Skip to contentSkip to conversation

input: atmel_mxt_ts: detect touchpad from devicetree

PATCH In review Emil Renner Berthing 2 participants 3 messages Oct 4, 2017 – Oct 9, 2017

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 input: atmel_mxt_ts: detect touchpad from devicetree

The Samsung Chromebook Plus (rk3399-gru-kevin) has two of these controllers. One for the touchscreen and one for the touchpad. However the touchpad doesn't have any associated gpio buttons, so it shows up as another touchscreen.

This patch fixes it by matching the "atmel,atmel_mxt_tp" compatibility string in the device tree.

---
 drivers/input/touchscreen/atmel_mxt_ts.c   | 8 ++++++--
 include/linux/platform_data/atmel_mxt_ts.h | 1 +
 2 files changed, 7 insertions(+), 2 deletions(-)
drivers/input/touchscreen/atmel_mxt_ts.c +6 -2
index 7659bc48f1db..4ea06f447d15 100644
@@ -1945,8 +1945,7 @@ static int mxt_initialize_input_device(struct mxt_data *data)
 		input_set_abs_params(input_dev, ABS_PRESSURE, 0, 255, 0, 0);
 	}
 
-	/* If device has buttons we assume it is a touchpad */
-	if (pdata->t19_num_keys) {
+	if (pdata->touchpad) {
 		mxt_set_up_as_touchpad(input_dev, data);
 		mt_flags |= INPUT_MT_POINTER;
 	} else {
@@ -2936,6 +2935,9 @@ static const struct mxt_platform_data *mxt_parse_dt(struct i2c_client *client)
 	if (!pdata)
 		return ERR_PTR(-ENOMEM);
 
+	if (of_device_is_compatible(np, "atmel,atmel_mxt_tp"))
+		pdata->touchpad = true;
+
 	if (of_find_property(np, "linux,gpio-keymap", &proplen)) {
 		pdata->t19_num_keys = proplen / sizeof(u32);
 
@@ -2986,6 +2988,7 @@ static struct mxt_acpi_platform_data samus_platform_data[] = {
 		.pdata	= {
 			.t19_num_keys	= ARRAY_SIZE(samus_touchpad_buttons),
 			.t19_keymap	= samus_touchpad_buttons,
+			.touchpad	= true,
 		},
 	},
 	{
@@ -3011,6 +3014,7 @@ static struct mxt_acpi_platform_data chromebook_platform_data[] = {
 		.pdata	= {
 			.t19_num_keys	= ARRAY_SIZE(chromebook_tp_buttons),
 			.t19_keymap	= chromebook_tp_buttons,
+			.touchpad	= true,
 		},
 	},
 	{
include/linux/platform_data/atmel_mxt_ts.h +1 -0
index 695035a8d7fb..acb595bba02d 100644
@@ -24,6 +24,7 @@ enum mxt_suspend_mode {
 struct mxt_platform_data {
 	unsigned long irqflags;
 	u8 t19_num_keys;
+	u8 touchpad;
 	const unsigned int *t19_keymap;
 	enum mxt_suspend_mode suspend_mode;
 };
Signature · 1 line
2.14.2
maintainer Reviewed-by on v1 1/1
#1 lore raw
Quoted 8 lines · Emil Renner Berthing

The Samsung Chromebook Plus (rk3399-gru-kevin) has two of these controllers. One for the touchscreen and one for the touchpad. However the touchpad doesn't have any associated gpio buttons, so it shows up as another touchscreen.

This patch fixes it by matching the "atmel,atmel_mxt_tp" compatibility string in the device tree.

Thanks. I think this makes sense, given there are already devices out there using "atmel_mxt_tp".

Is the hardcoded MXT_PIXELS_PER_MM going to cause issues on this new device?

