3, .LBB6_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
 ; CHECK-NOV-NEXT:    ld s0, 48(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s1, 40(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s2, 32(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s3, 24(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs0, 16(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs1, 8(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs2, 0(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    addi sp, sp, 64
 ; CHECK-NOV-NEXT:    ret
 ; 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
 ; CHECK-NOV-NEXT:    mv a2, a4
 ; CHECK-NOV-NEXT:    blt a3, a4, .LBB6_4
 ; CHECK-NOV-NEXT:  .LBB6_13: # %entry
 ; CHECK-NOV-NEXT:    mv a3, a4
 ; CHECK-NOV-NEXT:    blt a1, a4, .LBB6_5
 ; CHECK-NOV-NEXT:  .LBB6_14: # %entry
 ; CHECK-NOV-NEXT:    lui a3, 524288
 ; 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
 ;
 ; CHECK-V-LABEL: stest_f16i32:
 ; CHECK-V:       # %bb.0: # %entry
 ; CHECK-V-NEXT:    addi sp, sp, -48
 ; CHECK-V-NEXT:    .cfi_def_cfa_offset 48
 ; CHECK-V-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    .cfi_offset ra, -8
 ; CHECK-V-NEXT:    .cfi_offset s0, -16
 ; CHECK-V-NEXT:    .cfi_offset s1, -24
 ; CHECK-V-NEXT:    .cfi_offset s2, -32
 ; CHECK-V-NEXT:    csrr a1, vlenb
 ; CHECK-V-NEXT:    slli a1, a1, 1
 ; CHECK-V-NEXT:    sub sp, sp, a1
 ; CHECK-V-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x30, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 48 + 2 * vlenb
 ; CHECK-V-NEXT:    lhu s0, 24(a0)
 ; 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, 2, e64, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 3, e64, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    lui a0, 524288

 ; CHECK-NOV-NEXT:    .cfi_offset fs0, -48
 ; CHECK-NOV-NEXT:    .cfi_offset fs1, -56
 ; CHECK-NOV-NEXT:    .cfi_offset fs2, -64
 ; CHECK-NOV-NEXT:    lhu s1, 0(a1)
 ; CHECK-NOV-NEXT:    lhu s2, 24(a1)
 ; 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
 ; CHECK-NOV-NEXT:    srli a1, a1, 32
 ; 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
 ; CHECK-NOV-NEXT:  .LBB7_3: # %entry
 ; CHECK-NOV-NEXT:    bltu a3, a1, .LBB7_5
 ; CHECK-NOV-NEXT:  .LBB7_4: # %entry
 ; CHECK-NOV-NEXT:    mv a3, a1
 ; 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
 ; CHECK-NOV-NEXT:    ld s1, 40(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s2, 32(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s3, 24(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs0, 16(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs1, 8(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs2, 0(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    addi sp, sp, 64
 ; CHECK-NOV-NEXT:    ret
 ; 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
 ; CHECK-NOV-NEXT:    mv a2, a1
 ; CHECK-NOV-NEXT:    bgeu a3, a1, .LBB7_4
 ; CHECK-NOV-NEXT:    j .LBB7_5
 ;
 ; CHECK-V-LABEL: utesth_f16i32:

 ; CHECK-V-NEXT:    csrr a1, vlenb
 ; CHECK-V-NEXT:    slli a1, a1, 1
 ; CHECK-V-NEXT:    sub sp, sp, a1
 ; CHECK-V-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x30, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 48 + 2 * vlenb
 ; CHECK-V-NEXT:    lhu s0, 24(a0)
 ; 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, 2, e64, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 3, e64, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    li a0, -1

 ; CHECK-NOV-NEXT:    .cfi_offset fs0, -48
 ; CHECK-NOV-NEXT:    .cfi_offset fs1, -56
 ; CHECK-NOV-NEXT:    .cfi_offset fs2, -64
 ; CHECK-NOV-NEXT:    lhu s1, 0(a1)
 ; CHECK-NOV-NEXT:    lhu s2, 24(a1)
 ; 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
 ; CHECK-NOV-NEXT:    srli a2, a2, 32
 ; 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
 ; CHECK-NOV-NEXT:  .LBB8_3: # %entry
 ; CHECK-NOV-NEXT:    blt a3, a2, .LBB8_5
 ; CHECK-NOV-NEXT:  .LBB8_4: # %entry
 ; CHECK-NOV-NEXT:    mv a3, a2
 ; 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
 ; CHECK-NOV-NEXT:    and a2, a2, a3
 ; CHECK-NOV-NEXT:    sw a2, 12(s0)
 ; CHECK-NOV-NEXT:    sw a1, 8(s0)
 ; CHECK-NOV-NEXT:    sw a5, 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
 ; CHECK-NOV-NEXT:    ld s1, 40(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s2, 32(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s3, 24(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs0, 16(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs1, 8(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs2, 0(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    addi sp, sp, 64
 ; CHECK-NOV-NEXT:    ret
 ; 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
 ; CHECK-NOV-NEXT:    mv a1, a2
 ; CHECK-NOV-NEXT:    bge a3, a2, .LBB8_4
 ; CHECK-NOV-NEXT:    j .LBB8_5
 ;
 ; CHECK-V-LABEL: ustest_f16i32:

