Re: [PATCH 1/2] iio: adc: imx93: Make calibration properties configurable

From: Jonathan Cameron <jic23@kernel.org>
To: Andrej Picej <andrej.picej@norik.com>
Cc: haibo.chen@nxp.com, linux-iio@vger.kernel.org,
	devicetree@vger.kernel.org, lars@metafoo.de, shawnguo@kernel.org,
	s.hauer@pengutronix.de, kernel@pengutronix.de,
	festevam@gmail.com, imx@lists.linux.dev,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, robh@kernel.org,
	krzysztof.kozlowski+dt@linaro.org, conor+dt@kernel.org,
	upstream@lists.phytec.de
Subject: Re: [PATCH 1/2] iio: adc: imx93: Make calibration properties configurable
Date: Sun, 24 Mar 2024 14:02:33 +0000	[thread overview]
Message-ID: <20240324140233.577ec7d3@jic23-huawei> (raw)
In-Reply-To: <20240320100407.1639082-2-andrej.picej@norik.com>

On Wed, 20 Mar 2024 11:04:05 +0100
Andrej Picej <andrej.picej@norik.com> wrote:

> Make calibration properties:
>  - AVGEN: allow averaging of calibration time,

Confused. How do you average time?   Or is this enabling averaging of
ADC readings at calibration time?

>  - NRSMPL: select the number of averaging samples during calibration and

Assuming I read AVGEN right, just have a value of 1 in here mean AVGEN is
disabled.

>  - TSAMP: specifies the sample time of calibration conversions

Not sure what this means.  Is it acquisition time? Is it time after a mux
changes?  Anyhow, more info needed.
This is the only one I can see being possibly board related.  But if it
is and is needed for calibration, why not for normal read out?

> 
> configurable with device tree properties:
>  - nxp,calib-avg-en,
>  - nxp,calib-nr-samples and
>  - nxp,calib-t-samples.
> 
> Signed-off-by: Andrej Picej <andrej.picej@norik.com>
> ---
>  drivers/iio/adc/imx93_adc.c | 66 ++++++++++++++++++++++++++++++++++---
>  1 file changed, 61 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/iio/adc/imx93_adc.c b/drivers/iio/adc/imx93_adc.c
> index 4ccf4819f1f1..ad24105761ab 100644
> --- a/drivers/iio/adc/imx93_adc.c
> +++ b/drivers/iio/adc/imx93_adc.c
> @@ -18,6 +18,7 @@
>  #include <linux/platform_device.h>
>  #include <linux/pm_runtime.h>
>  #include <linux/regulator/consumer.h>
> +#include <linux/property.h>
>  
>  #define IMX93_ADC_DRIVER_NAME	"imx93-adc"
>  
> @@ -43,6 +44,9 @@
>  #define IMX93_ADC_MCR_MODE_MASK			BIT(29)
>  #define IMX93_ADC_MCR_NSTART_MASK		BIT(24)
>  #define IMX93_ADC_MCR_CALSTART_MASK		BIT(14)
> +#define IMX93_ADC_MCR_AVGEN_MASK		BIT(13)
> +#define IMX93_ADC_MCR_NRSMPL_MASK		GENMASK(12, 11)
> +#define IMX93_ADC_MCR_TSAMP_MASK		GENMASK(10, 9)
>  #define IMX93_ADC_MCR_ADCLKSE_MASK		BIT(8)
>  #define IMX93_ADC_MCR_PWDN_MASK			BIT(0)
>  #define IMX93_ADC_MSR_CALFAIL_MASK		BIT(30)
> @@ -145,7 +149,7 @@ static void imx93_adc_config_ad_clk(struct imx93_adc *adc)
>  
>  static int imx93_adc_calibration(struct imx93_adc *adc)
>  {
> -	u32 mcr, msr;
> +	u32 mcr, msr, value;
>  	int ret;
>  
>  	/* make sure ADC in power down mode */
> @@ -156,12 +160,64 @@ static int imx93_adc_calibration(struct imx93_adc *adc)
>  	mcr &= ~FIELD_PREP(IMX93_ADC_MCR_ADCLKSE_MASK, 1);
>  	writel(mcr, adc->regs + IMX93_ADC_MCR);
>  
> -	imx93_adc_power_up(adc);
> -
>  	/*
> -	 * TODO: we use the default TSAMP/NRSMPL/AVGEN in MCR,
> -	 * can add the setting of these bit if need in future.
> +	 * Set calibration settings:
> +	 * - AVGEN: allow averaging of calibration time,
> +	 * - NRSMPL: select the number of averaging samples during calibration,
> +	 * - TSAMP: specifies the sample time of calibration conversions.
>  	 */
> +	if (!device_property_read_u32(adc->dev, "nxp,calib-avg-en", &value)) {
> +		mcr &= ~IMX93_ADC_MCR_AVGEN_MASK;
> +		mcr |= FIELD_PREP(IMX93_ADC_MCR_AVGEN_MASK, value);
> +	}
> +
> +	if (!device_property_read_u32(adc->dev, "nxp,calib-nr-samples", &value)) {
Handle error for not present different from a failure to read or similar.
Not present isn't an error, just a fall back to defaults.
For other error codes we should fail the probe.
> +		switch (value) {
> +		case 16:
> +			value = 0x0;
Don't do this in place, meaning of value before this point different to what
you have in it going forwards. Use a different variable name to make that clear.
reg_val vs nr_samples perhaps?
> +			break;
> +		case 32:
> +			value = 0x1;
> +			break;
> +		case 128:
> +			value = 0x2;
> +			break;
> +		case 512:
> +			value = 0x3;
> +			break;
> +		default:
> +			dev_warn(adc->dev, "NRSMPL: wrong value, using default: 512\n");

Fail the probe rather than papering over a wrong value. I'd rather we got the DT fixed
quickly and if someone wanted another value, they really did want it.

We probably do want a default though for the property not being present (given it is new).
so take setting of this variable outside the if(!device_property_read_u32);
> +			value = 0x3;
> +		}
> +		mcr &= ~IMX93_ADC_MCR_NRSMPL_MASK;
> +		mcr |= FIELD_PREP(IMX93_ADC_MCR_NRSMPL_MASK, value);
> +	}
> +
> +	if (!device_property_read_u32(adc->dev, "nxp,calib-t-samples", &value)) {
> +		switch (value) {
> +		case 8:
> +			value = 0x1;
> +			break;
> +		case 16:
> +			value = 0x2;
> +			break;
> +		case 22:
> +			value = 0x0;
> +			break;
> +		case 32:
> +			value = 0x3;
> +			break;
> +		default:
> +			dev_warn(adc->dev, "TSAMP: wrong value, using default: 22\n");
> +			value = 0x0;
> +		}
> +		mcr &= ~IMX93_ADC_MCR_TSAMP_MASK;
> +		mcr |= FIELD_PREP(IMX93_ADC_MCR_TSAMP_MASK, value);
> +	}
> +
> +	writel(mcr, adc->regs + IMX93_ADC_MCR);
> +
> +	imx93_adc_power_up(adc);
>  
>  	/* run calibration */
>  	mcr = readl(adc->regs + IMX93_ADC_MCR);