Re: [PATCH 2/2] iio: xilinx-xadc: Make IRQ optional

[Lars-Peter Clausen <lars@metafoo.de>]

On 4/25/21 5:25 PM, Jonathan Cameron wrote:
> On Sun, 25 Apr 2021 12:26:38 +0200
> Lars-Peter Clausen <lars@metafoo.de> wrote:
>
>> In some setups the IRQ signal of the XADC might not be wired to the host
>> system. The driver currently requires that an interrupt is specified. Make
>> the interrupt optional so the driver can be used in such setups where the
>> interrupt is not connected.
>>
>> Since both buffered capture as well as events depend on the interrupt being
>> connected both are not available when the interrupt is not connected.
>>
>> Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
> Code looks fine to me but raises a question.  As there is no validate_trigger
> callback in this driver, nothing stops it being associated with for example
> an hrtimer trigger.  I don't think that will work so perhaps we should add the
> callback?  If it does work, then the presence or not of an irq should not
> change whether we register the buffer.

This is a good point. But the driver can actually work with a hrtimer 
trigger. You'd get aliasing problems and all that stuff, but if you only 
care about the instantaneous values it will work.

Without an IRQ we should only mask out the events and not register the 
triggers. Registering the buffer should be fine.

- Lars