Drivers broken? -Nvidia

General system Kernel, boot, graphics & hardware
1

Hi all! So this is a bit chaotic because I don’t really know what even happened.

Essentially, I do art and I bought an xp-pen tablet. I noticed whenever I connect it, Kwin dies. Screen goes black (occasionally requires restart via switching into CLI on another session-thing (tty3 etc)), desktop gone, and then recovers most of the time after a little. Because of this, I usually restart the computer and attach the xp-pen tablet while it’s off because if it starts with the tablet it doesn’t do this.

I know Kwin is the one that dies because I get a notification saying “Kwin restarted due to unexpected graphics reset” or something after it recovers.

Recently though I couldn’t really be bothered to do this and connected it, and waited for the screen to recover etc. I have memory problems so I can’t remember if this specifically is the cause, but I’m pretty sure it is.

What happened is that now my plasma widgets and all are messed up, and so are my panels, all the icons besides the virtual desktop list on it go to a small corner on the left side of the panel if I hover over them with my mouse. if I go into panel edit mode, it doesn’t show my desktop background anymore. the application launcher doesn’t show my user profile picture anymore, it’s like it’s unloaded.

for widgets, I have memory and CPU usage widgets from the preinstalled plasma widgets, and the moving elements don’t show anymore. I also noticed the temperature measurement widget doesn’t find my GPU anymore (Nvidia RTX 4070 Ti), which leads me to believe this may be an nvidia driver issue somehow.

more evidence is today I opened my system monitor to kill a process and it told me this:

Screenshot_20241003_145606
Screenshot_20241003_1456061027×161 49.6 KB

I use the nvidia-dkms drivers. ETA: I am on X11, not Wayland

How do I check my graphics drivers and what do I do with them to fix this? And is anybody else having a similar issue (would suggest an update may have done this…?)

2

There is an issue with nvidia on the latest kernel causing black screens.

3

That thread is for wayland specifically, I think, but I’m on X11 since wayland has never worked properly for me. Should’ve specified I was on X11, I’ll add that to my post. Thank you anyway!!

ETA: tested wayland, problem persists

4

When u say drawing tablet, do u mean one with a screen or one without a screen?

5

with a screen! not sure what the specific term for those is, sorry. the graphical reset happens because it adjusts to a second monitor, I’m not sure if the kwin crashes are specific to the graphical tablet. I have another monitor I can test that with, but up until this point I’ve been too lazy to dive into the cable hell under my desk

6

Wel it’s been a long time since I messed with X configuration, but a lot of problems of this nature can be solved by using the right configuration (typically through files under /etc/X11/xorg.conf.d/)

I believe you may also be able to control things such as your tablet’s default pressure sensitivity and such through that depending on which drivers are being used.

Drawing tablets other than wacom can be a bit tricky on linux unfortunately, but I did have a yiynova drawing tablet once with a screen which worked great so it should generally speaking be doable (but I had to jump through some hoops to make it work).

I don’t think your nvidia drivers necessarily have much to do with the problem. You could try using another desktop environment, see if it behaves better there, and you could try to switch to nouveau (or use your igpu instead) see if it will work that way.

7

No, it’s DEFINITELY my graphics drivers. My temperature widget no longer can read my GPU temperature, and the system monitor tells me hardware acceleration is unavailable, which to my knowledge is a definite confirmation that this is my graphics drivers. It even tells me to check the graphics drivers, and that visual glitches will appear, and what’s happening are definitely visual glitches. The clickable zone doesn’t move, but icons et cetera do.

I think my computer adjusting to the graphical tablet needing its own drivers may have started the problem, it also goes black when I switch to a normal monitor and Kwin crashes and restarts too, but I there is nothing wrong with my tablet per se. I can use it normally with no issues whatsoever, and it is marketed as Linux compatible, which is why I bought it.

I’m debating just fully reinstalling my nvidia drivers, but I am not confident this will not mess my computer up even more. I’ve so far been hoping an update to nvidia-dkms will fix the issue honestly

edit: more info

8

At this point, are you sure it isn’t the card or the motherboard being weird? I remember you had issues with games. Now you have issues with this.

9

I think whatever caused the games issue got wiped in the whole kernel-deleting-itself fiasco and rebuilt clean when I reinstalled it.

I’m honestly pretty sure the hardware is fine and this is just a radioactive combo of nvidia being itself and whatever beef software seems to have with me.

Frankly I’m not surprised at all nvidia stuff is causing problems, that’s just what it does

ETA: if you have any idea how I can verify my mobo and GPU are fine please share though. I am mildly terrified

10

This is all over the map. You need to resolve one issue before another. Get it working on Nvidia first. This includes X11 and Wayland. Deal with the games issue and or tablet stuff and widgets whatever … separately. :thinking:

Edit: Post your hardware output to start with. Post the url

inxi -Faz | eos-sendlog

Edit: There was just an update for nvdia-dkms and nvidia-utils. Lets see what you have first.

11

The games not working thing happened back in July- Steam games would glitch out madly. This was resolved when I was recommended to install the nvidia-dkms drivers, which did not fix the problem but caused my initram to be too big for its partition, resulting in my kernel deleting itself. Another user on this forum helped me out, and after my kernel was reinstalled the issue was gone.

winnyace was present on that topic, hence why they referenced this incident. The widgets are part of the visual glitches I’m experiencing right now, so I can’t really deal with those separately, and the tablet is mostly unrelated, it was more an unfortunate catalyst.