 ; CHECK-V-NEXT:    csrr a1, vlenb
 ; CHECK-V-NEXT:    slli a1, a1, 1
 ; CHECK-V-NEXT:    sub sp, sp, a1
 ; CHECK-V-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x30, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 48 + 2 * vlenb
 ; CHECK-V-NEXT:    lhu s0, 24(a0)
 ; 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, 2, e64, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 3, e64, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    li a0, -1

 ; CHECK-NOV-NEXT:    lhu s2, 0(a1)
 ; CHECK-NOV-NEXT:    lhu s3, 8(a1)
 ; CHECK-NOV-NEXT:    lhu s4, 16(a1)
 ; CHECK-NOV-NEXT:    lhu s5, 24(a1)
 ; CHECK-NOV-NEXT:    lhu s6, 32(a1)
 ; 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
 ; CHECK-NOV-NEXT:    addiw a7, a7, -1
 ; 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
 ; CHECK-NOV-NEXT:  .LBB15_3: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs3, rtz
 ; CHECK-NOV-NEXT:    bge a2, a7, .LBB15_21
 ; CHECK-NOV-NEXT:  .LBB15_4: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a4, fs2, rtz

 ; CHECK-NOV-NEXT:    bge a7, a4, .LBB15_28
 ; CHECK-NOV-NEXT:  .LBB15_11: # %entry
 ; CHECK-NOV-NEXT:    bge a7, a3, .LBB15_29
 ; CHECK-NOV-NEXT:  .LBB15_12: # %entry
 ; CHECK-NOV-NEXT:    bge a7, a2, .LBB15_30
 ; 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)
 ; CHECK-NOV-NEXT:    sh a4, 4(s0)
 ; CHECK-NOV-NEXT:    sh a5, 2(s0)
 ; CHECK-NOV-NEXT:    sh a6, 0(s0)
 ; CHECK-NOV-NEXT:    ld ra, 120(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s0, 112(sp) # 8-byte Folded Reload

 ; CHECK-NOV-NEXT:    fld fs4, 16(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs5, 8(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs6, 0(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    addi sp, sp, 128
 ; CHECK-NOV-NEXT:    ret
 ; 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
 ; CHECK-NOV-NEXT:    mv a1, a7
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs3, rtz
 ; CHECK-NOV-NEXT:    blt a2, a7, .LBB15_4
 ; CHECK-NOV-NEXT:  .LBB15_21: # %entry
 ; CHECK-NOV-NEXT:    mv a2, a7

 ; CHECK-NOV-NEXT:  .LBB15_29: # %entry
 ; CHECK-NOV-NEXT:    lui a3, 1048568
 ; CHECK-NOV-NEXT:    blt a7, a2, .LBB15_13
 ; CHECK-NOV-NEXT:  .LBB15_30: # %entry
 ; CHECK-NOV-NEXT:    lui a2, 1048568
 ; 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
 ;
 ; CHECK-V-LABEL: stest_f16i16:
 ; CHECK-V:       # %bb.0: # %entry
 ; CHECK-V-NEXT:    addi sp, sp, -80
 ; CHECK-V-NEXT:    .cfi_def_cfa_offset 80
 ; CHECK-V-NEXT:    sd ra, 72(sp) # 8-byte Folded Spill

 ; CHECK-V-NEXT:    lhu s0, 56(a0)
 ; CHECK-V-NEXT:    lhu s1, 48(a0)
 ; CHECK-V-NEXT:    lhu s2, 40(a0)
 ; CHECK-V-NEXT:    lhu s3, 32(a0)
 ; CHECK-V-NEXT:    lhu s4, 24(a0)
 ; 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, 2, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 3, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 4, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 5, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 6, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 7, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 7
 ; CHECK-V-NEXT:    lui a0, 8

