[BUG]

[Matzorator]

Laptop model

MSI Katana 17 B13VGK-825

Confirmation

• I'm confirming that EC version is supported in the supported devices list or in my fork. In other case use Support Request template

EC firmware version

17L5EMS1.115

Distro name and version

Fedora 43 6.18.13

Kernel version

6.18.13-200.fc43

Description

Since upgrading from kernel 6.18.12 to 6.18.13 on Fedora 43, the CPU fan
on my MSI Katana 17 runs constantly at ~2500 RPM and does not respond to
temperature changes anymore.

On kernel 6.18.12 the fan works normally (quiet at idle, speeds up under load).

Hardware: MSI Katana 17
Affected kernel: 6.18.13-200.fc43.x86_64
Working kernel: 6.18.12-200.fc43.x86_64

Relevant dmesg output:
module msi_ec: .gnu.linkonce.this_module section size must match the
kernel's built struct module size at run time

[Matzorator]

Base Board Information
Manufacturer: Micro-Star International Co., Ltd.
Product Name: MS-17L5
Version: REV:1.0
Serial Number: BSS-0123456789
Asset Tag: Default string

[Matzorator]

acpidump.zip

[Matzorator]

Hi, I attached the acpidump from my MSI Katana 17 (firmware 17L5EMS1.115).
Could you add support for this board? The fan runs constantly at ~2500 RPM
without working fan control.

Thanks in advance!

[glpnk]

Please attach EC dump

[glpnk]

How you installed module? DKMS?

[Matzorator]

Yes, installed via DKMS.

ec_sys module not found in kernel 6.18.12, cannot access /sys/kernel/debug/ec/ec0/io.

I already attached the acpidump in my previous comment. Is that sufficient or do you need something else?

[Matzorator]

Additional info: The fan does respond to high CPU load (RPM increases above 70°C),
but at idle it stays constantly at ~2553 RPM and does not spin down,
even at low temperatures around 52°C.

[Matzorator]

and another Additional fan behavior observations:

Fan turns ON when:

• CPU spike > 70°C or GPU reaches ~53°C sustained

Fan turns OFF when:

• GPU drops below 47°C

The fan does not respond to short CPU spikes (e.g. 82°C for 1-2 seconds),
only to sustained temperatures. GPU temperature seems to be the primary
sensor for the off-threshold.

[glpnk]

No EC dump no help. I can assume that your fan curve is in weird state but can't prove this. Try to reset EC if you don't know how to make dump

[Matzorator]

I already found the solution - a cold shutdown (poweroff, not reboot)
reset the EC and the fan works normally again.

Regarding the EC dump: ec_sys module is not available in kernel 6.18.12
on Fedora 43, so I cannot access /sys/kernel/debug/ec/ec0/io.

Is there another way to create the EC dump without ec_sys?

[glpnk]

https://github.com/BeardOverflow/msi-ec/blob/main/docs/device_support_guide.md#msi-ec-debug-mode

Paste it into comment, not in the file please