Is there a way to reliably filter out sensitive information such as mac address from that command? I fear I’m not very comfortable with posting information like that on the internet.

I will install the update and see if that does anything for me

12

I don’t know what information you are worried about? It’s just hardware. :thinking:

13

might be mistaking the command, gives me this warning which I recognise. Another thread I made I was told to also do this and I was told the physical mac address of my network card will be exposed.
Screenshot_20241003_220201

Genuinely sorry if I’m being ridiculous, I’m still pretty new to stuff like this, especially to nvidia and other problem magnet hardware so I don’t know my way in and out as much as I’d like

14

It’s just a providing you a warning to not send info you don’t want others to see because i can look at it and so can anyone else with the link. This is standard.

1 Like
15

https://0x0.st/XE-b.txt

here ya go

16

First your UEFI Bios is way out of date. UEFI: American Megatrends v: 0812 date: 02/24/2023
There are 17 newer updates plus one Beta update which makes 18 in total.

You do have the latest nvidia driver installed and it is rendering on the nvidia gpu although it looks like you are logged in under X11.

 Device-1: NVIDIA AD104 [GeForce RTX 4070 Ti] vendor: ASUSTeK driver: nvidia
    v: 560.35.03 alternate: nouveau,nvidia_drm non-free: 550.xx+
    status: current (as of 2024-09) arch: Lovelace code: AD1xx
    process: TSMC n4 (5nm) built: 2022+ pcie: gen: 4 speed: 16 GT/s lanes: 16
    ports: active: none off: HDMI-A-1,HDMI-A-2 empty: DP-1,DP-2,DP-3
    bus-ID: 01:00.0 chip-ID: 10de:2782 class-ID: 0300

 API: OpenGL v: 4.6.0 compat-v: 4.5 vendor: nvidia mesa v: 560.35.03
    glx-v: 1.4 direct-render: yes renderer: NVIDIA GeForce RTX 4070 Ti/PCIe/SSE2
    memory: 11.71 GiB

Does the system log in under Wayland? It should be working.
I would also suggest you update the UEFI Bios to the latest before the beta version.

Edit: The rest of the output looks okay although i do see it has a raid controller. But if it is booting and working then I’m not concerned.

17

Wayland works just the same as X11 does by now. It’s no longer broken on Nvidia like it was for ages but the visual glitches I’m getting are still there, tried that earlier. I’m convinced it’s something with the graphical drivers mainly because of the warning the system monitor is giving me regarding hardware acceleration not being available, father says he wouldn’t know what else would be causing that.

How do I update the BIOS? I remember when I first got my computer, I was told to update the BIOS to be able to enable some fancy RAM thing I have (forgot name, I think XMP? had an X in it), but my dad was like “oh that could go seriously wrong”, which scared me out of doing it because I saved well over half a year to be able to afford this thing and didn’t want to risk bricking it, especially because I am good at breaking things.

I sadly have no idea what a RAID controller is. I just assumed that was a normal thing when I read it in the log.

ETA: I fear I’m horrible to work with, this is because before now I was mostly on debian distros, which my father was able to help me with, but I’m on my own with Arch and have somehow managed to BS my way into being the tech savvy person in my friend groups as well as barely actually knowing anything. I have not yet managed to learn what I am doing and I apologise for this chaotic mess

18

You probably need to set up hardware acceleration which means going through the wiki page and installing some packages, setting up some environmental variables and verifying that everything is working. So I’m referencing VA-API and VDPAU. The bulk of it alraedy works with nvidia-utils but you need to make some adjustments from the Wiki page to verfy Vulkan and VA-API and VDPAU are working. It’s very minor stuff but it’s understanding how to do it.

https://wiki.archlinux.org/title/Hardware_video_acceleration#NVIDIA

I think you should update the UEFI Firmware (Bios)

Here is the page for the Bios files.

Here is a video showing how to use EZ Flash from with the UEFI settings screen.

https://www.asus.com/support/faq/1012815/

Edit: I know this is a lot of info. I would get the Bios updated properly first and then move on to the wiki page for the hardware acceleration stuff later. See how it works first once updated.

Edit: Make sure the computer isn’t turned off or touched while it’s flashing a new bios. Hopefully no power outages either. It has to complete and usually it reboots on it’s own when done.

19

If you need help after to set up the hardware acceleration stuff. Just ask first and i could guide you through it. One step at a time. :wink:

1 Like
20

Is there a reason hardware acceleration would break or get disabled, because this is the first time since I installed EOS that is has told me this. The rendering of the system monitor window is also broken now. maybe nvidia-utils has an issue?

Also I realise I never included a visual of the glitches, here they are:
Screenshot_20241003_225406
all panel icons move to the spot above the application launcher (including the icon for that) should I hover my mouse over them, but the clickable site doesn’t change.

I think I’m gonna need somebody to walk me through the UEFI update, I’m extremely anxious about doing something wrong and messing everything up, I don’t trust myself to get it right the first time with something so risky and I would really rather avoid doing this unless I’m over 90% certain this can fix my initial issue. I won’t be able to afford a new motherboard for a long while if I screw this one up and this PC is too important to risk that