diff mbox series

[4/7] HID: i2c-hid: goodix: Stop tying the reset line to the regulator

Message ID 20230206184744.4.I085b32b6140c7d1ac4e7e97b712bff9dd5962b62@changeid (mailing list archive)
State Not Applicable
Headers show
Series arm: qcom: Fix touchscreen voltage for sc7280-herobrine boards | expand

Commit Message

Doug Anderson Feb. 7, 2023, 2:48 a.m. UTC
In commit 18eeef46d359 ("HID: i2c-hid: goodix: Tie the reset line to
true state of the regulator"), we started tying the reset line of
Goodix touchscreens to the regulator.

The primary motivation for that patch was some pre-production hardware
(specifically sc7180-trogdor-homestar) where it was proposed to hook
the touchscreen's main 3.3V power rail to an always-on supply. In such
a case, when we turned "off" the touchscreen in Linux it was bad to
assert the "reset" GPIO because that was causing a power drain. The
patch accomplished that goal and did it in a general sort of way that
didn't require special properties to be added in the device tree for
homestar.

It turns out that the design of using an always-on power rail for the
touchscreen was rejected soon after the patch was written and long
before sc7180-trogdor-homestar went into production. The final design
of homestar actually fully separates the rail for the touchscreen and
the display panel and both can be powered off and on. That means that
the original motivation for the feature is gone.

There are 3 other users of the goodix i2c-hid driver in mainline.

I'll first talk about 2 of the other users in mainline: coachz and
mrbland. On both coachz and mrbland the touchscreen power and panel
power _are_ shared. That means that the patch to tie the reset line to
the true state of the regulator _is_ doing something on those
boards. Specifically, the patch reduced power consumption by tens of
mA in the case where we turned the touchscreen off but left the panel
on. Other than saving a small bit of power, the patch wasn't truly
necessary. That being said, even though a small bit of power was saved
in the state of "panel on + touchscreen off", that's not actually a
state we ever expect to be in, except perhaps for very short periods
of time at boot or during suspend/resume. Thus, the patch is truly not
necessary. It should be further noted that, as documented in the
original patch, the current code still didn't optimize power for every
corner case of the "shared rail" situation.

The last user in mainline was very recently added: evoker. Evoker is
actually the motivation for me removing this bit of code. It turns out
that for evoker we need to manage a second power rail for IO to the
touchscreen. Trying to fit the management of this IO rail into the
regulator notifiers turns out to be extremely hard. To avoid lockdep
splats you shouldn't enable/disable other regulators in regulator
notifiers and trying to find a way around this was going to be fairly
difficult.

Given the lack of any true motivation to tie the reset line to the
regulator, lets go back to the simpler days and remove the code. This
is, effectively, a revert of commit bdbc65eb77ee ("HID: i2c-hid:
goodix: Fix a lockdep splat"), commit 25ddd7cfc582 ("HID: i2c-hid:
goodix: Use the devm variant of regulator_register_notifier()"), and
commit 18eeef46d359 ("HID: i2c-hid: goodix: Tie the reset line to true
state of the regulator").

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

 drivers/hid/i2c-hid/i2c-hid-of-goodix.c | 88 ++++---------------------
 1 file changed, 13 insertions(+), 75 deletions(-)

