[AMDGPU] Stop hardcoding v_mov_b32 in VOPD. NFCI.

[rampitec]

Summary

There is no single possible reason to to it.

llvm/lib/Target/AMDGPU/VOP1Instructions.td

multiclass VOP1Inst <string opName, VOPProfile P,
   defvar should_mov_imm = !or(!eq(opName, "v_mov_b32"),
                               !eq(opName, "v_mov_b64"));
 
   let isMoveImm = should_mov_imm in {
     if !eq(VOPDOp, -1) then
       def _e32 : VOP1_Pseudo <opName, P>;
     else
       // Only for V_MOV_B32
-      def _e32 : VOP1_Pseudo <opName, P>, VOPD_Component<VOPDOp, "v_mov_b32">;
+      def _e32 : VOP1_Pseudo <opName, P>, VOPD_Component<VOPDOp, opName>;
     def _e64 : VOP3InstBase <opName, P, node>;
   }
 
   if P.HasExtSDWA then
     def _sdwa : VOP1_SDWA_Pseudo <opName, P>;
 
   if P.HasExtDPP then
     def _dpp : VOP1_DPP_Pseudo <opName, P>;