Convert the threaded IRQ registration to devm_request_threaded_irq() so
that the probe and remove paths can drop manual freeing of irqs.

No functionality change.

Suggested-by: Andy Shevchenko <andriy.shevchenko@intel.com>
Reviewed-by: David Lechner <dlechner@baylibre.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com>
---
 drivers/iio/accel/sca3000.c | 25 ++++++++-----------------
 1 file changed, 8 insertions(+), 17 deletions(-)

diff --git a/drivers/iio/accel/sca3000.c b/drivers/iio/accel/sca3000.c
index 43373c798714..e270f445bb35 100644
--- a/drivers/iio/accel/sca3000.c
+++ b/drivers/iio/accel/sca3000.c
@@ -1470,34 +1470,27 @@ static int sca3000_probe(struct spi_device *spi)
 		return ret;
 
 	if (spi->irq) {
-		ret = request_threaded_irq(spi->irq,
-					   NULL,
-					   &sca3000_event_handler,
-					   IRQF_TRIGGER_FALLING | IRQF_ONESHOT,
-					   "sca3000",
-					   indio_dev);
+		ret = devm_request_threaded_irq(dev, spi->irq, NULL,
+						&sca3000_event_handler,
+						IRQF_TRIGGER_FALLING | IRQF_ONESHOT,
+						"sca3000",
+						indio_dev);
 		if (ret)
 			return ret;
 	}
 	ret = sca3000_clean_setup(st);
 	if (ret)
-		goto error_free_irq;
+		return ret;
 
 	ret = sca3000_print_rev(indio_dev);
 	if (ret)
-		goto error_free_irq;
+		return ret;
 
 	ret = iio_device_register(indio_dev);
 	if (ret)
-		goto error_free_irq;
+		return ret;
 
 	return 0;
-
-error_free_irq:
-	if (spi->irq)
-		free_irq(spi->irq, indio_dev);
-
-	return ret;
 }
 
 static int sca3000_stop_all_interrupts(struct sca3000_state *st)
@@ -1527,8 +1520,6 @@ static void sca3000_remove(struct spi_device *spi)
 
 	/* Must ensure no interrupts can be generated after this! */
 	sca3000_stop_all_interrupts(st);
-	if (spi->irq)
-		free_irq(spi->irq, indio_dev);
 }
 
 static const struct spi_device_id sca3000_id[] = {
-- 
2.47.3

On Thu, Feb 05, 2026 at 05:12:08AM -0800, Harshit Mogalapalli wrote:
> Convert the threaded IRQ registration to devm_request_threaded_irq() so
> that the probe and remove paths can drop manual freeing of irqs.
> 
> No functionality change.

...

> -error_free_irq:
> -	if (spi->irq)
> -		free_irq(spi->irq, indio_dev);
> -
> -	return ret;

...

> -	if (spi->irq)
> -		free_irq(spi->irq, indio_dev);

Do we need an irq member to be in the struct after this patch?

-- 
With Best Regards,
Andy Shevchenko

Hi Andy,

On 05/02/26 21:53, Andy Shevchenko wrote:
> On Thu, Feb 05, 2026 at 05:12:08AM -0800, Harshit Mogalapalli wrote:
>> Convert the threaded IRQ registration to devm_request_threaded_irq() so
>> that the probe and remove paths can drop manual freeing of irqs.
>>
>> No functionality change.
> 
> ...
> 

Thanks for the reviews!

>> -error_free_irq:
>> -	if (spi->irq)
>> -		free_irq(spi->irq, indio_dev);
>> -
>> -	return ret;
> 
> ...
> 
>> -	if (spi->irq)
>> -		free_irq(spi->irq, indio_dev);
> 
> Do we need an irq member to be in the struct after this patch?

I probably didn't understand that question fully.

we still have a call to ret = devm_request_threaded_irq(dev, spi->irq, 
NULL,...) so we can't relaly get rid of the irq member I think, did I 
understand your question right ?

Thanks,
Harshit
> 

On Thu, Feb 05, 2026 at 10:09:45PM +0530, Harshit Mogalapalli wrote:
> On 05/02/26 21:53, Andy Shevchenko wrote:
> > On Thu, Feb 05, 2026 at 05:12:08AM -0800, Harshit Mogalapalli wrote:

...

> > > -	if (spi->irq)
> > > -		free_irq(spi->irq, indio_dev);
> > 
> > Do we need an irq member to be in the struct after this patch?
> 
> I probably didn't understand that question fully.
> 
> we still have a call to ret = devm_request_threaded_irq(dev, spi->irq,
> NULL,...) so we can't relaly get rid of the irq member I think,
> did I understand your question right ?

Yes. But now I realised that this is the external structure, and not the one
the driver defines. Sorry for the noise.

-- 
With Best Regards,
Andy Shevchenko

On 05/02/26 22:15, Andy Shevchenko wrote:
> On Thu, Feb 05, 2026 at 10:09:45PM +0530, Harshit Mogalapalli wrote:
>> On 05/02/26 21:53, Andy Shevchenko wrote:
>>> On Thu, Feb 05, 2026 at 05:12:08AM -0800, Harshit Mogalapalli wrote:
> 
> ...
> 
>>>> -	if (spi->irq)
>>>> -		free_irq(spi->irq, indio_dev);
>>>
>>> Do we need an irq member to be in the struct after this patch?
>>
>> I probably didn't understand that question fully.
>>
>> we still have a call to ret = devm_request_threaded_irq(dev, spi->irq,
>> NULL,...) so we can't relaly get rid of the irq member I think,
>> did I understand your question right ?
> 
> Yes. But now I realised that this is the external structure, and not the one
> the driver defines. Sorry for the noise.
> 
Np, yes this is part of spi_device a more central structure. Thanks for 
checking!


Regards,
Harshit