On Tue, Feb 26, 2013 at 09:44:12AM +0000, Lee Jones wrote: > Okay, I've just re-read this thread and it appears you're both talking > about different things.
> Linus is talking about a discussion with Bengt regarding how these > strange sysclkreq thingies work. When I first read the code they > appear to be equivalent to GPIO regulators, but in actual fact the > hardware logic is different on enable/disable. So they probably don't > belong in regulator code at all. > Where as, Mark is complaining about how the regulators are initialised > by lots of magic register writes during init. Although, comments are I'm actually complaining about both things. > inserted for each of the values, they're by no means exhaustive and > aren't really even helpful if you don't have the uber-s3cr3t design > specification. What he would like to see is that most of this stuff > being handled by the framework. Some of this stuff is clearly only > setting voltages and power-states and the like. Right, it's the voltages and so on - the stuff that's clearly already in the framework domain.
signature.asc
Description: Digital signature
