Arch Linux

AlexBeer

Member

Registered: 2023-08-24

Posts: 22

[Resolved] problems with Nvidia geforce rtx 3050

Hi,
I'm trying to make the below graphic card work with sway:

 lspci -k | grep -EA3 'VGA|3D|Display'
01:00.0 VGA compatible controller: NVIDIA Corporation GA106 [Geforce RTX 3050] (rev a1)
	Subsystem: ZOTAC International (MCO) Ltd. GA106 [Geforce RTX 3050]
	Kernel modules: nouveau

It's a bit strange that first time I logged in after  I set up following nouveau/wayland wiki, in renderer section there was something pointing at gbm, then I've installed nvidia pkg and can't login, even blacklisting all nvidia module and putting nouveau into initramfs, until I remove the nvidia package.

Logged-in llvmpipe is used as renderer

inxi -G
Graphics:
  Device-1: NVIDIA GA106 [Geforce RTX 3050] driver: N/A
  Display: wayland server: X.org v: 1.21.1.8 with: Xwayland v: 23.2.0
    compositor: sway v: 1.8.1 driver: N/A resolution: 3840x2160~60Hz
  API: OpenGL v: 4.5 Mesa 23.1.6 renderer: llvmpipe (LLVM 15.0.7 256 bits)

when I put in "/etc/profile":

export GBM_BACKEND=nvidia-drm
export __GLX_VENDOR_LIBRARY_NAME=nvidia

I get

[wlr] [EGL] command: eglQueryDevicesEXT, error: EGL_BAD_ALLOC (0X30003), ...
...

and can't login.

The biggest problem, though, is that when I wake computer from sleep screen get no input from gpu.

In "--system" section of journalctl I see:

systemd-modules-load[416]: Failed to find module 'nvidia-uvm'
...
:Shutting down.
: Using hardware watchdog 'iTCO_wdt', version 6, device /dev/watchdog0
: Watchdog running with a timeout of 10min.
kernel: watchdog: watchdog0: watchdog did not stop!
: Using hardware watchdog 'iTCO_wdt', version 6, device /dev/watchdog0
: Watchdog running with a timeout of 10min.
...
dbus-daemon[608]: [system] Activating via systemd: service name='org.freedesktop.resolve1' unit='dbus-org.freedesktop.resolve

...skipping 92 lines
kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS
) (20230331/psparse-529)
 kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALRE
ADY_EXISTS (20230331/dsfield-184)
 kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20230331/dswload2-477)
 kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS
) (20230331/psparse-529)
 kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALRE
ADY_EXISTS (20230331/dsfield-184)
 kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20230331/dswload2-477)
 kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS
) (20230331/psparse-529)
 kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALRE
ADY_EXISTS (20230331/dsfield-184)
kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20230331/dswload2-477)
kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS
) (20230331/psparse-529)
kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALRE
ADY_EXISTS (20230331/dsfield-184)
 kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20230331/dswload2-477)
kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS
) (20230331/psparse-529)
kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALRE
ADY_EXISTS (20230331/dsfield-184)
kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20230331/dswload2-477)
 kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS
) (20230331/psparse-529)
: nvidia: Process '/usr/bin/bash -c 'for i in $(cat /proc/driver/nvidia
/gpus/*/information | grep Minor | cut -d \  -f 4); do /usr/bin/mknod -Z -m 666 /dev/nvidia${i} c $(grep nvidia-frontend /proc/devices | cut -d \  
-f 1) ${i}; done'' failed with exit code 1.
: nvidia: Process '/usr/bin/bash -c 'for i in $(cat /proc/driver/nvidia
/gpus/*/information | grep Minor | cut -d \  -f 4); do /usr/bin/mknod -Z -m 666 /dev/nvidia${i} c $(grep nvidia-frontend /proc/devices | cut -d \  
-f 1) ${i}; done'' failed with exit code 1.

In "--user" section of journalctl I see:

Process 3140 (sway) of user 1000 dumped core.
                                              
                                              Stack trace of thread 3140:
                                              #0  0x00007f5a974156eb n/a (nvidia-drm_gbm.so + 0x156eb)
                                              #1  0x00007f5a974159b8 n/a (nvidia-drm_gbm.so + 0x159b8)
                                              #2  0x00007f5a9e109e19 n/a (libgbm.so.1 + 0x4e19)
                                              #3  0x00007f5a9e10bab1 gbm_create_device (libgbm.so.1 + 0x6ab1)
                                              #4  0x00007f5a9eb50428 wlr_egl_create_with_drm_fd (libwlroots.so.11 + 0x35428)
                                              #5  0x00007f5a9eb5078b wlr_gles2_renderer_create_with_drm_fd (libwlroots.so.11 + 0x3578b)
                                              #6  0x00007f5a9eb5089a n/a (libwlroots.so.11 + 0x3589a)
                                              #7  0x00005623de932a11 n/a (sway + 0x11a11)
                                              #8  0x00007f5a9e827cd0 n/a (libc.so.6 + 0x27cd0)
                                              #9  0x00007f5a9e827d8a __libc_start_main (libc.so.6 + 0x27d8a)
                                              #10 0x00005623de933fc5 n/a (sway + 0x12fc5)
                                              ELF object binary architecture: AMD x86-64
systemd-coredump[3174]: [] Process 3163 (sway) of user 1000 dumped core.
                                              
