Previously, the register selection in MacroAssembler:: decode_klass_not_null is misleading, better to refine it.

diff --git a/src/hotspot/cpu/riscv/macroAssembler_riscv.cpp b/src/hotspot/cpu/riscv/macroAssembler_riscv.cpp
index 1436bc02113..06186079cf8 100644
--- a/src/hotspot/cpu/riscv/macroAssembler_riscv.cpp
+++ b/src/hotspot/cpu/riscv/macroAssembler_riscv.cpp
@@ -3421,8 +3421,9 @@ void MacroAssembler::decode_klass_not_null(Register dst, Register src, Register
   mv(xbase, (uintptr_t)CompressedKlassPointers::base());
 
   if (CompressedKlassPointers::shift() != 0) {
- Register t = src == dst ? dst : t0;
+ Register t = dst == xbase ? t0 : dst;
     assert_different_registers(t, xbase);
+ // dst = (src << shift) + xbase
     shadd(dst, src, xbase, t, CompressedKlassPointers::shift());
   } else {
     add(dst, xbase, src);