 ; CHECK-NOV-NEXT:    lhu s2, 56(a1)
 ; CHECK-NOV-NEXT:    lhu s3, 48(a1)
 ; CHECK-NOV-NEXT:    lhu s4, 40(a1)
 ; CHECK-NOV-NEXT:    lhu s5, 32(a1)
 ; CHECK-NOV-NEXT:    lhu s6, 24(a1)
 ; 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
 ; CHECK-NOV-NEXT:    addiw a1, a1, -1
 ; 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
 ; CHECK-NOV-NEXT:  .LBB16_3: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a4, fs3, rtz
 ; CHECK-NOV-NEXT:    bgeu a3, a1, .LBB16_13
 ; CHECK-NOV-NEXT:  .LBB16_4: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a5, fs2, rtz

 ; CHECK-NOV-NEXT:    mv a7, a1
 ; CHECK-NOV-NEXT:  .LBB16_9: # %entry
 ; CHECK-NOV-NEXT:    sh a7, 14(s0)
 ; CHECK-NOV-NEXT:    sh a6, 12(s0)
 ; CHECK-NOV-NEXT:    sh a5, 10(s0)
 ; 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
 ; CHECK-NOV-NEXT:    ld s1, 104(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s2, 96(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s3, 88(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s4, 80(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s5, 72(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s6, 64(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s7, 56(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs0, 48(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs1, 40(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs2, 32(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs3, 24(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs4, 16(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs5, 8(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs6, 0(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    addi sp, sp, 128
 ; CHECK-NOV-NEXT:    ret
 ; 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
 ; CHECK-NOV-NEXT:    mv a2, a1
 ; CHECK-NOV-NEXT:    fcvt.lu.s a4, fs3, rtz
 ; CHECK-NOV-NEXT:    bltu a3, a1, .LBB16_4
 ; CHECK-NOV-NEXT:  .LBB16_13: # %entry
 ; CHECK-NOV-NEXT:    mv a3, a1

 ; CHECK-V-NEXT:    lhu s0, 56(a0)
 ; CHECK-V-NEXT:    lhu s1, 48(a0)
 ; CHECK-V-NEXT:    lhu s2, 40(a0)
 ; CHECK-V-NEXT:    lhu s3, 32(a0)
 ; CHECK-V-NEXT:    lhu s4, 24(a0)
 ; 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, 2, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 3, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 4, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 5, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 6, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 7, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 7
 ; CHECK-V-NEXT:    lui a0, 16

 ; CHECK-NOV-NEXT:    lhu s2, 56(a1)
 ; CHECK-NOV-NEXT:    lhu s3, 48(a1)
 ; CHECK-NOV-NEXT:    lhu s4, 40(a1)
 ; CHECK-NOV-NEXT:    lhu s5, 32(a1)
 ; CHECK-NOV-NEXT:    lhu s6, 24(a1)
 ; 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
 ; CHECK-NOV-NEXT:    addiw a3, a3, -1
 ; 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
 ; CHECK-NOV-NEXT:  .LBB17_3: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a4, fs3, rtz
 ; CHECK-NOV-NEXT:    bge a2, a3, .LBB17_13
 ; CHECK-NOV-NEXT:  .LBB17_4: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a5, fs2, rtz

 ; CHECK-NOV-NEXT:    mv a7, a3
 ; CHECK-NOV-NEXT:  .LBB17_9: # %entry
 ; CHECK-NOV-NEXT:    sgtz a3, a7
 ; CHECK-NOV-NEXT:    sgtz t0, a6
 ; CHECK-NOV-NEXT:    sgtz t1, a5
 ; 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
 ; CHECK-NOV-NEXT:    and a2, t3, a2
 ; CHECK-NOV-NEXT:    negw t2, t2
 ; CHECK-NOV-NEXT:    and a4, t2, a4
 ; CHECK-NOV-NEXT:    negw t1, t1
 ; CHECK-NOV-NEXT:    and a5, t1, a5

 ; CHECK-NOV-NEXT:    fld fs4, 16(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs5, 8(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs6, 0(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    addi sp, sp, 128
 ; CHECK-NOV-NEXT:    ret
 ; 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
 ; CHECK-NOV-NEXT:    mv a1, a3
 ; CHECK-NOV-NEXT:    fcvt.l.s a4, fs3, rtz
 ; CHECK-NOV-NEXT:    blt a2, a3, .LBB17_4
 ; CHECK-NOV-NEXT:  .LBB17_13: # %entry
 ; CHECK-NOV-NEXT:    mv a2, a3

