Arch Linux

justm3

Member

Registered: 2019-11-12

Posts: 43

module nct6775 reporting wrong values after resume from suspend

Hi,

I am using a B660 Pro Rs asrock motherboard with i3-12100 cpu.

The module nct6775 is loaded properly during boot and executing sensors reports correct motherboard temps and fans rpm.

However,
resuming from suspend state (sleep.target) the motherboard reports wrong temperature (lower) and likewise the respective fan rmp's (also lower).
Which could potentially lead to an unpleasant warm-up of the box since fans revs are following the wrong motherboard temperature.

dmesg, journalctl and lsmod report exactly the same before or after suspend. Same with or without the fix (see below).
Hence I display them only once.

To be honest I noticed the issue a day or two ago. So I am unaware of the situation with previous kernels.

$ uname -r
6.11.5-arch1-1

$ modinfo nct6775
filename:       /lib/modules/6.11.5-arch1-1/kernel/drivers/hwmon/nct6775.ko.zst
import_ns:      HWMON_NCT6775
license:        GPL
description:    Platform driver for NCT6775F and compatible chips
author:         Guenter Roeck <linux@roeck-us.net>
srcversion:     8AB9668D8D225EC0C9E367E
depends:        nct6775-core,hwmon-vid
retpoline:      Y
intree:         Y
name:           nct6775
vermagic:       6.11.5-arch1-1 SMP preempt mod_unload 
sig_id:         PKCS#7
signer:         Build time autogenerated kernel key
sig_key:        26:5B:4C:74:B5:2A:1C:69:CD:81:4A:65:46:35:E3:E1:71:30:AF:ED
sig_hashalgo:   sha512
signature:      30:66:02:31:00:DC:7C:63:4E:7D:FE:BC:CC:00:4C:67:CC:E2:1F:F7:
		B9:0D:43:0B:91:46:F3:1F:FA:37:2B:15:4C:AE:0C:D3:98:22:8C:73:
		12:37:55:EA:D7:32:71:1E:36:F1:A0:AF:68:02:31:00:CB:40:A1:BD:
		5E:03:49:AE:04:10:3B:84:AD:01:C2:14:4B:41:C9:50:C3:D8:17:AD:
		9A:DE:35:AE:A8:C7:62:28:36:80:17:02:24:D9:7C:1A:06:4B:83:BC:
		2A:38:B4:46
parm:           force_id:Override the detected device ID (ushort)
parm:           fan_debounce:Enable debouncing for fan RPM signal (ushort)

# dmesg | grep nct6775
[    5.256121] nct6775: Found NCT6798D or compatible chip at 0x2e:0x2a0

$ journalctl -b | grep nct6775
Οκτ 24 11:08:28 tarch kernel: nct6775: Found NCT6798D or compatible chip at 0x2e:0x2a0

$ lsmod | grep nct6775
hwmon_vid              12288  1 nct6775
nct6775                40960  0
nct6775_core           81920  1 nct6775

sensors reporting before going to sleep

$ sensors nct6798-isa-02a0
nct6798-isa-02a0
Adapter: ISA adapter
CPU Fan:            1161 RPM  (min =    0 RPM)
Chassis Front Fan1:  471 RPM  (min =    0 RPM)
Chassis Rear Fan:    475 RPM  (min =    0 RPM)
Chassis Front Fan2:  732 RPM  (min =    0 RPM)
Motherboard:         +29.0°C  (high = +80.0°C, hyst = +75.0°C)
                              (crit = +100.0°C)  sensor = thermistor

sensors after resume (without rmmod & modprobe the module)

$ sensors nct6798-isa-02a0
nct6798-isa-02a0
Adapter: ISA adapter
CPU Fan:            1164 RPM  (min =    0 RPM)
Chassis Front Fan1:  390 RPM  (min =    0 RPM)
Chassis Rear Fan:    399 RPM  (min =    0 RPM)
Chassis Front Fan2:  739 RPM  (min =    0 RPM)
Motherboard:         +17.0°C  (high = +80.0°C, hyst = +75.0°C)
                              (crit = +100.0°C)  sensor = thermistor

The instruction to go to sleep and resume differ by 10-15 seconds to avoid natural temp drop.

Fix
I am able to bypass the issue by scripting and enabling a systemd-service that removes(rmmod) and then probes(modprobe) the said module nct6775 during systemctl suspend.
Which makes my life a lot easier and conky refreshes values after resume (using openbox).

The module nct6775 loads automatically but I tried it manually and it messed up. No-go.

My questions are,
Why is this happening and the sensor needs to be probed every time the machine goes to & recovers from sleep?
Is it a bug (known)?
Should it be reported upsteam?

TIA

Last edited by justm3 (2024-10-24 09:52:07)

gromit

Package Maintainer (PM)

From: Germany

Registered: 2024-02-10

Posts: 784

Website

Re: module nct6775 reporting wrong values after resume from suspend

Could you try to test if this is an issue with other kernels aswell? You could for example downgrade the linux package by a few versions, check out linux-lts or similar.

It would also be interesting to see if the behaviour is the same for the latest mainline release:

sudo pacman -U https://pkgbuild.com/\~gromit/linux-bisection-kernels/linux-mainline-6.12rc4-1-x86_64.pkg.tar.zst

justm3

Member

Registered: 2019-11-12

Posts: 43

Re: module nct6775 reporting wrong values after resume from suspend

Thank you @gromit

Just tried both suggestions lts & mainline.

Unfortunately same issue.
The sensor has to be re-probed every time the machine goes to sleep.

As I said, a home-brewed service does the trick.
But the question remains.

gromit

Package Maintainer (PM)

From: Germany

Registered: 2024-02-10

Posts: 784

Website

Re: module nct6775 reporting wrong values after resume from suspend

I just asked about this because that would have made it a regression

I think you can then just follow the procedure described in https://docs.kernel.org/admin-guide/rep … ssues.html

$ ./scripts/get_maintainer.pl  drivers/hwmon/nct6775-core.c 
Jean Delvare <jdelvare@suse.com> (maintainer:HARDWARE MONITORING)
Guenter Roeck <linux@roeck-us.net> (maintainer:HARDWARE MONITORING)
linux-hwmon@vger.kernel.org (open list:HARDWARE MONITORING)
linux-kernel@vger.kernel.org (open list)

justm3

Member

Registered: 2019-11-12

Posts: 43

Re: module nct6775 reporting wrong values after resume from suspend

Well, pleased I can contribute.

Briefly reading the link you mentioned, it's quite a work to follow up.
TGIF

Take care.

gromit

Package Maintainer (PM)

From: Germany

Registered: 2024-02-10

Posts: 784

Website

Re: module nct6775 reporting wrong values after resume from suspend

It would also be cool if you could post the bugreport here once you have sent it so people can follow along