iio: adc: sc27xx: Add ADC data conversion timeout
diff mbox series

Message ID 208f9043fbd9b14b7a0c2743f68fa424388f5dcd.1541733478.git.baolin.wang@linaro.org
State New
Headers show
Series
  • iio: adc: sc27xx: Add ADC data conversion timeout
Related show

Commit Message

(Exiting) Baolin Wang Nov. 9, 2018, 3:25 a.m. UTC
From: Freeman Liu <freeman.liu@unisoc.com>

Sometimes the ADC controller met some problems, and it will not complete
the data conversion, that will can not wake up the read process any more
to block users. So we should add one maximum conversion time to avoid
this issue.

Signed-off-by: Freeman Liu <freeman.liu@unisoc.com>
Signed-off-by: Baolin Wang <baolin.wang@linaro.org>
---
 drivers/iio/adc/sc27xx_adc.c |   12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

Comments

Jonathan Cameron Nov. 11, 2018, 12:40 p.m. UTC | #1
On Fri,  9 Nov 2018 11:25:31 +0800
Baolin Wang <baolin.wang@linaro.org> wrote:

> From: Freeman Liu <freeman.liu@unisoc.com>
> 
> Sometimes the ADC controller met some problems, and it will not complete
> the data conversion, that will can not wake up the read process any more
> to block users. So we should add one maximum conversion time to avoid
> this issue.
> 
> Signed-off-by: Freeman Liu <freeman.liu@unisoc.com>
> Signed-off-by: Baolin Wang <baolin.wang@linaro.org>
Hi.

Patch looks good, but sounds like this is a fix so could I have a fixes
tag and a patch title that makes that clear?

I think we will want this one back ported to stable by the sound of it.

If it is just a theoretical issue then perhaps we don't need to bother.

Thanks,

Jonathan

> ---
>  drivers/iio/adc/sc27xx_adc.c |   12 +++++++++++-
>  1 file changed, 11 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/iio/adc/sc27xx_adc.c b/drivers/iio/adc/sc27xx_adc.c
> index 7940b23..f7f7a189 100644
> --- a/drivers/iio/adc/sc27xx_adc.c
> +++ b/drivers/iio/adc/sc27xx_adc.c
> @@ -52,6 +52,9 @@
>  /* Timeout (ms) for the trylock of hardware spinlocks */
>  #define SC27XX_ADC_HWLOCK_TIMEOUT	5000
>  
> +/* Timeout (ms) for ADC data conversion according to ADC datasheet */
> +#define SC27XX_ADC_RDY_TIMEOUT		100
> +
>  /* Maximum ADC channel number */
>  #define SC27XX_ADC_CHANNEL_MAX		32
>  
> @@ -223,7 +226,14 @@ static int sc27xx_adc_read(struct sc27xx_adc_data *data, int channel,
>  	if (ret)
>  		goto disable_adc;
>  
> -	wait_for_completion(&data->completion);
> +	ret = wait_for_completion_timeout(&data->completion,
> +				msecs_to_jiffies(SC27XX_ADC_RDY_TIMEOUT));
> +	if (!ret) {
> +		dev_err(data->dev, "read ADC data timeout\n");
> +		ret = -ETIMEDOUT;
> +	} else {
> +		ret = 0;
> +	}
>  
>  disable_adc:
>  	regmap_update_bits(data->regmap, data->base + SC27XX_ADC_CTL,
(Exiting) Baolin Wang Nov. 11, 2018, 3 p.m. UTC | #2
Hi Jonathan,

On 11 November 2018 at 20:40, Jonathan Cameron <jic23@kernel.org> wrote:
> On Fri,  9 Nov 2018 11:25:31 +0800
> Baolin Wang <baolin.wang@linaro.org> wrote:
>
>> From: Freeman Liu <freeman.liu@unisoc.com>
>>
>> Sometimes the ADC controller met some problems, and it will not complete
>> the data conversion, that will can not wake up the read process any more
>> to block users. So we should add one maximum conversion time to avoid
>> this issue.
>>
>> Signed-off-by: Freeman Liu <freeman.liu@unisoc.com>
>> Signed-off-by: Baolin Wang <baolin.wang@linaro.org>
> Hi.
>
> Patch looks good, but sounds like this is a fix so could I have a fixes
> tag and a patch title that makes that clear?
>
> I think we will want this one back ported to stable by the sound of it.
>
> If it is just a theoretical issue then perhaps we don't need to bother.

Thanks for your review. This patch is not one bug fix, just make the
code more robust in case of some bad cases. So I think we do not need
the stable tag.

>
> Thanks,
>
> Jonathan
>
>> ---
>>  drivers/iio/adc/sc27xx_adc.c |   12 +++++++++++-
>>  1 file changed, 11 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/iio/adc/sc27xx_adc.c b/drivers/iio/adc/sc27xx_adc.c
>> index 7940b23..f7f7a189 100644
>> --- a/drivers/iio/adc/sc27xx_adc.c
>> +++ b/drivers/iio/adc/sc27xx_adc.c
>> @@ -52,6 +52,9 @@
>>  /* Timeout (ms) for the trylock of hardware spinlocks */
>>  #define SC27XX_ADC_HWLOCK_TIMEOUT    5000
>>
>> +/* Timeout (ms) for ADC data conversion according to ADC datasheet */
>> +#define SC27XX_ADC_RDY_TIMEOUT               100
>> +
>>  /* Maximum ADC channel number */
>>  #define SC27XX_ADC_CHANNEL_MAX               32
>>
>> @@ -223,7 +226,14 @@ static int sc27xx_adc_read(struct sc27xx_adc_data *data, int channel,
>>       if (ret)
>>               goto disable_adc;
>>
>> -     wait_for_completion(&data->completion);
>> +     ret = wait_for_completion_timeout(&data->completion,
>> +                             msecs_to_jiffies(SC27XX_ADC_RDY_TIMEOUT));
>> +     if (!ret) {
>> +             dev_err(data->dev, "read ADC data timeout\n");
>> +             ret = -ETIMEDOUT;
>> +     } else {
>> +             ret = 0;
>> +     }
>>
>>  disable_adc:
>>       regmap_update_bits(data->regmap, data->base + SC27XX_ADC_CTL,
>
Jonathan Cameron Nov. 11, 2018, 4:03 p.m. UTC | #3
On Sun, 11 Nov 2018 23:00:18 +0800
Baolin Wang <baolin.wang@linaro.org> wrote:

