Modern nvidia driver or recent nouveau?

[Juanito]

Can you attach your xorg log after successfully modprobing the nouveau module - you have to start/restart x after loading the module.

Also, what does "dmesg | tail -20" give after loading the module?

[CardealRusso]

Can you attach your xorg log after successfully modprobing the nouveau module - you have to start/restart x after loading the module.

Also, what does "dmesg | tail -20" give after loading the module?

Code: [Select]

nouveau 0000:03:00.0: DRM: BIT table 'L' not found
nouveau 0000:03:00.0: DRM: TMDS table version 2.0
nouveau 0000:03:00.0: DRM: DCB version 4.1
nouveau 0000:03:00.0: DRM: DCB outp 00: 01000f32 04620030
nouveau 0000:03:00.0: DRM: DCB outp 01: 02011f52 00020010
nouveau 0000:03:00.0: DRM: DCB outp 02: 02822f66 04600020
nouveau 0000:03:00.0: DRM: DCB outp 03: 02022f62 00020020
nouveau 0000:03:00.0: DRM: DCB conn 00: 00001031
nouveau 0000:03:00.0: DRM: DCB conn 01: 00010161
nouveau 0000:03:00.0: DRM: DCB conn 02: 00020246
nouveau 0000:03:00.0: DRM: failed to create kernel channel, -22
nouveau 0000:03:00.0: DRM: MM: using COPY for buffer copies
nouveau 0000:03:00.0: DRM: allocated 1440x900 fb: 0x200000, bo 000000002aa9c8b7
fbcon: nouveaudrmfb (fb0) is primary device
Console: switching to colour frame buffer device 180x56
nouveau 0000:03:00.0: [drm] fb0: nouveaudrmfb frame buffer device
[drm] Initialized nouveau 1.3.1 20120801 for 0000:03:00.0 on minor 0

I guess this nouveau is just not compatible with my GPU (GTX 16xx)
I found this on xorg log:

Code: [Select]

[    24.241] (II) LoadModule: "vesa"
[    24.241] (II) Loading /usr/local/lib/xorg/modules/drivers/vesa_drv.so
[    24.241] (II) Module vesa: vendor="X.Org Foundation"
[    24.241]    compiled for 1.21.1.3, module version = 2.5.0
[    24.241]    Module class: X.Org Video Driver
[    24.241]    ABI class: X.Org Video Driver, version 25.2
[    24.241] (II) NOUVEAU driver
[    24.241] (II) NOUVEAU driver for NVIDIA chipset families :
[    24.241]    RIVA TNT            (NV04)
[    24.241]    RIVA TNT2           (NV05)
[    24.241]    GeForce 256         (NV10)
[    24.241]    GeForce 2           (NV11, NV15)
[    24.241]    GeForce 4MX         (NV17, NV18)
[    24.241]    GeForce 3           (NV20)
[    24.241]    GeForce 4Ti         (NV25, NV28)
[    24.241]    GeForce FX          (NV3x)
[    24.241]    GeForce 6           (NV4x)
[    24.241]    GeForce 7           (G7x)
[    24.241]    GeForce 8           (G8x)
[    24.242]    GeForce 9           (G9x)
[    24.242]    GeForce GTX 2xx/3xx (GT2xx)
[    24.242]    GeForce GTX 4xx/5xx (GFxxx)
[    24.242]    GeForce GTX 6xx/7xx (GKxxx)
[    24.242]    GeForce GTX 9xx     (GMxxx)
[    24.242]    GeForce GTX 10xx    (GPxxx)


[Juanito]

You didn’t go back far enough in dmesg (go back until you don’t see nouveau in the first column), but I’d guess you need to load firmware-nvidia before modprobe nouveau.

[aus9]

thanks to inxi output OP has Turing, post 11 and this link is very good timing
https://www.phoronix.com/news/NVIDIA-Turing-Firmware-2023

So should you have run into NVIDIA GeForce GTX 16 / RTX 20 series issues with the Nouveau driver, the new firmware is now in the Linux Firmware Git tree.

I have submitted a TCE called firmware-nvidia-turing.tcz thru the normal email system

it was a rush job so simple downloads and populate allegedly the correct dir with an tce.install script if needed.
It did pass submitqc --libs so fingers crossed eh?

