Re: [PATCH 1/2] ARM: dts: omap5-uevm: remove always_on, boot_on from smps10_out1

[Nishanth Menon <nm@ti.com>]

On Fri, Oct 11, 2013 at 1:54 AM, Kishon Vijay Abraham I <kishon@ti.com> wrote:
> On Friday 11 October 2013 12:23 PM, Kishon Vijay Abraham I wrote:
>> Hi,
>>
>> On Friday 11 October 2013 12:00 PM, Nishanth Menon wrote:
>>> On Fri, Oct 11, 2013 at 1:13 AM, Kishon Vijay Abraham I <kishon@ti.com> wrote:
>>>>
>>>>> regulator-boot-on indicates that PMIC enables it by default as part of
>>>>> OTP or some internal behavior -> Looking at the measurements done on
>>>>> uEVM and OTP information -> regulator-boot-on should be kept here.
>>>>
>>>> No. Actually I don’t want PMIC to enable it by default. I want the palmas-usb
>>>> driver to handle it.
>>>> Enabling it by default makes palmas-usb to detect VBUS interrupt. This should
>>>> ideally be detected only when you connect a host cable.
>>>> Btw I didn't exactly get why you want regulator-boot-on should be kept here.
>>>
>>> binding description states:
>>> - regulator-boot-on: bootloader/firmware enabled regulator
>>> Further info: include/linux/regulator/machine.h
>>> * @boot_on: Set if the regulator is enabled when the system is initially
>>> *           started.  If the regulator is not enabled by the hardware or
>>> *           bootloader then it will be enabled when the constraints are
>>> *           applied.
>>>
>>> What that means is that it is enabled by firmware/bootloader (in our
>>> case One Time Program {OTP} inside Palmas) when the system switches on
>>> even before the kernel starts. and we know SMPS10 is autoenabled by
>>> Palmas OTP configuration even before first instruction in A15
>>> executes.
>>
>> Not sure about that. Please note SMPS10 has two outputs OUT1 and OUT2 and I
>> tend to think that it might be OUT2 that's getting enabled by the OTP.
>>>
>>> I think you misunderstand this to mean that you'd like the regulator
>>> to be *switched on* automatically at kernel boot by regulator
>>> framework - there is no reasoning why we'd want such a binding since
>>> we'd expect drivers to do their job of requesting and enabling
>>> regulators on need..
>>
>> The comment you just quoted tells it enables the regulator if its not enabled
>> by hardware. "If the regulator is not enabled by the hardware or bootloader
>> then it will be enabled when the constraints are applied." At-least that's what
>> I understood from that comment.
>>
>> Also from our experiments it doesn't look like SMPS10_OUT1 is enabled by the
>> OTP and it gets enabled when we have *regulator-boot-on* constraints.
>
> btw.. I think this is the code in regulator fw that's responsible for enabling..
>
>         /* If the constraints say the regulator should be on at this point
>          * and we have control then make sure it is enabled.
>          */
>         if ((rdev->constraints->always_on || rdev->constraints->boot_on) &&
>             ops->enable) {
>                 ret = ops->enable(rdev);
>                 if (ret < 0) {
>                         rdev_err(rdev, "failed to enable\n");
>                         goto out;
>                 }
>         }

Drat, you are right, I did not really dig deep. thanks for correcting
my understanding here.

I propose the following change in binding as it seems completely
misleading to me.

diff --git a/Documentation/devicetree/bindings/regulator/regulator.txt
b/Documentation/devicetree/bindings/regulator/regulator.txt
index 2bd8f09..d999f096 100644
--- a/Documentation/devicetree/bindings/regulator/regulator.txt
+++ b/Documentation/devicetree/bindings/regulator/regulator.txt
@@ -8,7 +8,9 @@ Optional properties:
 - regulator-min-microamp: smallest current consumers may set
 - regulator-max-microamp: largest current consumers may set
 - regulator-always-on: boolean, regulator should never be disabled
-- regulator-boot-on: bootloader/firmware enabled regulator
+- regulator-boot-on: regulator is enabled when the system is initially started.
+  If the regulator is not enabled by the hardware or bootloader then it will be
+  enabled when the constraints are applied.
 - regulator-allow-bypass: allow the regulator to go into bypass mode
 - <name>-supply: phandle to the parent supply/regulator node
 - regulator-ramp-delay: ramp delay for regulator(in uV/uS)