[v2] iio: adc: ti_am335x_adc: Fix return value check of tiadc_request_dma()

Message ID 20230925134427.214556-1-w.egorov@phytec.de
State New
Headers
Series [v2] iio: adc: ti_am335x_adc: Fix return value check of tiadc_request_dma() |

Commit Message

Wadim Egorov Sept. 25, 2023, 1:44 p.m. UTC
  Fix wrong handling of a DMA request where the probing only failed
if -EPROPE_DEFER was returned. Instead, let us fail if a non -ENODEV
value is returned. This makes DMAs explicitly optional. Even if the
DMA request is unsuccessfully, the ADC can still work properly.
We do also handle the defer probe case by making use of dev_err_probe().

Fixes: f438b9da75eb ("drivers: iio: ti_am335x_adc: add dma support")
Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
---
v2:
  - Update description
  - Drop line break after Fixes tag
  - Move decision about optional DMA into probe/caller
---
 drivers/iio/adc/ti_am335x_adc.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
  

Comments

Jonathan Cameron Oct. 5, 2023, 2:09 p.m. UTC | #1
On Mon, 25 Sep 2023 15:44:27 +0200
Wadim Egorov <w.egorov@phytec.de> wrote:

> Fix wrong handling of a DMA request where the probing only failed
> if -EPROPE_DEFER was returned. Instead, let us fail if a non -ENODEV
> value is returned. This makes DMAs explicitly optional. Even if the
> DMA request is unsuccessfully, the ADC can still work properly.
> We do also handle the defer probe case by making use of dev_err_probe().
> 
> Fixes: f438b9da75eb ("drivers: iio: ti_am335x_adc: add dma support")
> Signed-off-by: Wadim Egorov <w.egorov@phytec.de>

+CC Bhavya,

Could you take a look at this given you had comments on v1.

Thanks,

Jonathan

> ---
> v2:
>   - Update description
>   - Drop line break after Fixes tag
>   - Move decision about optional DMA into probe/caller
> ---
>  drivers/iio/adc/ti_am335x_adc.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/iio/adc/ti_am335x_adc.c b/drivers/iio/adc/ti_am335x_adc.c
> index 8db7a01cb5fb..5f8795986995 100644
> --- a/drivers/iio/adc/ti_am335x_adc.c
> +++ b/drivers/iio/adc/ti_am335x_adc.c
> @@ -670,8 +670,10 @@ static int tiadc_probe(struct platform_device *pdev)
>  	platform_set_drvdata(pdev, indio_dev);
>  
>  	err = tiadc_request_dma(pdev, adc_dev);
> -	if (err && err == -EPROBE_DEFER)
> +	if (err && err != -ENODEV) {
> +		dev_err_probe(&pdev->dev, err, "DMA request failed\n");
>  		goto err_dma;
> +	}
>  
>  	return 0;
>
  
Wadim Egorov Nov. 24, 2023, 12:11 p.m. UTC | #2
Am 05.10.23 um 16:09 schrieb Jonathan Cameron:
> On Mon, 25 Sep 2023 15:44:27 +0200
> Wadim Egorov <w.egorov@phytec.de> wrote:
>
>> Fix wrong handling of a DMA request where the probing only failed
>> if -EPROPE_DEFER was returned. Instead, let us fail if a non -ENODEV
>> value is returned. This makes DMAs explicitly optional. Even if the
>> DMA request is unsuccessfully, the ADC can still work properly.
>> We do also handle the defer probe case by making use of dev_err_probe().
>>
>> Fixes: f438b9da75eb ("drivers: iio: ti_am335x_adc: add dma support")
>> Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
> +CC Bhavya,
>
> Could you take a look at this given you had comments on v1.

Bhavya, any comments on this?
If not, is there anything else that is blocking this patch?

