This is specific to the acpi and should be handled in the processor_idle.c file instead of the cpuidle core code.
Could be the function 'acpi_processor_cst_has_changed' the right place to set a dummy power value for the power in the new C-state ?
Thanks for your feedback. I think it wouldn't be wise to split the dummy power value logic over two places, but I could submit a patch that makes set_power_states globally accessible and calls it from acpi_processor_cst_has_changed instead.
However, I do not think this should really be ACPI specific. It applies to any cpuidle driver that wants to change its idle states at runtime. Currently only the ACPI one does, but the future might bring others that would run into the same problem. I also think that set_power_states fits much better into cpuidle_enable_device conceptually anyway (right next to poll_idle_init which also does state initialization).
Let me know what you think.