Hi,

This is an attempt to update sca3000 accelerometer driver to make use
of devm_ based helpers where needed. I have split it into 5 patches.

Patch 1 - some refactoring and simplification.
Patch 2 - switches request_threaded_irq() over to the devm helper
Patch 3 - Used devm_add_action_or_reset() for disabling interrupts.
(Ensured the ordering of teardown bits remain same)
Patch 4 - manage device registration with devm helper
Patch 5 - Make use of guard() in sca3000_stop_all_interrupts() function.

Thanks a lot to Jonathan for detailed review of V1.

During v1 review comments, Jonathan Cameron kindly provided some great
suggestions to improve this driver. One of them is making use of
gaurd(mutex) - autocleanup style, patch 5 does that for one function.

Yet to be addressed tasks:
1. We shouldn't be using the spi_device_id at all.
2. Modernize other fucntions to make use of autocleanup style locking
which simpifies the code and makes error paths cleaner.

I will be working on these two above tasks and will be sending a
different patches for those.

The series builds cleanly and I have performed static analysis with
smatch checker but haven't tested on actual hardware.

v1->v2 changes are documented in patches where necessary.

Thanks for your time.

Regards,
Harshit

Harshit Mogalapalli (5):
  iio: sca3000: cache SPI device ID in probe
  iio: sca3000: switch IRQ handling to devm helpers
  iio: sca3000: stop interrupts via devm_add_action_or_reset()
  iio: sca3000: manage device registration with devm helper
  iio: sca3000: use guard(mutex) to simplify return paths

 drivers/iio/accel/sca3000.c | 99 +++++++++++++++++--------------------
 1 file changed, 44 insertions(+), 55 deletions(-)

-- 
2.47.3

On Mon, Feb 02, 2026 at 11:40:07AM -0800, Harshit Mogalapalli wrote:
> Hi,
> 
> This is an attempt to update sca3000 accelerometer driver to make use
> of devm_ based helpers where needed. I have split it into 5 patches.
> 
> Patch 1 - some refactoring and simplification.
> Patch 2 - switches request_threaded_irq() over to the devm helper
> Patch 3 - Used devm_add_action_or_reset() for disabling interrupts.
> (Ensured the ordering of teardown bits remain same)
> Patch 4 - manage device registration with devm helper
> Patch 5 - Make use of guard() in sca3000_stop_all_interrupts() function.
> 
> Thanks a lot to Jonathan for detailed review of V1.
> 
> During v1 review comments, Jonathan Cameron kindly provided some great
> suggestions to improve this driver. One of them is making use of
> gaurd(mutex) - autocleanup style, patch 5 does that for one function.
> 
> Yet to be addressed tasks:
> 1. We shouldn't be using the spi_device_id at all.
> 2. Modernize other fucntions to make use of autocleanup style locking
> which simpifies the code and makes error paths cleaner.
> 
> I will be working on these two above tasks and will be sending a
> different patches for those.
> 
> The series builds cleanly and I have performed static analysis with
> smatch checker but haven't tested on actual hardware.

David already reviewed and v3 is anticipated, so I try to look into that one.

-- 
With Best Regards,
Andy Shevchenko

Hi Andy,

On 03/02/26 15:18, Andy Shevchenko wrote:
> On Mon, Feb 02, 2026 at 11:40:07AM -0800, Harshit Mogalapalli wrote:
>> Hi,
>>
>> This is an attempt to update sca3000 accelerometer driver to make use
>> of devm_ based helpers where needed. I have split it into 5 patches.
>>
>> Patch 1 - some refactoring and simplification.
>> Patch 2 - switches request_threaded_irq() over to the devm helper
>> Patch 3 - Used devm_add_action_or_reset() for disabling interrupts.
>> (Ensured the ordering of teardown bits remain same)
>> Patch 4 - manage device registration with devm helper
>> Patch 5 - Make use of guard() in sca3000_stop_all_interrupts() function.
>>
>> Thanks a lot to Jonathan for detailed review of V1.
>>
>> During v1 review comments, Jonathan Cameron kindly provided some great
>> suggestions to improve this driver. One of them is making use of
>> gaurd(mutex) - autocleanup style, patch 5 does that for one function.
>>
>> Yet to be addressed tasks:
>> 1. We shouldn't be using the spi_device_id at all.
>> 2. Modernize other fucntions to make use of autocleanup style locking
>> which simpifies the code and makes error paths cleaner.
>>
>> I will be working on these two above tasks and will be sending a
>> different patches for those.
>>
>> The series builds cleanly and I have performed static analysis with
>> smatch checker but haven't tested on actual hardware.
> 
> David already reviewed and v3 is anticipated, so I try to look into that one.
> 

Sure, I am working on v3.

Thank you very much.


Regards,
Harshit