Re: [PATCH v4 1/5] gpio-f7188x: Add GPIO support for Nuvoton NCT6116

[Hans de Goede <hdegoede@redhat.com>]

Hi Henning,

On 8/24/22 15:50, Henning Schild wrote:
> Am Wed, 24 Aug 2022 15:10:55 +0200
> schrieb simon.guinot@sequanux.org:
> 
>> On Tue, Aug 23, 2022 at 04:54:59PM +0200, Henning Schild wrote:
>>> Am Tue, 23 Aug 2022 17:47:38 +0300
>>> schrieb Andy Shevchenko <andriy.shevchenko@linux.intel.com>:  
>>
>> Hi Andy,
>>
>> Thanks for this new version. It is looking good to me.
>>
>>>   
>>>> On Tue, Aug 23, 2022 at 12:23:40PM +0200, Henning Schild wrote:  
>>>>> Add GPIO support for Nuvoton NCT6116 chip. Nuvoton SuperIO
>>>>> chips are very similar to the ones from Fintek. In other
>>>>> subsystems they also share drivers and are called a family of
>>>>> drivers.
>>>>>
>>>>> For the GPIO subsystem the only difference is that the direction
>>>>> bit is reversed and that there is only one data bit per pin. On
>>>>> the SuperIO level the logical device is another one.
>>>>>
>>>>> On a chip level we do not have a manufacturer ID to check and
>>>>> also no revision.    
>>>>
>>>> ...
>>>>   
>>>>> - * GPIO driver for Fintek Super-I/O F71869, F71869A, F71882,
>>>>> F71889 and F81866
>>>>> + * GPIO driver for Fintek and Nuvoton Super-I/O chips    
>>>>
>>>> I'm not sure it's good idea to drop it from here. It means reader
>>>> has to get this info in a hard way.
>>>>
>>>> ...  
>>>
>>> Let us see what others say. I wanted to keep this in line with what
>>> Kconfig says and the oneliner in the Kconfig was getting pretty
>>> longish. Hence i decided to shorten that. Other drivers also seem to
>>> not list all the possible chips in many places, it is all maint
>>> effort when a new chips is added and the list is in like 5 places.  
>>
>> I agree with you that we can drop this line. It was already incomplete
>> and the information is quite readable a few lines below in both the
>> define list and the chip enumeration.
>>
>>>   
>>>>> +#define gpio_dir_invert(type)	((type) == nct6116d)
>>>>> +#define gpio_data_single(type)	((type) == nct6116d)    
>>>>
>>>> What's prevents us to add a proper prefix to these? I don't like
>>>> the idea of them having "gpio" prefix.
>>>>
>>>> ...
>>>>   
>>>>> +		pr_info(DRVNAME ": Unsupported device
>>>>> 0x%04x\n", devid);
>>>>> +			pr_debug(DRVNAME ": Not a Fintek
>>>>> device at 0x%08x\n", addr);
>>>>> +	pr_info(DRVNAME ": Found %s at %#x\n",
>>>>> +		pr_info(DRVNAME ":   revision %d\n",    
>>>>
>>>> Can we, please, utilize pr_fmt()?
>>>>   
>>>>> +			(int)superio_inb(addr,
>>>>> SIO_FINTEK_DEVREV));    
>>>>
>>>> Explicit casting in printf() means wrong specifier in 99% of
>>>> cases. 
>>>
>>> For all the other comments i will wait for a second opinion. I
>>> specifically did not change existing code for more than the
>>> functional changes needed. And a bit of checkpatch.pl fixing.
>>> Beautification could be done on the way but would only cause
>>> inconsistency. That driver is what it is, if someone wants to
>>> overhaul the style ... that should be another patch. One likely not
>>> coming from me.  
>>
>> About the int cast, I think you can drop it while you are updating
>> this line. It is unneeded.
> 
> Ok two voices for doing that one fix along the way. I will send a v5
> and hope nobody insists on me fixing the other findings in code i never
> wrote.

You did not write it, but you are using it to do hw-enablement for
your company's products. So being asked to also some touch-ups
left and right while you are at it really is not unexpected IMHO.

Regards,

Hans