JDK-8341895 shows we have a single minor inefficiency with bit moves in copy_memory, which we can easily fix without reaching for a generic fix.
 https://github.com/openjdk/jdk/blob/8174cbd5cb797a80d48246a686897ef6fe64ed57/src/hotspot/cpu/aarch64/stubGenerator_aarch64.cpp#L1407

It can be similar to the optimization we do a few lines above:
https://github.com/openjdk/jdk/blob/8174cbd5cb797a80d48246a686897ef6fe64ed57/src/hotspot/cpu/aarch64/stubGenerator_aarch64.cpp#L1376-L1377

(And that optimizations should really be "shift > 0", not just implicit boolean conversion)