Comments

Dmitry Torokhov Feb. 7, 2023, 3:58 a.m. UTC | #1
On Mon, Feb 06, 2023 at 06:48:13PM -0800, Douglas Anderson wrote:
> In commit 18eeef46d359 ("HID: i2c-hid: goodix: Tie the reset line to
> true state of the regulator"), we started tying the reset line of
> Goodix touchscreens to the regulator.
> 
> The primary motivation for that patch was some pre-production hardware
> (specifically sc7180-trogdor-homestar) where it was proposed to hook
> the touchscreen's main 3.3V power rail to an always-on supply. In such
> a case, when we turned "off" the touchscreen in Linux it was bad to
> assert the "reset" GPIO because that was causing a power drain. The
> patch accomplished that goal and did it in a general sort of way that
> didn't require special properties to be added in the device tree for
> homestar.
> 
> It turns out that the design of using an always-on power rail for the
> touchscreen was rejected soon after the patch was written and long
> before sc7180-trogdor-homestar went into production. The final design
> of homestar actually fully separates the rail for the touchscreen and
> the display panel and both can be powered off and on. That means that
> the original motivation for the feature is gone.
> 
> There are 3 other users of the goodix i2c-hid driver in mainline.
> 
> I'll first talk about 2 of the other users in mainline: coachz and
> mrbland. On both coachz and mrbland the touchscreen power and panel
> power _are_ shared. That means that the patch to tie the reset line to
> the true state of the regulator _is_ doing something on those
> boards. Specifically, the patch reduced power consumption by tens of
> mA in the case where we turned the touchscreen off but left the panel
> on. Other than saving a small bit of power, the patch wasn't truly
> necessary. That being said, even though a small bit of power was saved
> in the state of "panel on + touchscreen off", that's not actually a
> state we ever expect to be in, except perhaps for very short periods
> of time at boot or during suspend/resume. Thus, the patch is truly not
> necessary. It should be further noted that, as documented in the
> original patch, the current code still didn't optimize power for every
> corner case of the "shared rail" situation.
> 
> The last user in mainline was very recently added: evoker. Evoker is
> actually the motivation for me removing this bit of code. It turns out
> that for evoker we need to manage a second power rail for IO to the
> touchscreen. Trying to fit the management of this IO rail into the
> regulator notifiers turns out to be extremely hard. To avoid lockdep
> splats you shouldn't enable/disable other regulators in regulator
> notifiers and trying to find a way around this was going to be fairly
> difficult.
> 
> Given the lack of any true motivation to tie the reset line to the
> regulator, lets go back to the simpler days and remove the code. This
> is, effectively, a revert of commit bdbc65eb77ee ("HID: i2c-hid:
> goodix: Fix a lockdep splat"), commit 25ddd7cfc582 ("HID: i2c-hid:
> goodix: Use the devm variant of regulator_register_notifier()"), and
> 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>

