Skip to contentSkip to conversation

Input: hynitron-cst816x: axis properties and gesture keys

PATCH Superseded Daniel Golle 3 participants 9 messages Oct 1

Subsystems (3)

v1 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

cover Input: hynitron-cst816x: axis properties and gesture keys

This series improves the hynitron-cst816x touchscreen driver. It makes the driver respect the common touchscreen axis properties, which are now documented in the binding as well, and it keeps gesture keys from getting stuck.

The gesture key is reported with the controller's touch flag as its value, so the key stays pressed once the controller stops reporting the gesture code. The controller has no lift event to offer, so the keys are now released on the first report carrying no gesture code, or 40 ms after the last one that carried a code, whichever comes first.

Daniel Golle (4):
dt-bindings: input: touchscreen: hynitron,cst816x: configure axes
Input: hynitron-cst816x: respect touchscreen DT properties
Input: hynitron-cst816x: release gesture keys
Input: hynitron-cst816x: time out gesture key release

 .../input/touchscreen/hynitron,cst816x.yaml   |  8 +++
 drivers/input/touchscreen/hynitron-cst816x.c  | 71 +++++++++++++++++--
 2 files changed, 74 insertions(+), 5 deletions(-)

base-commit: 6474fa070f2b8013b4b87350b775b8c3be6e8aac

Signature · 1 line
2.55.0
#1 lore raw

PATCH 1/4 dt-bindings: input: touchscreen: hynitron,cst816x: configure axes

Reference touchscreen.yaml and allow properties 'touchscreen-inverted-x', 'touchscreen-inverted-y' and 'touchscreen-swapped-x-y'.

---
 .../bindings/input/touchscreen/hynitron,cst816x.yaml      | 8 ++++++++
 1 file changed, 8 insertions(+)
Documentation/devicetree/bindings/input/touchscreen/hynitron,cst816x.yaml +8 -0
index 72d4da636881e..ba56bf61038c5 100644
@@ -36,6 +36,10 @@ properties:
       - description: Slide right gesture
       - description: Long press gesture
 
+  touchscreen-inverted-x: true
+  touchscreen-inverted-y: true
+  touchscreen-swapped-x-y: true
+
 required:
   - compatible
   - reg
@@ -43,6 +47,9 @@ required:
 
 additionalProperties: false
 
+allOf:
+  - $ref: touchscreen.yaml#
+
 examples:
   - |
     #include <dt-bindings/gpio/gpio.h>
@@ -59,6 +66,7 @@ examples:
             reset-gpios = <&gpio 17 GPIO_ACTIVE_LOW>;
             linux,keycodes = <KEY_UP>, <KEY_DOWN>, <KEY_LEFT>, <KEY_RIGHT>,
                              <BTN_TOOL_TRIPLETAP>;
+            touchscreen-swapped-x-y;
         };
     };
 
Signature · 1 line
2.55.0
#2 lore raw

Re: [PATCH 1/4] dt-bindings: input: touchscreen: hynitron,cst816x: configure axes

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider: - [Low] dt-bindings: incorrect use of additionalProperties with $ref --

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] dt-bindings: incorrect use of additionalProperties with $ref
--

commit 0f12b787d0fbdb216a571a789ed3dfa0d0a368db
Author: Daniel Golle <[email protected]>

dt-bindings: input: touchscreen: hynitron,cst816x: configure axes

This commit references touchscreen.yaml and allows the properties touchscreen-inverted-x, touchscreen-inverted-y, and touchscreen-swapped-x-y.

Documentation/devicetree/bindings/input/touchscreen/hynitron,cst816x.yaml properties: · line 41

Quoted 10 lines
Documentation/devicetree/bindings/input/touchscreen/hynitron,cst816x.yaml +3 -0
@@ -36,6 +36,10 @@ properties:
       - description: Slide right gesture
       - description: Long press gesture
 
+  touchscreen-inverted-x: true
+  touchscreen-inverted-y: true
+  touchscreen-swapped-x-y: true

[Severity: Low]
Looking at the properties block in Documentation/devicetree/bindings/input/touchscreen/hynitron,cst816x.yaml, are these redundant definitions a direct workaround for the improperly configured additionalProperties keyword below?