[aus9]

Hi

TCE has already hit the main repo but not sure if you are using a mirror, which takes longer. When it appears in Apps pulldown  choose Download only please

Please make a backup of your boot list and keep it somewhere safe in case my new fw package causes a bork etc

Can I suggest you do a manual edit of your boot list, if you have not renamed it, likely to be onboot.lst
so that the TCE has no mention of firmware-nvidia.tcz nor graphics-6.1.2-tinycore64.tcz
then order is important....someone correct me if i am wrong, for what I just wrote

firmware-nvidia-turing.tcz
nouveau-6.1.2-tinycore64.tcz
Xorg-7.7-3d.tcz

plus other TCEs that you regularly need
full reboot.

try loading inxi and rechecking if driver changes to nouveau

2) if it does....then as per Juanito advice

Code: [Select]

glxinfo | grep Accel
##############

If its glitchy.....or worse than before, do you know how to use an usb to get back and make changes?
if its glitchy....please save /var/log/Xorg.0.log and make a copy of dmesg via

Code: [Select]

dmesg > /tmp/dmesg.txt and keep those files please. We may need to see them

Good luck, bed time for me

[CardealRusso]

Hi

still the same

Code: [Select]

Graphics:
  Device-1: NVIDIA vendor: ASUSTeK driver: nouveau v: kernel arch: Turing
    pcie: speed: 2.5 GT/s lanes: 4 ports: active: HDMI-A-1 empty: DP-1,DVI-D-1
    bus-ID: 03:00.0 chip-ID: 10de:1f82 class-ID: 0300 temp: 37.0 C
  Display: server: X.Org v: 1.21.1.6 driver: X: loaded: modesetting
    unloaded: fbdev,vesa alternate: nouveau,nv dri: swrast gpu: nouveau
    display-ID: :0.0 screens: 1
  Screen-1: 0 s-res: 1440x900 s-dpi: 96 s-size: 381x238mm (15.00x9.37")
    s-diag: 449mm (17.69")
  Monitor-1: HDMI-A-1 mapped: HDMI-1 model: HDMI res: 1440x900 hz: 60
    dpi: 64 size: 575x323mm (22.64x12.72") modes: max: 1440x900 min: 720x400
  API: OpenGL v: 4.5 Mesa 22.3.7 renderer: llvmpipe (LLVM 15.0.7 256 bits)
    direct-render: Yes
tc@box:~$ glxinfo | grep Accel
    Accelerated: no


[CardealRusso]

Code: [Select]

[    24.039] (II) LoadModule: "nouveau"
[    24.039] (WW) Warning, couldn't open module nouveau
[    24.039] (EE) Failed to load module "nouveau" (module does not exist, 0)

Looks like I had a fallback,  nouvea module is not found, again.

Code: [Select]

...
firmware-nvidia-turing.tcz
nouveau-6.1.2-tinycore64.tcz
Xorg-7.7-3d.tcz

despite this error, nouveau appears to be in fact loading?

Code: [Select]

tc@box:~$ tce-status -i | grep nouveau
nouveau-6.1.2-tinycore64
tc@box:~$ sudo modprobe nouveau


[CardealRusso]

Got some results after playing with load order, but still, nothing

Code: [Select]

[    21.094] (II) LoadModule: "nouveau"
[    21.094] (II) Loading /usr/local/lib/xorg/modules/drivers/nouveau_drv.so
[    21.099] (II) Module nouveau: vendor="X.Org Foundation
...

[    21.117] (II) [drm] nouveau interface version: 1.3.1
[    21.117] (EE) Unknown chipset: NV167
[    21.118] (II) [drm] nouveau interface version: 1.3.1
[    21.118] (EE) Unknown chipset: NV167
[    21.118] (II) [drm] nouveau interface version: 1.3.1
[    21.118] (EE) Unknown chipset: NV167
[    21.118] (II) [drm] nouveau interface version: 1.3.1
[    21.118] (EE) Unknown chipset: NV167
[    21.118] (II) modeset(0): using drv /dev/dri/card0
[    21.118] (WW) Falling back to old probe method for fbdev



firmware-nvidia.tcz
firmware-nvidia-turing.tcz
graphics-6.1.2-tinycore64.tcz
nouveau-6.1.2-tinycore64.tcz
Xorg-7.7.tcz
Xorg-7.7-3d.tcz
inxi.tcz
nano.tcz
xf86-video-nouveau.tcz
xf86-video-nv.tcz
mesa-demos.tcz


