Follow-up from JDK-8325106: We hit an assert because we create a wrong JVMState in LibraryCallKit::arraycopy_restore_alloc_state for re-executing the newNullRestrictedArray intrinsic because the code does not account for the class argument on stack in addition to the size.