> On 10/20/2020 10:22 AM Geert Uytterhoeven <geert@xxxxxxxxxxxxxx> wrote:
> > Should input be enabled unconditionally, as recommended by the Hardware
> > Manual for backwards compatibility with existing software?
> > As per (errata?) commit ae9550f635533b1c ("gpio-rcar: Use OUTDT when
> > reading GPIOs configured as output"), the gpio-rcar driver does not use
> > the INDT register to read the status of a GPIO line when configured for
> > output.
>
> More investigation:
> 1. On SH7734, R-Car Gen1, R-Car Gen2, and RZ/G1, INDT only reflects
> the state of the input pin if the GPIO is configured for input
> (documented, verified on R-Car M2-W),
> 2. On R-Car Gen3 and RZ/G2, INDT always reflects the state of the
> input pin (documented, verified on R-Car H3/M3-W/M3-N),
> 3. On R-Car V3U, input enable can be configured explicitly through the
> new INEN register (documented, unverified).
>
> Given register accesses are slow, I think it makes sense to enable input
> explicitly and unconditionally on R-Car V3U, and always use INDT to read
> pin state on R-Car Gen3 and V3U.
That sounds like a good solution to me.
CU
Uli
