[PATCH v2 1/2] leds: pca955x: Clean up and optimize

Eddie James eajames at linux.ibm.com
Tue Apr 12 01:17:14 AEST 2022


On 4/8/22 06:11, Andy Shevchenko wrote:
> On Thu, Apr 7, 2022 at 10:42 PM Eddie James <eajames at linux.ibm.com> wrote:
>> Clean up the I2C access functions to avoid fetching the pca955x
>> driver data again. Optimize the probe to do at most 4 reads and
>> 4 writes of the LED selector regs, rather than 16 of each.
>> Rename some functions and variables to be more consistent and
>> descriptive.
> Separate patch.


Ack.



>
>> +               dev_err(&pca955x->client->dev,
>> +                       "%s: reg 0x%x, val 0x%x, err %d\n", __func__, n, val,
>> +                       ret);
> This can be indented better.
>
> I would add a temporary dev pointer variable and put this on one line.
>
>   struct device *dev = &pca955x->client->dev;
>
>                 dev_err(dev, "%s: reg 0x%x, val 0x%x, err %d\n",
> __func__, n, val, ret);


Hm, I still end up having to use two lines for the dev_err function, so 
not sure it's worth it?


>
> ...
>
>> +               dev_err(&pca955x->client->dev,
>> +                       "%s: reg 0x%x, val 0x%x, err %d\n", __func__, n, val,
>> +                       ret);
> Ditto.
>
> ...
>
>> +               dev_err(&pca955x->client->dev,
>> +                       "%s: reg 0x%x, val 0x%x, err %d\n", __func__, n, val,
>> +                       ret);
> Ditto.
>
> ...
>
>> +               dev_err(&pca955x->client->dev, "%s: reg 0x%x, err %d\n",
>>                          __func__, n, ret);
> Ditto.
>
> ...
>
>> +               dev_err(&pca955x->client->dev, "%s: reg 0x%x, err %d\n",
>>                          __func__, n, ret);
> Ditto.
>
> ...
>
>> +       struct pca955x_led *pca955x_led = container_of(led_cdev,
>> +                                                      struct pca955x_led,
>> +                                                      led_cdev);
> Is it used once? If more than once, consider a helper for that as well.


Ack.


Thanks for the review!

Eddie

>


More information about the openbmc mailing list