Index: llvm/lib/Target/RISCV/RISCVISelLowering.cpp
===================================================================
--- llvm/lib/Target/RISCV/RISCVISelLowering.cpp
+++ llvm/lib/Target/RISCV/RISCVISelLowering.cpp
@@ -352,7 +352,7 @@
 
   static const unsigned FPOpToExpand[] = {
       ISD::FSIN, ISD::FCOS,       ISD::FSINCOS,   ISD::FPOW,
-      ISD::FREM, ISD::FP16_TO_FP, ISD::FP_TO_FP16};
+      ISD::FREM};
 
   static const unsigned FPRndMode[] = {
       ISD::FCEIL, ISD::FFLOOR, ISD::FTRUNC, ISD::FRINT, ISD::FROUND,
@@ -427,6 +427,8 @@
     setLoadExtAction(ISD::EXTLOAD, MVT::f32, MVT::f16, Expand);
     setTruncStoreAction(MVT::f32, MVT::f16, Expand);
     setOperationAction(ISD::IS_FPCLASS, MVT::f32, Custom);
+    setOperationAction(ISD::FP_TO_FP16, MVT::f32, Custom);
+    setOperationAction(ISD::FP16_TO_FP, MVT::f32, Custom);
 
     if (Subtarget.hasStdExtZfa())
       setOperationAction(ISD::FNEARBYINT, MVT::f32, Legal);
@@ -461,6 +463,8 @@
     setLoadExtAction(ISD::EXTLOAD, MVT::f64, MVT::f16, Expand);
     setTruncStoreAction(MVT::f64, MVT::f16, Expand);
     setOperationAction(ISD::IS_FPCLASS, MVT::f64, Custom);
+    setOperationAction(ISD::FP_TO_FP16, MVT::f64, Custom);
+    setOperationAction(ISD::FP16_TO_FP, MVT::f64, Expand);
   }
 
   if (Subtarget.is64Bit()) {
@@ -4917,6 +4921,35 @@
   case ISD::FP_TO_SINT_SAT:
   case ISD::FP_TO_UINT_SAT:
     return lowerFP_TO_INT_SAT(Op, DAG, Subtarget);
+  case ISD::FP_TO_FP16: {
+    // Custom lower to ensure the libcall return is passed in an FPR on hard
+    // float ABIs.
+    assert(Subtarget.hasStdExtFOrZfinx() && "Unexpected custom 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:
Index: llvm/test/CodeGen/RISCV/calling-conv-half.ll
===================================================================
--- llvm/test/CodeGen/RISCV/calling-conv-half.ll
+++ llvm/test/CodeGen/RISCV/calling-conv-half.ll
@@ -83,7 +83,6 @@
 ; RV32-ILP32F-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-ILP32F-NEXT:    sw s0, 8(sp) # 4-byte Folded Spill
 ; RV32-ILP32F-NEXT:    mv s0, a0
-; RV32-ILP32F-NEXT:    fmv.x.w a0, fa0
 ; RV32-ILP32F-NEXT:    call __extendhfsf2@plt
 ; RV32-ILP32F-NEXT:    fcvt.w.s a0, fa0, rtz
 ; RV32-ILP32F-NEXT:    add a0, s0, a0
@@ -99,6 +98,7 @@
 ; RV64-LP64F-NEXT:    sd s0, 0(sp) # 8-byte Folded Spill
 ; RV64-LP64F-NEXT:    mv s0, a0
 ; RV64-LP64F-NEXT:    fmv.x.w a0, fa0
+; RV64-LP64F-NEXT:    fmv.w.x fa0, a0
 ; RV64-LP64F-NEXT:    call __extendhfsf2@plt
 ; RV64-LP64F-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-LP64F-NEXT:    addw a0, s0, a0
@@ -292,7 +292,6 @@
 ; RV32-ILP32F-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-ILP32F-NEXT:    sw s0, 8(sp) # 4-byte Folded Spill
 ; RV32-ILP32F-NEXT:    mv s0, a7
-; RV32-ILP32F-NEXT:    fmv.x.w a0, fa0
 ; RV32-ILP32F-NEXT:    call __extendhfsf2@plt
 ; RV32-ILP32F-NEXT:    fcvt.w.s a0, fa0, rtz
 ; RV32-ILP32F-NEXT:    add a0, s0, a0
@@ -308,6 +307,7 @@
 ; RV64-LP64F-NEXT:    sd s0, 0(sp) # 8-byte Folded Spill
 ; RV64-LP64F-NEXT:    mv s0, a7
 ; RV64-LP64F-NEXT:    fmv.x.w a0, fa0
+; RV64-LP64F-NEXT:    fmv.w.x fa0, a0
 ; RV64-LP64F-NEXT:    call __extendhfsf2@plt
 ; RV64-LP64F-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-LP64F-NEXT:    addw a0, s0, a0
@@ -602,7 +602,6 @@
 ; RV32-ILP32F-NEXT:    addi sp, sp, -16
 ; RV32-ILP32F-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-ILP32F-NEXT:    call callee_half_ret@plt
-; RV32-ILP32F-NEXT:    fmv.x.w a0, fa0
 ; RV32-ILP32F-NEXT:    call __extendhfsf2@plt
 ; RV32-ILP32F-NEXT:    fcvt.w.s a0, fa0, rtz
 ; RV32-ILP32F-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
@@ -615,6 +614,7 @@
 ; RV64-LP64F-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-LP64F-NEXT:    call callee_half_ret@plt
 ; RV64-LP64F-NEXT:    fmv.x.w a0, fa0
+; RV64-LP64F-NEXT:    fmv.w.x fa0, a0
 ; RV64-LP64F-NEXT:    call __extendhfsf2@plt
 ; RV64-LP64F-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-LP64F-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
Index: llvm/test/CodeGen/RISCV/copysign-casts.ll
===================================================================
--- llvm/test/CodeGen/RISCV/copysign-casts.ll
+++ llvm/test/CodeGen/RISCV/copysign-casts.ll
@@ -163,7 +163,7 @@
 ; RV32IFD-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32IFD-NEXT:    fsd fs0, 0(sp) # 8-byte Folded Spill
 ; RV32IFD-NEXT:    fmv.d fs0, fa0
-; RV32IFD-NEXT:    fmv.x.w a0, fa1
+; RV32IFD-NEXT:    fmv.s fa0, fa1
 ; RV32IFD-NEXT:    call __extendhfsf2@plt
 ; RV32IFD-NEXT:    fcvt.d.s fa5, fa0
 ; RV32IFD-NEXT:    fsgnj.d fa0, fs0, fa5
@@ -179,6 +179,7 @@
 ; RV64IFD-NEXT:    fsd fs0, 0(sp) # 8-byte Folded Spill
 ; RV64IFD-NEXT:    fmv.d fs0, fa0
 ; RV64IFD-NEXT:    fmv.x.w a0, fa1
+; RV64IFD-NEXT:    fmv.w.x fa0, a0
 ; RV64IFD-NEXT:    call __extendhfsf2@plt
 ; RV64IFD-NEXT:    fcvt.d.s fa5, fa0
 ; RV64IFD-NEXT:    fsgnj.d fa0, fs0, fa5
@@ -264,7 +265,7 @@
 ; RV32IF-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32IF-NEXT:    fsw fs0, 8(sp) # 4-byte Folded Spill
 ; RV32IF-NEXT:    fmv.s fs0, fa0
-; RV32IF-NEXT:    fmv.x.w a0, fa1
+; RV32IF-NEXT:    fmv.s fa0, fa1
 ; RV32IF-NEXT:    call __extendhfsf2@plt
 ; RV32IF-NEXT:    fsgnj.s fa0, fs0, fa0
 ; RV32IF-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
@@ -278,7 +279,7 @@
 ; RV32IFD-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32IFD-NEXT:    fsd fs0, 0(sp) # 8-byte Folded Spill
 ; RV32IFD-NEXT:    fmv.s fs0, fa0
-; RV32IFD-NEXT:    fmv.x.w a0, fa1
+; RV32IFD-NEXT:    fmv.s fa0, fa1
 ; RV32IFD-NEXT:    call __extendhfsf2@plt
 ; RV32IFD-NEXT:    fsgnj.s fa0, fs0, fa0
 ; RV32IFD-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
@@ -293,6 +294,7 @@
 ; RV64IFD-NEXT:    fsd fs0, 0(sp) # 8-byte Folded Spill
 ; RV64IFD-NEXT:    fmv.s fs0, fa0
 ; RV64IFD-NEXT:    fmv.x.w a0, fa1
+; RV64IFD-NEXT:    fmv.w.x fa0, a0
 ; RV64IFD-NEXT:    call __extendhfsf2@plt
 ; RV64IFD-NEXT:    fsgnj.s fa0, fs0, fa0
 ; RV64IFD-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
Index: llvm/test/CodeGen/RISCV/fp16-promote.ll
===================================================================
--- llvm/test/CodeGen/RISCV/fp16-promote.ll
+++ llvm/test/CodeGen/RISCV/fp16-promote.ll
@@ -15,8 +15,14 @@
 define float @test_fpextend_float(ptr %p) nounwind {
 ; CHECK-LABEL: test_fpextend_float:
 ; CHECK:       # %bb.0:
+; CHECK-NEXT:    addi sp, sp, -16
+; CHECK-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; CHECK-NEXT:    lhu a0, 0(a0)
-; CHECK-NEXT:    tail __extendhfsf2@plt
+; CHECK-NEXT:    fmv.w.x fa0, a0
+; CHECK-NEXT:    call __extendhfsf2@plt
+; CHECK-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
+; CHECK-NEXT:    addi sp, sp, 16
+; CHECK-NEXT:    ret
   %a = load half, ptr %p
   %r = fpext half %a to float
   ret float %r
@@ -28,6 +34,7 @@
 ; CHECK-NEXT:    addi sp, sp, -16
 ; CHECK-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; CHECK-NEXT:    lhu a0, 0(a0)
+; CHECK-NEXT:    fmv.w.x fa0, a0
 ; CHECK-NEXT:    call __extendhfsf2@plt
 ; CHECK-NEXT:    fcvt.d.s fa0, fa0
 ; CHECK-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
@@ -46,6 +53,7 @@
 ; CHECK-NEXT:    sw s0, 8(sp) # 4-byte Folded Spill
 ; CHECK-NEXT:    mv s0, a0
 ; CHECK-NEXT:    call __truncsfhf2@plt
+; CHECK-NEXT:    fmv.x.w a0, fa0
 ; CHECK-NEXT:    sh a0, 0(s0)
 ; CHECK-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; CHECK-NEXT:    lw s0, 8(sp) # 4-byte Folded Reload
@@ -64,6 +72,7 @@
 ; CHECK-NEXT:    sw s0, 8(sp) # 4-byte Folded Spill
 ; CHECK-NEXT:    mv s0, a0
 ; CHECK-NEXT:    call __truncdfhf2@plt
+; CHECK-NEXT:    fmv.x.w a0, fa0
 ; CHECK-NEXT:    sh a0, 0(s0)
 ; CHECK-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; CHECK-NEXT:    lw s0, 8(sp) # 4-byte Folded Reload
@@ -80,22 +89,25 @@