Regards,
Wadim

>
> Thanks,
>
> Jonathan
>
>> ---
>> v2:
>>    - Update description
>>    - Drop line break after Fixes tag
>>    - Move decision about optional DMA into probe/caller
>> ---
>>   drivers/iio/adc/ti_am335x_adc.c | 4 +++-
>>   1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/iio/adc/ti_am335x_adc.c b/drivers/iio/adc/ti_am335x_adc.c
>> index 8db7a01cb5fb..5f8795986995 100644
>> --- a/drivers/iio/adc/ti_am335x_adc.c
>> +++ b/drivers/iio/adc/ti_am335x_adc.c
>> @@ -670,8 +670,10 @@ static int tiadc_probe(struct platform_device *pdev)
>>   	platform_set_drvdata(pdev, indio_dev);
>>   
>>   	err = tiadc_request_dma(pdev, adc_dev);
>> -	if (err && err == -EPROBE_DEFER)
>> +	if (err && err != -ENODEV) {
>> +		dev_err_probe(&pdev->dev, err, "DMA request failed\n");
>>   		goto err_dma;
>> +	}
>>   
>>   	return 0;
>>
  
Bhavya Kapoor Nov. 27, 2023, 9:40 a.m. UTC | #3
On 05/10/23 7:39 pm, Jonathan Cameron wrote:
> On Mon, 25 Sep 2023 15:44:27 +0200
> Wadim Egorov <w.egorov@phytec.de> wrote:
>
>> Fix wrong handling of a DMA request where the probing only failed
>> if -EPROPE_DEFER was returned. Instead, let us fail if a non -ENODEV
>> value is returned. This makes DMAs explicitly optional. Even if the
>> DMA request is unsuccessfully, the ADC can still work properly.
>> We do also handle the defer probe case by making use of dev_err_probe().
>>
>> Fixes: f438b9da75eb ("drivers: iio: ti_am335x_adc: add dma support")
>> Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
> +CC Bhavya,
>
> Could you take a look at this given you had comments on v1.
>
> Thanks,
>
> Jonathan

Hi Jonathan, Patch Looks Good To Me and Should now work fine for every case.

Regards

~B-Kapoor

>
>> ---
>> v2:
>>    - Update description
>>    - Drop line break after Fixes tag
>>    - Move decision about optional DMA into probe/caller
>> ---
>>   drivers/iio/adc/ti_am335x_adc.c | 4 +++-
>>   1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/iio/adc/ti_am335x_adc.c b/drivers/iio/adc/ti_am335x_adc.c
>> index 8db7a01cb5fb..5f8795986995 100644
>> --- a/drivers/iio/adc/ti_am335x_adc.c
>> +++ b/drivers/iio/adc/ti_am335x_adc.c
>> @@ -670,8 +670,10 @@ static int tiadc_probe(struct platform_device *pdev)
>>   	platform_set_drvdata(pdev, indio_dev);
>>   
>>   	err = tiadc_request_dma(pdev, adc_dev);
>> -	if (err && err == -EPROBE_DEFER)
>> +	if (err && err != -ENODEV) {
>> +		dev_err_probe(&pdev->dev, err, "DMA request failed\n");
>>   		goto err_dma;
>> +	}
>>   
>>   	return 0;
>>
  
Bhavya Kapoor Nov. 27, 2023, 9:44 a.m. UTC | #4
LGTM !!🙂

On 24/11/23 5:41 pm, Wadim Egorov wrote:
>
> Am 05.10.23 um 16:09 schrieb Jonathan Cameron:
>> On Mon, 25 Sep 2023 15:44:27 +0200
>> Wadim Egorov <w.egorov@phytec.de> wrote:
>>
>>> Fix wrong handling of a DMA request where the probing only failed
>>> if -EPROPE_DEFER was returned. Instead, let us fail if a non -ENODEV
>>> value is returned. This makes DMAs explicitly optional. Even if the
>>> DMA request is unsuccessfully, the ADC can still work properly.
>>> We do also handle the defer probe case by making use of 
>>> dev_err_probe().
>>>
>>> Fixes: f438b9da75eb ("drivers: iio: ti_am335x_adc: add dma support")
>>> Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
Reviewed-by: Bhavya Kapoor <b-kapoor@ti.com>
>> +CC Bhavya,
>>
>> Could you take a look at this given you had comments on v1.
>
> Bhavya, any comments on this?
> If not, is there anything else that is blocking this patch?
>
> Regards,
> Wadim
>
>>
>> Thanks,
>>
>> Jonathan
>>
>>> ---
>>> v2:
>>>    - Update description
>>>    - Drop line break after Fixes tag
>>>    - Move decision about optional DMA into probe/caller
>>> ---
>>>   drivers/iio/adc/ti_am335x_adc.c | 4 +++-
>>>   1 file changed, 3 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/iio/adc/ti_am335x_adc.c 
>>> b/drivers/iio/adc/ti_am335x_adc.c
>>> index 8db7a01cb5fb..5f8795986995 100644
>>> --- a/drivers/iio/adc/ti_am335x_adc.c
>>> +++ b/drivers/iio/adc/ti_am335x_adc.c
>>> @@ -670,8 +670,10 @@ static int tiadc_probe(struct platform_device 
>>> *pdev)
>>>       platform_set_drvdata(pdev, indio_dev);
>>>         err = tiadc_request_dma(pdev, adc_dev);
>>> -    if (err && err == -EPROBE_DEFER)
>>> +    if (err && err != -ENODEV) {
>>> +        dev_err_probe(&pdev->dev, err, "DMA request failed\n");
>>>           goto err_dma;
>>> +    }
>>>         return 0;
  
