Hi,
On 11/2/21 16:52, Andy Shevchenko wrote:
> On Tue, Nov 2, 2021 at 4:59 PM Hans de Goede <hdegoede@xxxxxxxxxx> wrote:
>> On 11/2/21 15:34, Andy Shevchenko wrote:
>>> On Tue, Nov 2, 2021 at 11:50 AM Hans de Goede <hdegoede@xxxxxxxxxx> wrote:
>
> ...
>
>>>> + board_data = int3472_tps68470_get_board_data(dev_name(&client->dev));
>>>> + if (!board_data) {
>>>> + dev_err(&client->dev, "No board-data found for this laptop/tablet model\n");
>>>> + return -ENODEV;
>>>
>>> It's fine to use dev_err_probe() for known error codes.
>>>
>>>> + }
>>>
>>> ...
>>>
>>>> + cells[1].platform_data = (void *)board_data->tps68470_regulator_pdata;
>>>
>>> Do we need casting?
>>
>> Yes, the cast casts away a "const", the const is correct
>> since the data only ever gets read by the regulator driver,
>> but platform_data pointers are normally not const, so it
>> is either the cast, or loose the const on the definition
>> of the struct to which board_data->tps68470_regulator_pdata
>> points...
>>
>> So not good choice here really, only chosing between bad
>> options and I picked the lets do the cast "least worse"
>> option (at least to me). I'm open to changing this.
>
> Hmm... Okay, I was under the impression that MFD is using const for
> that field...
Nope, I just double-checked and it is a plain "void *"
Regards,
Hans
