bug: kernel regression in 6.12 removes the charging threshold

[okarin69]

Subject of the issue

With the arrival linux kernel 6.12, in my asus laptop the /sys/class/power_supply/BAT_0/charge_control_end_threshold file has disappered and charging threshold does not work. But when I switch to linux-lts kernel it reappers and the charging threshold works properly.

Hope this information is helpful.

Your system

For example,

OS: Endeavour OS x86_64
Host: Asus Vivobook Pro 16 2022
Kernel: Linux 6.12.1

Version

1.0 (Latest)

Steps to reproduce

Switch to latest stable Kernel 6.12.1 or newer. => Error: Charge threshold file does not exist

Going back to linux- 6.6.63-lts => Working properly

Expected behaviour

No response

Actual behaviour

No response

[tshakalekholoane]

I cannot find anything to suggest this caused by a change in the kernel. The code that handles this hasn't been updated for half a decade. (Not withstanding any changes made upstream by your distribution).

Can you post the exact error you are getting from running the program?

(Also, the threshold is at one of /sys/class/power_supply/BAT{0,1,C,T} not /sys/class/power_supply/BAT_0/charge_control_end_threshold so that could also be the source of your problem).

[pepa65]

What is the output of ls /sys/class/power_supply/ and ls /sys/class/power_supply/BAT0 on the problematic kernel?

It could be that somehow this was not included in the built of that particular kernel, but the output would shed some light.

[okarin69]

Output for ls /sys/class/power_supply/

BAT0

Output for ls /sys/class/power_supply/BAT0

alarm           charge_full_design  device        power          subsystem   voltage_min_design
capacity        charge_now          hwmon1        present        technology  voltage_now
capacity_level  current_now         manufacturer  serial_number  type
charge_full     cycle_count         model_name    status         uevent
 

The above output is got linux 6.12.1.
The charge_control_end_threshold file is present in LTS kernel 6.6.63. Everything was working properly until linux-6.11.9.

The output when sudo ./bat threshold 80 is run is Charging threshold setting not found. again in linux 6.12.1

[pepa65]

I would make inquiries of the people who build the kernel. Unfortunately, there is nothing this project can do about it...

[okarin69]

Yes I understand, and you guys are very well suited to know what's going on in the kernel or if there is a change in the process. Hope the word gets out and the problem is solved.

[pepa65]

Do you have something like /boot/config-6.12.1 on your system? You could check the output of grep ASUS /boot/config-6.12.1 and lsmod |grep asus.

[okarin69]

grep ASUS /boot/config-6.12.1
Output: grep: /boot/config-6.12.1: No such file or directory

lsmod |grep asus
Output:

asus_nb_wmi            28672  0
asus_wmi               98304  1 asus_nb_wmi
ledtrig_audio          12288  3 snd_ctl_led,snd_hda_codec_generic,asus_wmi
sparse_keymap          12288  1 asus_wmi
platform_profile       12288  1 asus_wmi
rfkill                 40960  8 asus_wmi,bluetooth,cfg80211
i8042                  57344  1 asus_wmi
video                  77824  3 asus_wmi,i915,nvidia_modeset
wmi                    45056  3 video,asus_wmi,wmi_bmof

[pepa65]

At least we know the module is loaded..! I should maybe have asked for ls /boot first to see in a config is available there.

[okarin69]

Here is the output for ls /boot

drwx------    - root  1 Jan  1970  efi
drwxr-xr-x    - root 25 Nov 05:59  grub
.rw------- 252M root 28 Nov 01:34  initramfs-linux-fallback.img
.rw------- 204M root 28 Nov 01:34  initramfs-linux-lts-fallback.img
.rw------- 131M root 28 Nov 01:34  initramfs-linux-lts.img
.rw------- 175M root 28 Nov 01:34  initramfs-linux.img
.rw-r--r-- 8.1M root 12 Nov 22:50  intel-ucode.img
.rw-r--r--  14M root 24 Nov 08:56  vmlinuz-linux
.rw-r--r--  13M root 25 Nov 04:15  vmlinuz-linux-lts

[pepa65]

OK, so no config information in that directory. I think we would be looking for entries CONFIG_ASUS_NB_WMI (maybe CONFIG_ASUS_LAPTOP and CONFIG_SENSORS_ASUS_WMI).

[okarin69]

How do I look up this information?
CONFIG_ASUS_NB_WMI (maybe CONFIG_ASUS_LAPTOP and CONFIG_SENSORS_ASUS_WMI).

[pepa65]

Endeavour OS is Arch, right? Your kernel config might be in /proc/config.gz, try ls /proc/config.gz to see if it's there. if it is, you can do zgrep ASUS /proc/config.gz (if you have zgrep, which most likely you do).

[okarin69]

Here's the output:

CONFIG_USB_PEGASUS=m
CONFIG_TABLET_USB_PEGASUS=m
CONFIG_SENSORS_ASUS_WMI=m
CONFIG_SENSORS_ASUS_EC=m
CONFIG_HID_ASUS=m
CONFIG_ASUS_LAPTOP=m
CONFIG_ASUS_WIRELESS=m
CONFIG_ASUS_WMI=m
CONFIG_ASUS_NB_WMI=m
CONFIG_ASUS_TF103C_DOCK=m

[pepa65]

This looks like a similar problem: https://bugzilla.kernel.org/show_bug.cgi?id=219075

[Biasio]

Same here, for me also bluetooth controller stopped working. I think it's something related to the last kernel

[pepa65]

@Biasio What distro and kernel are you on?

[pepa65]

The issue has already been reported and a fix been found:
https://bugzilla.kernel.org/show_bug.cgi?id=219517

[okarin69]

@pepa65 Does this fix reflect in the kernel in later stages? In linux-6.12.1 I am still facing the issue.

[pepa65]

I trust it will reflect in future kernels. But this was already a regression, so who can tell...

[okarin69]

Since the program is working correctly and the problem is with the kernel, I am closing this issue.

What I have found is that the asus_nb_wmi driver is not working properly. Hope we get a fix soon. Until then LTS kernel it is.

[tshakalekholoane]

Let's keep it open for now to avoid duplicate bug reports (see #106).