Nvidia Drivers unstable/unusable

seth · 2025-01-05 21:24:55

when nvidia is driving the monitors etc, it forces a different nomenclature to devices such as monitors

Yes, that is unfortunately normal - why is that a problem?

also picom stops working when its intel driving although

You'll have to provide details one that (error messages, coredumps, picom config) though there's a good chance this was just because of the xf86-video-intel induced

[     6.435] (II) IGLX: Loaded and initialized swrast
[     6.435] (II) GLX: Initialized DRISWRAST GL provider for screen 0

So, where are we with this after losing the dated and misconfigure xf86* drivers?

Pena · 2025-01-06 19:29:49

seth wrote:

when nvidia is driving the monitors etc, it forces a different nomenclature to devices such as monitors
Yes, that is unfortunately normal - why is that a problem?

Uhm nothing (i didn't know if it was standard in every instance)

seth wrote:

also picom stops working when its intel driving although
You'll have to provide details one that (error messages, coredumps, picom config) though there's a good chance this was just because of the xf86-video-intel induced

I think this one will go for another issue, as even if it was related to it. It might be a sidequest, and im 80% running on nvidia its just when im not plugged in that i use the integrated graphics

[     6.435] (II) IGLX: Loaded and initialized swrast
[     6.435] (II) GLX: Initialized DRISWRAST GL provider for screen 0

So, where are we with this after losing the dated and misconfigure xf86* drivers?
Should i uninstall all of them ? Right now i have. for the xf86-video-vmware i will need to remove also virtualbox-guest-utils which requires me to know if it was just a fluke or coicidence(ie do i need to also install virtualbox-guest-utils on the host? or just the target vm?)

local/xf86-video-dummy 0.4.1-2 (xorg-drivers)
    X.org dummy video driver
local/xf86-video-nouveau 1.0.18-1 (xorg-drivers)
    Open Source 3D acceleration driver for nVidia cards
local/xf86-video-vesa 2.6.0-2 (xorg-drivers xorg)
    X.org vesa video driver
local/xf86-video-vmware 13.4.0-3 (xorg-drivers)
    X.org vmware video driver

Heres journalctl(its still running intel graphics as far as i am aware):
http://0x0.st/8iwS.txt
and heres Xorg:
http://0x0.st/8iwQ.txt

seth · 2025-01-07 10:07:55

Remove dummy, nouveau and vesa - you can keep the vmware driver.

You're running the system in regular prime config on the intel chip with the modesetting driver (not software GL), you can "prime-run superturboturkeypncher™" to invoke the nvidia GPU for heavier games.
Since you've all outputs on the intel chip this is typically what you'd want on that hardware.

Do you currently experience any problems?
The journal isn't complete (certainly lacks its head) but has

jan 06 19:15:05 micron kernel: NVRM: GPU at PCI:0000:01:00: GPU-8ec817dd-35d1-e7c3-5ba1-c017c9200b60
jan 06 19:15:05 micron kernel: NVRM: Xid (PCI:0000:01:00): 62, pid='<unknown>', name=<unknown>, 0001f96a 0001f940 0001f91d 00002559 00002548 0001f919 00000011 00000000

what's an "Internal micro-controller halt (newer drivers)" which might be down to an early lightdm start or just that nouveau got loaded, but doesn't manifest in errors on the xorg level.

Pena · 2025-01-07 19:30:42

seth wrote:

Remove dummy, nouveau and vesa - you can keep the vmware driver.
You're running the system in regular prime config on the intel chip with the modesetting driver (not software GL), you can "prime-run superturboturkeypncher™" to invoke the nvidia GPU for heavier games.
Since you've all outputs on the intel chip this is typically what you'd want on that hardware.
Do you currently experience any problems?
The journal isn't complete (certainly lacks its head) but has
jan 06 19:15:05 micron kernel: NVRM: GPU at PCI:0000:01:00: GPU-8ec817dd-35d1-e7c3-5ba1-c017c9200b60
jan 06 19:15:05 micron kernel: NVRM: Xid (PCI:0000:01:00): 62, pid='<unknown>', name=<unknown>, 0001f96a 0001f940 0001f91d 00002559 00002548 0001f919 00000011 00000000
what's an "Internal micro-controller halt (newer drivers)" which might be down to an early lightdm start or just that nouveau got loaded, but doesn't manifest in errors on the xorg level.