 ; CHECK-NEXT:    addi sp, sp, -32
 ; CHECK-NEXT:    sw ra, 28(sp) # 4-byte Folded Spill
 ; CHECK-NEXT:    sw s0, 24(sp) # 4-byte Folded Spill
-; CHECK-NEXT:    sw s1, 20(sp) # 4-byte Folded Spill
-; CHECK-NEXT:    fsd fs0, 8(sp) # 8-byte Folded Spill
+; CHECK-NEXT:    fsd fs0, 16(sp) # 8-byte Folded Spill
+; CHECK-NEXT:    fsd fs1, 8(sp) # 8-byte Folded Spill
 ; CHECK-NEXT:    mv s0, a0
-; CHECK-NEXT:    lhu s1, 0(a0)
-; CHECK-NEXT:    lhu a0, 0(a1)
+; CHECK-NEXT:    lhu a0, 0(a0)
+; CHECK-NEXT:    lhu a1, 0(a1)
+; CHECK-NEXT:    fmv.w.x fs0, a0
+; CHECK-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NEXT:    call __extendhfsf2@plt
-; CHECK-NEXT:    fmv.s fs0, fa0
-; CHECK-NEXT:    mv a0, s1
+; CHECK-NEXT:    fmv.s fs1, fa0
+; CHECK-NEXT:    fmv.s fa0, fs0
 ; CHECK-NEXT:    call __extendhfsf2@plt
-; CHECK-NEXT:    fadd.s fa0, fa0, fs0
+; CHECK-NEXT:    fadd.s fa0, fa0, fs1
 ; CHECK-NEXT:    call __truncsfhf2@plt
+; CHECK-NEXT:    fmv.x.w a0, fa0
 ; CHECK-NEXT:    sh a0, 0(s0)
 ; CHECK-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; CHECK-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
-; CHECK-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
-; CHECK-NEXT:    fld fs0, 8(sp) # 8-byte Folded Reload
+; CHECK-NEXT:    fld fs0, 16(sp) # 8-byte Folded Reload
+; CHECK-NEXT:    fld fs1, 8(sp) # 8-byte Folded Reload
 ; CHECK-NEXT:    addi sp, sp, 32
 ; CHECK-NEXT:    ret
   %a = load half, ptr %p
@@ -111,22 +123,25 @@
 ; CHECK-NEXT:    addi sp, sp, -32
 ; CHECK-NEXT:    sw ra, 28(sp) # 4-byte Folded Spill
 ; CHECK-NEXT:    sw s0, 24(sp) # 4-byte Folded Spill
-; CHECK-NEXT:    sw s1, 20(sp) # 4-byte Folded Spill
-; CHECK-NEXT:    fsd fs0, 8(sp) # 8-byte Folded Spill
+; CHECK-NEXT:    fsd fs0, 16(sp) # 8-byte Folded Spill
+; CHECK-NEXT:    fsd fs1, 8(sp) # 8-byte Folded Spill
 ; CHECK-NEXT:    mv s0, a0
-; CHECK-NEXT:    lhu s1, 0(a0)
-; CHECK-NEXT:    lhu a0, 0(a1)
+; CHECK-NEXT:    lhu a0, 0(a0)
+; CHECK-NEXT:    lhu a1, 0(a1)
+; CHECK-NEXT:    fmv.w.x fs0, a0
+; CHECK-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NEXT:    call __extendhfsf2@plt
-; CHECK-NEXT:    fmv.s fs0, fa0
-; CHECK-NEXT:    mv a0, s1
+; CHECK-NEXT:    fmv.s fs1, fa0
+; CHECK-NEXT:    fmv.s fa0, fs0
 ; CHECK-NEXT:    call __extendhfsf2@plt
-; CHECK-NEXT:    fmul.s fa0, fa0, fs0
+; CHECK-NEXT:    fmul.s fa0, fa0, fs1
 ; CHECK-NEXT:    call __truncsfhf2@plt
+; CHECK-NEXT:    fmv.x.w a0, fa0
 ; CHECK-NEXT:    sh a0, 0(s0)
 ; CHECK-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; CHECK-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
-; CHECK-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
-; CHECK-NEXT:    fld fs0, 8(sp) # 8-byte Folded Reload
+; CHECK-NEXT:    fld fs0, 16(sp) # 8-byte Folded Reload
+; CHECK-NEXT:    fld fs1, 8(sp) # 8-byte Folded Reload
 ; CHECK-NEXT:    addi sp, sp, 32
 ; CHECK-NEXT:    ret
   %a = load half, ptr %p
Index: llvm/test/CodeGen/RISCV/fpclamptosat.ll
===================================================================
--- llvm/test/CodeGen/RISCV/fpclamptosat.ll
+++ llvm/test/CodeGen/RISCV/fpclamptosat.ll
@@ -355,7 +355,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    call __fixsfdi@plt
 ; RV32-NEXT:    lui a2, 524288
@@ -397,6 +396,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-NEXT:    lui a1, 524288
@@ -429,7 +429,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    call __fixunssfdi@plt
 ; RV32-NEXT:    sltiu a2, a0, -1
@@ -448,6 +447,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; RV64-NEXT:    li a1, -1
@@ -474,7 +474,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    call __fixsfdi@plt
 ; RV32-NEXT:    beqz a1, .LBB8_2
@@ -508,6 +507,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-NEXT:    li a1, -1
@@ -891,7 +891,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    fcvt.w.s a0, fa0, rtz
 ; RV32-NEXT:    lui a1, 8
@@ -916,6 +915,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-NEXT:    lui a1, 8
@@ -949,7 +949,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    fcvt.wu.s a0, fa0, rtz
 ; RV32-NEXT:    lui a1, 16
@@ -969,6 +968,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; RV64-NEXT:    lui a1, 16
@@ -995,7 +995,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    fcvt.w.s a0, fa0, rtz
 ; RV32-NEXT:    lui a1, 16
@@ -1018,6 +1017,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-NEXT:    lui a1, 16
@@ -1666,7 +1666,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 32
 ; RV32-NEXT:    sw ra, 28(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    addi a0, sp, 8
 ; RV32-NEXT:    call __fixsfti@plt
@@ -1729,6 +1728,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    call __fixsfti@plt
 ; RV64-NEXT:    li a2, -1
@@ -1779,7 +1779,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 32
 ; RV32-NEXT:    sw ra, 28(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    addi a0, sp, 8
 ; RV32-NEXT:    call __fixunssfti@plt
@@ -1808,6 +1807,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    call __fixunssfti@plt
 ; RV64-NEXT:    snez a1, a1
@@ -1831,7 +1831,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 32
 ; RV32-NEXT:    sw ra, 28(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    addi a0, sp, 8
 ; RV32-NEXT:    call __fixsfti@plt
@@ -1886,6 +1885,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    call __fixsfti@plt
 ; RV64-NEXT:    slti a2, a1, 1
@@ -2256,7 +2256,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    call __fixsfdi@plt
 ; RV32-NEXT:    lui a2, 524288
@@ -2308,6 +2307,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-NEXT:    lui a1, 524288
@@ -2338,7 +2338,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    call __fixunssfdi@plt
 ; RV32-NEXT:    seqz a1, a1
@@ -2355,6 +2354,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; RV64-NEXT:    li a1, -1
@@ -2380,7 +2380,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    call __fixsfdi@plt
 ; RV32-NEXT:    slti a2, a1, 0
@@ -2405,6 +2404,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-NEXT:    li a1, -1
@@ -2776,7 +2776,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    fcvt.w.s a0, fa0, rtz
 ; RV32-NEXT:    lui a1, 8
@@ -2801,6 +2800,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-NEXT:    lui a1, 8
@@ -2832,7 +2832,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    fcvt.wu.s a0, fa0, rtz
 ; RV32-NEXT:    lui a1, 16
@@ -2852,6 +2851,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; RV64-NEXT:    sext.w a0, a0
@@ -2878,7 +2878,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    fcvt.w.s a0, fa0, rtz
 ; RV32-NEXT:    lui a1, 16
@@ -2901,6 +2900,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    fcvt.l.s a0, fa0, rtz
 ; RV64-NEXT:    lui a1, 16
@@ -3606,7 +3606,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 32
 ; RV32-NEXT:    sw ra, 28(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    addi a0, sp, 8
 ; RV32-NEXT:    call __fixsfti@plt
@@ -3699,6 +3698,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    call __fixsfti@plt
 ; RV64-NEXT:    li a2, -1
@@ -3756,7 +3756,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 32
 ; RV32-NEXT:    sw ra, 28(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    addi a0, sp, 8
 ; RV32-NEXT:    call __fixunssfti@plt
@@ -3786,6 +3785,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    call __fixunssfti@plt
 ; RV64-NEXT:    snez a2, a1
@@ -3812,7 +3812,6 @@
 ; RV32-NEXT:    .cfi_def_cfa_offset 32
 ; RV32-NEXT:    sw ra, 28(sp) # 4-byte Folded Spill
 ; RV32-NEXT:    .cfi_offset ra, -4
-; RV32-NEXT:    fmv.x.w a0, fa0
 ; RV32-NEXT:    call __extendhfsf2@plt
 ; RV32-NEXT:    addi a0, sp, 8
 ; RV32-NEXT:    call __fixsfti@plt
@@ -3854,6 +3853,7 @@
 ; RV64-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64-NEXT:    .cfi_offset ra, -8