> ---
> 
>  drivers/hid/i2c-hid/i2c-hid-of-goodix.c | 88 ++++---------------------
>  1 file changed, 13 insertions(+), 75 deletions(-)
> 
> diff --git a/drivers/hid/i2c-hid/i2c-hid-of-goodix.c b/drivers/hid/i2c-hid/i2c-hid-of-goodix.c
> index 29c6cb174032..584d833dc0aa 100644
> --- a/drivers/hid/i2c-hid/i2c-hid-of-goodix.c
> +++ b/drivers/hid/i2c-hid/i2c-hid-of-goodix.c
> @@ -26,28 +26,28 @@ struct i2c_hid_of_goodix {
>  	struct i2chid_ops ops;
>  
>  	struct regulator *vdd;
> -	struct notifier_block nb;
>  	struct gpio_desc *reset_gpio;
>  	const struct goodix_i2c_hid_timing_data *timings;
>  };
>  
> -static void goodix_i2c_hid_deassert_reset(struct i2c_hid_of_goodix *ihid_goodix,
> -					  bool regulator_just_turned_on)
> +static int goodix_i2c_hid_power_up(struct i2chid_ops *ops)
>  {
> -	if (regulator_just_turned_on && ihid_goodix->timings->post_power_delay_ms)
> +	struct i2c_hid_of_goodix *ihid_goodix =
> +		container_of(ops, struct i2c_hid_of_goodix, ops);
> +	int ret;
> +
> +	ret = regulator_enable(ihid_goodix->vdd);
> +	if (ret)
> +		return ret;
> +
> +	if (ihid_goodix->timings->post_power_delay_ms)
>  		msleep(ihid_goodix->timings->post_power_delay_ms);
>  
>  	gpiod_set_value_cansleep(ihid_goodix->reset_gpio, 0);
>  	if (ihid_goodix->timings->post_gpio_reset_delay_ms)
>  		msleep(ihid_goodix->timings->post_gpio_reset_delay_ms);
> -}
> -
> -static int goodix_i2c_hid_power_up(struct i2chid_ops *ops)
> -{
> -	struct i2c_hid_of_goodix *ihid_goodix =
> -		container_of(ops, struct i2c_hid_of_goodix, ops);
>  
> -	return regulator_enable(ihid_goodix->vdd);
> +	return 0;
>  }
>  
>  static void goodix_i2c_hid_power_down(struct i2chid_ops *ops)
> @@ -55,42 +55,14 @@ static void goodix_i2c_hid_power_down(struct i2chid_ops *ops)
>  	struct i2c_hid_of_goodix *ihid_goodix =
>  		container_of(ops, struct i2c_hid_of_goodix, ops);
>  
> +	gpiod_set_value_cansleep(ihid_goodix->reset_gpio, 1);
>  	regulator_disable(ihid_goodix->vdd);
>  }
>  
> -static int ihid_goodix_vdd_notify(struct notifier_block *nb,
> -				    unsigned long event,
> -				    void *ignored)
> -{
> -	struct i2c_hid_of_goodix *ihid_goodix =
> -		container_of(nb, struct i2c_hid_of_goodix, nb);
> -	int ret = NOTIFY_OK;
> -
> -	switch (event) {
> -	case REGULATOR_EVENT_PRE_DISABLE:
> -		gpiod_set_value_cansleep(ihid_goodix->reset_gpio, 1);
> -		break;
> -
> -	case REGULATOR_EVENT_ENABLE:
> -		goodix_i2c_hid_deassert_reset(ihid_goodix, true);
> -		break;
> -
> -	case REGULATOR_EVENT_ABORT_DISABLE:
> -		goodix_i2c_hid_deassert_reset(ihid_goodix, false);
> -		break;
> -
> -	default:
> -		ret = NOTIFY_DONE;
> -		break;
> -	}
> -
> -	return ret;
> -}
> -
>  static int i2c_hid_of_goodix_probe(struct i2c_client *client)
>  {
>  	struct i2c_hid_of_goodix *ihid_goodix;
> -	int ret;
> +
>  	ihid_goodix = devm_kzalloc(&client->dev, sizeof(*ihid_goodix),
>  				   GFP_KERNEL);
>  	if (!ihid_goodix)
> @@ -111,40 +83,6 @@ static int i2c_hid_of_goodix_probe(struct i2c_client *client)
>  
>  	ihid_goodix->timings = device_get_match_data(&client->dev);
>  
> -	/*
> -	 * We need to control the "reset" line in lockstep with the regulator
> -	 * actually turning on an off instead of just when we make the request.
> -	 * This matters if the regulator is shared with another consumer.
> -	 * - If the regulator is off then we must assert reset. The reset
> -	 *   line is active low and on some boards it could cause a current
> -	 *   leak if left high.
> -	 * - If the regulator is on then we don't want reset asserted for very
> -	 *   long. Holding the controller in reset apparently draws extra
> -	 *   power.
> -	 */
> -	ihid_goodix->nb.notifier_call = ihid_goodix_vdd_notify;
> -	ret = devm_regulator_register_notifier(ihid_goodix->vdd, &ihid_goodix->nb);
> -	if (ret)
> -		return dev_err_probe(&client->dev, ret,
> -			"regulator notifier request failed\n");
> -
> -	/*
> -	 * If someone else is holding the regulator on (or the regulator is
> -	 * an always-on one) we might never be told to deassert reset. Do it
> -	 * now... and temporarily bump the regulator reference count just to
> -	 * make sure it is impossible for this to race with our own notifier!
> -	 * We also assume that someone else might have _just barely_ turned
> -	 * the regulator on so we'll do the full "post_power_delay" just in
> -	 * case.
> -	 */
> -	if (ihid_goodix->reset_gpio && regulator_is_enabled(ihid_goodix->vdd)) {
> -		ret = regulator_enable(ihid_goodix->vdd);
> -		if (ret)
> -			return ret;
> -		goodix_i2c_hid_deassert_reset(ihid_goodix, true);
> -		regulator_disable(ihid_goodix->vdd);
> -	}
> -
>  	return i2c_hid_core_probe(client, &ihid_goodix->ops, 0x0001, 0);
>  }
>  
> -- 
> 2.39.1.519.gcb327c4b5f-goog
>
Matthias Kaehlcke Feb. 7, 2023, 6:31 p.m. UTC | #2
On Mon, Feb 06, 2023 at 06:48:13PM -0800, Douglas Anderson wrote:
> In commit 18eeef46d359 ("HID: i2c-hid: goodix: Tie the reset line to
> true state of the regulator"), we started tying the reset line of
> Goodix touchscreens to the regulator.
> 
> The primary motivation for that patch was some pre-production hardware
> (specifically sc7180-trogdor-homestar) where it was proposed to hook
> the touchscreen's main 3.3V power rail to an always-on supply. In such
> a case, when we turned "off" the touchscreen in Linux it was bad to
> assert the "reset" GPIO because that was causing a power drain. The
> patch accomplished that goal and did it in a general sort of way that
> didn't require special properties to be added in the device tree for
> homestar.
> 
> It turns out that the design of using an always-on power rail for the
> touchscreen was rejected soon after the patch was written and long
> before sc7180-trogdor-homestar went into production. The final design
> of homestar actually fully separates the rail for the touchscreen and
> the display panel and both can be powered off and on. That means that
> the original motivation for the feature is gone.
> 
> There are 3 other users of the goodix i2c-hid driver in mainline.
> 
> I'll first talk about 2 of the other users in mainline: coachz and
> mrbland. On both coachz and mrbland the touchscreen power and panel
> power _are_ shared. That means that the patch to tie the reset line to
> the true state of the regulator _is_ doing something on those
> boards. Specifically, the patch reduced power consumption by tens of
> mA in the case where we turned the touchscreen off but left the panel
> on. Other than saving a small bit of power, the patch wasn't truly
> necessary. That being said, even though a small bit of power was saved
> in the state of "panel on + touchscreen off", that's not actually a
> state we ever expect to be in, except perhaps for very short periods
> of time at boot or during suspend/resume. Thus, the patch is truly not
> necessary. It should be further noted that, as documented in the
> original patch, the current code still didn't optimize power for every
> corner case of the "shared rail" situation.
> 
> The last user in mainline was very recently added: evoker. Evoker is
> actually the motivation for me removing this bit of code. It turns out
> that for evoker we need to manage a second power rail for IO to the
> touchscreen. Trying to fit the management of this IO rail into the
> regulator notifiers turns out to be extremely hard. To avoid lockdep
> splats you shouldn't enable/disable other regulators in regulator
> notifiers and trying to find a way around this was going to be fairly
> difficult.
> 
> Given the lack of any true motivation to tie the reset line to the
> regulator, lets go back to the simpler days and remove the code. This
> is, effectively, a revert of commit bdbc65eb77ee ("HID: i2c-hid:
> goodix: Fix a lockdep splat"), commit 25ddd7cfc582 ("HID: i2c-hid:
> goodix: Use the devm variant of regulator_register_notifier()"), and
> 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: Matthias Kaehlcke <mka@chromium.org>
diff mbox series

