Index | Thread | Search

From:
Chris Narkiewicz <hello@ezaquarii.com>
Subject:
Re: imt touchpad scaling and acceleration
To:
tech@openbsd.org
Date:
Sun, 1 Jun 2025 13:05:05 +0100

Download raw body.

Thread
On Sun, Jun 01, 2025 at 02:17:03AM +0200, Thomas Dettbarn wrote:
> In it, there is a function hidmt_get_resolution(). (Roughly line 68).
> 
> If you want to edit the kernel sources, you could add some printf() to see
> the values of
> h->logical_maximum/h->logical_minimum and
> h->physical_maximum/h->physical_minimum
> You might see those values 1162 and 634 somewhere.

I'm happy about the resolution, but I'm unhappy about acceleration.
The cursor moves very slowly initially, until a threshold is crossed,
when it violently accelerates, overshooting whatever control I'm aiming
for.

I'm going to experiment with that piece a bit.

> BUT I think that you should look into the options provided by
> xinput/xrandr/xset first.

I tired to alter acceleration using xinput, but it seems to have
no effect. Accel profile is flat (1:1, no accel), acceleration is 1x,
I also disabled velicity scaling.

Device '/dev/wsmouse1':
	Device Enabled (183):	1
	Coordinate Transformation Matrix (184):	1.000000, 0.000000, 0.000000, 0.000000, 1.000000, 0.000000, 0.000000, 0.000000, 1.000000
	Device Accel Profile (300):	0
	Device Accel Constant Deceleration (301):	1.000000
	Device Accel Adaptive Deceleration (302):	1.000000
	Device Accel Velocity Scaling (303):	1.000000

I also tried to alter mouse1 props (this is my touchpad). Even some exagerated
values yield no effect:

mouse1.type=touchpad
mouse1.rawmode=0  (tried 1 as well)
mouse1.scale=0,3840,0,2160,0,1,1  (altered resolution and accel thresholds)
mouse1.reverse_scrolling=0
mouse1.scaling=4.000    (bumped scaling 4x)

The only parameter that has some impact is mouse.scaling (no number after mouse),
but this doesn't impact the acceleration THRESHOLD which I'm after. It correctly
alters cursor movement after accel kicks in.

Thank you for suggestions so far. I'm going to do more experiments.


Best regards,
Chris Narkiewicz