[1/2] dt-bindings: HID: i2c-hid: elan: Introduce bindings for Elan eKTH6915

Like many i2c-hid touchscreen controllers, the Elan eKTH6915
controller has a reset gpio. For the Goodix GT7375P touchscreen the
decision was to add a new binding rather than trying to add a new GPIO
to the existing i2c-hid binding. We'll follow the lead and do it here,
too.

SIDE NOTE: the Elan eKTH6915 is a touchscreen _controller_ that's
included as a part on some touchscreens. The reset line isn't truly
necessary for the functioning of the touchscreen, so it's possible
that some designs won't have it hooked up and will just guarantee the
power sequencing requirements with RLC circuits. Thus, we'll mark the
reset gpio as optional.

Note that if the reset GPIO isn't used there's actually no true need
to use the "elan,ekth6915" compatible instead of the "hid-over-i2c" on
Linux. However:
- Officially using just "hid-over-i2c" for this device violates the
  existing "hid-over-i2c" bindings. The bindings say that you're not
  supposed to use "post-power-on-delay-ms" without specifying a more
  specific compatible. Currently the Linux driver doesn't enforce
  this, but it violates the bindings to just use
  "hid-over-i2c". ...and if you're going to add a more specific
  compatible anyway, might as well do it right.
- Using this compatible means we don't need to specify
  "hid-descr-addr" since it's inferred from the compatible.
- Using this compatible means that the regulator names match the names
  on the Elan datasheet (vcc33 / vccio) vs the generic hid-over-i2c
  (vdd / vddl).

Signed-off-by: Douglas Anderson <dianders@chromium.org>
---

 .../bindings/input/elan,ekth6915.yaml         | 65 +++++++++++++++++++
 1 file changed, 65 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/input/elan,ekth6915.yaml

On Thu, May 19, 2022 at 04:00:03PM -0700, Douglas Anderson wrote:
> Like many i2c-hid touchscreen controllers, the Elan eKTH6915 has a
> reset GPIO hooked up to it. According to the datasheet, the way we're
> supposed to turn the touchscreen on is:
> 
> 1. Turn on the 3.3V supply.
> 2. Turn on the IO supply. It's OK if this is hardwired to the 3.3V
>    supply, but if it's not then it must be turned on _after_ the 3.3V
>    supply.
> 3. Wait >= 1 ms.
> 4. Deassert the reset GPIO (reset GPIO is active low, so there would
>    be a leakage path if this was deasserted _before_ the IO supply).
> 5. Wait 300 ms.
> 
> Much of the above can be handled by the generic i2c-hid-of driver, but
> the "reset" GPIO is not supported by that driver. Thus we'll do the
> same as we did for Goodix and add a new tiny driver that uses the
> i2c-hid core.
> 
> NOTE: support for this new touchscreen could theorically fit into the
> Goodix driver. I've made it a separate driver because the Elan driver
> supports _two_ regulators and it's unclear exactly how that would fit
> in with commit 18eeef46d359 ("HID: i2c-hid: goodix: Tie the reset line
> to true state of the regulator").
> 
> Signed-off-by: Douglas Anderson <dianders@chromium.org>

Reviewed-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>

> +
> +	return i2c_hid_core_probe(client, &ihid_elan->ops, 0x0001, 0);

Maybe #define for the descriptor address, or rename
elan_i2c_hid_timing_data into elan_i2c_hid_chip_data and put it there?

Thanks.

On Thu, May 19, 2022 at 04:00:02PM -0700, Douglas Anderson wrote:
> Like many i2c-hid touchscreen controllers, the Elan eKTH6915
> controller has a reset gpio. For the Goodix GT7375P touchscreen the
> decision was to add a new binding rather than trying to add a new GPIO
> to the existing i2c-hid binding. We'll follow the lead and do it here,
> too.
> 
> SIDE NOTE: the Elan eKTH6915 is a touchscreen _controller_ that's
> included as a part on some touchscreens. The reset line isn't truly
> necessary for the functioning of the touchscreen, so it's possible
> that some designs won't have it hooked up and will just guarantee the
> power sequencing requirements with RLC circuits. Thus, we'll mark the
> reset gpio as optional.
> 
> Note that if the reset GPIO isn't used there's actually no true need
> to use the "elan,ekth6915" compatible instead of the "hid-over-i2c" on
> Linux. However:
> - Officially using just "hid-over-i2c" for this device violates the
>   existing "hid-over-i2c" bindings. The bindings say that you're not
>   supposed to use "post-power-on-delay-ms" without specifying a more
>   specific compatible. Currently the Linux driver doesn't enforce
>   this, but it violates the bindings to just use
>   "hid-over-i2c". ...and if you're going to add a more specific
>   compatible anyway, might as well do it right.
> - Using this compatible means we don't need to specify
>   "hid-descr-addr" since it's inferred from the compatible.
> - Using this compatible means that the regulator names match the names
>   on the Elan datasheet (vcc33 / vccio) vs the generic hid-over-i2c
>   (vdd / vddl).
> 
> Signed-off-by: Douglas Anderson <dianders@chromium.org>

Reviewed-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>

Thanks.