Patch

diff --git a/drivers/hid/i2c-hid/i2c-hid-of-goodix.c b/drivers/hid/i2c-hid/i2c-hid-of-goodix.c
index 29c6cb174032..584d833dc0aa 100644
--- a/drivers/hid/i2c-hid/i2c-hid-of-goodix.c
+++ b/drivers/hid/i2c-hid/i2c-hid-of-goodix.c
@@ -26,28 +26,28 @@  struct i2c_hid_of_goodix {
 	struct i2chid_ops ops;
 
 	struct regulator *vdd;
-	struct notifier_block nb;
 	struct gpio_desc *reset_gpio;
 	const struct goodix_i2c_hid_timing_data *timings;
 };
 
-static void goodix_i2c_hid_deassert_reset(struct i2c_hid_of_goodix *ihid_goodix,
-					  bool regulator_just_turned_on)
+static int goodix_i2c_hid_power_up(struct i2chid_ops *ops)
 {
-	if (regulator_just_turned_on && ihid_goodix->timings->post_power_delay_ms)
+	struct i2c_hid_of_goodix *ihid_goodix =
+		container_of(ops, struct i2c_hid_of_goodix, ops);
+	int ret;
+
+	ret = regulator_enable(ihid_goodix->vdd);
+	if (ret)
+		return ret;
+
+	if (ihid_goodix->timings->post_power_delay_ms)
 		msleep(ihid_goodix->timings->post_power_delay_ms);
 
 	gpiod_set_value_cansleep(ihid_goodix->reset_gpio, 0);
 	if (ihid_goodix->timings->post_gpio_reset_delay_ms)
 		msleep(ihid_goodix->timings->post_gpio_reset_delay_ms);
-}
-
-static int goodix_i2c_hid_power_up(struct i2chid_ops *ops)
-{
-	struct i2c_hid_of_goodix *ihid_goodix =
-		container_of(ops, struct i2c_hid_of_goodix, ops);
 