Documentation/devicetree/bindings/input/touchscreen/hynitron,cst816x.yaml required: · line 51

Quoted 10 lines
Documentation/devicetree/bindings/input/touchscreen/hynitron,cst816x.yaml +3 -0
+
 required:
   - compatible
   - reg
@@ -43,6 +47,9 @@ required:
 
 additionalProperties: false
 
+allOf:
+  - $ref: touchscreen.yaml#

[Severity: Low]
Looking at the allOf block in the same file, does the combination of
additionalProperties: false and a $ref create a constraint collision that restricts validation exclusively to locally defined properties, breaking proper schema inheritance?

If a valid generic touchscreen property (e.g., touchscreen-size-x) not explicitly listed in the local schema is defined, running make dtbs_check will incorrectly reject it. Would changing additionalProperties: false to
unevaluatedProperties: false address this when inheriting from another schema?

Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/d18629ede3f2fd5a08d0cbdaf82f4254e89500ed.1790851590.git.daniel@makrotopia.org?part=1
#3 lore raw

Re: [PATCH 1/4] dt-bindings: input: touchscreen: hynitron,cst816x: configure axes

In reply to Daniel Golle
Quoted 3 lines · Daniel Golle

Reference touchscreen.yaml and allow properties 'touchscreen-inverted-x', 'touchscreen-inverted-y' and 'touchscreen-swapped-x-y'.

This is a useless commit message, it just says what the diff does. Please provide a justification for why this is the right thing to do.
pw-bot: changes-requested

Quoted 41 lines
---
 .../bindings/input/touchscreen/hynitron,cst816x.yaml      | 8 ++++++++
 1 file changed, 8 insertions(+)
Documentation/devicetree/bindings/input/touchscreen/hynitron,cst816x.yaml +8 -0
index 72d4da636881e..ba56bf61038c5 100644
@@ -36,6 +36,10 @@ properties:
       - description: Slide right gesture
       - description: Long press gesture
 
+  touchscreen-inverted-x: true
+  touchscreen-inverted-y: true
+  touchscreen-swapped-x-y: true
+
 required:
   - compatible
   - reg
@@ -43,6 +47,9 @@ required:
 
 additionalProperties: false
 
+allOf:
+  - $ref: touchscreen.yaml#
+
 examples:
   - |
     #include <dt-bindings/gpio/gpio.h>
@@ -59,6 +66,7 @@ examples:
             reset-gpios = <&gpio 17 GPIO_ACTIVE_LOW>;
             linux,keycodes = <KEY_UP>, <KEY_DOWN>, <KEY_LEFT>, <KEY_RIGHT>,
                              <BTN_TOOL_TRIPLETAP>;
+            touchscreen-swapped-x-y;
         };
     };
 
Signature · 1 line
2.55.0
#4 lore raw

PATCH 2/4 Input: hynitron-cst816x: respect touchscreen DT properties

In reply to Daniel Golle

Parse common touchscreen properties and use touchscreen_report_pos() instead of reporting raw absolute position.

This makes the driver respect the touchscreen-inverted-x, touchscreen-inverted-y and touchscreen-swapped-x-y properties

---
 drivers/input/touchscreen/hynitron-cst816x.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
drivers/input/touchscreen/hynitron-cst816x.c +5 -2
index 47d9cd7412d1b..865c5a767ba01 100644
@@ -11,6 +11,7 @@
 #include <linux/gpio/consumer.h>
 #include <linux/i2c.h>
 #include <linux/input.h>
+#include <linux/input/touchscreen.h>
 #include <linux/unaligned.h>
 #include <linux/interrupt.h>
 #include <linux/module.h>
@@ -31,6 +32,7 @@ struct cst816x_priv {
 	struct input_dev *input;
 	unsigned int keycode[CST816X_NUM_KEYS];
 	unsigned int keycodemax;
+	struct touchscreen_properties prop;
 };
 
 static int cst816x_parse_keycodes(struct device *dev, struct cst816x_priv *priv)
@@ -142,6 +144,7 @@ static int cst816x_register_input(struct cst816x_priv *priv)
 	input_set_abs_params(priv->input, ABS_X, 0, 240, 0, 0);
 	input_set_abs_params(priv->input, ABS_Y, 0, 240, 0, 0);
 	input_set_capability(priv->input, EV_KEY, BTN_TOUCH);
+	touchscreen_parse_properties(priv->input, false, &priv->prop);
 
 	priv->input->keycode = priv->keycode;
 	priv->input->keycodesize = sizeof(priv->keycode[0]);
@@ -173,8 +176,8 @@ static irqreturn_t cst816x_irq_cb(int irq, void *cookie)
 	if (!cst816x_process_touch(priv, &tch))
 		return IRQ_HANDLED;
 
