[RESEND][PATCH] iio: simplify with regmap_set_bits(), regmap_clear_bits()

[Trevor Gamblin]

On 2024-06-13 1:20 p.m., Jonathan Cameron wrote:
> On Tue, 11 Jun 2024 12:52:06 -0400
> Trevor Gamblin <tgamblin at baylibre.com> wrote:
>
>> Simplify the way regmap is accessed in iio drivers.
>>
>> Instead of using regmap_update_bits() and passing the mask twice, use
>> regmap_set_bits().
>>
>> Instead of using regmap_update_bits() and passing val = 0, use
>> regmap_clear_bits().
>>
>> Suggested-by: Uwe Kleine-König <u.kleine-koenig at baylibre.com>
>> Signed-off-by: Trevor Gamblin <tgamblin at baylibre.com>
> Looks like a good change in general. However...
>
> The problem with a change like this is it results in
> non trivial backporting if we need to due to a fix in related
> code.
>
> As such, whilst it will obviously generate a lot of patches, I'd
> like this split up into a series so each patch touches only one driver.
> Fine to keep a single patch for the multiple module for a single
> device cases though.
>
> Also some very long lines that need the line breaks put back.

Thank you. I'll resubmit as a series of separate patches with those 
fixes soon.

Trevor

>
> Jonathan
>
>
>> diff --git a/drivers/iio/accel/msa311.c b/drivers/iio/accel/msa311.c
>> index b8ddbfd98f11..40e605c57adb 100644
>> --- a/drivers/iio/accel/msa311.c
>> +++ b/drivers/iio/accel/msa311.c
>> @@ -1034,10 +1034,8 @@ static int msa311_chip_init(struct msa311_priv *msa311)
>>   				     "failed to unmap map0/map1 interrupts\n");
>>   
>>   	/* Disable all axes by default */
>> -	err = regmap_update_bits(msa311->regs, MSA311_ODR_REG,
>> -				 MSA311_GENMASK(F_X_AXIS_DIS) |
>> -				 MSA311_GENMASK(F_Y_AXIS_DIS) |
>> -				 MSA311_GENMASK(F_Z_AXIS_DIS), 0);
>> +	err = regmap_clear_bits(msa311->regs, MSA311_ODR_REG,
>> +				MSA311_GENMASK(F_X_AXIS_DIS) | MSA311_GENMASK(F_Y_AXIS_DIS) | MSA311_GENMASK(F_Z_AXIS_DIS));
> Too long
>
>> diff --git a/drivers/iio/adc/cpcap-adc.c b/drivers/iio/adc/cpcap-adc.c
>> index b6c4ef70484e..8fabf748c36b 100644
>> --- a/drivers/iio/adc/cpcap-adc.c
>> +++ b/drivers/iio/adc/cpcap-adc.c
>> @@ -385,9 +385,8 @@ static irqreturn_t cpcap_adc_irq_thread(int irq, void *data)
>> @@ -424,23 +423,17 @@ static void cpcap_adc_setup_calibrate(struct cpcap_adc *ddata,
>>   	if (error)
>>   		return;
>>   
>> -	error = regmap_update_bits(ddata->reg, CPCAP_REG_ADCC2,
>> -				   CPCAP_BIT_ATOX_PS_FACTOR |
>> -				   CPCAP_BIT_ADC_PS_FACTOR1 |
>> -				   CPCAP_BIT_ADC_PS_FACTOR0,
>> -				   0);
>> +	error = regmap_clear_bits(ddata->reg, CPCAP_REG_ADCC2,
>> +				  CPCAP_BIT_ATOX_PS_FACTOR | CPCAP_BIT_ADC_PS_FACTOR1 | CPCAP_BIT_ADC_PS_FACTOR0);
> That one is over 100!
>>   	if (error)
>> diff --git a/drivers/iio/gyro/mpu3050-core.c b/drivers/iio/gyro/mpu3050-core.c
>> index a791ba3a693a..ff1c81553045 100644
>> --- a/drivers/iio/gyro/mpu3050-core.c
>> +++ b/drivers/iio/gyro/mpu3050-core.c
>>   
>> @@ -513,12 +513,8 @@ static irqreturn_t mpu3050_trigger_handler(int irq, void *p)
>>   				 "FIFO overflow! Emptying and resetting FIFO\n");
>>   			fifo_overflow = true;
>>   			/* Reset and enable the FIFO */
>> -			ret = regmap_update_bits(mpu3050->map,
>> -						 MPU3050_USR_CTRL,
>> -						 MPU3050_USR_CTRL_FIFO_EN |
>> -						 MPU3050_USR_CTRL_FIFO_RST,
>> -						 MPU3050_USR_CTRL_FIFO_EN |
>> -						 MPU3050_USR_CTRL_FIFO_RST);
>> +			ret = regmap_set_bits(mpu3050->map, MPU3050_USR_CTRL,
>> +					      MPU3050_USR_CTRL_FIFO_EN | MPU3050_USR_CTRL_FIFO_RST);
> Keep the line break to stay under 80 chars.
>
>> @@ -997,11 +991,8 @@ static int mpu3050_drdy_trigger_set_state(struct iio_trigger *trig,
>>   			return ret;
>>   
>>   		/* Reset and enable the FIFO */
>> -		ret = regmap_update_bits(mpu3050->map, MPU3050_USR_CTRL,
>> -					 MPU3050_USR_CTRL_FIFO_EN |
>> -					 MPU3050_USR_CTRL_FIFO_RST,
>> -					 MPU3050_USR_CTRL_FIFO_EN |
>> -					 MPU3050_USR_CTRL_FIFO_RST);
>> +		ret = regmap_set_bits(mpu3050->map, MPU3050_USR_CTRL,
>> +				      MPU3050_USR_CTRL_FIFO_EN | MPU3050_USR_CTRL_FIFO_RST);
> For IIO stuff try and stay under 80 chars unless there is a strong
> readability argument for going longer.  Here there isn't one.
>
>>   		if (ret)
>>   			return ret;
>>