 ; RV64-NEXT:    fmv.x.w a0, fa0
+; RV64-NEXT:    fmv.w.x fa0, a0
 ; RV64-NEXT:    call __extendhfsf2@plt
 ; RV64-NEXT:    call __fixsfti@plt
 ; RV64-NEXT:    mv a2, a1
Index: llvm/test/CodeGen/RISCV/half-fpext.ll
===================================================================
--- llvm/test/CodeGen/RISCV/half-fpext.ll
+++ llvm/test/CodeGen/RISCV/half-fpext.ll
@@ -11,8 +11,6 @@
 ; 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
@@ -22,8 +20,6 @@
 ; 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
@@ -31,7 +27,12 @@
 ;
 ; RV32ID-ILP32-LABEL: half_to_float:
 ; RV32ID-ILP32:       # %bb.0:
-; RV32ID-ILP32-NEXT:    tail __extendhfsf2@plt
+; RV32ID-ILP32-NEXT:    addi sp, sp, -16
+; RV32ID-ILP32-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
+; RV32ID-ILP32-NEXT:    call __extendhfsf2@plt
+; RV32ID-ILP32-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
+; RV32ID-ILP32-NEXT:    addi sp, sp, 16
+; RV32ID-ILP32-NEXT:    ret
 ;
 ; RV64ID-LP64-LABEL: half_to_float:
 ; RV64ID-LP64:       # %bb.0:
@@ -44,13 +45,23 @@
 ;
 ; RV32ID-ILP32D-LABEL: half_to_float:
 ; RV32ID-ILP32D:       # %bb.0:
-; RV32ID-ILP32D-NEXT:    fmv.x.w a0, fa0
-; RV32ID-ILP32D-NEXT:    tail __extendhfsf2@plt
+; RV32ID-ILP32D-NEXT:    addi sp, sp, -16
+; RV32ID-ILP32D-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
+; RV32ID-ILP32D-NEXT:    call __extendhfsf2@plt
+; RV32ID-ILP32D-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
+; RV32ID-ILP32D-NEXT:    addi sp, sp, 16
+; RV32ID-ILP32D-NEXT:    ret
 ;
 ; RV64ID-LP64D-LABEL: half_to_float:
 ; RV64ID-LP64D:       # %bb.0:
+; RV64ID-LP64D-NEXT:    addi sp, sp, -16
+; RV64ID-LP64D-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64ID-LP64D-NEXT:    fmv.x.w a0, fa0
-; RV64ID-LP64D-NEXT:    tail __extendhfsf2@plt
+; RV64ID-LP64D-NEXT:    fmv.w.x fa0, a0
+; RV64ID-LP64D-NEXT:    call __extendhfsf2@plt
+; RV64ID-LP64D-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
+; RV64ID-LP64D-NEXT:    addi sp, sp, 16
+; RV64ID-LP64D-NEXT:    ret
   %1 = fpext half %a to float
   ret float %1
 }
@@ -60,8 +71,6 @@
 ; 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
@@ -72,11 +81,7 @@
 ; 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
@@ -112,7 +117,6 @@
 ; RV32ID-ILP32D:       # %bb.0:
 ; RV32ID-ILP32D-NEXT:    addi sp, sp, -16
 ; RV32ID-ILP32D-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
-; RV32ID-ILP32D-NEXT:    fmv.x.w a0, fa0
 ; RV32ID-ILP32D-NEXT:    call __extendhfsf2@plt
 ; RV32ID-ILP32D-NEXT:    fcvt.d.s fa0, fa0
 ; RV32ID-ILP32D-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
@@ -124,6 +128,7 @@
 ; RV64ID-LP64D-NEXT:    addi sp, sp, -16
 ; RV64ID-LP64D-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64ID-LP64D-NEXT:    fmv.x.w a0, fa0
+; RV64ID-LP64D-NEXT:    fmv.w.x fa0, a0
 ; RV64ID-LP64D-NEXT:    call __extendhfsf2@plt
 ; RV64ID-LP64D-NEXT:    fcvt.d.s fa0, fa0
 ; RV64ID-LP64D-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
Index: llvm/test/CodeGen/RISCV/half-fptrunc.ll
===================================================================
--- llvm/test/CodeGen/RISCV/half-fptrunc.ll
+++ llvm/test/CodeGen/RISCV/half-fptrunc.ll
@@ -52,6 +52,7 @@
 ; RV32ID-ILP32D-NEXT:    addi sp, sp, -16
 ; RV32ID-ILP32D-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32ID-ILP32D-NEXT:    call __truncsfhf2@plt
+; RV32ID-ILP32D-NEXT:    fmv.x.w a0, fa0
 ; RV32ID-ILP32D-NEXT:    lui a1, 1048560
 ; RV32ID-ILP32D-NEXT:    or a0, a0, a1
 ; RV32ID-ILP32D-NEXT:    fmv.w.x fa0, a0
@@ -64,6 +65,7 @@
 ; RV64ID-LP64D-NEXT:    addi sp, sp, -16
 ; RV64ID-LP64D-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64ID-LP64D-NEXT:    call __truncsfhf2@plt
+; RV64ID-LP64D-NEXT:    fmv.x.w a0, fa0
 ; RV64ID-LP64D-NEXT:    lui a1, 1048560
 ; RV64ID-LP64D-NEXT:    or a0, a0, a1
 ; RV64ID-LP64D-NEXT:    fmv.w.x fa0, a0
@@ -120,6 +122,7 @@
 ; RV32ID-ILP32D-NEXT:    addi sp, sp, -16
 ; RV32ID-ILP32D-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32ID-ILP32D-NEXT:    call __truncdfhf2@plt
+; RV32ID-ILP32D-NEXT:    fmv.x.w a0, fa0
 ; RV32ID-ILP32D-NEXT:    lui a1, 1048560
 ; RV32ID-ILP32D-NEXT:    or a0, a0, a1
 ; RV32ID-ILP32D-NEXT:    fmv.w.x fa0, a0
@@ -132,6 +135,7 @@
 ; RV64ID-LP64D-NEXT:    addi sp, sp, -16
 ; RV64ID-LP64D-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64ID-LP64D-NEXT:    call __truncdfhf2@plt
+; RV64ID-LP64D-NEXT:    fmv.x.w a0, fa0
 ; RV64ID-LP64D-NEXT:    lui a1, 1048560
 ; RV64ID-LP64D-NEXT:    or a0, a0, a1
 ; RV64ID-LP64D-NEXT:    fmv.w.x fa0, a0
Index: llvm/test/CodeGen/RISCV/libcall-tail-calls.ll
===================================================================
--- llvm/test/CodeGen/RISCV/libcall-tail-calls.ll
+++ llvm/test/CodeGen/RISCV/libcall-tail-calls.ll
@@ -120,10 +120,10 @@
 ; RV32IFD-ILP32D:       # %bb.0:
 ; RV32IFD-ILP32D-NEXT:    addi sp, sp, -16
 ; RV32IFD-ILP32D-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
-; RV32IFD-ILP32D-NEXT:    fmv.x.w a0, fa0
 ; RV32IFD-ILP32D-NEXT:    call __extendhfsf2@plt
 ; RV32IFD-ILP32D-NEXT:    call sinf@plt
 ; RV32IFD-ILP32D-NEXT:    call __truncsfhf2@plt
+; RV32IFD-ILP32D-NEXT:    fmv.x.w a0, fa0
 ; RV32IFD-ILP32D-NEXT:    lui a1, 1048560
 ; RV32IFD-ILP32D-NEXT:    or a0, a0, a1
 ; RV32IFD-ILP32D-NEXT:    fmv.w.x fa0, a0
@@ -135,10 +135,10 @@
 ; RV32IF-ILP32F:       # %bb.0:
 ; RV32IF-ILP32F-NEXT:    addi sp, sp, -16
 ; RV32IF-ILP32F-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
-; RV32IF-ILP32F-NEXT:    fmv.x.w a0, fa0
 ; RV32IF-ILP32F-NEXT:    call __extendhfsf2@plt
 ; RV32IF-ILP32F-NEXT:    call sinf@plt
 ; RV32IF-ILP32F-NEXT:    call __truncsfhf2@plt
+; RV32IF-ILP32F-NEXT:    fmv.x.w a0, fa0
 ; RV32IF-ILP32F-NEXT:    lui a1, 1048560
 ; RV32IF-ILP32F-NEXT:    or a0, a0, a1
 ; RV32IF-ILP32F-NEXT:    fmv.w.x fa0, a0
@@ -177,9 +177,11 @@
 ; RV64IFD-LP64D-NEXT:    addi sp, sp, -16
 ; RV64IFD-LP64D-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64IFD-LP64D-NEXT:    fmv.x.w a0, fa0
+; RV64IFD-LP64D-NEXT:    fmv.w.x fa0, a0
 ; RV64IFD-LP64D-NEXT:    call __extendhfsf2@plt
 ; RV64IFD-LP64D-NEXT:    call sinf@plt
 ; RV64IFD-LP64D-NEXT:    call __truncsfhf2@plt
+; RV64IFD-LP64D-NEXT:    fmv.x.w a0, fa0
 ; RV64IFD-LP64D-NEXT:    lui a1, 1048560
 ; RV64IFD-LP64D-NEXT:    or a0, a0, a1
 ; RV64IFD-LP64D-NEXT:    fmv.w.x fa0, a0
@@ -192,9 +194,11 @@
 ; RV64IF-LP64F-NEXT:    addi sp, sp, -16
 ; RV64IF-LP64F-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64IF-LP64F-NEXT:    fmv.x.w a0, fa0
+; RV64IF-LP64F-NEXT:    fmv.w.x fa0, a0
 ; RV64IF-LP64F-NEXT:    call __extendhfsf2@plt
 ; RV64IF-LP64F-NEXT:    call sinf@plt
 ; RV64IF-LP64F-NEXT:    call __truncsfhf2@plt
+; RV64IF-LP64F-NEXT:    fmv.x.w a0, fa0
 ; RV64IF-LP64F-NEXT:    lui a1, 1048560
 ; RV64IF-LP64F-NEXT:    or a0, a0, a1
 ; RV64IF-LP64F-NEXT:    fmv.w.x fa0, a0
Index: llvm/test/CodeGen/RISCV/rvv/fpclamptosat_vec.ll
===================================================================
--- llvm/test/CodeGen/RISCV/rvv/fpclamptosat_vec.ll
+++ llvm/test/CodeGen/RISCV/rvv/fpclamptosat_vec.ll
@@ -369,17 +369,17 @@
 ; CHECK-NOV-NEXT:    lhu s3, 8(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 16(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.l.s s2, fs2, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.l.s s1, fs2, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    lui a1, 524288
@@ -387,7 +387,7 @@
 ; CHECK-NOV-NEXT:    bge a0, a4, .LBB6_10
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs1, rtz
-; CHECK-NOV-NEXT:    bge s2, a4, .LBB6_11
+; CHECK-NOV-NEXT:    bge s1, a4, .LBB6_11
 ; CHECK-NOV-NEXT:  .LBB6_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    bge a2, a4, .LBB6_12
@@ -398,14 +398,14 @@
 ; CHECK-NOV-NEXT:  .LBB6_5: # %entry
