LTS Kernel update places block on wifi

General Kernel
1

Upgrading from linux-cachyos-lts-6.12.69-3 → 6.18.10-2 : Caused me to lose the ability to power on my wireless card.

My system (working side):

Kernel: 6.12.69-3-cachyos-lts arch: x86_64 bits: 64
Machine:
Type: Laptop System: Micro-Star product: Vector A16 HX A8WHG v: REV:1.0
Mobo: Micro-Star model: MS-15MM v: REV:1.0 serial:
Firmware: UEFI vendor: American Megatrends LLC. v: E15MMAMS.106
date: 11/10/2025
CPU:
Info: 16-core model: AMD Ryzen 9 8940HX with Radeon Graphics bits: 64
type: MT MCP cache: L2: 16 MiB
Speed (MHz): avg: 1491 min/max: 400/5385 cores: 1: 1491 2: 1491 3: 1491
4: 1491 5: 1491 6: 1491 7: 1491 8: 1491 9: 1491 10: 1491 11: 1491 12: 1491
13: 1491 14: 1491 15: 1491 16: 1491 17: 1491 18: 1491 19: 1491 20: 1491
21: 1491 22: 1491 23: 1491 24: 1491 25: 1491 26: 1491 27: 1491 28: 1491
29: 1491 30: 1491 31: 1491 32: 1491
Graphics:
Device-1: NVIDIA GB205M [GeForce RTX 5070 Ti Mobile] driver: nvidia
v: 590.48.01
Device-2: Advanced Micro Devices [AMD/ATI] Raphael driver: amdgpu
v: kernel
…
Network:
Device-1: Realtek RTL8125 2.5GbE driver: r8169
IF: enp4s0 state: down mac:
Device-2: MEDIATEK MT7925 802.11be 160MHz 2x2 PCIe Wireless Network
Adapter [Filogic 360] driver: mt7925e
IF: wlan0 state: up mac:
Bluetooth:
Device-1: Foxconn / Hon Hai Wireless_Device driver: btusb type: USB
Report: btmgmt ID: hci0 state: up address: bt-v: 5.4

Output of rfkill –output-all on new breaking kernel looks like this:
❯ rfkill list --output-all
0: hci0: Bluetooth
Soft blocked: no
Hard blocked: no
1: phy0: Wireless LAN
Soft blocked: no
Hard blocked: yes

I tried unplugging and holding the power for 30s to reset the state of the card, rmmod mt7925e and then modprobe mt7925e didn’t fix either. Have tried to turn it on with both network manager and iwctl, have restarted networkmanger.service and wpa_supplicant.service to no avail.

100% of the time I’ve been blocked on latest LTS and on cachyos-znver4/linux-cachyos 6.19.0-2 kernels, despite multiple reboots and unplug attempts trying to see if it was hardware state issues. However, it works 100% of the time on the kernel above. Is this a regression of some kind?

Note:
I see in the following in the starup logs of the non-working kernels:
❯ journalctl -b -1 | grep rfkill
Feb 13 22:01:20 vector systemd[1]: Listening on Load/Save RF Kill Switch Status /dev/rfkill Watch.
Feb 13 22:01:22 vector NetworkManager[1178]: [1771038082.3645] rfkill1: found Wi-Fi radio killswitch (at /sys/devices/pci0000:00/0000:00:03.2/0000:05:00.0/ieee80211/phy0/rfkill1) (driver mt7925e)
Feb 13 22:01:22 vector NetworkManager[1178]: [1771038082.3648] manager[0x557b4b2bd990]: rfkill: Wi-Fi hardware radio set enabled
Feb 13 22:01:22 vector NetworkManager[1178]: [1771038082.3648] manager[0x557b4b2bd990]: rfkill: WWAN hardware radio set enabled
Feb 13 22:01:22 vector NetworkManager[1178]: [1771038082.3853] manager: rfkill: Wi-Fi enabled by radio killswitch; enabled by state file
Feb 13 22:01:22 vector NetworkManager[1178]: [1771038082.3853] manager: rfkill: WWAN enabled by radio killswitch; enabled by state file
Feb 13 22:01:22 vector NetworkManager[1178]: [1771038082.6496] manager: rfkill: Wi-Fi now disabled by radio killswitch
Feb 13 22:01:27 vector systemd[1]: systemd-rfkill.service: Deactivated successfully.

But on the working system I see:
Feb 13 22:08:36 vector systemd[1]: Listening on Load/Save RF Kill Switch Status /dev/rfkill Watch.
Feb 13 22:08:38 vector NetworkManager[1180]: [1771038518.4131] rfkill1: found Wi-Fi radio killswitch (at /sys/devices/pci0000:00/0000:00:03.2/0000:05:00.0/ieee80211/phy0/rfkill1) (driver mt7925e)
Feb 13 22:08:38 vector NetworkManager[1180]: [1771038518.4134] manager[0x64ae5865b9c0]: rfkill: Wi-Fi hardware radio set enabled
Feb 13 22:08:38 vector NetworkManager[1180]: [1771038518.4134] manager[0x64ae5865b9c0]: rfkill: WWAN hardware radio set enabled
Feb 13 22:08:38 vector NetworkManager[1180]: [1771038518.4354] manager: rfkill: Wi-Fi enabled by radio killswitch; enabled by state file
Feb 13 22:08:38 vector NetworkManager[1180]: [1771038518.4355] manager: rfkill: WWAN enabled by radio killswitch; enabled by state file
Feb 13 22:08:43 vector systemd[1]: systemd-rfkill.service: Deactivated successfully.

Thanks, any help welcome.

2

For me, it was VirtualBox. I uninstalled and rebooted, and it self-resolved. It has something to do with how it requires specific kernel module(s) - I’m unable to speak to the details, but this comment led me to believe I can try pacman -Rs on everything virtualbox related.

I don’t need VirtualBox, but of course, this isn’t the best experience. I just ran through the normal Cachy Update as I do once a week. I saw an error thrown during the upgrade, but I assumed it wouldn’t break anything and didn’t pay much attention. I assumed it would gracefully rollback anything that ‘breaks stuff’ - whoops on my part for sure.

3

Yeah, there were no errors on my update, cachyos (non-lts) already didn’t work in the same way for me, but lts did on the 6.12 version of the kernel with the same version of firmware. Something changed in kernel or in rfkill.service that hard blocks this.
I at first thought it was the standard reboot / drain power on the EC and it will be fixed (this laptop has no hotkey / button to hardware enable network) but no dice multiple times.
It’s also been reported in fedora issues by someone else, but that guy bought a dongle (which I have also done) and there are no updates.