2022-02-18 14:53:40

by Billy Tsai

[permalink] [raw]
Subject: [PATCH] iio: adc: aspeed: Add divider flag to fix incorrect voltage reading.

The formula for the ADC sampling period in ast2400/ast2500 is:
ADC clock period = PCLK * 2 * (ADC0C[31:17] + 1) * (ADC0C[9:0])
When ADC0C[9:0] is set to 0 the sampling voltage will be lower than
expected, because the hardware may not have enough time to
charge/discharge to a stable voltage.

Reported-by: Konstantin Klubnichkin <[email protected]>
Signed-off-by: Billy Tsai <[email protected]>
---
drivers/iio/adc/aspeed_adc.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/iio/adc/aspeed_adc.c b/drivers/iio/adc/aspeed_adc.c
index a957cad1bfab..ffae64f39221 100644
--- a/drivers/iio/adc/aspeed_adc.c
+++ b/drivers/iio/adc/aspeed_adc.c
@@ -539,7 +539,9 @@ static int aspeed_adc_probe(struct platform_device *pdev)
data->clk_scaler = devm_clk_hw_register_divider(
&pdev->dev, clk_name, clk_parent_name, scaler_flags,
data->base + ASPEED_REG_CLOCK_CONTROL, 0,
- data->model_data->scaler_bit_width, 0, &data->clk_lock);
+ data->model_data->scaler_bit_width,
+ data->model_data->need_prescaler ? CLK_DIVIDER_ONE_BASED : 0,
+ &data->clk_lock);
if (IS_ERR(data->clk_scaler))
return PTR_ERR(data->clk_scaler);

--
2.25.1


2022-02-21 09:49:19

by Billy Tsai

[permalink] [raw]
Subject: Re: [PATCH] iio: adc: aspeed: Add divider flag to fix incorrect voltage reading.

Hi Jonathan,

On 2022/2/19, 11:53 PM, "Jonathan Cameron" <[email protected]> wrote:

On Fri, 18 Feb 2022 16:57:08 +0800
Billy Tsai <[email protected]> wrote:

> > The formula for the ADC sampling period in ast2400/ast2500 is:
> > ADC clock period = PCLK * 2 * (ADC0C[31:17] + 1) * (ADC0C[9:0])
> > When ADC0C[9:0] is set to 0 the sampling voltage will be lower than
> > expected, because the hardware may not have enough time to
> > charge/discharge to a stable voltage.
> >
> > Reported-by: Konstantin Klubnichkin <[email protected]>
> > Signed-off-by: Billy Tsai <[email protected]>
> Hi Billy,

> Fixes tag?

> Also, would be good to call out in the patch description that
> CLK_DIVIDER_ONE_BASED rules at 0 as a valid value and hence
> avoids the ADC0C[9:0] value of 0 that is causing problems.

> That may be obvious to people who make frequent use of clk dividers
> but it's not locally obvious when looking at this patch.

> Otherwise looks good to me.

> Thanks,

> Jonathan

I will add fixes tag and add more description about the CLK_DIVIDER_ONE_BASED flag in v2.

Thanks,

Billy

> > ---
> > drivers/iio/adc/aspeed_adc.c | 4 +++-
> > 1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/iio/adc/aspeed_adc.c b/drivers/iio/adc/aspeed_adc.c
> > index a957cad1bfab..ffae64f39221 100644
> > --- a/drivers/iio/adc/aspeed_adc.c
> > +++ b/drivers/iio/adc/aspeed_adc.c
> > @@ -539,7 +539,9 @@ static int aspeed_adc_probe(struct platform_device *pdev)
> > data->clk_scaler = devm_clk_hw_register_divider(
> > &pdev->dev, clk_name, clk_parent_name, scaler_flags,
> > data->base + ASPEED_REG_CLOCK_CONTROL, 0,
> > - data->model_data->scaler_bit_width, 0, &data->clk_lock);
> > + data->model_data->scaler_bit_width,
> > + data->model_data->need_prescaler ? CLK_DIVIDER_ONE_BASED : 0,
> > + &data->clk_lock);
> > if (IS_ERR(data->clk_scaler))
> > return PTR_ERR(data->clk_scaler);
> >


2022-02-21 10:03:20

by Jonathan Cameron

[permalink] [raw]
Subject: Re: [PATCH] iio: adc: aspeed: Add divider flag to fix incorrect voltage reading.

On Fri, 18 Feb 2022 16:57:08 +0800
Billy Tsai <[email protected]> wrote:

> The formula for the ADC sampling period in ast2400/ast2500 is:
> ADC clock period = PCLK * 2 * (ADC0C[31:17] + 1) * (ADC0C[9:0])
> When ADC0C[9:0] is set to 0 the sampling voltage will be lower than
> expected, because the hardware may not have enough time to
> charge/discharge to a stable voltage.
>
> Reported-by: Konstantin Klubnichkin <[email protected]>
> Signed-off-by: Billy Tsai <[email protected]>
Hi Billy,

Fixes tag?

Also, would be good to call out in the patch description that
CLK_DIVIDER_ONE_BASED rules at 0 as a valid value and hence
avoids the ADC0C[9:0] value of 0 that is causing problems.

That may be obvious to people who make frequent use of clk dividers
but it's not locally obvious when looking at this patch.

Otherwise looks good to me.

Thanks,

Jonathan

> ---
> drivers/iio/adc/aspeed_adc.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/iio/adc/aspeed_adc.c b/drivers/iio/adc/aspeed_adc.c
> index a957cad1bfab..ffae64f39221 100644
> --- a/drivers/iio/adc/aspeed_adc.c
> +++ b/drivers/iio/adc/aspeed_adc.c
> @@ -539,7 +539,9 @@ static int aspeed_adc_probe(struct platform_device *pdev)
> data->clk_scaler = devm_clk_hw_register_divider(
> &pdev->dev, clk_name, clk_parent_name, scaler_flags,
> data->base + ASPEED_REG_CLOCK_CONTROL, 0,
> - data->model_data->scaler_bit_width, 0, &data->clk_lock);
> + data->model_data->scaler_bit_width,
> + data->model_data->need_prescaler ? CLK_DIVIDER_ONE_BASED : 0,
> + &data->clk_lock);
> if (IS_ERR(data->clk_scaler))
> return PTR_ERR(data->clk_scaler);
>