 ; CHECK-NOV-NEXT:    bge a1, a2, .LBB6_15
 ; CHECK-NOV-NEXT:  .LBB6_6: # %entry
-; CHECK-NOV-NEXT:    bge a1, s2, .LBB6_16
+; CHECK-NOV-NEXT:    bge a1, s1, .LBB6_16
 ; CHECK-NOV-NEXT:  .LBB6_7: # %entry
 ; CHECK-NOV-NEXT:    blt a1, a0, .LBB6_9
 ; CHECK-NOV-NEXT:  .LBB6_8: # %entry
 ; CHECK-NOV-NEXT:    lui a0, 524288
 ; CHECK-NOV-NEXT:  .LBB6_9: # %entry
 ; CHECK-NOV-NEXT:    sw a0, 12(s0)
-; CHECK-NOV-NEXT:    sw s2, 8(s0)
+; CHECK-NOV-NEXT:    sw s1, 8(s0)
 ; CHECK-NOV-NEXT:    sw a2, 4(s0)
 ; CHECK-NOV-NEXT:    sw a3, 0(s0)
 ; CHECK-NOV-NEXT:    ld ra, 56(sp) # 8-byte Folded Reload
@@ -421,9 +421,9 @@
 ; CHECK-NOV-NEXT:  .LBB6_10: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a4
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs1, rtz
-; CHECK-NOV-NEXT:    blt s2, a4, .LBB6_2
+; CHECK-NOV-NEXT:    blt s1, a4, .LBB6_2
 ; CHECK-NOV-NEXT:  .LBB6_11: # %entry
-; CHECK-NOV-NEXT:    mv s2, a4
+; CHECK-NOV-NEXT:    mv s1, a4
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    blt a2, a4, .LBB6_3
 ; CHECK-NOV-NEXT:  .LBB6_12: # %entry
@@ -437,9 +437,9 @@
 ; CHECK-NOV-NEXT:    blt a1, a2, .LBB6_6
 ; CHECK-NOV-NEXT:  .LBB6_15: # %entry
 ; CHECK-NOV-NEXT:    lui a2, 524288
-; CHECK-NOV-NEXT:    blt a1, s2, .LBB6_7
+; CHECK-NOV-NEXT:    blt a1, s1, .LBB6_7
 ; CHECK-NOV-NEXT:  .LBB6_16: # %entry
-; CHECK-NOV-NEXT:    lui s2, 524288
+; CHECK-NOV-NEXT:    lui s1, 524288
 ; CHECK-NOV-NEXT:    bge a1, a0, .LBB6_8
 ; CHECK-NOV-NEXT:    j .LBB6_9
 ;
@@ -463,13 +463,14 @@
 ; CHECK-V-NEXT:    lhu s1, 16(a0)
 ; CHECK-V-NEXT:    lhu s2, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 2, e64, m2, tu, ma
@@ -478,7 +479,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 3, e64, m2, tu, ma
@@ -487,7 +488,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
@@ -546,17 +547,17 @@
 ; CHECK-NOV-NEXT:    lhu s3, 16(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 8(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.lu.s s2, fs2, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.lu.s s1, fs2, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    li a1, -1
@@ -564,7 +565,7 @@
 ; CHECK-NOV-NEXT:    bgeu a0, a1, .LBB7_6
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a2, fs1, rtz
-; CHECK-NOV-NEXT:    bgeu s2, a1, .LBB7_7
+; CHECK-NOV-NEXT:    bgeu s1, a1, .LBB7_7
 ; CHECK-NOV-NEXT:  .LBB7_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    bgeu a2, a1, .LBB7_8
@@ -575,7 +576,7 @@
 ; CHECK-NOV-NEXT:  .LBB7_5: # %entry
 ; CHECK-NOV-NEXT:    sw a3, 12(s0)
 ; CHECK-NOV-NEXT:    sw a2, 8(s0)
-; CHECK-NOV-NEXT:    sw s2, 4(s0)
+; CHECK-NOV-NEXT:    sw s1, 4(s0)
 ; CHECK-NOV-NEXT:    sw a0, 0(s0)
 ; CHECK-NOV-NEXT:    ld ra, 56(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s0, 48(sp) # 8-byte Folded Reload
@@ -590,9 +591,9 @@
 ; CHECK-NOV-NEXT:  .LBB7_6: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a1
 ; CHECK-NOV-NEXT:    fcvt.lu.s a2, fs1, rtz
-; CHECK-NOV-NEXT:    bltu s2, a1, .LBB7_2
+; CHECK-NOV-NEXT:    bltu s1, a1, .LBB7_2
 ; CHECK-NOV-NEXT:  .LBB7_7: # %entry
-; CHECK-NOV-NEXT:    mv s2, a1
+; CHECK-NOV-NEXT:    mv s1, a1
 ; CHECK-NOV-NEXT:    fcvt.lu.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    bltu a2, a1, .LBB7_3
 ; CHECK-NOV-NEXT:  .LBB7_8: # %entry
@@ -620,13 +621,14 @@
 ; CHECK-V-NEXT:    lhu s1, 16(a0)
 ; CHECK-V-NEXT:    lhu s2, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 2, e64, m2, tu, ma
@@ -635,7 +637,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 3, e64, m2, tu, ma
@@ -644,7 +646,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
@@ -700,17 +702,17 @@
 ; CHECK-NOV-NEXT:    lhu s3, 16(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 8(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.l.s s2, fs2, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.l.s s1, fs2, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    li a2, -1
@@ -718,7 +720,7 @@
 ; CHECK-NOV-NEXT:    bge a0, a2, .LBB8_6
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs1, rtz
-; CHECK-NOV-NEXT:    bge s2, a2, .LBB8_7
+; CHECK-NOV-NEXT:    bge s1, a2, .LBB8_7
 ; CHECK-NOV-NEXT:  .LBB8_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    bge a1, a2, .LBB8_8
@@ -729,12 +731,12 @@
 ; CHECK-NOV-NEXT:  .LBB8_5: # %entry
 ; CHECK-NOV-NEXT:    sgtz a2, a3
 ; CHECK-NOV-NEXT:    sgtz a4, a1
-; CHECK-NOV-NEXT:    sgtz a5, s2
+; CHECK-NOV-NEXT:    sgtz a5, s1
 ; CHECK-NOV-NEXT:    sgtz a6, a0
 ; CHECK-NOV-NEXT:    negw a6, a6
 ; CHECK-NOV-NEXT:    and a0, a6, a0
 ; CHECK-NOV-NEXT:    negw a5, a5
-; CHECK-NOV-NEXT:    and a5, a5, s2
+; CHECK-NOV-NEXT:    and a5, a5, s1
 ; CHECK-NOV-NEXT:    negw a4, a4
 ; CHECK-NOV-NEXT:    and a1, a4, a1
 ; CHECK-NOV-NEXT:    negw a2, a2
@@ -756,9 +758,9 @@
 ; CHECK-NOV-NEXT:  .LBB8_6: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a2
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs1, rtz
-; CHECK-NOV-NEXT:    blt s2, a2, .LBB8_2
+; CHECK-NOV-NEXT:    blt s1, a2, .LBB8_2
 ; CHECK-NOV-NEXT:  .LBB8_7: # %entry
-; CHECK-NOV-NEXT:    mv s2, a2
+; CHECK-NOV-NEXT:    mv s1, a2
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    blt a1, a2, .LBB8_3
 ; CHECK-NOV-NEXT:  .LBB8_8: # %entry
@@ -786,13 +788,14 @@
 ; CHECK-V-NEXT:    lhu s1, 16(a0)
 ; CHECK-V-NEXT:    lhu s2, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 2, e64, m2, tu, ma
@@ -801,7 +804,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 3, e64, m2, tu, ma
@@ -810,7 +813,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
@@ -1231,29 +1234,29 @@
 ; CHECK-NOV-NEXT:    lhu s7, 40(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 48(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs6, fa0
-; CHECK-NOV-NEXT:    mv a0, s7
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s7
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs5, fa0
-; CHECK-NOV-NEXT:    mv a0, s6
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s6
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs4, fa0
-; CHECK-NOV-NEXT:    mv a0, s5
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s5
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs3, fa0
-; CHECK-NOV-NEXT:    mv a0, s4
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s4
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.l.s s2, fs6, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.l.s s1, fs6, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    lui a7, 8
@@ -1261,7 +1264,7 @@
 ; CHECK-NOV-NEXT:    bge a0, a7, .LBB15_18
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs5, rtz
-; CHECK-NOV-NEXT:    bge s2, a7, .LBB15_19
+; CHECK-NOV-NEXT:    bge s1, a7, .LBB15_19
 ; CHECK-NOV-NEXT:  .LBB15_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs4, rtz
 ; CHECK-NOV-NEXT:    bge a1, a7, .LBB15_20
@@ -1293,14 +1296,14 @@
 ; CHECK-NOV-NEXT:  .LBB15_13: # %entry
 ; CHECK-NOV-NEXT:    bge a7, a1, .LBB15_31
 ; CHECK-NOV-NEXT:  .LBB15_14: # %entry
-; CHECK-NOV-NEXT:    bge a7, s2, .LBB15_32
+; CHECK-NOV-NEXT:    bge a7, s1, .LBB15_32
 ; CHECK-NOV-NEXT:  .LBB15_15: # %entry
 ; CHECK-NOV-NEXT:    blt a7, a0, .LBB15_17
 ; CHECK-NOV-NEXT:  .LBB15_16: # %entry
 ; CHECK-NOV-NEXT:    lui a0, 1048568
 ; CHECK-NOV-NEXT:  .LBB15_17: # %entry
 ; CHECK-NOV-NEXT:    sh a0, 14(s0)
-; CHECK-NOV-NEXT:    sh s2, 12(s0)
+; CHECK-NOV-NEXT:    sh s1, 12(s0)
 ; CHECK-NOV-NEXT:    sh a1, 10(s0)
 ; CHECK-NOV-NEXT:    sh a2, 8(s0)
 ; CHECK-NOV-NEXT:    sh a3, 6(s0)
@@ -1328,9 +1331,9 @@
 ; CHECK-NOV-NEXT:  .LBB15_18: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a7
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs5, rtz
-; CHECK-NOV-NEXT:    blt s2, a7, .LBB15_2
+; CHECK-NOV-NEXT:    blt s1, a7, .LBB15_2
 ; CHECK-NOV-NEXT:  .LBB15_19: # %entry
-; CHECK-NOV-NEXT:    mv s2, a7
+; CHECK-NOV-NEXT:    mv s1, a7