-	input_report_abs(priv->input, ABS_X, tch.abs_x);
-	input_report_abs(priv->input, ABS_Y, tch.abs_y);
+	touchscreen_report_pos(priv->input, &priv->prop,
+			       tch.abs_x, tch.abs_y, false);
 
 	if (tch.gest)
 		input_report_key(priv->input,
Signature · 1 line
2.55.0
#5 lore raw

PATCH 3/4 Input: hynitron-cst816x: release gesture keys

In reply to Daniel Golle

The gesture key is reported with the value of the touch flag, so it stays pressed once the controller stops reporting the gesture code while the finger is still down. Report the press on its own and release the keys the input core still holds down on the first report that carries no gesture code.

Fixes c87a819bec86 ("Input: add driver for Hynitron CST816x series")
---
 drivers/input/touchscreen/hynitron-cst816x.c | 17 +++++++++++++++--
 1 file changed, 15 insertions(+), 2 deletions(-)
drivers/input/touchscreen/hynitron-cst816x.c +15 -2
index 865c5a767ba01..e2093518e8af1 100644
@@ -168,6 +168,18 @@ static void cst816x_reset(struct cst816x_priv *priv)
 	msleep(100);
 }
 
+static void cst816x_release_keys(struct cst816x_priv *priv)
+{
+	unsigned int i;
+
+	for (i = 0; i < priv->keycodemax; i++) {
+		if (!test_bit(priv->keycode[i], priv->input->key))
+			continue;
+
+		input_report_key(priv->input, priv->keycode[i], 0);
+	}
+}
+
 static irqreturn_t cst816x_irq_cb(int irq, void *cookie)
 {
 	struct cst816x_priv *priv = cookie;
@@ -181,8 +193,9 @@ static irqreturn_t cst816x_irq_cb(int irq, void *cookie)
 
 	if (tch.gest)
 		input_report_key(priv->input,
-				 priv->keycode[cst816x_gest_idx(tch.gest)],
-				 tch.active);
+				 priv->keycode[cst816x_gest_idx(tch.gest)], 1);
+	else
+		cst816x_release_keys(priv);
 
 	input_report_key(priv->input, BTN_TOUCH, tch.active);
 
Signature · 1 line
2.55.0
#6 lore raw

Re: [PATCH 3/4] Input: hynitron-cst816x: release gesture keys

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider: - [Medium] input: OOB read in cst816x_release_keys() due to unvalidated DT keycodes --

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] input: OOB read in cst816x_release_keys() due to unvalidated DT keycodes --

commit cb8390c95078c2f0098d2be4f937c758b29ad76d
Author: Daniel Golle <[email protected]>

Input: hynitron-cst816x: release gesture keys

The gesture key is reported with the value of the touch flag, so it stays pressed once the controller stops reporting the gesture code while the finger is still down. This commit reports the press on its own and releases the keys the input core still holds down when the first report without a gesture code arrives.

Quoted 4 lines
drivers/input/touchscreen/hynitron-cst816x.c +0 -0
index 865c5a767ba01..e2093518e8af1 100644

[ ... ]

drivers/input/touchscreen/hynitron-cst816x.c static void cst816x_reset(struct cst816x_priv *priv) · line 176

Quoted 10 lines
drivers/input/touchscreen/hynitron-cst816x.c +6 -0
@@ -168,6 +168,18 @@ static void cst816x_reset(struct cst816x_priv *priv)
 	msleep(100);
 }
 
+static void cst816x_release_keys(struct cst816x_priv *priv)
+{
+	unsigned int i;
+
+	for (i = 0; i < priv->keycodemax; i++) {
+		if (!test_bit(priv->keycode[i], priv->input->key))

[Severity: Medium]
Can this read out of bounds on the priv->input->key bitmap?

If a keycode provided in the device tree property exceeds KEY_MAX, input_set_capability() safely ignores it during registration, but the invalid large value remains in the priv->keycode array.

When cst816x_release_keys() iterates over this array, it directly calls test_bit(priv->keycode[i], priv->input->key). Since test_bit() doesn't perform bounds checking like the input core's event handlers do, passing an oversized keycode could result in a memory read past the end of the bitmap structure during touch interrupts.