Graphics:
  Device-1: NVIDIA vendor: ASUSTeK driver: nouveau v: kernel arch: Turing
    pcie: speed: 2.5 GT/s lanes: 4 ports: active: HDMI-A-1 empty: DP-1,DVI-D-1
    bus-ID: 03:00.0 chip-ID: 10de:1f82 class-ID: 0300 temp: 37.0 C
  Display: server: X.Org v: 1.21.1.6 driver: X:
    loaded: modesetting,nouveau,nv unloaded: fbdev,vesa dri: swrast gpu: nouveau
    display-ID: :0.0 screens: 1
  Screen-1: 0 s-res: 1440x900 s-dpi: 96 s-size: 381x238mm (15.00x9.37")
    s-diag: 449mm (17.69")
  Monitor-1: HDMI-A-1 mapped: HDMI-1 model: HDMI res: 1440x900 hz: 60
    dpi: 64 size: 575x323mm (22.64x12.72") modes: max: 1440x900 min: 720x400
  API: OpenGL v: 4.5 Mesa 22.3.7 renderer: llvmpipe (LLVM 15.0.7 256 bits)
    direct-render: Yes


Accelerated: no

[aus9]

oops my overconfidence I forgot to check if those fw were a match to your actual device.

nouveau 0000:03:00.0: NVIDIA TU117 (167000a1) # from dmesg
[    21.117] (EE) Unknown chipset: NV167 # from xorg log

my fw is for 102 & 116....sorry about that.

lets start afresh?

NV160 family (Turing)
Code name    Official Name
NV162 (TU102)    NVIDIA Titan RTX, GeForce RTX 2080 Ti
NV164 (TU104)    GeForce RTX (2070 Super, 2080, 2080 Super)
NV166 (TU106)    GeForce RTX (2060, 2060 Super, 2070)
NV168 (TU116)    GeForce GTX (1650 Super, 1660, 1660 Ti, 1660 Super)
NV167 (TU117)    GeForce GTX 1650

I will get the info file for my latest fw updated later, to not mislead any more members during to my poor research

can you confirm you have the 1650?

Code: [Select]

lspci -k | grep -A 2 -E "(VGA|3D)"
if it is....firmware for 117 has not changed since 2020
https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/log/nvidia/tu117/gr?id=2e92a49f90f73c8edc44b25c6e669d5e70893c90

and TU117 was in the firmware-nvidia which you have already tried.
##########################################

@Juanito

its above my pay scale, would you think the link below might help?
https://github.com/NVIDIA/open-gpu-kernel-modules

[aus9]

otherwise, maybe OP could look at the info file for sway-dot and sway and see if things work better on sway?

sway can not work with closed source nvidia driver, I know OP does not have it,
but in theory can work with open source nouveau. YMMV

[CardealRusso]

oops my overconfidence I forgot to check if those fw were a match to your actual device.

nouveau 0000:03:00.0: NVIDIA TU117 (167000a1) # from dmesg
[    21.117] (EE) Unknown chipset: NV167 # from xorg log

can you confirm you have the 1650?

Yes, it is gtx 1650 (the gddr6 version, not the gddr5, but i guess they are just the same thing, just changed the memory thing)

Code: [Select]

03:00.0 VGA compatible controller: NVIDIA Corporation Device 1f82 (rev a1)
        Subsystem: ASUSTeK Computer Inc. Device 8773
        Kernel driver in use: nouveau


[aus9]

Hi Rich or Juanito

can the following be added to the info file for
firmware-nvidia-turing

only for chipsets
NV162 (TU102)    NVIDIA Titan RTX, GeForce RTX 2080 Ti
NV168 (TU116)    GeForce GTX (1650 Super, 1660, 1660 Ti, 1660 Super)

[CardealRusso]

Everything seems to be compatible. Am I doing something wrong? Is there anything i need to do other than simply organize the packages?

[aus9]

Am I doing something wrong?

I mislead you. if you have not already guessed, please remove from your boot list
firmware-nvidia-turing.tcz but keep firmware-nvidia.tcz

I am not sure what your current boot list is....from top to bottom IMHO it should show
in generic terms

firmware
kernel module
video driver....but I claim that driver is in Xorg-7.7-3d

recently you appear to have

firmware-nvidia.tcz
firmware-nvidia-turing.tcz
graphics-6.1.2-tinycore64.tcz
nouveau-6.1.2-tinycore64.tcz
Xorg-7.7.tcz
Xorg-7.7-3d.tcz
inxi.tcz
nano.tcz
xf86-video-nouveau.tcz
xf86-video-nv.tcz
mesa-demos.tcz

Maybe wait for expert advice....but me thinks xf86* TCEs should be before Xorg to have a benefit if any.
Also Xorg-7.7 is a dep of the 3d TCE

I do not believe graphics TCE has any benefit to you either
http://tinycorelinux.net/14.x/x86_64/tcz/graphics-6.1.2-tinycore64.tcz.list

[aus9]

correction.

the graphics TCE has drm modules so please keep it in boot list