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs4, rtz
 ; CHECK-NOV-NEXT:    blt a1, a7, .LBB15_3
 ; CHECK-NOV-NEXT:  .LBB15_20: # %entry
@@ -1373,9 +1376,9 @@
 ; CHECK-NOV-NEXT:    blt a7, a1, .LBB15_14
 ; CHECK-NOV-NEXT:  .LBB15_31: # %entry
 ; CHECK-NOV-NEXT:    lui a1, 1048568
-; CHECK-NOV-NEXT:    blt a7, s2, .LBB15_15
+; CHECK-NOV-NEXT:    blt a7, s1, .LBB15_15
 ; CHECK-NOV-NEXT:  .LBB15_32: # %entry
-; CHECK-NOV-NEXT:    lui s2, 1048568
+; CHECK-NOV-NEXT:    lui s1, 1048568
 ; CHECK-NOV-NEXT:    bge a7, a0, .LBB15_16
 ; CHECK-NOV-NEXT:    j .LBB15_17
 ;
@@ -1411,13 +1414,14 @@
 ; CHECK-V-NEXT:    lhu s5, 16(a0)
 ; CHECK-V-NEXT:    lhu s6, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s6
+; CHECK-V-NEXT:    fmv.w.x fa0, s6
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 2, e32, m2, tu, ma
@@ -1426,7 +1430,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s5
+; CHECK-V-NEXT:    fmv.w.x fa0, s5
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 3, e32, m2, tu, ma
@@ -1435,7 +1439,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s4
+; CHECK-V-NEXT:    fmv.w.x fa0, s4
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e32, m2, tu, ma
@@ -1444,7 +1448,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s3
+; CHECK-V-NEXT:    fmv.w.x fa0, s3
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 5, e32, m2, tu, ma
@@ -1453,7 +1457,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 4
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 6, e32, m2, tu, ma
@@ -1462,7 +1466,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 5
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 7, e32, m2, tu, ma
@@ -1471,7 +1475,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 6
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
@@ -1555,29 +1559,29 @@
 ; CHECK-NOV-NEXT:    lhu s7, 16(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 8(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs6, fa0
-; CHECK-NOV-NEXT:    mv a0, s7
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s7
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs5, fa0
-; CHECK-NOV-NEXT:    mv a0, s6
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s6
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs4, fa0
-; CHECK-NOV-NEXT:    mv a0, s5
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s5
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs3, fa0
-; CHECK-NOV-NEXT:    mv a0, s4
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s4
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.lu.s s2, fs6, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.lu.s s1, fs6, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    lui a1, 16
@@ -1585,7 +1589,7 @@
 ; CHECK-NOV-NEXT:    bgeu a0, a1, .LBB16_10
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a2, fs5, rtz
-; CHECK-NOV-NEXT:    bgeu s2, a1, .LBB16_11
+; CHECK-NOV-NEXT:    bgeu s1, a1, .LBB16_11
 ; CHECK-NOV-NEXT:  .LBB16_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a3, fs4, rtz
 ; CHECK-NOV-NEXT:    bgeu a2, a1, .LBB16_12
@@ -1612,7 +1616,7 @@
 ; CHECK-NOV-NEXT:    sh a4, 8(s0)
 ; CHECK-NOV-NEXT:    sh a3, 6(s0)
 ; CHECK-NOV-NEXT:    sh a2, 4(s0)
-; CHECK-NOV-NEXT:    sh s2, 2(s0)
+; CHECK-NOV-NEXT:    sh s1, 2(s0)
 ; CHECK-NOV-NEXT:    sh a0, 0(s0)
 ; CHECK-NOV-NEXT:    ld ra, 120(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s0, 112(sp) # 8-byte Folded Reload
@@ -1635,9 +1639,9 @@
 ; CHECK-NOV-NEXT:  .LBB16_10: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a1
 ; CHECK-NOV-NEXT:    fcvt.lu.s a2, fs5, rtz
-; CHECK-NOV-NEXT:    bltu s2, a1, .LBB16_2
+; CHECK-NOV-NEXT:    bltu s1, a1, .LBB16_2
 ; CHECK-NOV-NEXT:  .LBB16_11: # %entry
-; CHECK-NOV-NEXT:    mv s2, a1
+; CHECK-NOV-NEXT:    mv s1, a1
 ; CHECK-NOV-NEXT:    fcvt.lu.s a3, fs4, rtz
 ; CHECK-NOV-NEXT:    bltu a2, a1, .LBB16_3
 ; CHECK-NOV-NEXT:  .LBB16_12: # %entry
@@ -1693,13 +1697,14 @@
 ; CHECK-V-NEXT:    lhu s5, 16(a0)
 ; CHECK-V-NEXT:    lhu s6, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s6
+; CHECK-V-NEXT:    fmv.w.x fa0, s6
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 2, e32, m2, tu, ma
@@ -1708,7 +1713,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s5
+; CHECK-V-NEXT:    fmv.w.x fa0, s5
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 3, e32, m2, tu, ma
@@ -1717,7 +1722,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s4
+; CHECK-V-NEXT:    fmv.w.x fa0, s4
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e32, m2, tu, ma
@@ -1726,7 +1731,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s3
+; CHECK-V-NEXT:    fmv.w.x fa0, s3
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 5, e32, m2, tu, ma
@@ -1735,7 +1740,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 4
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 6, e32, m2, tu, ma
@@ -1744,7 +1749,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 5
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 7, e32, m2, tu, ma
@@ -1753,7 +1758,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 6
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
@@ -1833,29 +1838,29 @@
 ; CHECK-NOV-NEXT:    lhu s7, 16(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 8(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs6, fa0
-; CHECK-NOV-NEXT:    mv a0, s7
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s7
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs5, fa0
-; CHECK-NOV-NEXT:    mv a0, s6
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s6
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs4, fa0
-; CHECK-NOV-NEXT:    mv a0, s5
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s5
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs3, fa0
-; CHECK-NOV-NEXT:    mv a0, s4
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s4
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.l.s s2, fs6, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.l.s s1, fs6, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    lui a3, 16
@@ -1863,7 +1868,7 @@
 ; CHECK-NOV-NEXT:    bge a0, a3, .LBB17_10
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs5, rtz
-; CHECK-NOV-NEXT:    bge s2, a3, .LBB17_11
+; CHECK-NOV-NEXT:    bge s1, a3, .LBB17_11
 ; CHECK-NOV-NEXT:  .LBB17_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs4, rtz
 ; CHECK-NOV-NEXT:    bge a1, a3, .LBB17_12
@@ -1890,12 +1895,12 @@
 ; CHECK-NOV-NEXT:    sgtz t2, a4
 ; CHECK-NOV-NEXT:    sgtz t3, a2
 ; CHECK-NOV-NEXT:    sgtz t4, a1
-; CHECK-NOV-NEXT:    sgtz t5, s2
+; CHECK-NOV-NEXT:    sgtz t5, s1
 ; CHECK-NOV-NEXT:    sgtz t6, a0
 ; CHECK-NOV-NEXT:    negw t6, t6
 ; CHECK-NOV-NEXT:    and a0, t6, a0
 ; CHECK-NOV-NEXT:    negw t5, t5
-; CHECK-NOV-NEXT:    and t5, t5, s2
+; CHECK-NOV-NEXT:    and t5, t5, s1
 ; CHECK-NOV-NEXT:    negw t4, t4
 ; CHECK-NOV-NEXT:    and a1, t4, a1
 ; CHECK-NOV-NEXT:    negw t3, t3
@@ -1937,9 +1942,9 @@
 ; CHECK-NOV-NEXT:  .LBB17_10: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a3
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs5, rtz
-; CHECK-NOV-NEXT:    blt s2, a3, .LBB17_2
+; CHECK-NOV-NEXT:    blt s1, a3, .LBB17_2
 ; CHECK-NOV-NEXT:  .LBB17_11: # %entry
-; CHECK-NOV-NEXT:    mv s2, a3
+; CHECK-NOV-NEXT:    mv s1, a3
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs4, rtz
 ; CHECK-NOV-NEXT:    blt a1, a3, .LBB17_3
 ; CHECK-NOV-NEXT:  .LBB17_12: # %entry
@@ -1995,13 +2000,14 @@
 ; CHECK-V-NEXT:    lhu s5, 16(a0)
 ; CHECK-V-NEXT:    lhu s6, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s6
+; CHECK-V-NEXT:    fmv.w.x fa0, s6
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 2, e32, m2, tu, ma
@@ -2010,7 +2016,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s5
+; CHECK-V-NEXT:    fmv.w.x fa0, s5
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 3, e32, m2, tu, ma