Quoted 6 lines
drivers/input/touchscreen/hynitron-cst816x.c +6 -0
+			continue;
+
+		input_report_key(priv->input, priv->keycode[i], 0);
+	}
+}
+
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/d18629ede3f2fd5a08d0cbdaf82f4254e89500ed.1790851590.git.daniel@makrotopia.org?part=3
#7 lore raw

PATCH 4/4 Input: hynitron-cst816x: time out gesture key release

In reply to Daniel Golle

The controller reports a gesture code once and can then stay quiet until the next touch, which leaves the gesture key pressed for as long as the panel is left alone. Release the gesture keys 40 ms after the last report that carried a gesture code, so that the release comes from whichever happens first, the timeout or a report without a gesture code. The 40 ms matches TSC200X_PENUP_TIME_MS, chosen there for the same want of a lift event from the controller, and has to stay above the interval between consecutive reports of a held gesture.

---
 drivers/input/touchscreen/hynitron-cst816x.c | 51 ++++++++++++++++++--
 1 file changed, 48 insertions(+), 3 deletions(-)
drivers/input/touchscreen/hynitron-cst816x.c +48 -3
index e2093518e8af1..30f8ea314ed2c 100644
@@ -15,9 +15,12 @@
 #include <linux/unaligned.h>
 #include <linux/interrupt.h>
 #include <linux/module.h>
+#include <linux/spinlock.h>
+#include <linux/timer.h>
 
 #define CST816X_RD_REG		0x01
 #define CST816X_NUM_KEYS	5
+#define CST816X_GEST_HOLD_MS	40
 
 struct cst816x_touch {
 	u8 gest;
@@ -32,6 +35,9 @@ struct cst816x_priv {
 	struct input_dev *input;
 	unsigned int keycode[CST816X_NUM_KEYS];
 	unsigned int keycodemax;
+	struct timer_list release_timer;
+	/* lock keeps the timeout out of the middle of a report */
+	spinlock_t lock;
 	struct touchscreen_properties prop;
 };
 
@@ -168,8 +174,9 @@ static void cst816x_reset(struct cst816x_priv *priv)
 	msleep(100);
 }
 
-static void cst816x_release_keys(struct cst816x_priv *priv)
+static bool cst816x_release_keys(struct cst816x_priv *priv)
 {
+	bool released = false;
 	unsigned int i;
 
 	for (i = 0; i < priv->keycodemax; i++) {
@@ -177,33 +184,65 @@ static void cst816x_release_keys(struct cst816x_priv *priv)
 			continue;
 
 		input_report_key(priv->input, priv->keycode[i], 0);
+		released = true;
 	}
+
+	return released;
+}
+
+static void cst816x_release_timeout(struct timer_list *t)
+{
+	struct cst816x_priv *priv = timer_container_of(priv, t, release_timer);
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	if (cst816x_release_keys(priv))
+		input_sync(priv->input);
+
+	spin_unlock_irqrestore(&priv->lock, flags);
 }
 
 static irqreturn_t cst816x_irq_cb(int irq, void *cookie)
 {
 	struct cst816x_priv *priv = cookie;
 	struct cst816x_touch tch;
+	unsigned long flags;
 
 	if (!cst816x_process_touch(priv, &tch))
 		return IRQ_HANDLED;
 
+	spin_lock_irqsave(&priv->lock, flags);
+
 	touchscreen_report_pos(priv->input, &priv->prop,
 			       tch.abs_x, tch.abs_y, false);
 
-	if (tch.gest)
+	if (tch.gest) {
 		input_report_key(priv->input,
 				 priv->keycode[cst816x_gest_idx(tch.gest)], 1);
-	else
+		mod_timer(&priv->release_timer,
+			  jiffies + msecs_to_jiffies(CST816X_GEST_HOLD_MS));
+	} else {
+		timer_delete(&priv->release_timer);
 		cst816x_release_keys(priv);
+	}
 
 	input_report_key(priv->input, BTN_TOUCH, tch.active);
 
 	input_sync(priv->input);
 
+	spin_unlock_irqrestore(&priv->lock, flags);
+
 	return IRQ_HANDLED;
 }
 
+static void cst816x_timer_stop(void *data)
+{
+	struct cst816x_priv *priv = data;
+
+	timer_delete_sync(&priv->release_timer);
+}
+
 static int cst816x_probe(struct i2c_client *client)
 {
 	struct device *dev = &client->dev;
@@ -215,6 +254,8 @@ static int cst816x_probe(struct i2c_client *client)
 		return -ENOMEM;
 
 	priv->client = client;
+	spin_lock_init(&priv->lock);
+	timer_setup(&priv->release_timer, cst816x_release_timeout, 0);
 
 	priv->reset = devm_gpiod_get_optional(dev, "reset", GPIOD_OUT_HIGH);
 	if (IS_ERR(priv->reset))
@@ -232,6 +273,10 @@ static int cst816x_probe(struct i2c_client *client)
 	if (error)
 		return dev_err_probe(dev, error, "input register failed\n");
 
