Re: [PATCH v5 1/2] iio:imu: inv_mpu6050: support more interrupt types

[Martin Kelly <mkelly@xevo.com>]

On 04/17/2018 07:10 AM, Jean-Baptiste Maneyrol wrote:
> 
> 
> On 15/04/2018 21:05, Martin Kelly wrote:
>> On 04/15/2018 10:43 AM, Jonathan Cameron wrote:
>>> On Wed, 11 Apr 2018 09:42:14 -0700
>>> Martin Kelly <mkelly@xevo.com> wrote:
>>>
>>>> On 04/11/2018 12:01 AM, Jean-Baptiste Maneyrol wrote:
>>>>> This is OK for me.
>>>>>
>>>>> Jonathan will tell us about EBUSY error code for sure if it is not 
>>>>> correct.
>>>>>
>>>>> JB
>>>>
>>>> Sounds good; once we hear from Jonathan, I will submit the next 
>>>> revision.
>>> Optimists.  I can never make my mind up on some of the error codes.
>>>
>>> It's not totally silly so I'm happy with EBUSY or ENODEV as you wish.
>>>
>>> Jonathan
>>>
>>
>> OK, then I will defer to Jean-Baptiste on this. Shall we go with ENODEV?
> 
> After looking into the irq kernel code, I think perhaps the best value 
> should be EINVAL. What it really means, is that configuration is missing 
> in the dts file where it shouldn't. Incorrect value seams more 
> meaningful in this case.
> 
> Do you agree?
> 
> JB

Yes, EINVAL is fine with me. I didn't use it because there was a note in 
the aforementioned document saying that it's best to use a more specific 
error code. However, none of these error codes have a clear case for 
their use, and it's all a bit muddled, as Jonathan mentioned. I'll send 
a revision with EINVAL later this week when I'm back to my desk with 
hardware to test.