Another confirmed freeze on September 24: kernel 7.2.5-100.fc43, NVIDIA open 580.178.04, BIOS 306, i3/X11.
An earlier overnight incident left the laptop unreachable by SSH; Xorg logs show NVIDIA driving the internal panel before that event. This does not prove every CPU stopped.
Freezes regularly happen unattended overnight on a stationary desk. I’ve experienced them across Ubuntu 22.04/24.04/26.04 LTS, Fedora and CachyOS, kernels/BIOS versions, performance profiles, dGPU-only and Dynamic/Hybrid, Xorg and Wayland. They also occur with the original ASUS charger and Lenovo/Dell adapters of different ratings/voltages. Trying modprobe.blacklist=i915 through GRUB did not prevent them. Cable movement sometimes coincided with a freeze, but it is not necessary or a reliable trigger.
The latest CachyOS incident (7.2.7, BIOS 311) stopped the screen clock and local input/VT switching. The journal ends without a nearby panic, Xid, OOM, MCE or NVMe timeout. Performance was the last recorded profile. Brief profile/EPP changes did not reproduce it.
Three archived Fedora boots contain skipped BERT records, but their payloads were lost. A separate MCE followed a clean warm reboot; I am not attributing the freezes to it. Bounded BERT/pstore/RAS and power-event capture is now installed. I have no known-good kernel and no vanilla upstream reproduction.
Which minimal Fedora kernel/debug setup would you prioritize to capture this without keeping the machine artificially busy? Is there an existing Fedora report I should join?
very old kernel + old Fedora release, on a fairly recent platform.
Fedora KDE 44 + kernel 7.2.5 + Intel Xe driver is rock solid on my Intel 255H.
I suffered freezes too on older kernel versions, but they are long gone starting kernel 7.1
You have Nvidia on top of it, but I’m sure it’s the kernel.
Also worth trying the KDE edition, I’ve found it to be more solid than the workstation (Gnome) one.
ty for answering, but as I mentioned, I’ve tried multiple distros, like ubuntu 24/26LTS, Arch based ones (Cachy) with latest kernels, latest LTS and non-LTS as well - and it didn’t help
for example now I use CachyOS with 7.2.7 and still got this issue
consider upgrading to f44 to get the 615.* nvidia drivers. F43 will remain on the 580.* drivers branch and will no longer be supported in late Nov/ early Dec 2026.
Asus has diagnostic tools. If you are in a hot, humid environment, you should consider cleaning and applying contact enhancer (now that cars have computers and bus connectors, contact enhancer is available at auto parts stores).
Thanks, everyone. I’m already on NVIDIA 615.71.09 with kernel 7.2.7 on CachyOS, and it froze again overnight. I haven’t tested Fedora 44, though.
@andym Did your 255H use Xe by default, or did you explicitly switch from i915? Mine currently uses i915.
@gnwiii Which ASUS diagnostic would you recommend for an intermittent idle freeze? ASUS support has asked for a Windows test; I haven’t completed that yet.
Heikki Krogerus also suggested testing with ucsi_acpi unloaded. I haven’t tried it yet and am checking the effects on USB-C charging and the dock first.
No, it was i915 by default. But back then I was experiencing some flickering issues or the system could hang completely once or twice a day, with the image frozen and no longer updating (only power off pressed 5+ seconds helped).
so I started troubleshooting and investigating possible video subsystem issues, and eventually found out that modern Intel devices (like our laptops) are supposed to use more modern Intel XE driver which is specifically designed for them; i915 is obsolete technology.
