Re: [PATCH v2 3/3] bus: stm32_firewall: Use scoped allocation to simplify cleanup

[Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>]

On 09/01/2026 13:34, Gatien CHEVALLIER wrote:
> 
> 
> On 1/9/26 12:55, Krzysztof Kozlowski wrote:
>> On 09/01/2026 11:37, Gatien CHEVALLIER wrote:
>>> On 1/5/26 15:37, Krzysztof Kozlowski wrote:
>>>> Allocate the memory with scoped/cleanup.h to reduce error handling and
>>>> make the code a bit simpler.
>>>>
>>>> Suggested-by: Jonathan Cameron <jonathan.cameron@huawei.com>
>>>> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
>>>>
>>>> ---
>>>>
>>>> Changes in v2:
>>>> 1. New patch
>>>> ---
>>>>    drivers/bus/stm32_firewall.c | 10 +++-------
>>>>    1 file changed, 3 insertions(+), 7 deletions(-)
>>>>
>>>> diff --git a/drivers/bus/stm32_firewall.c b/drivers/bus/stm32_firewall.c
>>>> index fae881cea9a0..92414a4c7bb1 100644
>>>> --- a/drivers/bus/stm32_firewall.c
>>>> +++ b/drivers/bus/stm32_firewall.c
>>>> @@ -240,7 +240,6 @@ EXPORT_SYMBOL_GPL(stm32_firewall_controller_unregister);
>>>>    
>>>>    int stm32_firewall_populate_bus(struct stm32_firewall_controller *firewall_controller)
>>>>    {
>>>> -	struct stm32_firewall *firewalls;
>>>>    	struct device *parent;
>>>>    	unsigned int i;
>>>>    	int len;
>>>> @@ -257,15 +256,14 @@ int stm32_firewall_populate_bus(struct stm32_firewall_controller *firewall_contr
>>>>    		if (len <= 0)
>>>>    			return -EINVAL;
>>>>    
>>>> -		firewalls = kcalloc(len, sizeof(*firewalls), GFP_KERNEL);
>>>> +		struct stm32_firewall *firewalls __free(kfree) =
>>>> +			kcalloc(len, sizeof(*firewalls), GFP_KERNEL);
>>>
>>> Hello,
>>>
>>> I'd prefer to keep declarations separated.
>>
>> Separated how? This is the coding style of cleanup (which is quite
>> explicit around other styles)? Or you meant one line without wrapping?
>>
> 
> I found odd to declare and assign a variable mid-loop but I found
> multiple occurrences of the same kind of cleanup.
> 
> You can drop my comment, I'm fine with how it is. Thanks for the
> cleanup.
> 

Sure, just to let you know - it is fair to reject cleanup.h entirely for
code you maintain (like netdev did). I only disagree with sort of
semi-cleanup.h without following its declaration and initialization
rules, because it is more error-prone, leads to bugs and often does not
make the code easier to read/simpler.

Best regards,
Krzysztof