-	return regulator_enable(ihid_goodix->vdd);
+	return 0;
 }
 
 static void goodix_i2c_hid_power_down(struct i2chid_ops *ops)
@@ -55,42 +55,14 @@  static void goodix_i2c_hid_power_down(struct i2chid_ops *ops)
 	struct i2c_hid_of_goodix *ihid_goodix =
 		container_of(ops, struct i2c_hid_of_goodix, ops);
 
+	gpiod_set_value_cansleep(ihid_goodix->reset_gpio, 1);
 	regulator_disable(ihid_goodix->vdd);
 }
 
-static int ihid_goodix_vdd_notify(struct notifier_block *nb,
-				    unsigned long event,
-				    void *ignored)
-{
-	struct i2c_hid_of_goodix *ihid_goodix =
-		container_of(nb, struct i2c_hid_of_goodix, nb);
-	int ret = NOTIFY_OK;
-
-	switch (event) {
-	case REGULATOR_EVENT_PRE_DISABLE:
-		gpiod_set_value_cansleep(ihid_goodix->reset_gpio, 1);
-		break;
-
-	case REGULATOR_EVENT_ENABLE:
-		goodix_i2c_hid_deassert_reset(ihid_goodix, true);
-		break;
-
-	case REGULATOR_EVENT_ABORT_DISABLE:
-		goodix_i2c_hid_deassert_reset(ihid_goodix, false);
-		break;
-
-	default:
-		ret = NOTIFY_DONE;
-		break;
-	}
-
-	return ret;
-}
-
 static int i2c_hid_of_goodix_probe(struct i2c_client *client)
 {
 	struct i2c_hid_of_goodix *ihid_goodix;
-	int ret;
+
 	ihid_goodix = devm_kzalloc(&client->dev, sizeof(*ihid_goodix),
 				   GFP_KERNEL);
 	if (!ihid_goodix)
@@ -111,40 +83,6 @@  static int i2c_hid_of_goodix_probe(struct i2c_client *client)
 
 	ihid_goodix->timings = device_get_match_data(&client->dev);
 
-	/*
-	 * We need to control the "reset" line in lockstep with the regulator
-	 * actually turning on an off instead of just when we make the request.
-	 * This matters if the regulator is shared with another consumer.
-	 * - If the regulator is off then we must assert reset. The reset
-	 *   line is active low and on some boards it could cause a current
-	 *   leak if left high.
-	 * - If the regulator is on then we don't want reset asserted for very
-	 *   long. Holding the controller in reset apparently draws extra
-	 *   power.
-	 */
-	ihid_goodix->nb.notifier_call = ihid_goodix_vdd_notify;
-	ret = devm_regulator_register_notifier(ihid_goodix->vdd, &ihid_goodix->nb);
-	if (ret)
-		return dev_err_probe(&client->dev, ret,
-			"regulator notifier request failed\n");
-
-	/*
-	 * If someone else is holding the regulator on (or the regulator is
-	 * an always-on one) we might never be told to deassert reset. Do it
-	 * now... and temporarily bump the regulator reference count just to
-	 * make sure it is impossible for this to race with our own notifier!
-	 * We also assume that someone else might have _just barely_ turned
-	 * the regulator on so we'll do the full "post_power_delay" just in
-	 * case.
-	 */
-	if (ihid_goodix->reset_gpio && regulator_is_enabled(ihid_goodix->vdd)) {
-		ret = regulator_enable(ihid_goodix->vdd);
-		if (ret)
-			return ret;
-		goodix_i2c_hid_deassert_reset(ihid_goodix, true);
-		regulator_disable(ihid_goodix->vdd);
-	}
-
 	return i2c_hid_core_probe(client, &ihid_goodix->ops, 0x0001, 0);
 }