Re: [PATCH 1/2] Input: atmel_mxt_ts: Add wake-up support

Hi Dmitry,

On Thu, 24 Jun 2021 at 02:41, Dmitry Torokhov [off-list ref] wrote:

Hi Loic,

On Wed, Jun 23, 2021 at 03:56:36PM +0200, Loic Poulain wrote:

quoted

Add capability for the touchscreen to wakeup the host on touch event.

Signed-off-by: Loic Poulain <redacted>
---
 drivers/input/touchscreen/atmel_mxt_ts.c | 18 ++++++++++++++----
 1 file changed, 14 insertions(+), 4 deletions(-)

diff --git a/drivers/input/touchscreen/atmel_mxt_ts.c b/drivers/input/touchscreen/atmel_mxt_ts.c
index 05de92c..807f449 100644
--- a/drivers/input/touchscreen/atmel_mxt_ts.c
+++ b/drivers/input/touchscreen/atmel_mxt_ts.c

@@ -3223,6 +3223,8 @@ static int mxt_probe(struct i2c_client *client, const struct i2c_device_id *id)
              return error;
      }

+     device_set_wakeup_capable(&client->dev, true);

We do not want to make the touch controller be wakeup source
unconditionally. I2C core recognized "wakeup-source" in device tree,
other platforms may employ different techniques setting I2C_CLIENT_WAKE
when registering I2C devices to mark them as wakeup capable/enabled.

Contrary to device_init_wakeup(), used in some other input drivers,
device_set_wakeup_capable() does not enable the device as a wakeup
source but just sets it as wakeup capable, and it's up to the user or
distro policy to enable it as a wakeup source or not. It's a quite
common way to do, and it does not change the behavior of this driver.
The I2C_CLIENT_WAKE forces enabling wakeup source, which is maybe not
what we want by default for a touchscreen. remote-wakeup enabling is a
device configuration not a hardware property. Thoughts?

I should probably also add dev_pm_set_wake_irq() for auto-enabling
wake on suspend instead of doing it manually.

Regards,
Loic