Re: [PATCH] regulator: Support different config and dev of_nodes in regulator_register

[Tim Bird <tim.bird@sonymobile.com>]

On 02/06/2015 03:49 AM, Mark Brown wrote:
> On Thu, Feb 05, 2015 at 04:52:40PM -0800, Bjorn Andersson wrote:
>> On Thu 05 Feb 16:32 PST 2015, Mark Brown wrote:
>>> On Thu, Feb 05, 2015 at 02:08:54PM -0800, Bjorn Andersson wrote:
> 
>>>> However this only works for the non-supply regulator properties - and
>>>> this is where Tim's patch is trying to sort out.
> 
>>> No, this works completely fine for supply properties - to repeat what I
>>> said in reply to the original patch the supply is a supply to the chip
>>> not to an individual IP on the chip.
> 
>> It does make some sense to consider the vbus-supply being connected to
>> the block, rather than directly to the vbus-switch. So it would work for
>> Tim's use case.
> 
> Like I say if we do that then we don't have consistency in how we map a
> schematic into a DT binding - you have to dig into the binding of each
> device and figure out if the supply is viewed as being for blocks or for
> the chip as a whole and we've got the potential for problems in the
> binding if we figure out that a supply is actually used by other blocks
> later on and don't want to break existing DTs.

OK - the light bulb finally went on for me on this one.
So a chip can have multiple supplies (I saw examples of this 
poking around in other source), and the details of
internal routing in the chip don't have to be expressed in
DT at all (in fact shouldn't, for the reason you mention).

Thanks - I will implement along these lines.
 -- Tim