Re: [PATCH 1/2] dt-bindings: iio: adc: Add TI ADS1018/ADS1118

["Kurt Borja" <kuurtb@gmail.com>]

On Fri Nov 21, 2025 at 5:40 PM -05, David Lechner wrote:
> On 11/21/25 2:56 PM, Kurt Borja wrote:
>> Hi Krzysztof,
>> 
>> On Fri Nov 21, 2025 at 2:10 PM -05, Krzysztof Kozlowski wrote:
>>> On 21/11/2025 18:16, Kurt Borja wrote:
>>>> Add documentation for Texas Instruments ADS1018 and ADS1118
>>>> analog-to-digital converters.
>>>>
>>>> Signed-off-by: Kurt Borja <kuurtb@gmail.com>
>>>
>>> You did not test it before sending, so no full review but few nits to
>>> save you one round of reviews:
>> 
>> My bad! I will fix the errors. Thanks!
>> 
>
> ...
>
>>>> +  interrupts:
>>>> +    description: DOUT/DRDY (Data Out/Data Ready) line.
>>>> +    maxitems: 1
>>>> +
>>>> +  drdy-gpios:
>>>> +    description:
>>>> +      Extra GPIO line connected to DOUT/DRDY (Data Out/Data Ready). This allows
>>>> +      distinguishing between latched and real DRDY IRQs.
>>>
>>> I have feeling that you miss proper handling of IRQs (e.g. active level)
>>> on your board.
>> 
>> Can you elaborate? Should I specify active level here?
>> 
>>>
> The problem is not about the levels. It is rather that the behavior of the
> interrupt when disabled/masked is different on different interrupt controllers.
>
> On some controllers, if an event happens while disabled/masked, it "remembers"
> that and will trigger the interrupt as soon as it is enabled even if the
> condition doesn't exist anymore. Not a great hardware design IMHO, but there
> is real hardware that does this.

I can attest to that. My hardware (RPI 5) does this haha.

>
> I think a better wording would be to leave out "latched" and say that
> it is needed to be able to distinguish between an interrupt triggered
> by the DRDY signal vs. an interrupt triggered by SPI data.

I'll take this wording, thanks.


-- 
 ~ Kurt