How can I diagnose system getting non responsive on Debian/Ubuntu/LM?

[Levent]

I am having a issue where when I am playing rimworld (or any other game for that matter, FTL being the easiest to replicate the issue with) whole computer locks up and will not response to anything but a hard shutdown. I have checked dmesg syslog and I cannot find anything. Logs simply stop and start showing logs of system booting up. I have tested CPU, RAM and GPU both individually and all at once for total of 4 days. System is stable no errors whatsoever. This issue does not happen on Windows 10. Clean install of LM20 also doesnt fix it. So any ideas on further troubleshooting?

 

Device is a shitty HP 15 laptop that I got brand new almost 2 years ago.

N3060, 4G ram, 850 Evo. I am running LM20 cinnamon with latest LTS kernel, latest bios.

[Nayr438]

Intel Graphics are known to be somewhat problematic with 3D acceleration, with the only workaround on the Arch Wiki being to disable 3D acceleration...

 

How responsive is the system normally when working with other applications like gimp/krita, watching youtube, working with libreoffice, etc...

Does it regularly peg out a core?

 

You said you checked dmesg, what did you check?

• dmesg | grep i915
• dmesg | grep error

[Levent]

26 minutes ago, Nayr438 said:

Intel Graphics are known to be somewhat problematic with 3D acceleration, with the only workaround on the Arch Wiki being to disable 3D acceleration...

 

How responsive is the system normally when working with other applications like gimp/krita, watching youtube, working with libreoffice, etc...

Does it regularly peg out a core?

 

You said you checked dmesg, what did you check?

• dmesg | grep i915
• dmesg | grep error

CPU gets maxed out while doing the most basic tasks due it being an absolute garbage. Everything is a waiting game with this device, youtube playback at 720.30 with AVC is totally fine no dropped frames once page loads. Libreoffice is also fine, doesnt cause much of a load.

 

I check errors in dmesg, havent really checked intel gpu driver errors but checking now doesnt really yield anything to go with. I even checked all dmesg logs but nothing but start up logs are there.

 

Weird thing is I cannot replicate this issue easily, for example sometimes this happens after rimworld has been running for 8 hours or sometimes in just 30 minutes. Same thing goes with FTL, sometimes I would be playing for 5 minutes and sometimes an hour. 

Spoiler

levent@Levent-Laptop:~$ dmesg | grep i915
[    3.700028] i915 0000:00:02.0: vgaarb: deactivate vga console
[    3.702207] i915 0000:00:02.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=io+mem:owns=io+mem
[    3.723166] [drm] Initialized i915 1.6.0 20190822 for 0000:00:02.0 on minor 0
[    3.924873] fbcon: i915drmfb (fb0) is primary device
[    3.924882] i915 0000:00:02.0: fb0: i915drmfb frame buffer device
[   15.280921] snd_hda_intel 0000:00:1b.0: bound 0000:00:02.0 (ops i915_audio_component_bind_ops [i915])
levent@Levent-Laptop:~$ dmesg | grep error
[    8.550120] EXT4-fs (sda2): re-mounted. Opts: errors=remount-ro
[   39.765571] traps: csd-keyboard[1129] trap int3 ip:7fbad65ea0d5 sp:7ffeac06c9a0 error:0 in libglib-2.0.so.0.6400.3[7fbad65ae000+84000]

csd-keyboard error is interesting and has been around for a while but I dont think its related.

[Levent]

I may have figured it out but not entirely sure what did the trick. First off, I switched to linux kernel 5.8, then changed my grub settings at the line GRUB_CMDLINE_LINUX_DEFAULT to "quiet splash intel_idle.max_cstate=1 consoleblank=0 acpi=force" (bold ones are the added lines). I proceded to update grub (sudo update-grub) then voila, been playing ftl for 8 hours and not a single problem so far.