 ; CHECK-V-NEXT:    lhu s0, 56(a0)
 ; CHECK-V-NEXT:    lhu s1, 48(a0)
 ; CHECK-V-NEXT:    lhu s2, 40(a0)
 ; CHECK-V-NEXT:    lhu s3, 32(a0)
 ; CHECK-V-NEXT:    lhu s4, 24(a0)
 ; 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, 2, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 3, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 4, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 5, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 6, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 7, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 7
 ; CHECK-V-NEXT:    lui a0, 16

 ; CHECK-NOV-NEXT:    sd s0, 16(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    sd s1, 8(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    sd s2, 0(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    .cfi_offset ra, -8
 ; CHECK-NOV-NEXT:    .cfi_offset s0, -16
 ; 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
 ; CHECK-NOV-NEXT:    li a0, -1
 ; CHECK-NOV-NEXT:    srli a3, a0, 1
 ; CHECK-NOV-NEXT:    beqz a1, .LBB24_3
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    slti a4, a1, 0

 ; CHECK-V-NEXT:    sd s0, 16(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s1, 8(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s2, 0(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    .cfi_offset ra, -8
 ; CHECK-V-NEXT:    .cfi_offset s0, -16
 ; 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
 ; CHECK-V-NEXT:    srli a3, a2, 1
 ; CHECK-V-NEXT:    beqz a1, .LBB24_3
 ; CHECK-V-NEXT:  # %bb.1: # %entry
 ; CHECK-V-NEXT:    slti a4, a1, 0
 ; CHECK-V-NEXT:    bnez s1, .LBB24_4

 ; CHECK-NOV-NEXT:    sd s0, 16(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    sd s1, 8(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    sd s2, 0(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    .cfi_offset ra, -8
 ; CHECK-NOV-NEXT:    .cfi_offset s0, -16
 ; 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
 ; CHECK-NOV-NEXT:    snez a2, s2
 ; CHECK-NOV-NEXT:    addi a2, a2, -1
 ; CHECK-NOV-NEXT:    and a2, a2, s1
 ; CHECK-NOV-NEXT:    addi a1, a1, -1
 ; CHECK-NOV-NEXT:    and a1, a1, a0

 ; CHECK-V-NEXT:    sd s0, 16(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s1, 8(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s2, 0(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    .cfi_offset ra, -8
 ; CHECK-V-NEXT:    .cfi_offset s0, -16
 ; 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
 ; CHECK-V-NEXT:    snez a2, s2
 ; CHECK-V-NEXT:    addi a2, a2, -1
 ; CHECK-V-NEXT:    and a2, a2, s1
 ; CHECK-V-NEXT:    addi a1, a1, -1
 ; CHECK-V-NEXT:    and a0, a1, a0

 ; CHECK-NOV-NEXT:    sd s0, 16(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    sd s1, 8(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    sd s2, 0(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    .cfi_offset ra, -8
 ; CHECK-NOV-NEXT:    .cfi_offset s0, -16
 ; 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
 ; CHECK-NOV-NEXT:    blez s1, .LBB26_2
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    li a2, 1
 ; CHECK-NOV-NEXT:  .LBB26_2: # %entry
 ; CHECK-NOV-NEXT:    slti a3, a1, 1

 ; CHECK-V-NEXT:    sd s0, 16(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s1, 8(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s2, 0(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    .cfi_offset ra, -8
 ; CHECK-V-NEXT:    .cfi_offset s0, -16
 ; 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
 ; CHECK-V-NEXT:    blez s1, .LBB26_2
 ; CHECK-V-NEXT:  # %bb.1: # %entry
 ; CHECK-V-NEXT:    li a2, 1
 ; CHECK-V-NEXT:  .LBB26_2: # %entry
 ; CHECK-V-NEXT:    slti a4, a1, 1