+	error = devm_add_action_or_reset(dev, cst816x_timer_stop, priv);
+	if (error)
+		return error;
+
 	error = devm_request_threaded_irq(dev, client->irq,
 					  NULL, cst816x_irq_cb, IRQF_ONESHOT,
 					  dev_driver_string(dev), priv);
Signature · 1 line
2.55.0
#8 lore raw

Re: [PATCH 4/4] Input: hynitron-cst816x: time out gesture key release

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider: - [Medium] input: spurious key release due to timer callback race --

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] input: spurious key release due to timer callback race
--

commit 84ec9fce5971ed83d872b86322e5cd6170da5358
Author: Daniel Golle <[email protected]>

Input: hynitron-cst816x: time out gesture key release

This commit adds a 40ms timer to release gesture keys if a new touch event with a gesture code is not reported in time. It helps prevent keys from remaining stuck in the pressed state when the panel stops sending events.

Quoted 4 lines
drivers/input/touchscreen/hynitron-cst816x.c +0 -0
index e2093518e8af1..30f8ea314ed2c 100644

[ ... ]

Quoted 1 line
drivers/input/touchscreen/hynitron-cst816x.c +0 -0
@@ -177,33 +184,65 @@ static void cst816x_release_keys(struct cst816x_priv *priv)

[ ... ]

drivers/input/touchscreen/hynitron-cst816x.c static void cst816x_release_keys(struct cst816x_priv *priv) · line 228

Quoted 12 lines
drivers/input/touchscreen/hynitron-cst816x.c +12 -0
+static void cst816x_release_timeout(struct timer_list *t)
+{
+	struct cst816x_priv *priv = timer_container_of(priv, t, release_timer);
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	if (cst816x_release_keys(priv))
+		input_sync(priv->input);
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+}

[Severity: Medium]
Could this unconditionally release keys even if the timer was just concurrently re-armed by the IRQ handler?

If the timer softirq and the threaded IRQ handler (cst816x_irq_cb) run concurrently on different CPUs, the timer callback might spin waiting for the lock while the IRQ handler is reporting a new key press and re-arming the timer.

Once the IRQ handler drops the lock, the timer callback acquires it and unconditionally executes cst816x_release_keys(), undoing the IRQ handler's action and causing a spurious release of gesture keys.

Would it be safer to check if the timer was re-enqueued (e.g., via timer_pending()) while waiting for the lock before releasing the keys?

Quoted 11 lines

static irqreturn_t cst816x_irq_cb(int irq, void *cookie) {

drivers/input/touchscreen/hynitron-cst816x.c +3 -0
 	struct cst816x_priv *priv = cookie;
 	struct cst816x_touch tch;
+	unsigned long flags;
 
 	if (!cst816x_process_touch(priv, &tch))
 		return IRQ_HANDLED;
 
+	spin_lock_irqsave(&priv->lock, flags);
+

drivers/input/touchscreen/hynitron-cst816x.c static void cst816x_release_keys(struct cst816x_priv *priv) · line 224

Quoted 10 lines
drivers/input/touchscreen/hynitron-cst816x.c +3 -2
 	touchscreen_report_pos(priv->input, &priv->prop,
 			       tch.abs_x, tch.abs_y, false);
 
-	if (tch.gest)
+	if (tch.gest) {
 		input_report_key(priv->input,
 				 priv->keycode[cst816x_gest_idx(tch.gest)], 1);
-	else
+		mod_timer(&priv->release_timer,
+			  jiffies + msecs_to_jiffies(CST816X_GEST_HOLD_MS));

[Severity: Medium]
Does re-arming the timer here race with the timer callback executing on another CPU?

Because the timer callback does not check if the timer was re-armed, the newly reported key press here could be immediately reversed by the racing timer callback once the IRQ handler releases the lock.

Quoted 4 lines
drivers/input/touchscreen/hynitron-cst816x.c +3 -0
+	} else {
+		timer_delete(&priv->release_timer);
 		cst816x_release_keys(priv);
+	}
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/d18629ede3f2fd5a08d0cbdaf82f4254e89500ed.1790851590.git.daniel@makrotopia.org?part=4