You are not logged in.
Started using Arch yesterday, coming from Windows 11. I'm familiar with Ubuntu servers, but still learning Arch and how Linux works as a Desktop OS. My system is fully updated as of this post.
Whenever I scroll anything with my mouse, a web page, document, gui terminal, between 5 and 6 lines (at default scrolling speed magnitude) are scrolled evenly between each *click* of the scroll wheel on my mouse. It seems that my mouse has great precision on the scroll wheel, but I do not want this behavior. I would much prefer that each click is it's own distinct scroll event with nothing in between, like how it works on Windows. I've tested a few different DE's to make sure this problem wasn't just a Plasma on Wayland thing, it happens on all of the following:
* Plasma (Wayland)
* Plasma (X)
* Hyprland (Wayland)
* Gnome (Wayland)
* Gnome (X)
From the searching I did, it seems like `libinput` is to blame here, but I can't find any config options that help with this specific issue. Turning down scrolling speed doesn't help, as that just decreases the speed in general, not the weird events between each scroll *click*.
Last edited by GalaxyAngel69 (2025-09-03 11:12:20)
Offline
Offline
Thanks for the pointer! After doing some digging from there (didn't know what quirks were for a bit), I found this guide to disable it for my mouse specifically
https://gist.github.com/djalexkidd/b004 … a95a72dd4e
But it seems that as of August 2nd, there's a bug with libinput that prevents hi-res scrolling from being disabled for my mouse.
https://gitlab.freedesktop.org/libinput … ssues/1160
It's apparently been fixed, but the new version isn't out, so I'm gonna try building libinput and I'll report back.
Offline
SOLUTION (sort of)
I was able to kill the kernel module from loading entirely by adding the file
/etc/modprobe.d/hires_sux.confcontaining
install hid_logitech_hidpp /usr/bin/trueThis was the only fix that actually worked for me, out of all the things I've found today. It seems like the support for Logitech mice are a bit messy at the moment, this doesn't seem like this should be the way to fix this.
Offline
Offline
I've personally been hit by this and have asked the maintainer to allow a module parameter to just disable the hires enablement but he's rather adamant that he wants bug reports and exclusion lists for devices where it doesn't work...
Note that you will lose some other useful functionality like battery reporting by blacklisting the module.
FWIW in my case the libinput quirks file works fine... even with current libinput, weird, but I have a GS903... but this highres code in the kernel module has been very brittle ever since it got introduced
for reference: https://bugzilla.kernel.org/show_bug.cgi?id=216885 https://bugzilla.kernel.org/show_bug.cgi?id=216997
Last edited by V1del (2025-09-03 13:38:46)
Offline
he's rather adamant that he wants bug reports and exclusion lists for devices where it doesn't work
would be against adding any module parameter to work-around driver bugs
These are not the same things.
If there's a bug in the driver, fix it.
But if this relates to the necessity of black/white/graylisting because of weird issues w/ random mouse firmware iterations, you're not gonna "fix" anything - you maintain quirks
In that case link him https://en.wikipedia.org/wiki/False_dilemma - there's nothing wrong with having *both* and that's super-common in the kernel.
One could even make the quirk print a super-annoying warning with a link to where report this.
Offline
FWIW if the history I've been involved with is an indication so far it seems only the G502 and my G903 seem to be this notably affected by this. so adding an additional quirk for the G503 might not be beyond measure. But I'll try to ping him again
Offline