A bug in the AMDGPU Linux kernel graphics driver leading to some laptop displays freezing after periods of use may finally be close to being resolved. Given the length and quantity of bug reports and one of the problematic commits being tracked back to 2017, it's a heavy hitting issue for some Linux users. With the help of Claude Code, it looks like a fix is on the way to the Linux kernel.

A Phoronix reader who has a Lenovo ThinkPad T14 Gen1 AMD laptop where its display will "freeze about once per week of use" wrote in that this years-old problem looks about to be resolved. Laptop displays freezing with AMD Radeon graphics has been a known issue for some select laptops, especially after long-term use/uptime and one of the workarounds to now was disabling the Panel Self Refresh (PSR) power-saving feature.

The active bug ticket tracking this issue has been "7840U/780M: `*ERROR* [CRTC:83:crtc-1] flip_done timed out`, leads to frozen display and need to hard reboot". That report is with a Framework Laptop 13 using an AMD Ryzen 7 7840U SoC but, again, it happens for different laptops and AMD Ryzen SoCs. In that report the issue is summed up as:

"Machine is a Framework 13 7840U running Fedora with kernel 6.13.11-200.fc41.x86_64, but note this has also been reported on a Thinkpad with the same APU, starting at least under kernel-6.13.10-200.fc41.

After about 10 hours of runtime which included periods of s2idle sleep, the internal display stopped updating. The external monitor however kept responding at least for a few minutes while I was trying to diagnose, then later also became unresponsive after I tried to disable the laptop's internal display under GNOME settings. The idea was to toggle that off/on to see if that worked around."

There have similarly been other duplicate bug reports around laptop display freezes with Radeon graphics.

For those going through that very lengthy bug report, as of yesterday are some new patches around consolidating DCN vblank/page-flip handling and other display code changes. After previous prior fixes didn't pan out, it appears these latest patches should fix the issue. The patches have yet to be upstreamed to the mainline Linux kernel but if all goes well hopefully these fixes will be upstreamed soon.

Making this long-running display issue all the more intriguing is that getting the work over the finish line involved some "vibe debugging" with Claude Code as well as Claude assisting in the patches.