Re: [PATCH v3 2/5] mfd: sec: Add support for S2MPS15 PMIC

[Krzysztof Kozlowski <k.kozlowski@samsung.com>]

W dniu 28.10.2015 o 17:46, Lee Jones pisze:
> On Wed, 28 Oct 2015, Krzysztof Kozlowski wrote:
> 
>> On 26.10.2015 23:34, Lee Jones wrote:
>>> On Mon, 26 Oct 2015, Alim Akhtar wrote:
>>>
>>>> From: Thomas Abraham <thomas.ab@samsung.com>
>>>>
>>>> Add support for S2MPS15 PMIC which is similar to S2MPS11 PMIC. The S2MPS15
>>>> PMIC supports 27 LDO regulators, 10 buck regulators, RTC, three 32.768KHz
>>>> clock outputs and battery charger. This patch adds initial support for
>>>> LDO and buck regulators of S2MPS15 device.
>>>>
>>>> Signed-off-by: Thomas Abraham <thomas.ab@samsung.com>
>>>> Signed-off-by: Alim Akhtar <alim.akhtar@samsung.com>
>>>> [Alim: Added s2mps15_devs like rtc and clk and related changes]
>>>> Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
>>>> ---
>>>>  drivers/mfd/sec-core.c              |   31 +++++++
>>>>  drivers/mfd/sec-irq.c               |    8 ++
>>>>  include/linux/mfd/samsung/core.h    |    1 +
>>>>  include/linux/mfd/samsung/s2mps15.h |  158 +++++++++++++++++++++++++++++++++++
>>>>  4 files changed, 198 insertions(+)
>>>>  create mode 100644 include/linux/mfd/samsung/s2mps15.h
>>>
>>> I replied to the previous set and won't be reviewing this one until
>>> all of the open points are solved.
>>
>> The naming and compatibles used by the driver are confusing but how it
>> was at beginning. Beside the confusion, the names are correct:
>>
>> 1. Main mfd driver:
>>  - compatible: samsung,s2mps1*-pmic
>>  - driver name: sec_pmic
>>
>> 2. Regulator driver:
>>  - no compatible (because it always searches for "regulators" subnode of
>> its parent... that is the convention/legacy behaviour)
>>  - driver name: s2mps1*-pmic
>>
>> I hope that explains your concerns.
> 
> It explains *why*, but doesn't ease my concerns in any way.
> 
> Unfortunately I've only just realised the disparity we have between
> MFD and the Regulator subsystem, which is annoying because it's now
> almost impossible to rectify.  
> 
> We should have taken one of two views.  Either a) The MFD is the PMIC
> device which encompasses regulator control.  In which case the MFD
> and it's corresponding compatible string would be named *-pmic and the
> regulator driver would be called *-regulator. Or b) The MFD could be
> considered a normal MFD and be named after the model number, then the
> regulator 'could' be named *-pmic.
> 
> However, with reference to b), how much other Power Management does
> the regulator driver do besides control regulators?  I suspect not
> much.  Therefore my preference would be for a).  My second choice
> would be a mixuture of the two where nothing gets named *-pmic.  The
> last option on my list would be the current situation where we seem to
> be calling both the MFD (PMIC) itself and the Regulator driver
> *-pmic, which is not good.

Starting from the description of device-family. This is called "Power
Management IC" but it is rather a "Power Deliver/Distribute IC". There
isn't any logic inside except enable/disable/configure/set low power
mode for regulators.

However in the same time the IC comes (always) with:
 - 32kHz clocks,
 - RTC,
 - backup battery charger (no driver for it),
 - reset for SoC,
 - shutdown on thermal alert (also no driver for this control).

The solution a) seems fine to me. Make sense and it looks entirely
backward compatible - only driver names will be modified.

Best regards,
Krzysztof