@@ -2019,7 +2025,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s4
+; CHECK-V-NEXT:    fmv.w.x fa0, s4
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e32, m2, tu, ma
@@ -2028,7 +2034,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s3
+; CHECK-V-NEXT:    fmv.w.x fa0, s3
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 5, e32, m2, tu, ma
@@ -2037,7 +2043,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 4
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 6, e32, m2, tu, ma
@@ -2046,7 +2052,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 5
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 7, e32, m2, tu, ma
@@ -2055,7 +2061,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 6
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
@@ -2924,11 +2930,12 @@
 ; CHECK-NOV-NEXT:    .cfi_offset s1, -24
 ; CHECK-NOV-NEXT:    .cfi_offset s2, -32
 ; CHECK-NOV-NEXT:    mv s2, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a0
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixsfti@plt
 ; CHECK-NOV-NEXT:    mv s0, a0
 ; CHECK-NOV-NEXT:    mv s1, a1
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixsfti@plt
 ; CHECK-NOV-NEXT:    mv a2, a0
@@ -3005,11 +3012,12 @@
 ; CHECK-V-NEXT:    .cfi_offset s1, -24
 ; CHECK-V-NEXT:    .cfi_offset s2, -32
 ; CHECK-V-NEXT:    mv s2, a1
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixsfti@plt
 ; CHECK-V-NEXT:    mv s0, a0
 ; CHECK-V-NEXT:    mv s1, a1
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixsfti@plt
 ; CHECK-V-NEXT:    li a2, -1
@@ -3097,11 +3105,12 @@
 ; CHECK-NOV-NEXT:    .cfi_offset s1, -24
 ; CHECK-NOV-NEXT:    .cfi_offset s2, -32
 ; CHECK-NOV-NEXT:    mv s0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a0
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixunssfti@plt
 ; CHECK-NOV-NEXT:    mv s1, a0
 ; CHECK-NOV-NEXT:    mv s2, a1
-; CHECK-NOV-NEXT:    mv a0, s0
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s0
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixunssfti@plt
 ; CHECK-NOV-NEXT:    snez a1, a1
@@ -3131,11 +3140,12 @@
 ; CHECK-V-NEXT:    .cfi_offset s1, -24
 ; CHECK-V-NEXT:    .cfi_offset s2, -32
 ; CHECK-V-NEXT:    mv s0, a1
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixunssfti@plt
 ; CHECK-V-NEXT:    mv s1, a0
 ; CHECK-V-NEXT:    mv s2, a1
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixunssfti@plt
 ; CHECK-V-NEXT:    snez a1, a1
@@ -3176,11 +3186,12 @@
 ; CHECK-NOV-NEXT:    .cfi_offset s1, -24
 ; CHECK-NOV-NEXT:    .cfi_offset s2, -32
 ; CHECK-NOV-NEXT:    mv s2, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a0
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixsfti@plt
 ; CHECK-NOV-NEXT:    mv s0, a0
 ; CHECK-NOV-NEXT:    mv s1, a1
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixsfti@plt
 ; CHECK-NOV-NEXT:    mv a2, s1
@@ -3236,11 +3247,12 @@
 ; CHECK-V-NEXT:    .cfi_offset s1, -24
 ; CHECK-V-NEXT:    .cfi_offset s2, -32
 ; CHECK-V-NEXT:    mv s2, a1
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixsfti@plt
 ; CHECK-V-NEXT:    mv s0, a0
 ; CHECK-V-NEXT:    mv s1, a1
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixsfti@plt
 ; CHECK-V-NEXT:    mv a2, s1
@@ -3653,17 +3665,17 @@
 ; CHECK-NOV-NEXT:    lhu s3, 8(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 16(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.l.s s2, fs2, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.l.s s1, fs2, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    lui a1, 524288
@@ -3671,7 +3683,7 @@
 ; CHECK-NOV-NEXT:    bge a0, a4, .LBB33_10
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs1, rtz
-; CHECK-NOV-NEXT:    bge s2, a4, .LBB33_11
+; CHECK-NOV-NEXT:    bge s1, a4, .LBB33_11
 ; CHECK-NOV-NEXT:  .LBB33_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    bge a2, a4, .LBB33_12
@@ -3682,14 +3694,14 @@
 ; 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
@@ -3705,9 +3717,9 @@
 ; CHECK-NOV-NEXT:  .LBB33_10: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a4
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs1, rtz
-; CHECK-NOV-NEXT:    blt s2, a4, .LBB33_2
+; CHECK-NOV-NEXT:    blt s1, a4, .LBB33_2
 ; CHECK-NOV-NEXT:  .LBB33_11: # %entry
-; CHECK-NOV-NEXT:    mv s2, a4
+; CHECK-NOV-NEXT:    mv s1, a4
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    blt a2, a4, .LBB33_3
 ; CHECK-NOV-NEXT:  .LBB33_12: # %entry
@@ -3721,9 +3733,9 @@
 ; CHECK-NOV-NEXT:    blt a1, a2, .LBB33_6
 ; CHECK-NOV-NEXT:  .LBB33_15: # %entry
 ; CHECK-NOV-NEXT:    lui a2, 524288
-; CHECK-NOV-NEXT:    blt a1, s2, .LBB33_7
+; CHECK-NOV-NEXT:    blt a1, s1, .LBB33_7
 ; CHECK-NOV-NEXT:  .LBB33_16: # %entry
-; CHECK-NOV-NEXT:    lui s2, 524288
+; CHECK-NOV-NEXT:    lui s1, 524288
 ; CHECK-NOV-NEXT:    bge a1, a0, .LBB33_8
 ; CHECK-NOV-NEXT:    j .LBB33_9
 ;
@@ -3747,13 +3759,14 @@
 ; CHECK-V-NEXT:    lhu s1, 16(a0)
 ; CHECK-V-NEXT:    lhu s2, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 2, e64, m2, tu, ma
@@ -3762,7 +3775,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 3, e64, m2, tu, ma
@@ -3771,7 +3784,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
@@ -3828,17 +3841,17 @@
 ; CHECK-NOV-NEXT:    lhu s3, 16(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 8(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.lu.s s2, fs2, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.lu.s s1, fs2, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    li a1, -1
@@ -3846,7 +3859,7 @@
 ; CHECK-NOV-NEXT:    bgeu a0, a1, .LBB34_6
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a2, fs1, rtz
-; CHECK-NOV-NEXT:    bgeu s2, a1, .LBB34_7
+; CHECK-NOV-NEXT:    bgeu s1, a1, .LBB34_7
 ; CHECK-NOV-NEXT:  .LBB34_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    bgeu a2, a1, .LBB34_8
@@ -3857,7 +3870,7 @@
 ; CHECK-NOV-NEXT:  .LBB34_5: # %entry
 ; CHECK-NOV-NEXT:    sw a3, 12(s0)
 ; CHECK-NOV-NEXT:    sw a2, 8(s0)
-; CHECK-NOV-NEXT:    sw s2, 4(s0)
+; CHECK-NOV-NEXT:    sw s1, 4(s0)
 ; CHECK-NOV-NEXT:    sw a0, 0(s0)
 ; CHECK-NOV-NEXT:    ld ra, 56(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s0, 48(sp) # 8-byte Folded Reload
@@ -3872,9 +3885,9 @@
 ; CHECK-NOV-NEXT:  .LBB34_6: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a1
 ; CHECK-NOV-NEXT:    fcvt.lu.s a2, fs1, rtz
-; CHECK-NOV-NEXT:    bltu s2, a1, .LBB34_2
+; CHECK-NOV-NEXT:    bltu s1, a1, .LBB34_2
 ; CHECK-NOV-NEXT:  .LBB34_7: # %entry
-; CHECK-NOV-NEXT:    mv s2, a1
+; CHECK-NOV-NEXT:    mv s1, a1
 ; CHECK-NOV-NEXT:    fcvt.lu.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    bltu a2, a1, .LBB34_3
 ; CHECK-NOV-NEXT:  .LBB34_8: # %entry
@@ -3902,13 +3915,14 @@
 ; CHECK-V-NEXT:    lhu s1, 16(a0)
 ; CHECK-V-NEXT:    lhu s2, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 2, e64, m2, tu, ma
@@ -3917,7 +3931,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 3, e64, m2, tu, ma
@@ -3926,7 +3940,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
@@ -3981,17 +3995,17 @@
 ; CHECK-NOV-NEXT:    lhu s3, 8(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 16(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.l.s s2, fs2, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.l.s s1, fs2, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    li a2, -1
@@ -3999,7 +4013,7 @@
 ; CHECK-NOV-NEXT:    bge a0, a2, .LBB35_6
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs1, rtz
-; CHECK-NOV-NEXT:    bge s2, a2, .LBB35_7
+; CHECK-NOV-NEXT:    bge s1, a2, .LBB35_7
 ; CHECK-NOV-NEXT:  .LBB35_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    bge a1, a2, .LBB35_8
@@ -4014,9 +4028,9 @@
 ; CHECK-NOV-NEXT:    sgtz a3, a1
 ; CHECK-NOV-NEXT:    negw a3, a3
 ; CHECK-NOV-NEXT:    and a1, a3, a1
-; CHECK-NOV-NEXT:    sgtz a3, s2