 ; CHECK-NOV-NEXT:    .cfi_offset fs0, -48
 ; CHECK-NOV-NEXT:    .cfi_offset fs1, -56
 ; CHECK-NOV-NEXT:    .cfi_offset fs2, -64
 ; CHECK-NOV-NEXT:    lhu s1, 24(a1)
 ; CHECK-NOV-NEXT:    lhu s2, 0(a1)
 ; 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
 ; CHECK-NOV-NEXT:    addiw a4, a1, -1
 ; 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
 ; CHECK-NOV-NEXT:  .LBB33_3: # %entry
 ; CHECK-NOV-NEXT:    bge a3, a4, .LBB33_13
 ; CHECK-NOV-NEXT:  .LBB33_4: # %entry
 ; CHECK-NOV-NEXT:    bge a1, a3, .LBB33_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
 ; CHECK-NOV-NEXT:    ld s0, 48(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s1, 40(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s2, 32(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s3, 24(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs0, 16(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs1, 8(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs2, 0(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    addi sp, sp, 64
 ; CHECK-NOV-NEXT:    ret
 ; 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
 ; CHECK-NOV-NEXT:    mv a2, a4
 ; CHECK-NOV-NEXT:    blt a3, a4, .LBB33_4
 ; CHECK-NOV-NEXT:  .LBB33_13: # %entry
 ; CHECK-NOV-NEXT:    mv a3, a4
 ; CHECK-NOV-NEXT:    blt a1, a4, .LBB33_5
 ; CHECK-NOV-NEXT:  .LBB33_14: # %entry
 ; CHECK-NOV-NEXT:    lui a3, 524288
 ; 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
 ;
 ; CHECK-V-LABEL: stest_f16i32_mm:
 ; CHECK-V:       # %bb.0: # %entry
 ; CHECK-V-NEXT:    addi sp, sp, -48
 ; CHECK-V-NEXT:    .cfi_def_cfa_offset 48
 ; CHECK-V-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    .cfi_offset ra, -8
 ; CHECK-V-NEXT:    .cfi_offset s0, -16
 ; CHECK-V-NEXT:    .cfi_offset s1, -24
 ; CHECK-V-NEXT:    .cfi_offset s2, -32
 ; CHECK-V-NEXT:    csrr a1, vlenb
 ; CHECK-V-NEXT:    slli a1, a1, 1
 ; CHECK-V-NEXT:    sub sp, sp, a1
 ; CHECK-V-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x30, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 48 + 2 * vlenb
 ; CHECK-V-NEXT:    lhu s0, 24(a0)
 ; 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, 2, e64, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 3, e64, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    lui a0, 524288

 ; CHECK-NOV-NEXT:    .cfi_offset fs0, -48
 ; CHECK-NOV-NEXT:    .cfi_offset fs1, -56
 ; CHECK-NOV-NEXT:    .cfi_offset fs2, -64
 ; CHECK-NOV-NEXT:    lhu s1, 0(a1)
 ; CHECK-NOV-NEXT:    lhu s2, 24(a1)
 ; 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
 ; CHECK-NOV-NEXT:    srli a1, a1, 32
 ; 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
 ; CHECK-NOV-NEXT:  .LBB34_3: # %entry
 ; CHECK-NOV-NEXT:    bltu a3, a1, .LBB34_5
 ; CHECK-NOV-NEXT:  .LBB34_4: # %entry
 ; CHECK-NOV-NEXT:    mv a3, a1
 ; 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
 ; CHECK-NOV-NEXT:    ld s1, 40(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s2, 32(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s3, 24(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs0, 16(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs1, 8(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs2, 0(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    addi sp, sp, 64
 ; CHECK-NOV-NEXT:    ret
 ; 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
 ; CHECK-NOV-NEXT:    mv a2, a1
 ; CHECK-NOV-NEXT:    bgeu a3, a1, .LBB34_4
 ; CHECK-NOV-NEXT:    j .LBB34_5
 ;
 ; CHECK-V-LABEL: utesth_f16i32_mm:

 ; CHECK-V-NEXT:    csrr a1, vlenb
 ; CHECK-V-NEXT:    slli a1, a1, 1
 ; CHECK-V-NEXT:    sub sp, sp, a1
 ; CHECK-V-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x30, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 48 + 2 * vlenb
 ; CHECK-V-NEXT:    lhu s0, 24(a0)
 ; 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, 2, e64, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 3, e64, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    li a0, -1