                                              Stack trace of thread 3163:
                                              #0  0x00007f994ac156eb n/a (nvidia-drm_gbm.so + 0x156eb)
                                              #1  0x00007f994ac159b8 n/a (nvidia-drm_gbm.so + 0x159b8)
                                              #2  0x00007f99518b9e19 n/a (libgbm.so.1 + 0x4e19)
                                              #3  0x00007f99518bbab1 gbm_create_device (libgbm.so.1 + 0x6ab1)
                                              #4  0x00007f9952300428 wlr_egl_create_with_drm_fd (libwlroots.so.11 + 0x35428)
                                              #5  0x00007f995230078b wlr_gles2_renderer_create_with_drm_fd (libwlroots.so.11 + 0x3578b)
                                              #6  0x00007f995230089a n/a (libwlroots.so.11 + 0x3589a)
                                              #7  0x0000557b11cfda11 n/a (sway + 0x11a11)
                                              #8  0x00007f9951e27cd0 n/a (libc.so.6 + 0x27cd0)
                                              #9  0x00007f9951e27d8a __libc_start_main (libc.so.6 + 0x27d8a)
                                              #10 0x0000557b11cfefc5 n/a (sway + 0x12fc5)
                                              ELF object binary architecture: AMD x86-64
systemd-coredump[3271]: [] Process 3260 (sway) of user 1000 dumped core.
                                              
                                              Stack trace of thread 3260:
                                              #0  0x00007f97354156eb n/a (nvidia-drm_gbm.so + 0x156eb)
                                              #1  0x00007f97354159b8 n/a (nvidia-drm_gbm.so + 0x159b8)
                                              #2  0x00007f973bfb6e19 n/a (libgbm.so.1 + 0x4e19)
                                              #3  0x00007f973bfb8ab1 gbm_create_device (libgbm.so.1 + 0x6ab1)
                                              #4  0x00007f973c9fe428 wlr_egl_create_with_drm_fd (libwlroots.so.11 + 0x35428)
                                              #5  0x00007f973c9fe78b wlr_gles2_renderer_create_with_drm_fd (libwlroots.so.11 + 0x3578b)
                                              #6  0x00007f973c9fe89a n/a (libwlroots.so.11 + 0x3589a)
                                              #7  0x0000555c8b526a11 n/a (sway + 0x11a11)
                                              #8  0x00007f973c627cd0 n/a (libc.so.6 + 0x27cd0)
                                              #9  0x00007f973c627d8a __libc_start_main (libc.so.6 + 0x27d8a)
                                              #10 0x0000555c8b527fc5 n/a (sway + 0x12fc5)
                                              ELF object binary architecture: AMD x86-64

Don't know if these log extract have something useful.

Any help?

Best Regards.

Last edited by AlexBeer (2023-08-28 18:51:01)

Scimmia

Fellow

Registered: 2012-09-01

Posts: 13,694

Re: [Resolved] problems with Nvidia geforce rtx 3050

I'm confused, why would you put

export GBM_BACKEND=nvidia-drm
export __GLX_VENDOR_LIBRARY_NAME=nvidia

If you're using nouveau/mesa, not nvidia?

AlexBeer

Member

Registered: 2023-08-24

Posts: 22

Re: [Resolved] problems with Nvidia geforce rtx 3050

You are right I'm using nouveau: I've read a sentence in the wiki (" To force GBM as a backend, set the following environment variables: ") but didn't realize was related to nvidia driver. Well, one less problem. Sorry I'm a little bit confused: with nouveau llvmpipe is the right renderer?

seth

Member

From: Don't DM me only for attention

Registered: 2012-09-03

Posts: 74,307

Re: [Resolved] problems with Nvidia geforce rtx 3050

Yes, why are you trying to use nouveau w/ an Ampère chip?
https://nouveau.freedesktop.org/FeatureMatrix.html (You've an NV170)

