input: goodix: Add option to power off the controller during suspend

From: Ondřej Jirman <megi@xff.cz>

For whatever reason the controller is not suspended on Pinephone
by the default procedure. It consumes quite a bit of power (~40mW)
during system sleep, and more when the screen is touched.

Let's power off the controller during system sleep instead.

Signed-off-by: Ondrej Jirman <megi@xff.cz>

input: goodix: Don't disable regulators during suspend

It does no harm to disable them, but on Pinephone we have other
peripherals attached to the same power supplies, whose drivers
will not reference the regulator, so powering down the regulators
from Goodix driver will break those other devices.

Until those drivers gain the regulator support, don't disable
the regulators in Goodix driver.

Signed-off-by: Ondrej Jirman <megi@xff.cz>
Signed-off-by: Jan Jasper de Kroon <jajadekroon@gmail.com>
---
 drivers/input/touchscreen/goodix.c | 23 +++++++++++++++++++++++
 drivers/input/touchscreen/goodix.h |  1 +
 2 files changed, 24 insertions(+)

Hi,

On 3/16/23 16:22, Jan Jasper de Kroon wrote:
> From: Ondřej Jirman <megi@xff.cz>
> 
> The Goodix touchscreen controller used in PinePhone is not properly
> suspended during system sleep, leading to high power consumption. This
> commit modifies the driver to hold the controller in reset during
> system sleep, reducing power consumption by around 40mW.
> 
> The original patch also disabled the regulators during system sleep, but
> this could cause issues with other peripherals attached to the same power
> supplies. Hence, this commit only holds the controller in reset during
> system sleep and does not attempt to power it off.
> 
> A separate commit adds a new device tree property "goodix-hold-in-reset"
> to control the hold-in-reset behavior of the controller during system sleep.
> 
> Signed-off-by: Ondrej Jirman <megi@xff.cz>
> Signed-off-by: Jan Jasper de Kroon <jajadekroon@gmail.com>
> ---
> Changes from v3 to v4:
> - Renamed the hold-in-reset-in-suspend function to 
>   goodix-hold-in-reset to prevent potential naming conflicts with other 
>   functions in the codebase. No functional changes were made.
> 
> Changes from v2 to v3:
> - Some patch style cleaning up, to confirm to submission standard.
> 
> Changes from v1 to v2:
> - Rewrote the commit message and function name to better reflect it's 
>   behavior.
>  drivers/input/touchscreen/goodix.c | 23 +++++++++++++++++++++++
>  drivers/input/touchscreen/goodix.h |  1 +
>  2 files changed, 24 insertions(+)
> 
> diff --git a/drivers/input/touchscreen/goodix.c b/drivers/input/touchscreen/goodix.c
> index b348172f19c3..9f86bdfbfc73 100644
> --- a/drivers/input/touchscreen/goodix.c
> +++ b/drivers/input/touchscreen/goodix.c
> @@ -1284,6 +1284,7 @@ static void goodix_disable_regulators(void *arg)
>  
>  static int goodix_ts_probe(struct i2c_client *client)
>  {
> +	struct device_node *np = client->dev.of_node;
>  	struct goodix_ts_data *ts;
>  	const char *cfg_name;
>  	int error;
> @@ -1303,6 +1304,7 @@ static int goodix_ts_probe(struct i2c_client *client)
>  	i2c_set_clientdata(client, ts);
>  	init_completion(&ts->firmware_loading_complete);
>  	ts->contact_size = GOODIX_CONTACT_SIZE;
> +	ts->goodix_hold_in_reset = of_property_read_bool(np, "goodix-hold-in-reset");

I only noticed this just now, sorry. Can you please use
"device_property_read_bool(&client-dev, "goodix-hold-in-reset");
instead, this way this will also work with non device-tree (aka OF)
platforms.

Also with devicetree properties if you prefix them by the vendor
name the vendor-prefix should end with a comma, so you get:

"goodix,hold-in-reset"

And I think this name is not very descriptive, this really should be something like:

"goodix,hold-in-reset-during-suspend"

to make what it does more clear.

I assume that this is going to get set inside a dts file, in that case this
new property should be documented in:

Documentation/devicetree/bindings/input/touchscree /goodix.yaml

and these devicetree-bindings changes must be in a separate patch
(and the dt-bindings maintainers should be Cc-ed on that patch).

Regards,

Hans




>  
>  	error = goodix_get_gpio_config(ts);
>  	if (error)
> @@ -1410,6 +1412,13 @@ static int goodix_suspend(struct device *dev)
>  	if (ts->load_cfg_from_disk)
>  		wait_for_completion(&ts->firmware_loading_complete);
>  
> +	if (ts->goodix_hold_in_reset) {
> +		goodix_free_irq(ts);
> +		goodix_irq_direction_output(ts, 0);
> +		gpiod_direction_output(ts->gpiod_rst, 0);
> +		return 0;
> +	}
> +
>  	/* We need gpio pins to suspend/resume */
>  	if (ts->irq_pin_access_method == IRQ_PIN_ACCESS_NONE) {
>  		disable_irq(client->irq);
> @@ -1455,6 +1464,20 @@ static int goodix_resume(struct device *dev)
>  	u8 config_ver;
>  	int error;
>  
> +	if (ts->goodix_hold_in_reset) {
> +		error = goodix_reset(ts);
> +		if (error) {
> +			dev_err(dev, "Controller reset failed.\n");
> +			return error;
> +		}
> +
> +		error = goodix_request_irq(ts);
> +		if (error)
> +			return error;
> +
> +		return 0;
> +	}
> +
>  	if (ts->irq_pin_access_method == IRQ_PIN_ACCESS_NONE) {
>  		enable_irq(client->irq);
>  		return 0;
> diff --git a/drivers/input/touchscreen/goodix.h b/drivers/input/touchscreen/goodix.h
> index 87797cc88b32..7be7eb1a8f27 100644
> --- a/drivers/input/touchscreen/goodix.h
> +++ b/drivers/input/touchscreen/goodix.h
> @@ -104,6 +104,7 @@ struct goodix_ts_data {
>  	u8 main_clk[GOODIX_MAIN_CLK_LEN];
>  	int bak_ref_len;
>  	u8 *bak_ref;
> +	bool goodix_hold_in_reset;
>  };
>  
>  int goodix_i2c_read(struct i2c_client *client, u16 reg, u8 *buf, int len);