> Hi Jonathan,
> 
> On 11 November 2018 at 20:40, Jonathan Cameron <jic23@kernel.org> wrote:
> > On Fri,  9 Nov 2018 11:25:31 +0800
> > Baolin Wang <baolin.wang@linaro.org> wrote:
> >  
> >> From: Freeman Liu <freeman.liu@unisoc.com>
> >>
> >> Sometimes the ADC controller met some problems, and it will not complete
> >> the data conversion, that will can not wake up the read process any more
> >> to block users. So we should add one maximum conversion time to avoid
> >> this issue.
> >>
> >> Signed-off-by: Freeman Liu <freeman.liu@unisoc.com>
> >> Signed-off-by: Baolin Wang <baolin.wang@linaro.org>  
> > Hi.
> >
> > Patch looks good, but sounds like this is a fix so could I have a fixes
> > tag and a patch title that makes that clear?
> >
> > I think we will want this one back ported to stable by the sound of it.
> >
> > If it is just a theoretical issue then perhaps we don't need to bother.  
> 
> Thanks for your review. This patch is not one bug fix, just make the
> code more robust in case of some bad cases. So I think we do not need
> the stable tag.
Thanks.

Applied to the togreg branch of iio.git and pushed out as testing for
the autobuilders to play with it.

Jonathan

> 
> >
> > Thanks,
> >
> > Jonathan
> >  
> >> ---
> >>  drivers/iio/adc/sc27xx_adc.c |   12 +++++++++++-
> >>  1 file changed, 11 insertions(+), 1 deletion(-)
> >>
> >> diff --git a/drivers/iio/adc/sc27xx_adc.c b/drivers/iio/adc/sc27xx_adc.c
> >> index 7940b23..f7f7a189 100644
> >> --- a/drivers/iio/adc/sc27xx_adc.c
> >> +++ b/drivers/iio/adc/sc27xx_adc.c
> >> @@ -52,6 +52,9 @@
> >>  /* Timeout (ms) for the trylock of hardware spinlocks */
> >>  #define SC27XX_ADC_HWLOCK_TIMEOUT    5000
> >>
> >> +/* Timeout (ms) for ADC data conversion according to ADC datasheet */
> >> +#define SC27XX_ADC_RDY_TIMEOUT               100
> >> +
> >>  /* Maximum ADC channel number */
> >>  #define SC27XX_ADC_CHANNEL_MAX               32
> >>
> >> @@ -223,7 +226,14 @@ static int sc27xx_adc_read(struct sc27xx_adc_data *data, int channel,
> >>       if (ret)
> >>               goto disable_adc;
> >>
> >> -     wait_for_completion(&data->completion);
> >> +     ret = wait_for_completion_timeout(&data->completion,
> >> +                             msecs_to_jiffies(SC27XX_ADC_RDY_TIMEOUT));
> >> +     if (!ret) {
> >> +             dev_err(data->dev, "read ADC data timeout\n");
> >> +             ret = -ETIMEDOUT;
> >> +     } else {
> >> +             ret = 0;
> >> +     }
> >>
> >>  disable_adc:
> >>       regmap_update_bits(data->regmap, data->base + SC27XX_ADC_CTL,  
> >  
> 
> 
>

Patch
diff mbox series

diff --git a/drivers/iio/adc/sc27xx_adc.c b/drivers/iio/adc/sc27xx_adc.c
index 7940b23..f7f7a189 100644
--- a/drivers/iio/adc/sc27xx_adc.c
+++ b/drivers/iio/adc/sc27xx_adc.c
@@ -52,6 +52,9 @@ 
 /* Timeout (ms) for the trylock of hardware spinlocks */
 #define SC27XX_ADC_HWLOCK_TIMEOUT	5000
 
+/* Timeout (ms) for ADC data conversion according to ADC datasheet */
+#define SC27XX_ADC_RDY_TIMEOUT		100
+
 /* Maximum ADC channel number */
 #define SC27XX_ADC_CHANNEL_MAX		32
 
@@ -223,7 +226,14 @@  static int sc27xx_adc_read(struct sc27xx_adc_data *data, int channel,
 	if (ret)
 		goto disable_adc;
 
-	wait_for_completion(&data->completion);
+	ret = wait_for_completion_timeout(&data->completion,
+				msecs_to_jiffies(SC27XX_ADC_RDY_TIMEOUT));
+	if (!ret) {
+		dev_err(data->dev, "read ADC data timeout\n");
+		ret = -ETIMEDOUT;
+	} else {
+		ret = 0;
+	}
 
 disable_adc:
 	regmap_update_bits(data->regmap, data->base + SC27XX_ADC_CTL,