 ; CHECK-NOV-NEXT:    .cfi_offset fs0, -48
 ; CHECK-NOV-NEXT:    .cfi_offset fs1, -56
 ; CHECK-NOV-NEXT:    .cfi_offset fs2, -64
 ; CHECK-NOV-NEXT:    lhu s1, 24(a1)
 ; CHECK-NOV-NEXT:    lhu s2, 0(a1)
 ; 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
 ; CHECK-NOV-NEXT:    srli a2, a2, 32
 ; 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
 ; CHECK-NOV-NEXT:  .LBB35_3: # %entry
 ; CHECK-NOV-NEXT:    blt a3, a2, .LBB35_5
 ; CHECK-NOV-NEXT:  .LBB35_4: # %entry
 ; CHECK-NOV-NEXT:    mv a3, a2
 ; CHECK-NOV-NEXT:  .LBB35_5: # %entry
 ; CHECK-NOV-NEXT:    sgtz a2, a3
 ; CHECK-NOV-NEXT:    negw a2, a2
 ; CHECK-NOV-NEXT:    and a2, a2, a3
 ; 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
 ; CHECK-NOV-NEXT:    sw a0, 12(s0)
 ; CHECK-NOV-NEXT:    sw a3, 8(s0)
 ; CHECK-NOV-NEXT:    sw a1, 4(s0)
 ; CHECK-NOV-NEXT:    sw a2, 0(s0)
 ; CHECK-NOV-NEXT:    ld ra, 56(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s0, 48(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s1, 40(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s2, 32(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s3, 24(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs0, 16(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs1, 8(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs2, 0(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    addi sp, sp, 64
 ; CHECK-NOV-NEXT:    ret
 ; 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
 ; CHECK-NOV-NEXT:    mv a1, a2
 ; CHECK-NOV-NEXT:    bge a3, a2, .LBB35_4
 ; CHECK-NOV-NEXT:    j .LBB35_5
 ;
 ; CHECK-V-LABEL: ustest_f16i32_mm:

 ; CHECK-V-NEXT:    csrr a1, vlenb
 ; CHECK-V-NEXT:    slli a1, a1, 1
 ; CHECK-V-NEXT:    sub sp, sp, a1
 ; CHECK-V-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x30, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 48 + 2 * vlenb
 ; CHECK-V-NEXT:    lhu s0, 24(a0)
 ; 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, 2, e64, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 3, e64, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    li a0, -1

 ; CHECK-NOV-NEXT:    lhu s2, 0(a1)
 ; CHECK-NOV-NEXT:    lhu s3, 8(a1)
 ; CHECK-NOV-NEXT:    lhu s4, 16(a1)
 ; CHECK-NOV-NEXT:    lhu s5, 24(a1)
 ; CHECK-NOV-NEXT:    lhu s6, 32(a1)
 ; 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
 ; CHECK-NOV-NEXT:    addiw a7, a7, -1
 ; 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
 ; CHECK-NOV-NEXT:  .LBB42_3: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs3, rtz
 ; CHECK-NOV-NEXT:    bge a2, a7, .LBB42_21
 ; CHECK-NOV-NEXT:  .LBB42_4: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a4, fs2, rtz

 ; CHECK-NOV-NEXT:    bge a7, a4, .LBB42_28
 ; CHECK-NOV-NEXT:  .LBB42_11: # %entry
 ; CHECK-NOV-NEXT:    bge a7, a3, .LBB42_29
 ; CHECK-NOV-NEXT:  .LBB42_12: # %entry
 ; CHECK-NOV-NEXT:    bge a7, a2, .LBB42_30
 ; 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)
 ; CHECK-NOV-NEXT:    sh a4, 4(s0)
 ; CHECK-NOV-NEXT:    sh a5, 2(s0)
 ; CHECK-NOV-NEXT:    sh a6, 0(s0)
 ; CHECK-NOV-NEXT:    ld ra, 120(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s0, 112(sp) # 8-byte Folded Reload

 ; CHECK-NOV-NEXT:    fld fs4, 16(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs5, 8(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs6, 0(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    addi sp, sp, 128
 ; CHECK-NOV-NEXT:    ret
 ; 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
 ; CHECK-NOV-NEXT:    mv a1, a7
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs3, rtz
 ; CHECK-NOV-NEXT:    blt a2, a7, .LBB42_4
 ; CHECK-NOV-NEXT:  .LBB42_21: # %entry
 ; CHECK-NOV-NEXT:    mv a2, a7

 ; CHECK-NOV-NEXT:  .LBB42_29: # %entry
 ; CHECK-NOV-NEXT:    lui a3, 1048568
 ; CHECK-NOV-NEXT:    blt a7, a2, .LBB42_13
 ; CHECK-NOV-NEXT:  .LBB42_30: # %entry
 ; CHECK-NOV-NEXT:    lui a2, 1048568
 ; 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
 ;
 ; CHECK-V-LABEL: stest_f16i16_mm:
 ; CHECK-V:       # %bb.0: # %entry
 ; CHECK-V-NEXT:    addi sp, sp, -80
 ; CHECK-V-NEXT:    .cfi_def_cfa_offset 80
 ; CHECK-V-NEXT:    sd ra, 72(sp) # 8-byte Folded Spill

