On Tue, Dec 10, 2013 at 11:59:18AM +0530, Sachin Kamat wrote: > Added PMIC node to Arndale-Octa board. Is there a bootloader posted for this anywhere yet? > + regulators { > + ldo1_reg: LDO1 { > + regulator-name = "vdd_ldo1"; So, there is a label on the schematic for PVDD_LDO1 but it's not entirely obvious why that's there as throughout the rest of the schematic it's got the rather more constructive name PVDD_ALIVE_1V0. We should really be using the more readable versions of the names in the DT since the whole point in overriding the driver default names is to make it easier to associate the kernel diagnostics with the schematic, what the DT is doing here serves no purpose. The same comment applies to all the LDO supplies. > + buck2_reg: BUCK2 { > + regulator-name = "vdd_arm"; > + regulator-min-microvolt = <800000>; > + regulator-max-microvolt = <1500000>; > + regulator-always-on; > + regulator-boot-on; > + }; Why are you specifying boot-on? It doesn't seem terribly plausible that the cores won't be powered on boot, or that we can't tell what the state is by reading the registers. Same for all the DCDCs. > + buck7_reg: BUCK7 { > + regulator-name = "vdd_1.0v_ldo"; > + regulator-min-microvolt = <800000>; > + regulator-max-microvolt = <1500000>; > + regulator-always-on; > + regulator-boot-on; > + }; This doesn't correspond to the schematic which says that BUCK7 is supplying VIN_LLDO_1V4 (the input for some of the LDOs). The voltage range specified there doesn't seem terribly sensible either, are you sure you haven't just got the underlying voltage range for the DCDC? I can't see why you'd want to raise the voltage over 1.4V (that's rather defeating the point) and while it'd be nice to have the core be able to vary supply to track the headroom needed by the LDOs none of the LDOs were configured with variable voltage so even if we implemented that it'd never be used. Similar issues apply to the other bucks being used to drop the battery voltage for LDO supplies.
Attachment:
signature.asc
Description: Digital signature