Well the integrated graphics from intel are kinda getting old, even with me repasting everything and removing dust every year(sometimes even more than once), I suffer a noticeable fatigue on "simple" games like minecraft, nothing feels extremely out of the ordinary i just feel like its just overall old... Basically I just use intel when im not plugged.

Also i just tested it and prime-run is still ultra instable Flickering on tecnic launcher(mod launcher), and then a complete freeze

seth · 2025-01-07 20:11:18

prime-run is still ultra instable Flickering on tecnic launcher(mod launcher), and then a complete freeze

Please don't paraphrase, https://bbs.archlinux.org/viewtopic.php?id=57855

1st make sure that prime-run works as expected at all

prime-run glxinfo -B

Then check

prime-run glxgears

Then elaborate on "ultra instable Flickering" and "tecnic launcher" and if you get "a complete freeze", see how complete that actually is (can you still switch the VT or at least reboot w/ https://wiki.archlinux.org/title/Keyboa … el_(SysRq) )
Do you run some compositor (picom) or compositing WM (gnome/kde/xfce/…)?
Same behavior in an uncomposited openbox session?

Unless you can disable the IGP in the UFI there is no way to not use it, because that's where your outputs are.
The only thing you can do is to effectively prime-run everything (incl. X11)

Pena · 2025-01-08 19:52:10

seth wrote:

prime-run is still ultra instable Flickering on tecnic launcher(mod launcher), and then a complete freeze
Please don't paraphrase, https://bbs.archlinux.org/viewtopic.php?id=57855
1st make sure that prime-run works as expected at all
prime-run glxinfo -B
Then check
prime-run glxgears

Well, i "can't" it first displayed the env variable of the Display something akin to " name of display: :0 " and then stalled (like it hit a blocking function), i wanted to check if it was regular behaviour so i ctrl+c, it didn't appear to kill the process so i closed the terminal window, a few seconds later the OS froze making me unable to switch to a TTY (control+ alt + f<0..9>) I will retry it though after i read the link you posted about Kernel keyboard shortcuts

seth wrote:

Then elaborate on "ultra instable Flickering" and "tecnic launcher" and if you get "a complete freeze", see how complete that actually is (can you still switch the VT or at least reboot w/ https://wiki.archlinux.org/title/Keyboa … el_(SysRq) )

The screen goes completely dark other than the backlight, the same reaction as when a WM fails and shutdowns abruptly, and im unable to switch to a tty (control+ alt + f<0..9>), tippically i try to always shutdown as soon as possible as im scared of it having repercutions on the firmware. EDIT: i've learned about the RIESUB system and enabling it for a session so i could test it out and after rebooting, setting sysctl to enable the code, and the control+printscreen+b nothing happens when im "frozen" which is unlike when everything is fine(causing a seriously quick reboot). STRANGELY, the shutdown button seems to work(not bruteforcing it) ie i can see (grub? ) doing its thing making shure everything unloads(without any problems)

When i say flickering i mean like the following pictu... Oh... Apparently it also flickers in intel so i guess its a tecnic problem?, basically the loading bar alternated between having text and no text akin to having a OpenGL which updates the drawing of the text and loading bar too slow causing some frames to not have either, but now im unshure if that was always a thing and just me noticing when im paying attention to it or an actual problem that just appeared now

seth wrote:

Do you run some compositor (picom) or compositing WM (gnome/kde/xfce/…)?
Same behavior in an uncomposited openbox session?

I only use openbox, although picom is installed and worked for year(s) (when i used to launch with nvidia and optimus),
AND... i know you are going to call me an idiot but apparently the picom thing was intentional, now that i see what i wrote i probably configured it so due to batery life/stability concerns
On /etc/xdg/openbox/autostart:

polybar main &
redshift -l XXXXXXprivacy_reasons:XXXXXX &
xrootgif -q -a /home/micron/sav/Configs/wallpaper3.gif &
yakuake &
setxkbmap corvo &
xscreensaver -no-splash &
clipman &
if [[ `optimus-manager --status | grep "Current GPU mode :"` = `echo "Current GPU mode : nvidia"` ]]; then
 picom --config sav/Configs/Scripts/picom.conf &
fi

seth wrote:

Unless you can disable the IGP in the UFI there is no way to not use it, because that's where your outputs are.
The only thing you can do is to effectively prime-run everything (incl. X11)

Before nvidia broke something about the kernel was incompatible with the integrated graphics so i was forced to use NVIDIA(which typically wasn't a problem as i only noticed the problem when i was on the car or at a weird spot with my laptop). Does Optimus force usage from the bootloader onwards or...?

