On Mon, Jun 11, 2012 at 02:06:29PM +0200, Guennadi Liakhovetski wrote: > On Mon, 11 Jun 2012, Mark Brown wrote: > > The stubs return NULL and we could decide in future to use NULL for > > something else (though that seems unlikely). > Ah, sure, forgot about that, sorry. In fact, I was wondering, whether it's > indeed a good idea to return success from functions like > regulator_set_voltage() or regulator_get_voltage() when regulator is even > not configured? Isn't it confusing for drivers to get a success from Anything that really cares about set_voltage() should have a hard dependency on the API, it's mainly done so that cpufreq type drivers can work easily (since they don't *really* care about the voltages normally, generally if nothing happens in the regulator they're just less effective). > set_voltage(voltage > 0) and then get 0V back from get_voltage()? There's not a use case I can think of for this, though.
Attachment:
signature.asc
Description: Digital signature
