Re: [RESEND PATCH v6 3/3] input: pm8xxx-vibrator: add new SPMI vibrator support

[Fenglin Wu <quic_fenglinw@quicinc.com>]

On 10/9/2023 12:01 PM, Fenglin Wu wrote:
> 
> 
> On 10/1/2023 12:17 AM, Dmitry Torokhov wrote:
>> On Mon, Sep 25, 2023 at 10:54:45AM +0800, Fenglin Wu wrote:
>>>
>>>
>>> On 9/24/2023 3:07 AM, Dmitry Baryshkov wrote:
>>>>> +
>>>>> +       switch (vib->data->hw_type) {
>>>>> +       case SSBI_VIB:
>>>>>                   mask = SSBI_VIB_DRV_LEVEL_MASK;
>>>>>                   shift = SSBI_VIB_DRV_SHIFT;
>>>>> +               break;
>>>>> +       case SPMI_VIB:
>>>>> +               mask = SPMI_VIB_DRV_LEVEL_MASK;
>>>>> +               shift = SPMI_VIB_DRV_SHIFT;
>>>>> +               break;
>>>>> +       case SPMI_VIB_GEN2:
>>>>> +               mask = SPMI_VIB_GEN2_DRV_MASK;
>>>>> +               shift = SPMI_VIB_GEN2_DRV_SHIFT;
>>>>> +               break;
>>>>> +       default:
>>>>> +               return -EINVAL;
>>>> Could you please move the switch to the previous patch? Then it would
>>>> be more obvious that you are just adding the SPMI_VIB_GEN2 here.
>>>>
>>>> Other than that LGTM.
>>>
>>> Sure, I can move the switch to the previous refactoring patch.
>>
>> Actually, the idea of having a const "reg" or "chip", etc. structure is
>> to avoid this kind of runtime checks based on hardware type and instead
>> use common computation. I believe you need to move mask and shift into
>> the chip-specific structure and avoid defining hw_type.
>>
> 
> Actually, the main motivation for adding 'hw_type' is to avoid reading 
> 'reg_base' from DT for SSBI_VIB. It can also help to simplify the 
> 'pm8xxx_vib_data' structure and make following code logic more 
> straightforward and easier to understand(check hw_type instead of 
> checking specific constant reg/mask value), it has been used in 
> following places:
> 
>    1) Avoid reading 'reg_base' from DT for SSBI_VIB.
>    2) Only do manual-mode-mask-write for SSBI_VIB. Previously, it was 
> achieved by giving a valid 'drv_en_manual_mask' value only for SSBI_VIB, 
> having hw_type make it more straightforward.
>    3) Not writing VIB_EN register for SSBI_VIB. A similar strategy was 
> used previously, only write VIB_EN register when 'enable_mask' is valid, 
>   checking hw_type make it more straightforward.
>    4) To achieve different drive step size for SPMI_VIB （100mV per 
> step) and SPMI_VIB_GEN2 (1mV per step).
>    5) Do different VIB_DRV mask and shift assignment for SPMI_VIB and 
> SPMI_VIB_GEN2
>    6) Only write VIB_DRV2 for SPMI_VIB_GEN2.
> 

Hi Dmitry,

Can you please help to comment if this looks good for you?
I actually have pushed a V7 to address your last comment before you made 
this one.
V7 change: 
https://lore.kernel.org/linux-arm-msm/20230927-pm8xxx-vibrator-v7-1-b5d8c92ce818@quicinc.com/, 
just want to know how to move forward.
Thanks

Fenglin

> 
>> Thanks.
>>