+; CHECK-NOV-NEXT:    sgtz a3, s1
 ; CHECK-NOV-NEXT:    negw a3, a3
-; CHECK-NOV-NEXT:    and a3, a3, s2
+; CHECK-NOV-NEXT:    and a3, a3, s1
 ; CHECK-NOV-NEXT:    sgtz a4, a0
 ; CHECK-NOV-NEXT:    negw a4, a4
 ; CHECK-NOV-NEXT:    and a0, a4, a0
@@ -4037,9 +4051,9 @@
 ; CHECK-NOV-NEXT:  .LBB35_6: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a2
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs1, rtz
-; CHECK-NOV-NEXT:    blt s2, a2, .LBB35_2
+; CHECK-NOV-NEXT:    blt s1, a2, .LBB35_2
 ; CHECK-NOV-NEXT:  .LBB35_7: # %entry
-; CHECK-NOV-NEXT:    mv s2, a2
+; CHECK-NOV-NEXT:    mv s1, a2
 ; CHECK-NOV-NEXT:    fcvt.l.s a3, fs0, rtz
 ; CHECK-NOV-NEXT:    blt a1, a2, .LBB35_3
 ; CHECK-NOV-NEXT:  .LBB35_8: # %entry
@@ -4067,13 +4081,14 @@
 ; CHECK-V-NEXT:    lhu s1, 16(a0)
 ; CHECK-V-NEXT:    lhu s2, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 2, e64, m2, tu, ma
@@ -4082,7 +4097,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 3, e64, m2, tu, ma
@@ -4091,7 +4106,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
@@ -4500,29 +4515,29 @@
 ; CHECK-NOV-NEXT:    lhu s7, 40(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 48(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs6, fa0
-; CHECK-NOV-NEXT:    mv a0, s7
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s7
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs5, fa0
-; CHECK-NOV-NEXT:    mv a0, s6
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s6
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs4, fa0
-; CHECK-NOV-NEXT:    mv a0, s5
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s5
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs3, fa0
-; CHECK-NOV-NEXT:    mv a0, s4
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s4
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.l.s s2, fs6, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.l.s s1, fs6, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    lui a7, 8
@@ -4530,7 +4545,7 @@
 ; CHECK-NOV-NEXT:    bge a0, a7, .LBB42_18
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs5, rtz
-; CHECK-NOV-NEXT:    bge s2, a7, .LBB42_19
+; CHECK-NOV-NEXT:    bge s1, a7, .LBB42_19
 ; CHECK-NOV-NEXT:  .LBB42_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs4, rtz
 ; CHECK-NOV-NEXT:    bge a1, a7, .LBB42_20
@@ -4562,14 +4577,14 @@
 ; CHECK-NOV-NEXT:  .LBB42_13: # %entry
 ; CHECK-NOV-NEXT:    bge a7, a1, .LBB42_31
 ; CHECK-NOV-NEXT:  .LBB42_14: # %entry
-; CHECK-NOV-NEXT:    bge a7, s2, .LBB42_32
+; CHECK-NOV-NEXT:    bge a7, s1, .LBB42_32
 ; CHECK-NOV-NEXT:  .LBB42_15: # %entry
 ; CHECK-NOV-NEXT:    blt a7, a0, .LBB42_17
 ; CHECK-NOV-NEXT:  .LBB42_16: # %entry
 ; CHECK-NOV-NEXT:    lui a0, 1048568
 ; CHECK-NOV-NEXT:  .LBB42_17: # %entry
 ; CHECK-NOV-NEXT:    sh a0, 14(s0)
-; CHECK-NOV-NEXT:    sh s2, 12(s0)
+; CHECK-NOV-NEXT:    sh s1, 12(s0)
 ; CHECK-NOV-NEXT:    sh a1, 10(s0)
 ; CHECK-NOV-NEXT:    sh a2, 8(s0)
 ; CHECK-NOV-NEXT:    sh a3, 6(s0)
@@ -4597,9 +4612,9 @@
 ; CHECK-NOV-NEXT:  .LBB42_18: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a7
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs5, rtz
-; CHECK-NOV-NEXT:    blt s2, a7, .LBB42_2
+; CHECK-NOV-NEXT:    blt s1, a7, .LBB42_2
 ; CHECK-NOV-NEXT:  .LBB42_19: # %entry
-; CHECK-NOV-NEXT:    mv s2, a7
+; CHECK-NOV-NEXT:    mv s1, a7
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs4, rtz
 ; CHECK-NOV-NEXT:    blt a1, a7, .LBB42_3
 ; CHECK-NOV-NEXT:  .LBB42_20: # %entry
@@ -4642,9 +4657,9 @@
 ; CHECK-NOV-NEXT:    blt a7, a1, .LBB42_14
 ; CHECK-NOV-NEXT:  .LBB42_31: # %entry
 ; CHECK-NOV-NEXT:    lui a1, 1048568
-; CHECK-NOV-NEXT:    blt a7, s2, .LBB42_15
+; CHECK-NOV-NEXT:    blt a7, s1, .LBB42_15
 ; CHECK-NOV-NEXT:  .LBB42_32: # %entry
-; CHECK-NOV-NEXT:    lui s2, 1048568
+; CHECK-NOV-NEXT:    lui s1, 1048568
 ; CHECK-NOV-NEXT:    bge a7, a0, .LBB42_16
 ; CHECK-NOV-NEXT:    j .LBB42_17
 ;
@@ -4680,13 +4695,14 @@
 ; CHECK-V-NEXT:    lhu s5, 16(a0)
 ; CHECK-V-NEXT:    lhu s6, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s6
+; CHECK-V-NEXT:    fmv.w.x fa0, s6
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 2, e32, m2, tu, ma
@@ -4695,7 +4711,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s5
+; CHECK-V-NEXT:    fmv.w.x fa0, s5
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 3, e32, m2, tu, ma
@@ -4704,7 +4720,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s4
+; CHECK-V-NEXT:    fmv.w.x fa0, s4
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e32, m2, tu, ma
@@ -4713,7 +4729,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s3
+; CHECK-V-NEXT:    fmv.w.x fa0, s3
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 5, e32, m2, tu, ma
@@ -4722,7 +4738,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 4
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 6, e32, m2, tu, ma
@@ -4731,7 +4747,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 5
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 7, e32, m2, tu, ma
@@ -4740,7 +4756,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 6
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
@@ -4822,31 +4838,31 @@
 ; CHECK-NOV-NEXT:    lhu s7, 16(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 8(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs5, fa0
-; CHECK-NOV-NEXT:    mv a0, s7
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s7
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs6, fa0
-; CHECK-NOV-NEXT:    mv a0, s6
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s6
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs4, fa0
-; CHECK-NOV-NEXT:    mv a0, s5
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s5
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs3, fa0
-; CHECK-NOV-NEXT:    mv a0, s4
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s4
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.lu.s s3, fs6, rtz
-; CHECK-NOV-NEXT:    fcvt.lu.s s2, fs5, rtz
-; CHECK-NOV-NEXT:    sext.w s2, s2
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fcvt.lu.s s2, fs6, rtz
+; CHECK-NOV-NEXT:    fcvt.lu.s a0, fs5, rtz
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    sext.w s1, a0
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    sext.w a0, a0
@@ -4857,10 +4873,10 @@
 ; CHECK-NOV-NEXT:    mv a0, a1
 ; CHECK-NOV-NEXT:  .LBB43_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a3, fs4, rtz
-; CHECK-NOV-NEXT:    sext.w a2, s3
-; CHECK-NOV-NEXT:    bltu s2, a1, .LBB43_4
+; CHECK-NOV-NEXT:    sext.w a2, s2
+; CHECK-NOV-NEXT:    bltu s1, a1, .LBB43_4
 ; CHECK-NOV-NEXT:  # %bb.3: # %entry
-; CHECK-NOV-NEXT:    mv s2, a1
+; CHECK-NOV-NEXT:    mv s1, a1
 ; CHECK-NOV-NEXT:  .LBB43_4: # %entry
 ; CHECK-NOV-NEXT:    fcvt.lu.s a4, fs3, rtz
 ; CHECK-NOV-NEXT:    sext.w a3, a3
@@ -4897,7 +4913,7 @@
 ; CHECK-NOV-NEXT:    sh a4, 8(s0)
 ; CHECK-NOV-NEXT:    sh a3, 6(s0)
 ; CHECK-NOV-NEXT:    sh a2, 4(s0)
-; CHECK-NOV-NEXT:    sh s2, 2(s0)
+; CHECK-NOV-NEXT:    sh s1, 2(s0)
 ; CHECK-NOV-NEXT:    sh a0, 0(s0)
 ; CHECK-NOV-NEXT:    ld ra, 120(sp) # 8-byte Folded Reload
 ; CHECK-NOV-NEXT:    ld s0, 112(sp) # 8-byte Folded Reload
@@ -4958,13 +4974,14 @@
 ; CHECK-V-NEXT:    lhu s5, 16(a0)
 ; CHECK-V-NEXT:    lhu s6, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s6
+; CHECK-V-NEXT:    fmv.w.x fa0, s6
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 2, e32, m2, tu, ma
@@ -4973,7 +4990,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s5
+; CHECK-V-NEXT:    fmv.w.x fa0, s5
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 3, e32, m2, tu, ma
@@ -4982,7 +4999,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s4
+; CHECK-V-NEXT:    fmv.w.x fa0, s4
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e32, m2, tu, ma
