Bug 85116

Created attachment 107965 [details]
dmesg output after hang + "sudo chvt 1"

This is a cross-post of this bug that I just logged on the kernel bugtracker: https://bugzilla.kernel.org/show_bug.cgi?id=86451

At first I wasn't sure whether the hang had anything to do with the graphics chip so I posted it to the kernel bugtracker, but after seeing the dmesg output it seems more likely that intel DRM might be acting up here. I'll reattach the dmesg output here for posterity.

Here's the original message of the kernel bug I logged (of course a bit out of date since I did get some output in dmesg in the end):

First of: I have no idea if "Process Management" is the right topic to put this under, since I don't know what's causing this. I'm also having a hard time debugging this as the kernel doesn't print out anything that would signify that it knows something is wrong. Please change it if need be.

I recently upgraded from 3.15.8 to 3.17.1 and found that everything works fine, except that after a while, the system partially freezes.

I say partially because I can still connect with ssh and do a lot of things, such as edit text, run htop, reboot, ...

However, the input devices (such as the keyboard and the mouse) don't respond anymore, and the screen appears frozen as well. Sometimes the mouse still responds (and the cursor moves across the screen) but the keyboard presumably doesn't.

Neither dmesg nor Xorg.0.log show anything out of the ordinary (no errors at all).

I couldn't use my keyboard to perform ctrl+alt+f1 to switch to the VT and try something there because the keyboard is unresponsive. So I thought I was being clever and logged in via ssh then executed "sudo chvt 1". This seemed to hang for quite a long time (a few minutes at least) and to my surprise did finish after a while. However, there was no VT greeting me, the screen stayed unresponsive.

How do I proceed to debug this issue?