[RISCV][VLOPT] Add vmv.x.s and vfmv.f.s to isVectorOpUsedAsScalarOp

[michaelmaitland]

There is no diff compared to the precommit, but we will see a diff in #120345

[llvmbot]

@llvm/pr-subscribers-backend-risc-v

Author: Michael Maitland (michaelmaitland)

Changes

---

Full diff: https://github.com/llvm/llvm-project/pull/121588.diff

2 Files Affected:

• (modified) llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp (+3)
• (modified) llvm/test/CodeGen/RISCV/rvv/vl-opt.mir (+13)

diff --git a/llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp b/llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp
index 85ea5a23e8f293..14b0c9d641d2d7 100644
--- a/llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp
+++ b/llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp
@@ -840,6 +840,9 @@ static bool isVectorOpUsedAsScalarOp(MachineOperand &MO) {
   case RISCV::VFWREDOSUM_VS:
   case RISCV::VFWREDUSUM_VS:
     return MO.getOperandNo() == 3;
+  case RISCV::VMV_X_S:
+  case RISCV::VFMV_F_S:
+    return MO.getOperandNo() == 1;
   default:
     return false;
   }
diff --git a/llvm/test/CodeGen/RISCV/rvv/vl-opt.mir b/llvm/test/CodeGen/RISCV/rvv/vl-opt.mir
index 3f966b036589fd..2df317c4e975ad 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vl-opt.mir
+++ b/llvm/test/CodeGen/RISCV/rvv/vl-opt.mir
@@ -110,4 +110,17 @@ body: |
     %y:vr = PseudoVADD_VV_M1 $noreg, %x, $noreg, 1, 3 /* e8 */, 0
     %z:vr = PseudoVADD_VV_M1 $noreg, %x, $noreg, -1, 3 /* e8 */, 0
 ...
+---
+name: vmv_is_used_as_scalar_op
+body: |
+  bb.0:
+    liveins: $x1
+    ; CHECK-LABEL: name: vmv_is_used_as_scalar_op
+    ; CHECK: liveins: $x1
+    ; CHECK-NEXT: {{  $}}
+    ; CHECK-NEXT: %x:vr = PseudoVADD_VV_MF4 $noreg, $noreg, $noreg, 1, 3 /* e8 */, 0 /* tu, mu */
+    ; CHECK-NEXT: %y:gpr = PseudoVMV_X_S %x, 3 /* e8 */
+    %x:vr = PseudoVADD_VV_MF4 $noreg, $noreg, $noreg, -1, 3 /* e8 */, 0 /* tu, mu */
+    %y:gpr = PseudoVMV_X_S %x, 3 /* e8 */
+...
 

[lukel97]

LGTM

[lukel97]

I don't think the test is necessary since there's no VL being reduced