I think you need to add this to Documentation/devicetree/bindings/input/atmel,maxtouch.txt

Quoted 61 lines
---
 drivers/input/touchscreen/atmel_mxt_ts.c   | 8 ++++++--
 include/linux/platform_data/atmel_mxt_ts.h | 1 +
 2 files changed, 7 insertions(+), 2 deletions(-)
drivers/input/touchscreen/atmel_mxt_ts.c +6 -2
index 7659bc48f1db..4ea06f447d15 100644
@@ -1945,8 +1945,7 @@ static int mxt_initialize_input_device(struct mxt_data *data)
 		input_set_abs_params(input_dev, ABS_PRESSURE, 0, 255, 0, 0);
 	}
 
-	/* If device has buttons we assume it is a touchpad */
-	if (pdata->t19_num_keys) {
+	if (pdata->touchpad) {
 		mxt_set_up_as_touchpad(input_dev, data);
 		mt_flags |= INPUT_MT_POINTER;
 	} else {
@@ -2936,6 +2935,9 @@ static const struct mxt_platform_data *mxt_parse_dt(struct i2c_client *client)
 	if (!pdata)
 		return ERR_PTR(-ENOMEM);
 
+	if (of_device_is_compatible(np, "atmel,atmel_mxt_tp"))
+		pdata->touchpad = true;
+
 	if (of_find_property(np, "linux,gpio-keymap", &proplen)) {
 		pdata->t19_num_keys = proplen / sizeof(u32);
 
@@ -2986,6 +2988,7 @@ static struct mxt_acpi_platform_data samus_platform_data[] = {
 		.pdata	= {
 			.t19_num_keys	= ARRAY_SIZE(samus_touchpad_buttons),
 			.t19_keymap	= samus_touchpad_buttons,
+			.touchpad	= true,
 		},
 	},
 	{
@@ -3011,6 +3014,7 @@ static struct mxt_acpi_platform_data chromebook_platform_data[] = {
 		.pdata	= {
 			.t19_num_keys	= ARRAY_SIZE(chromebook_tp_buttons),
 			.t19_keymap	= chromebook_tp_buttons,
+			.touchpad	= true,
 		},
 	},
 	{
include/linux/platform_data/atmel_mxt_ts.h +1 -0
index 695035a8d7fb..acb595bba02d 100644
@@ -24,6 +24,7 @@ enum mxt_suspend_mode {
 struct mxt_platform_data {
 	unsigned long irqflags;
 	u8 t19_num_keys;
+	u8 touchpad;
 	const unsigned int *t19_keymap;
 	enum mxt_suspend_mode suspend_mode;
 };
Signature · 1 line
2.14.2
#2 lore raw
Quoted 7 lines · Nick Dyer
Quoted 6 lines · Emil Renner Berthing

The Samsung Chromebook Plus (rk3399-gru-kevin) has two of these controllers. One for the touchscreen and one for the touchpad. However the touchpad doesn't have any associated gpio buttons, so it shows up as another touchscreen.

Quoted 8 lines
Quoted 2 lines

This patch fixes it by matching the "atmel,atmel_mxt_tp" compatibility string in the device tree.

Thanks. I think this makes sense, given there are already devices out there using "atmel_mxt_tp".

Is the hardcoded MXT_PIXELS_PER_MM going to cause issues on this new device?

I'm not exactly sure what kind of issues a wrong value would give, but the touchpad feels totally usable in X with libinput.
I just checked google's chromeos kernel and it seems to use the same hard-coded value.

Quoted 2 lines

I think you need to add this to Documentation/devicetree/bindings/input/atmel,maxtouch.txt

Speaking of issues I was just about to complain that the "click button" in the touchpad
didn't work, but only touch. It turns out I completely misunderstood what those "gpio buttons" was all about. Adding

+ linux,gpio-keymap = <KEY_RESERVED

+                                    KEY_RESERVED
+                                    KEY_RESERVED
+                                    BTN_LEFT>;

to the device tree fixes exactly that, and hence the current logic in the driver is fine for detecting the touchpad.

I'll submit another patch to add this to the device tree rather than this misunderstood patch.

Quoted 1 line

Thank you for the review, and sorry about the noise.

/Emil