Re: [PATCH v2 resend 1/2] gpio: tps68470: Fix tps68470_gpio_get() reading from the wrong register

[Andy Shevchenko <andy.shevchenko@xxxxxxxxx>]

On Wed, Jan 25, 2023 at 1:23 PM Andy Shevchenko
<andy.shevchenko@xxxxxxxxx> wrote:
> On Wed, Jan 25, 2023 at 1:05 PM Hans de Goede <hdegoede@xxxxxxxxxx> wrote:
> >
> > For the regular GPIO pins the value should be read from TPS68470_REG_GPDI,
> > so that the actual value of the pin is read, rather than the value the pin
> > would output when put in output mode.
>
> It really depends. I think it's a wrong perception and brings nothing
> to software. If we output, we know what we program, so reading back
> returns us what we _assume_ should be on the pin under normal
> circumstances. The difference is OD/OS/OE/OC cases where we output
> only one possible value.
>
> > Fixes: 275b13a65547 ("gpio: Add support for TPS68470 GPIOs")
>
> Is it really fixing anything? Can we have more?
>
> P.S. Before doing this, I would have a clarification in the
> documentation. Sorry that I have had no time to respond to my series
> regarding that. But it seems we have a strong disagreement in the
> area.

Just realized that what we are missing is BIDI settings for the
direction. That would solve a lot of these kinds of grey areas. Linus?
Bart?


-- 
With Best Regards,
Andy Shevchenko