Jonathan Cameron Dec. 4, 2023, 9:33 a.m. UTC | #5
On Mon, 27 Nov 2023 15:14:29 +0530
Bhavya Kapoor <b-kapoor@ti.com> wrote:

> LGTM !!🙂
> 
> On 24/11/23 5:41 pm, Wadim Egorov wrote:
> >
> > Am 05.10.23 um 16:09 schrieb Jonathan Cameron:  
> >> On Mon, 25 Sep 2023 15:44:27 +0200
> >> Wadim Egorov <w.egorov@phytec.de> wrote:
> >>  
> >>> Fix wrong handling of a DMA request where the probing only failed
> >>> if -EPROPE_DEFER was returned. Instead, let us fail if a non -ENODEV
> >>> value is returned. This makes DMAs explicitly optional. Even if the
> >>> DMA request is unsuccessfully, the ADC can still work properly.
> >>> We do also handle the defer probe case by making use of 
> >>> dev_err_probe().
> >>>
> >>> Fixes: f438b9da75eb ("drivers: iio: ti_am335x_adc: add dma support")
> >>> Signed-off-by: Wadim Egorov <w.egorov@phytec.de>  
> Reviewed-by: Bhavya Kapoor <b-kapoor@ti.com>

Applied to the fixes-togreg branch of iio.git and marked for stable.

> >> +CC Bhavya,
> >>
> >> Could you take a look at this given you had comments on v1.  
> >
> > Bhavya, any comments on this?
> > If not, is there anything else that is blocking this patch?
> >
> > Regards,
> > Wadim
> >  
> >>
> >> Thanks,
> >>
> >> Jonathan
> >>  
> >>> ---
> >>> v2:
> >>>    - Update description
> >>>    - Drop line break after Fixes tag
> >>>    - Move decision about optional DMA into probe/caller
> >>> ---
> >>>   drivers/iio/adc/ti_am335x_adc.c | 4 +++-
> >>>   1 file changed, 3 insertions(+), 1 deletion(-)
> >>>
> >>> diff --git a/drivers/iio/adc/ti_am335x_adc.c 
> >>> b/drivers/iio/adc/ti_am335x_adc.c
> >>> index 8db7a01cb5fb..5f8795986995 100644
> >>> --- a/drivers/iio/adc/ti_am335x_adc.c
> >>> +++ b/drivers/iio/adc/ti_am335x_adc.c
> >>> @@ -670,8 +670,10 @@ static int tiadc_probe(struct platform_device 
> >>> *pdev)
> >>>       platform_set_drvdata(pdev, indio_dev);
> >>>         err = tiadc_request_dma(pdev, adc_dev);
> >>> -    if (err && err == -EPROBE_DEFER)
> >>> +    if (err && err != -ENODEV) {
> >>> +        dev_err_probe(&pdev->dev, err, "DMA request failed\n");
> >>>           goto err_dma;
> >>> +    }
> >>>         return 0;
  

Patch

diff --git a/drivers/iio/adc/ti_am335x_adc.c b/drivers/iio/adc/ti_am335x_adc.c
index 8db7a01cb5fb..5f8795986995 100644
--- a/drivers/iio/adc/ti_am335x_adc.c
+++ b/drivers/iio/adc/ti_am335x_adc.c
@@ -670,8 +670,10 @@  static int tiadc_probe(struct platform_device *pdev)
 	platform_set_drvdata(pdev, indio_dev);
 
 	err = tiadc_request_dma(pdev, adc_dev);
-	if (err && err == -EPROBE_DEFER)
+	if (err && err != -ENODEV) {
+		dev_err_probe(&pdev->dev, err, "DMA request failed\n");
 		goto err_dma;
+	}
 
 	return 0;