FS#80143 - [linux] Keyboard backlight breakage
Attached to Project:
Arch Linux
Opened by Polarian (Polarian) - Thursday, 02 November 2023, 19:22 GMT
Last edited by Toolybird (Toolybird) - Wednesday, 22 November 2023, 01:25 GMT
Opened by Polarian (Polarian) - Thursday, 02 November 2023, 19:22 GMT
Last edited by Toolybird (Toolybird) - Wednesday, 22 November 2023, 01:25 GMT
|
Details
Description:
Laptops with Clevo keyboards (other keyboards have not been tested, I do not have the hardware to test them) will have their backlight permanently enabled and can not be disabled, nor can the brightness be changed. Possible reasons could be a breakage of the detection of the keyboard shortcuts provided by the hardware, or simply linux no longer detecting the hardware properly. Clevo is used by Tuxedo Computers, which are reasonably common, I am using a custom build from NovaCustom which also use Clevo as their ODM. I am yet to check https://github.com/tuxedocomputers/tuxedo-keyboard which provides a kernel module for the controlling of the keyboard, this might fix the issue (NovaCustom has their own fork as well: https://github.com/wessel-novacustom/clevo-keyboard/). I should check this BEFORE submitting an issue however controls worked just fine in 6.5.9.arch2-1 (I downgraded the kernel version again and verified this is the case), and also 6.6.arch1-1 has been signed off, so I wanted to make this potential issue known before it is merged into the repository and those with Clevo keyboards could have a bad time. Additional info: * package version: 6.6.arch1-1 Steps to reproduce: - Update your kernel to 6.6.arch1-1, which is within [core-testing] - Boot the new kernel - Keyboard backlight will be permanently enabled, can't be disabled via keybinds, nor does brightness controls work. |
This task depends upon
Closed by Toolybird (Toolybird)
Wednesday, 22 November 2023, 01:25 GMT
Reason for closing: Upstream
Additional comments about closing: If still an issue, please report upstream. You can always ask for bisection/compilation help in the forum.
Wednesday, 22 November 2023, 01:25 GMT
Reason for closing: Upstream
Additional comments about closing: If still an issue, please report upstream. You can always ask for bisection/compilation help in the forum.
"Dasharo coreboot firmware laptop users don't need to and should not use this application. This also applies for users of other coreboot firmware distributions. The keyboard backlight control is already included in the firmware for that firmware version."
This would not fix my situation, the firmware already controls the hardware here, so it does appear 6.6 has breaking changes. I could contact NovaCustom and ask them about the situation if needed? Because this might be an edge case only effecting NovaCustom laptops.
If someone has a Tuxedo laptop it would be useful to know if they encounter the same issue (as it has the same ODM and similar hardware).
Thank you,
Polarian
Therefore it sounds like an upstream kernel regression. General debugging advice [1]. It's a lot of work...but git-bisect will identify the culprit (offending commit) which you can then report upstream. Please let us know what you find out.
> so I wanted to make this potential issue known before it is merged into the repository
Thanks, but a specific keyboard backlight problem, while annoying, is not going to hold up a new kernel when such a minuscule percentage of the userbase is likely to be impacted.
[1] https://wiki.archlinux.org/title/Kernel#Debugging_regressions
It was more to "make it known", others weren't having issues :)
> Therefore it sounds like an upstream kernel regression. General debugging advice [1]. It's a lot of work...but git-bisect will identify the culprit (offending commit) which you can then report upstream. Please let us know what you find out.
Never reported issues to Linux before... this is bound to be fun.
Thanks for the advice,
Polarian
I have tested `6.6.1-arch1-1` for this issue, and it does appear to have been a regression which has been fixed. However the fix for it now breaks backlight so that it is always off.
I haven't had the time to find the offending commit, and probably will not find the time, so I will wait until `6.6.2` to release and see if someone else has picked up on the issue, hopefully by then I will have time to find the offending commit myself if someone else has not done already.
Thank you,
Polarian