 ; CHECK-V-NEXT:    lhu s0, 56(a0)
 ; CHECK-V-NEXT:    lhu s1, 48(a0)
 ; CHECK-V-NEXT:    lhu s2, 40(a0)
 ; CHECK-V-NEXT:    lhu s3, 32(a0)
 ; CHECK-V-NEXT:    lhu s4, 24(a0)
 ; 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, 2, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 3, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 4, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 5, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 6, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 7, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 7
 ; CHECK-V-NEXT:    lui a0, 8

 ; CHECK-NOV-NEXT:    lhu s2, 56(a1)
 ; CHECK-NOV-NEXT:    lhu s3, 48(a1)
 ; CHECK-NOV-NEXT:    lhu s4, 40(a1)
 ; CHECK-NOV-NEXT:    lhu s5, 32(a1)
 ; CHECK-NOV-NEXT:    lhu s6, 24(a1)
 ; 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
 ; CHECK-NOV-NEXT:    lui a1, 16
 ; CHECK-NOV-NEXT:    addiw a1, a1, -1
 ; CHECK-NOV-NEXT:    bltu a0, a1, .LBB43_2
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; 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
 ; CHECK-NOV-NEXT:    bltu a2, a1, .LBB43_6
 ; CHECK-NOV-NEXT:  # %bb.5: # %entry
 ; CHECK-NOV-NEXT:    mv a2, a1
 ; CHECK-NOV-NEXT:  .LBB43_6: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a5, fs2, rtz

 ; CHECK-NOV-NEXT:    mv a7, a1
 ; CHECK-NOV-NEXT:  .LBB43_14: # %entry
 ; CHECK-NOV-NEXT:    sh a7, 14(s0)
 ; CHECK-NOV-NEXT:    sh a6, 12(s0)
 ; CHECK-NOV-NEXT:    sh a5, 10(s0)
 ; 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
 ; CHECK-NOV-NEXT:    ld s1, 104(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s2, 96(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s3, 88(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s4, 80(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s5, 72(sp) # 8-byte Folded Reload

 ; CHECK-V-NEXT:    lhu s0, 56(a0)
 ; CHECK-V-NEXT:    lhu s1, 48(a0)
 ; CHECK-V-NEXT:    lhu s2, 40(a0)
 ; CHECK-V-NEXT:    lhu s3, 32(a0)
 ; CHECK-V-NEXT:    lhu s4, 24(a0)
 ; 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, 2, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 3, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 4, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 5, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 6, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 7, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 7
 ; CHECK-V-NEXT:    lui a0, 16

 ; CHECK-NOV-NEXT:    lhu s2, 0(a1)
 ; CHECK-NOV-NEXT:    lhu s3, 8(a1)
 ; CHECK-NOV-NEXT:    lhu s4, 16(a1)
 ; CHECK-NOV-NEXT:    lhu s5, 24(a1)
 ; CHECK-NOV-NEXT:    lhu s6, 32(a1)
 ; 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
 ; CHECK-NOV-NEXT:    addiw a3, a3, -1
 ; 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
 ; CHECK-NOV-NEXT:  .LBB44_3: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a4, fs3, rtz
 ; CHECK-NOV-NEXT:    bge a2, a3, .LBB44_13
 ; CHECK-NOV-NEXT:  .LBB44_4: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a5, fs2, rtz

 ; CHECK-NOV-NEXT:    negw a7, a7
 ; CHECK-NOV-NEXT:    and a4, a7, a4
 ; CHECK-NOV-NEXT:    sgtz a7, a2
 ; CHECK-NOV-NEXT:    negw a7, a7
 ; CHECK-NOV-NEXT:    and a2, a7, a2
 ; 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
 ; CHECK-NOV-NEXT:    sh a0, 14(s0)
 ; CHECK-NOV-NEXT:    sh a7, 12(s0)
 ; CHECK-NOV-NEXT:    sh a1, 10(s0)
 ; CHECK-NOV-NEXT:    sh a2, 8(s0)
 ; CHECK-NOV-NEXT:    sh a4, 6(s0)

