Re: [PATCH v2 7/7] iio: adc: ti-ads1015: Add static assert to test if shifted realbits fit into storagebits

[Marek Vasut <marex@denx.de>]

On 3/20/22 14:10, Jonathan Cameron wrote:
> On Sat, 12 Mar 2022 22:28:22 +0100
> Marek Vasut <marex@denx.de> wrote:
> 
>> On 3/12/22 13:19, Andy Shevchenko wrote:
>>> On Fri, Mar 11, 2022 at 8:50 PM Marek Vasut <marex@denx.de> wrote:
>>>>
>>>> Add compile-time static_assert wrapper to verify that shifted realbits
>>>> fit into storagebits. The macro is implemented in a more generic way so
>>>> it can be used to verify other values if required.
>>>
>>> Thanks! I think we may leave it to maintainers to decide if it is
>>> worth adding or not.
>>
>> Right, that's why I placed it as 7/7, since the macro is ... not pretty.
> 
> It's ugly but that's all wrapped up in the macro so I'll take it and
> see what blows up :)
> 
> Longer term, maybe we'd should add a general runtime check in the IIO core?

I wonder whether something like that could be made completely generic, 
but, let's do that in the next step.

> Not quite so nice as catching at compile time but would catch all such issues
> the moment anyone actually tries out a driver with whatever device they've
> just added support for.
> 
> I'll let this sit a little longer for additional review before picking it up
> (missed this cycle anyway so lots of time).
> 
> 
> A few other comments on this driver whilst we are here on things we should
> cleanup at somepoint.
> 1) Move over to the read_avail callback rather than having the attribute
>     groups.  It may well be slightly more code but then makes them available
>     to in kernel users.   Slowly moving all drivers with _available for standard
>     ABI over to the callback is on the todo list but it will take a while...
> 2) Possibly introduced a static const array of
>     struct ad1015_chip_info {
> 	all the per device stuff currently handled in code in probe
>     };
> Then simply access that directly using the the device_match_data.

I found a small bug in the tla2024 integration, the attrs assignment is 
wrong, so I'll send 1 and 2 above alongside V3 shortly.