You are not logged in.
Pages: 1
I've been facing this issue for more than 2 years now on my desktop which has an Intel i3-4150 and an AMD RX 560 GPU and I thought it was due to bad hardware but now I'm facing the same issue on my ThinkPad E495 with an integrated Vega GPU.
My monitor sometimes doesn't wake up from sleep. No matter what I do, the screen remains powered off. I have to forcibly reboot my system and sometimes end up losing my work. The trouble is that meaningful logs were never really generated on my desktop about this but I've started facing this issue on my laptop as well. I was, however, able to get some meaningful logs on it.
I was using the linux-hardened package initially on my laptop but that turned out to infeasible because session just froze after working for a day or two. I switched to linux-lts and the freezing problem went away but this problem wasn't resolved.
These are the logs after I wake up my laptop. The laptop was connected to an external monitor as well. Neither the laptop screen, nor the monitor powered up after I woke up my laptop from sleep.
I'm using the latest linux-lts 5.10.23-1-lts kernel right now.
Is this issue because of bugs in the AMDGPU driver? Or is it possible that it might be because of some security hardening I've done? I've enabled the lockdown=integrity kernel parameter, if that's helpful.
Offline
What have you researched yourself on this? I do not have AMD, but just a 3 second Google of "amdgpu_dm_backlight_update_status" leads me to your issue that you describe, for instance: https://gitlab.freedesktop.org/drm/amd/-/issues/1251, so there has to be more details out there and possibly a workaround/fix?
Offline
Like I said, meaningful logs still aren't even generated on my desktop when this issue happens. I got my laptop just a few months ago and this is the first time I've seen any meaningful logs which I posted here.
Honestly, I feel pretty distraught at this point. All I've found are links with endless dmesg error logs and people claiming they're seeing this too. A comment in the redhat bugzilla also claims to have been suffering from this issue for more than 2 years now. Another comment claims that this suspend/resume issue is present on his desktop as well.
https://bugzilla.redhat.com/show_bug.cgi?id=1881889
https://gitlab.freedesktop.org/drm/amd/-/issues/1337
https://gitlab.freedesktop.org/drm/amd/-/issues/1251
This is the closest thing I've found that claims to fix this issue but the author of the patches himself says these are hacks rather than fixing the root cause. This basically means that even if I upgrade to latest kernel right now, this issue would still be there.
https://gitlab.freedesktop.org/drm/amd/ … ote_841963
Apparently, I'd have to compile my own kernel with CONFIG_HSA_AMD=n and then hope that the issue is fixed.
I'd always heard that one should avoid using Nvidia hardware when using Linux so I did that. Now there are these critical issues in AMD hardware as well. This is not some edge case exotic issue. Suspend/Resume is basically broken. I've not been confident about suspending my desktop, and now my laptop as well, with my web browser and text editor open with unsaved work because I don't know if it will come back up and then I'd have to risk losing data. I'm not sure if the situation will become better if I restrict my hardware to Intel only.
Offline
You are just blabbering and complaining here. This is open-source, volunteer-driven software. If you don't like the service, either find something else, or roll up your sleeves and help fix the problem(s).
My "3 second Google search" comment got you to find a possible workaround with a kernel setting, but now it's too much effort again to try?
Offline
Thanks for all your help, you've been really helpful. I'm sorry for "blabbering" and being apprehensive about compiling my own kernel with patches that the author calls hacks. I guess I'll keep in mind that I have to be ready to compile my own kernels if I use Linux or shut up and not complain or raise issues or make support threads if something doesn't work as expected.
Offline
Pages: 1