https://bugs.kde.org/show_bug.cgi?id=525668

--- Comment #2 from roc <[email protected]> ---
## Additional technical evidence

I would like to add the following low-level observations from an affected
S3/resume cycle.

### Input device identity remains unchanged

Before and after the failure, the ALPS device remains:

AlpsPS/2 ALPS GlidePoint

Kernel input device:

/dev/input/event13

udev:

/devices/platform/i8042/serio1/input/input9/event13

Relevant udev properties:

ID_INPUT=1
ID_INPUT_TOUCHPAD=1
ID_BUS=i8042
ID_PATH=platform-i8042-serio-1

The device does not disappear or get re-enumerated during the failure.

### KWin device state

While clicks are not working, querying KWin gives:

name = AlpsPS/2 ALPS GlidePoint
enabled = true

Therefore the KWin input device is still present and enabled.

### libinput

While the failure is active:

sudo libinput debug-events --device /dev/input/event13

continues to emit normal POINTER_MOTION events.

This means the input device continues producing motion events after resume.

The failure is therefore not a complete loss of the device or pointer motion.

### Cross-device failure

The most important observation is that the failure is not limited to the ALPS
device.

At the same time:

* the internal ALPS touchpad can move the pointer but cannot click
* an external USB mouse can move the pointer but cannot click

Thus pointer-button functionality is lost across multiple pointer devices.

### Suspend/resume sequence

The problem occurs after actual S3/deep suspend:

PM: suspend entry (deep)
ACPI: PM: Low-level resume complete
PM: suspend exit

PowerDevil also reports:

PrepareForSleep(true=prepare)
PrepareForSleep(false=resume)

### KWin/libinput warning

At an affected resume, KWin reports:

Libinput: event13 - AlpsPS/2 ALPS GlidePoint: client bug: event processing
lagging behind by 3158ms, your system is too slow

KWin simultaneously reported similar event-processing lag for the Power Button
devices:

event1 - Power Button: client bug: event processing lagging behind by 1610ms
event3 - Power Button: client bug: event processing lagging behind by 1607ms

### Recovery behavior

The KDE "Touchpad Toggle" action is sufficient to restore pointer clicking
system-wide:

1. Touchpad Off
2. Touchpad On

After this, both the internal touchpad and external USB mouse can click again.

This is notable because toggling the touchpad restores clicking on a separate
USB mouse as well. This suggests that the operation is affecting shared
pointer/input state rather than simply restarting the ALPS device.

### psmouse test

Reloading the ALPS/PS2 kernel driver:

sudo modprobe -r psmouse && sudo modprobe psmouse

does not provide a permanent solution.

The problem can return after subsequent S3 resume or boot.

### Kernel comparison

The issue has been observed on:

linux-cachyos 7.2.4-3

and also on:

linux-cachyos-lts 6.18.50-3

It is much more frequent on 7.2.4-3, but it is not completely absent on
6.18.50-3.

### Current interpretation

The evidence currently points away from:

* disappearance of the ALPS device
* loss of the psmouse/i8042 device
* KWin disabling the touchpad
* complete libinput failure
* an ALPS-only hardware problem

The common failure is pointer-button handling/routing after S3 resume.

The key distinction is:

pointer motion continues
+
multiple pointer devices lose clicking
+
the devices remain present
+
KWin reports the ALPS device enabled=true
+
libinput continues receiving motion
+
KDE Touchpad Off → On restores clicking globally

This makes KWin pointer-button state/event routing after resume a strong area
for investigation, while leaving open the possibility of a lower-level
interaction between resume, libinput and KWin.

I am providing these observations to help narrow the fault domain and avoid
treating this as a conventional "touchpad not detected" problem.

-- 
You are receiving this mail because:
You are watching all bug changes.

Reply via email to