Last edited by Pena (2025-01-08 20:58:15)

Pena · 2025-01-08 21:03:44

Going to separate posts as the other was groing a lot
I managed to fetch a rare instance aparently killing the prime-run instance midway(before the system "collapses/freezes" as previously described) i entered in a weird state where it would finally respond without "issues":

name of display: :0
X Error of failed request:  BadValue (integer parameter out of range for operation)
  Major opcode of failed request:  152 (GLX)
  Minor opcode of failed request:  24 (X_GLXCreateNewContext)
  Value in failed request:  0x0
  Serial number of failed request:  62
  Current serial number in output stream:  63

seth · 2025-01-09 12:31:29

Well, i "can't" it first displayed the env variable of the Display something akin to " name of display: :0 " and then stalled

The GPU doesn't respond and stalls.
Can you access it w/ nvidia-smi ?

Pena · 2025-01-09 22:35:03

I think once i do a prime-run its becomes unavailable but right now its being recognized

Thu Jan  9 22:34:14 2025       
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 565.77                 Driver Version: 565.77         CUDA Version: 12.7     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA GeForce GTX 1050        Off |   00000000:01:00.0 Off |                  N/A |
| N/A   48C    P0             N/A / 5001W |       5MiB /   4096MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+
                                                                                         
+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI        PID   Type   Process name                              GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|    0   N/A  N/A      2832      G   /usr/lib/Xorg                                   4MiB |
+-----------------------------------------------------------------------------------------+

Pena · 2025-01-09 22:38:10

that temp reading doesn't make any sense. im activating the extreme cooling feature(to see if it does anything about the temp)

Pena · 2025-01-09 22:46:34

well with the fans going full blast.. and me h0lding the pc in the air things d0 g0 d0wn alt0ugh everything is ultra h0t i will d0 s0me maintenance t0 check if s0mething wr0ng 0r check f0r dust, etc... Also when i run prime-run with the fans on full blast instead of going to a complete black screen i get a "_" simbol top left similar to what was happening under optimus (im still unable to tty1..etc...)

Last edited by Pena (2025-01-09 23:09:53)

seth · 2025-01-10 11:10:35

nb. the GPU is in P0 (should be P8) and draws 5001W (probably harvesting cosmic rays)

Test this w/ some alternative software stack (live distro) but this looks like it god fried (unless the cooler detached from the die)

Pena · 2025-01-11 17:00:39

also how can i check whos really responsible for displaying my second monitor(my laptop has a hdmi out connector)?
I've used lshw -c display and theres nvdia but also intel

seth · 2025-01-11 20:37:08

It'll show up in Xorg logs, you could also

for OUT in /sys/class/drm/card*; do echo $OUT; edid-decode $OUT/edid; echo "================="; done

But https://aur.archlinux.org/packages/edid-decode-git is gone and edid-decode isn't in the v4l-utils package yet.

The typical optimus setup is the eDP on the intel chip and the external outputs on the nvidia one, but that's not a given.

Arch Linux

#26 2025-01-05 21:24:55

Re: Nvidia Drivers unstable/unusable

#27 2025-01-06 19:29:49

Re: Nvidia Drivers unstable/unusable

#28 2025-01-07 10:07:55

Re: Nvidia Drivers unstable/unusable

#29 2025-01-07 19:30:42

Re: Nvidia Drivers unstable/unusable

#30 2025-01-07 20:11:18

Re: Nvidia Drivers unstable/unusable

#31 2025-01-08 19:52:10

Re: Nvidia Drivers unstable/unusable

#32 2025-01-08 21:03:44

Re: Nvidia Drivers unstable/unusable

#33 2025-01-09 12:31:29

Re: Nvidia Drivers unstable/unusable

#34 2025-01-09 22:35:03

Re: Nvidia Drivers unstable/unusable

#35 2025-01-09 22:38:10

Re: Nvidia Drivers unstable/unusable

#36 2025-01-09 22:46:34

Re: Nvidia Drivers unstable/unusable

#37 2025-01-10 11:10:35

Re: Nvidia Drivers unstable/unusable

#38 2025-01-11 17:00:39

Re: Nvidia Drivers unstable/unusable

#39 2025-01-11 20:37:08

Re: Nvidia Drivers unstable/unusable

Board footer