Greetings Salvatore, greetings. Thanks for your reply.
Salvatore Bonaccorso - 29.07.26, 17:13:35 CEST: > On Tue, Jul 28, 2026 at 11:51:03PM +0200, Martin Steigerwald wrote: > > I see this issue on upstream vanilla kernel, self-compiled, on a > > ThinkPad T14 AMD Gen 5. So I think I can confirm that this is an > > upstream kernel issue. […] > Is this issue still present as well with 7.1.5? Sorry I forgot to mention. Yes it is. Running 7.1.5 currently. > Can you additionally test 7.2~rc5-1~exp1 in experimental to verify if > it affects still mainline? I could look into it. Given a mixture of stability issues with the Linux kernel on various laptops I am a bit hesitant about release candidate kernels. Also this is not the only issue I am having with Linux on this laptop. (But the other issues have been happening before 7.1 already and are unrelated.) Have been using them before, but my general impression is that the general kernel quality deteriorated in the last years. So I usually wait for a vanilla release or even a .1 or .2 one. So maybe I wait a bit longer before I test 7.2. > If you reliably can reproduce the problem and assert the problem then > it would be helpful if you can bisect the problem (between 7.0.y and > 7.1.y). As I explained, I cannot reproduce the issue with absolute reliability. Sometimes shutdown and reboot does work. Most of the times it does not, but sometimes it does. Honestly written I do not feel like investing both the time and the risk to do a bisect on my main laptop. A bisect can lead into a state between rc1 and older release kernel. Filesystems, NVME and block layer could be in whatever unknown state and this laptop holds 3,5 TB of data. Which is backed up of course, but the effort to play back the backup from hard disk… I bet you get the idea. Also I may need multiple attempts to be certain whether the issue is actually really gone. This is the main reason I did not report this with upstream kernel community. They almost always ask for a bisect. Which can be very labor and time intensive way to identify an issue and can be risky as well. I know cause I did a bisect at least once. I wished the kernel could be more helpful instead of just not working correctly without any helpful error message. But usually it is just that: Some kind of crash without indication as to why. I am sorry if this sounds like a bit of a resignation. But actually it is. I am very surprised that this issue has not been caught before 7.1 release. This bug is happening on main stream hardware. So I am sorry but I think I cannot be of much more help here. I may try 7.2 before vanilla release. Until then I just use the work-around to forcefully switch off the laptop and be done with it I bet. (BTW I do not want to turn this into a discussion, but I wanted to justify declining your request for a bisect.) Best Regards, -- Martin
