Re: [PATCH] extcon: maxim: Fix missing IRQF_ONESHOT as only threaded handler

[Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>]

On 18/04/2021 16:41, Guangqing Zhu wrote:
> 
> 
> On 16/04/2021 16:43, Krzysztof Kozlowski wrote:
>> On 15/04/2021 13:36, zhuguangqing83@gmail.com wrote:
>>> From: Guangqing Zhu <zhuguangqing83@gmail.com>
>>>
>>> Coccinelle noticed:
>>>    1. drivers/extcon/extcon-max14577.c:699:8-33: ERROR: Threaded IRQ with
>>> no primary handler requested without IRQF_ONESHOT
>>>    2. drivers/extcon/extcon-max77693.c:1143:8-33: ERROR: Threaded IRQ with
>>> no primary handler requested without IRQF_ONESHOT
>>>    3. drivers/extcon/extcon-max77843.c:907:8-33: ERROR: Threaded IRQ with
>>> no primary handler requested without IRQF_ONESHOT
>>>    4. drivers/extcon/extcon-max8997.c:665:8-28: ERROR: Threaded IRQ with
>>> no primary handler requested without IRQF_ONESHOT
>>>
>>> Signed-off-by: Guangqing Zhu <zhuguangqing83@gmail.com>
>>> ---
>>>   drivers/extcon/extcon-max14577.c | 2 +-
>>>   drivers/extcon/extcon-max77693.c | 2 +-
>>>   drivers/extcon/extcon-max77843.c | 3 ++-
>>>   drivers/extcon/extcon-max8997.c  | 2 +-
>>>   4 files changed, 5 insertions(+), 4 deletions(-)
>>>
>>> diff --git a/drivers/extcon/extcon-max14577.c b/drivers/extcon/extcon-max14577.c
>>> index ace523924e58..af15a9e00ee9 100644
>>> --- a/drivers/extcon/extcon-max14577.c
>>> +++ b/drivers/extcon/extcon-max14577.c
>>> @@ -698,7 +698,7 @@ static int max14577_muic_probe(struct platform_device *pdev)
>>>   
>>>   		ret = devm_request_threaded_irq(&pdev->dev, virq, NULL,
>>>   				max14577_muic_irq_handler,
>>> -				IRQF_NO_SUSPEND,
>>> +				IRQF_NO_SUSPEND | IRQF_ONESHOT,
>>
>> The same with all other patches for IRQF_ONESHOT which are send recently:
>> 1. On what board did you test it?
> 
> I didn't test it.
> 
>> 2. Is this just blind patch from Coccinelle without investigation
>> whether it is needed (hint: it's not needed here, it does not use
>> default primary handler).
> 
> I found the error notice from Coccinelle and I saw the code. Maybe
> I'm mistaken, I think it's needed here. Because handler == NULL and
> thread_fn != NULL, it use irq_default_primary_handler() in
> request_threaded_irq().

No, the primary handler is nested, not default one. Otherwise it would
have absolutely never worked. Therefore you are not fixing anything,
except Coccinelle report.


Best regards,
Krzysztof