@@ -4991,7 +5008,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s3
+; CHECK-V-NEXT:    fmv.w.x fa0, s3
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 5, e32, m2, tu, ma
@@ -5000,7 +5017,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 4
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 6, e32, m2, tu, ma
@@ -5009,7 +5026,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 5
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 7, e32, m2, tu, ma
@@ -5018,7 +5035,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 6
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.lu.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
@@ -5097,29 +5114,29 @@
 ; CHECK-NOV-NEXT:    lhu s7, 40(a1)
 ; CHECK-NOV-NEXT:    lhu a1, 48(a1)
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs6, fa0
-; CHECK-NOV-NEXT:    mv a0, s7
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s7
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs5, fa0
-; CHECK-NOV-NEXT:    mv a0, s6
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s6
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs4, fa0
-; CHECK-NOV-NEXT:    mv a0, s5
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s5
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs3, fa0
-; CHECK-NOV-NEXT:    mv a0, s4
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s4
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs2, fa0
-; CHECK-NOV-NEXT:    mv a0, s3
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s3
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs1, fa0
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fmv.s fs0, fa0
-; CHECK-NOV-NEXT:    fcvt.l.s s2, fs6, rtz
-; CHECK-NOV-NEXT:    mv a0, s1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s1
+; CHECK-NOV-NEXT:    fcvt.l.s s1, fs6, rtz
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-NOV-NEXT:    lui a3, 16
@@ -5127,7 +5144,7 @@
 ; CHECK-NOV-NEXT:    bge a0, a3, .LBB44_10
 ; CHECK-NOV-NEXT:  # %bb.1: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs5, rtz
-; CHECK-NOV-NEXT:    bge s2, a3, .LBB44_11
+; CHECK-NOV-NEXT:    bge s1, a3, .LBB44_11
 ; CHECK-NOV-NEXT:  .LBB44_2: # %entry
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs4, rtz
 ; CHECK-NOV-NEXT:    bge a1, a3, .LBB44_12
@@ -5166,9 +5183,9 @@
 ; CHECK-NOV-NEXT:    sgtz a7, a1
 ; CHECK-NOV-NEXT:    negw a7, a7
 ; CHECK-NOV-NEXT:    and a1, a7, a1
-; CHECK-NOV-NEXT:    sgtz a7, s2
+; CHECK-NOV-NEXT:    sgtz a7, s1
 ; CHECK-NOV-NEXT:    negw a7, a7
-; CHECK-NOV-NEXT:    and a7, a7, s2
+; CHECK-NOV-NEXT:    and a7, a7, s1
 ; CHECK-NOV-NEXT:    sgtz t0, a0
 ; CHECK-NOV-NEXT:    negw t0, t0
 ; CHECK-NOV-NEXT:    and a0, t0, a0
@@ -5201,9 +5218,9 @@
 ; CHECK-NOV-NEXT:  .LBB44_10: # %entry
 ; CHECK-NOV-NEXT:    mv a0, a3
 ; CHECK-NOV-NEXT:    fcvt.l.s a1, fs5, rtz
-; CHECK-NOV-NEXT:    blt s2, a3, .LBB44_2
+; CHECK-NOV-NEXT:    blt s1, a3, .LBB44_2
 ; CHECK-NOV-NEXT:  .LBB44_11: # %entry
-; CHECK-NOV-NEXT:    mv s2, a3
+; CHECK-NOV-NEXT:    mv s1, a3
 ; CHECK-NOV-NEXT:    fcvt.l.s a2, fs4, rtz
 ; CHECK-NOV-NEXT:    blt a1, a3, .LBB44_3
 ; CHECK-NOV-NEXT:  .LBB44_12: # %entry
@@ -5259,13 +5276,14 @@
 ; CHECK-V-NEXT:    lhu s5, 16(a0)
 ; CHECK-V-NEXT:    lhu s6, 0(a0)
 ; CHECK-V-NEXT:    lhu a0, 8(a0)
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; CHECK-V-NEXT:    vmv.s.x v8, a0
 ; CHECK-V-NEXT:    addi a0, sp, 16
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s6
+; CHECK-V-NEXT:    fmv.w.x fa0, s6
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 2, e32, m2, tu, ma
@@ -5274,7 +5292,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v8, v10, 1
 ; CHECK-V-NEXT:    vs2r.v v8, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s5
+; CHECK-V-NEXT:    fmv.w.x fa0, s5
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 3, e32, m2, tu, ma
@@ -5283,7 +5301,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 2
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s4
+; CHECK-V-NEXT:    fmv.w.x fa0, s4
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 4, e32, m2, tu, ma
@@ -5292,7 +5310,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 3
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s3
+; CHECK-V-NEXT:    fmv.w.x fa0, s3
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 5, e32, m2, tu, ma
@@ -5301,7 +5319,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 4
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 6, e32, m2, tu, ma
@@ -5310,7 +5328,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 5
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s1
+; CHECK-V-NEXT:    fmv.w.x fa0, s1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 7, e32, m2, tu, ma
@@ -5319,7 +5337,7 @@
 ; CHECK-V-NEXT:    vl2r.v v10, (a0) # Unknown-size Folded Reload
 ; CHECK-V-NEXT:    vslideup.vi v10, v8, 6
 ; CHECK-V-NEXT:    vs2r.v v10, (a0) # Unknown-size Folded Spill
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    fcvt.l.s a0, fa0, rtz
 ; CHECK-V-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
@@ -6288,11 +6306,12 @@
 ; CHECK-NOV-NEXT:    .cfi_offset s1, -24
 ; CHECK-NOV-NEXT:    .cfi_offset s2, -32
 ; CHECK-NOV-NEXT:    mv s2, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a0
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixsfti@plt
 ; CHECK-NOV-NEXT:    mv s0, a0
 ; CHECK-NOV-NEXT:    mv s1, a1
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixsfti@plt
 ; CHECK-NOV-NEXT:    mv a2, a0
@@ -6393,11 +6412,12 @@
 ; CHECK-V-NEXT:    .cfi_offset s1, -24
 ; CHECK-V-NEXT:    .cfi_offset s2, -32
 ; CHECK-V-NEXT:    mv s2, a1
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixsfti@plt
 ; CHECK-V-NEXT:    mv s0, a0
 ; CHECK-V-NEXT:    mv s1, a1
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixsfti@plt
 ; CHECK-V-NEXT:    li a2, -1
@@ -6507,12 +6527,12 @@
 ; CHECK-NOV-NEXT:    .cfi_offset s1, -24
 ; CHECK-NOV-NEXT:    .cfi_offset s2, -32
 ; CHECK-NOV-NEXT:    mv s0, a0
-; CHECK-NOV-NEXT:    mv a0, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a1
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixunssfti@plt
 ; CHECK-NOV-NEXT:    mv s1, a0
 ; CHECK-NOV-NEXT:    mv s2, a1
-; CHECK-NOV-NEXT:    mv a0, s0
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s0
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixunssfti@plt
 ; CHECK-NOV-NEXT:    snez a2, a1
@@ -6549,12 +6569,12 @@
 ; CHECK-V-NEXT:    .cfi_offset s1, -24
 ; CHECK-V-NEXT:    .cfi_offset s2, -32
 ; CHECK-V-NEXT:    mv s0, a0
-; CHECK-V-NEXT:    mv a0, a1
+; CHECK-V-NEXT:    fmv.w.x fa0, a1
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixunssfti@plt
 ; CHECK-V-NEXT:    mv s1, a0
 ; CHECK-V-NEXT:    mv s2, a1
-; CHECK-V-NEXT:    mv a0, s0
+; CHECK-V-NEXT:    fmv.w.x fa0, s0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixunssfti@plt
 ; CHECK-V-NEXT:    snez a2, a1
@@ -6602,11 +6622,12 @@
 ; CHECK-NOV-NEXT:    .cfi_offset s1, -24
 ; CHECK-NOV-NEXT:    .cfi_offset s2, -32
 ; CHECK-NOV-NEXT:    mv s2, a1
+; CHECK-NOV-NEXT:    fmv.w.x fa0, a0
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixsfti@plt
 ; CHECK-NOV-NEXT:    mv s0, a0
 ; CHECK-NOV-NEXT:    mv s1, a1
-; CHECK-NOV-NEXT:    mv a0, s2
+; CHECK-NOV-NEXT:    fmv.w.x fa0, s2
 ; CHECK-NOV-NEXT:    call __extendhfsf2@plt
 ; CHECK-NOV-NEXT:    call __fixsfti@plt
 ; CHECK-NOV-NEXT:    mv a2, a1
@@ -6659,11 +6680,12 @@
 ; CHECK-V-NEXT:    .cfi_offset s1, -24
 ; CHECK-V-NEXT:    .cfi_offset s2, -32
 ; CHECK-V-NEXT:    mv s2, a1
+; CHECK-V-NEXT:    fmv.w.x fa0, a0
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixsfti@plt
 ; CHECK-V-NEXT:    mv s0, a0
 ; CHECK-V-NEXT:    mv s1, a1
-; CHECK-V-NEXT:    mv a0, s2
+; CHECK-V-NEXT:    fmv.w.x fa0, s2
 ; CHECK-V-NEXT:    call __extendhfsf2@plt
 ; CHECK-V-NEXT:    call __fixsfti@plt
 ; CHECK-V-NEXT:    mv a2, a1