 ; CHECK-NOV-NEXT:    fld fs4, 16(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs5, 8(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    fld fs6, 0(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    addi sp, sp, 128
 ; CHECK-NOV-NEXT:    ret
 ; 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
 ; CHECK-NOV-NEXT:    mv a1, a3
 ; CHECK-NOV-NEXT:    fcvt.l.s a4, fs3, rtz
 ; CHECK-NOV-NEXT:    blt a2, a3, .LBB44_4
 ; CHECK-NOV-NEXT:  .LBB44_13: # %entry
 ; CHECK-NOV-NEXT:    mv a2, a3

 ; CHECK-V-NEXT:    lhu s0, 56(a0)
 ; CHECK-V-NEXT:    lhu s1, 48(a0)
 ; CHECK-V-NEXT:    lhu s2, 40(a0)
 ; CHECK-V-NEXT:    lhu s3, 32(a0)
 ; CHECK-V-NEXT:    lhu s4, 24(a0)
 ; 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, 2, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 3, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 4, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 5, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 6, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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, 7, e32, m2, tu, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; 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
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 7
 ; CHECK-V-NEXT:    lui a0, 16

 ; CHECK-NOV-NEXT:    sd s0, 16(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    sd s1, 8(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    sd s2, 0(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    .cfi_offset ra, -8
 ; CHECK-NOV-NEXT:    .cfi_offset s0, -16
 ; 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
 ; CHECK-NOV-NEXT:    li a0, -1
 ; CHECK-NOV-NEXT:    srli a3, a0, 1
 ; CHECK-NOV-NEXT:    mv a4, a2
 ; CHECK-NOV-NEXT:    bgez a1, .LBB51_15
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry

 ; CHECK-V-NEXT:    sd s0, 16(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s1, 8(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s2, 0(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    .cfi_offset ra, -8
 ; CHECK-V-NEXT:    .cfi_offset s0, -16
 ; 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
 ; CHECK-V-NEXT:    srli a3, a2, 1
 ; CHECK-V-NEXT:    mv a4, a0
 ; CHECK-V-NEXT:    bgez a1, .LBB51_15
 ; CHECK-V-NEXT:  # %bb.1: # %entry
 ; CHECK-V-NEXT:    bgeu a0, a3, .LBB51_16

 ; CHECK-NOV-NEXT:    sd s0, 16(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    sd s1, 8(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    sd s2, 0(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    .cfi_offset ra, -8
 ; CHECK-NOV-NEXT:    .cfi_offset s0, -16
 ; 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
 ; CHECK-NOV-NEXT:    addi a2, a2, -1
 ; CHECK-NOV-NEXT:    and a0, a2, a0
 ; CHECK-NOV-NEXT:    addi a1, a1, -1
 ; CHECK-NOV-NEXT:    seqz a1, a1
 ; CHECK-NOV-NEXT:    addi a1, a1, -1

 ; CHECK-V-NEXT:    sd s0, 16(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s1, 8(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s2, 0(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    .cfi_offset ra, -8
 ; CHECK-V-NEXT:    .cfi_offset s0, -16
 ; 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
 ; CHECK-V-NEXT:    addi a2, a2, -1
 ; CHECK-V-NEXT:    and a0, a2, a0
 ; CHECK-V-NEXT:    addi a1, a1, -1
 ; CHECK-V-NEXT:    seqz a1, a1
 ; CHECK-V-NEXT:    addi a1, a1, -1

 ; CHECK-NOV-NEXT:    sd s0, 16(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    sd s1, 8(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    sd s2, 0(sp) # 8-byte Folded Spill
 ; CHECK-NOV-NEXT:    .cfi_offset ra, -8
 ; CHECK-NOV-NEXT:    .cfi_offset s0, -16
 ; 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
 ; CHECK-NOV-NEXT:    blez a1, .LBB53_2
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    li a2, 1
 ; CHECK-NOV-NEXT:  .LBB53_2: # %entry
 ; CHECK-NOV-NEXT:    mv a3, s1

 ; CHECK-V-NEXT:    sd s0, 16(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s1, 8(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    sd s2, 0(sp) # 8-byte Folded Spill
 ; CHECK-V-NEXT:    .cfi_offset ra, -8
 ; CHECK-V-NEXT:    .cfi_offset s0, -16
 ; 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
 ; CHECK-V-NEXT:    blez a1, .LBB53_2
 ; CHECK-V-NEXT:  # %bb.1: # %entry
 ; CHECK-V-NEXT:    li a2, 1
 ; CHECK-V-NEXT:  .LBB53_2: # %entry
 ; CHECK-V-NEXT:    mv a3, s1