---

How to upload text · How to boot w/o GUI · Disable Windows Fast-Start! · Fix your xinitrc

uchiha_vk

Member

Registered: 2023-08-26

Posts: 5

Re: [Resolved] problems with Nvidia geforce rtx 3050

I had the same problem, happened when i tried running Hyprland  on nvidia. I tried debugging the source, It loads intel mesa for GLES and tries to use nvidia_drm (which return bad_alloc). I'm not an expert, but seem Wayland itself is have this problem with nvidia drivers. To be noted this happens only with muxless nvidia laptops .   i really wanted Hyprland running nvidia, using nouveau has no problem (i can use my multi-monitor setup). If you want to have multi monitor and use nvidia NVENC for like obs recording, (https://wiki.archlinux.org/title/nouvea … and_tricks) this will be helpful (this is what i now settled for).

seth

Member

From: Don't DM me only for attention

Registered: 2012-09-03

Posts: 74,307

Re: [Resolved] problems with Nvidia geforce rtx 3050

https://wiki.archlinux.org/title/NVIDIA … de_setting is a hard requirement for wayland and if it loads intel mesa, this is most likely an optimus system and that comes w/ it's own problems.
Please don't paraphrase, https://bbs.archlinux.org/viewtopic.php?id=57855
Please post your complete system journal for the boot:

sudo journalctl -b | curl -F 'file=@-' 0x0.st

---

How to upload text · How to boot w/o GUI · Disable Windows Fast-Start! · Fix your xinitrc

AlexBeer

Member

Registered: 2023-08-24

Posts: 22

Re: [Resolved] problems with Nvidia geforce rtx 3050

I'm trying with nvidia driver since nouveau provide little support for my card (hence fallback to llvmpipe when using nouveau?). After vulkan-intel installation:

inxi -G
Graphics:
  Device-1: NVIDIA GA106 [Geforce RTX 3050] driver: nvidia v: 535.104.05
  Display: wayland server: X.org v: 1.21.1.8 with: Xwayland v: 23.2.0
    compositor: sway v: 1.8.1 driver: gpu: nvidia,nvidia-nvswitch
    resolution: 3840x2160~60Hz
  API: OpenGL v: 4.6.0 NVIDIA 535.104.05 renderer: NVIDIA GeForce RTX
    3050/PCIe/SSE2

I can, waiting some minutes, bring screen up after sleep.

BUT: I have heavy flickering and dark shapes in the screen: this could be due to lack of support of nvidia driver by sway?

Any hope for nouveau support for my card?

Any hope for better results with nvidia-open driver? (I would try it but repo is unreachable!?)

sudo journalctl -b | curl -F 'file=@-' 0x0.st

Cool.
I've uploaded log after latest configuration changes: http://0x0.st/H9xI.txt

seth

Member

From: Don't DM me only for attention

Registered: 2012-09-03

Posts: 74,307

Re: [Resolved] problems with Nvidia geforce rtx 3050

So there's no intel IGP and no sign of anything like it being used.
You're now using the nvidia driver and nvidia_drm.modeset=1 is there and the simpledrm device isn't. All good here.

I can, waiting some minutes, bring screen up after sleep.

BUT: I have heavy flickering and dark shapes in the screen: this could be due to lack of support of nvidia driver by sway?

Is this all about waking up from S3 now?
https://download.nvidia.com/XFree86/Lin … ement.html

I would try it but repo is unreachable!?

Please don't paraphrase, https://bbs.archlinux.org/viewtopic.php?id=57855
nvidia-open is in the "extra" repo like most packages; but it was updated on 2023-08-25, don't try to run partial updates.

---

How to upload text · How to boot w/o GUI · Disable Windows Fast-Start! · Fix your xinitrc

AlexBeer

Member

Registered: 2023-08-24

Posts: 22

Re: [Resolved] problems with Nvidia geforce rtx 3050

Sorry replying late.

So there's no intel IGP and no sign of anything like it being used.
You're now using the nvidia driver and nvidia_drm.modeset=1 is there and the simpledrm device isn't. All good here.

Ok. Firefox work pretty well, for other pkgs I may need to search in aur for better wayland support.

Is this all about waking up from S3 now?
https://download.nvidia.com/XFree86/Lin … ement.html

The flickering is solved using sway-nvidia from aur: also using rofi-lbonn-wayland for dmenu replacement.

nvidia-open is in the "extra" repo like most packages; but it was updated on 2023-08-25, don't try to run partial updates.

sorry I'm used to apt, just need to sync local pacman db

seth

Member

From: Don't DM me only for attention

Registered: 2012-09-03

Posts: 74,307

Re: [Resolved] problems with Nvidia geforce rtx 3050

Soo… is there a pending problem wrt the original post…?

---

How to upload text · How to boot w/o GUI · Disable Windows Fast-Start! · Fix your xinitrc