This patch makes the assumption that the code doing the cmos choice has CONFIG_USE_OPTION_TABLE enabled.

Is this always the case?