linux-kernel - Re: [PATCH v4 2/2] input: touchscreen: goodix: Hold controller in reset during suspend

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <5657267e-9d57-4d55-5b1e-d46fed90ff59@redhat.com>
Date:   Thu, 16 Mar 2023 18:25:37 +0100
From:   Hans de Goede <hdegoede@...hat.com>
To:     Jan Jasper de Kroon <jajadekroon@...il.com>
Cc:     dmitry.torokhov@...il.com, hadess@...ess.net,
        linux-input@...r.kernel.org, linux-kernel@...r.kernel.org,
        megi@....cz
Subject: Re: [PATCH v4 2/2] input: touchscreen: goodix: Hold controller in
 reset during suspend

Hi,

On 3/16/23 16:22, Jan Jasper de Kroon wrote:
> From: Ondřej Jirman <megi@....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@....cz>
> Signed-off-by: Jan Jasper de Kroon <jajadekroon@...il.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);