Bug 1225296 - nouveau/GT 218 from 2009: Dual monitor not recognised after upgrade 15.5 to 15.6 (kernel driver regression?)
Summary: nouveau/GT 218 from 2009: Dual monitor not recognised after upgrade 15.5 to 1...
Status: IN_PROGRESS
Alias: None
Product: openSUSE Distribution
Classification: openSUSE
Component: Kernel:Drivers (show other bugs)
Version: Leap 15.6
Hardware: x86 Other
: P5 - None : Normal (vote)
Target Milestone: ---
Assignee: Kernel Bugs
QA Contact: E-mail List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2024-05-26 11:35 UTC by Martin Schmidt
Modified: 2024-06-14 17:38 UTC (History)
3 users (show)

See Also:
Found By: ---
Services Priority:
Business Priority:
Blocker: ---
Marketing QA Status: ---
IT Deployment: ---


Attachments
Xorg.0.log (31.18 KB, text/plain)
2024-05-27 09:54 UTC, Martin Schmidt
Details
dmesg for kernel 5.14 (66.40 KB, text/plain)
2024-06-06 14:26 UTC, Martin Schmidt
Details
dmesg for kernel 6.4 (75.87 KB, text/plain)
2024-06-06 14:26 UTC, Martin Schmidt
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Martin Schmidt 2024-05-26 11:35:18 UTC
Moin, moin,
I am using openSUSE 15.5 on an old intel based x86 PC, equipped with a simple nvidia gforce and two monitors. The monitors are using a displayport connnection from the graphics card to the monitors via an Y-shaped cable. Both monitors are recognised correctly and can be configured to form a single graphical work space, when the system is up and running.  

After upgrading to openSUSE 15.6 this does not work any more. When booting, monitor 1 shows some firmware screen and the openSUSE boot screen, where the OS version (several subversions of 15.6 and frozen snapshots of 15.5) can be selected. Using the default option (15.6 from harddrive), the output of the boot process is moving to monitor 2, but monitor 1 is going to a "power saving sleep mode" and stays inactive frome here on. The boot process is finished correctly with monitor 2 as the only graphical device. Monitor 1 cannot be activated again.

Switching off monitor 2 before booting, the system comes up without any valid output device. Monitor 1 goes to sleep after the system boot started. I did not check the system status with a remote ssh-connection yet. 

Switching off monitor 1 before booting, lets the system come up completely with monitor 2. Also the initial hardware screen is diplayerd there. 

In turn, booting from a read-only snapshot with openSUSE 15.5, both monitors are used as before. The nouveau-dri drivers are used, no propriatary nvidia driver is installed. 

Some upgrade history:
- just to test, if the few graphic card cores could be used in a cuda environment, cuda was installed under 15.5. This failed and cuda was deinstalled. However, when upgrading later from 15.5 to 15.6 the cuda development environment was reinstalled automatically together with nvidia drivers. As a result the nvidia drivers allowed only for a single low resolution vga monitor. May be, this happened, since the nvidia repository was not deleted and there was some leftover from the cuda test. 
- with the ascii-yast the nvidia drivers (gforce003/004) and cuda could be removed and the nouveau-dri drivers where reinstalled - together with dependencies. This allows for a single monitor OS_15.6 system with high resolution, but the second monitor (monitor 1) is lost during the boot process as described above.

Before I try to install from scratch just with the hope that the error goes away, I am reporting this strange behaviour here. May be, this helps to streamline the distribution. Any hints are very welcome - especially how to modify the bootloader options.
Cheers,
Martin Schmidt
Comment 1 Stefan Dirsch 2024-05-26 17:51:55 UTC
Hmm. This looks like a regression in nouveau kernel driver in kernel 6.4 of Leap 15.6. Unfortunately we can't support nouveau driver. I suggest to install proprietary driver instead. If you already have a modern card (Turing+ from 2018+), you could even install the new open source kernel driver from NVIDIA.

Output of 

  hwinfo --gfxcard
  xrandr -q 

would be useful. And /var/log/Xorg.*.log or ~/.local/share/xorg/Xorg.*.log.
Comment 2 Martin Schmidt 2024-05-27 09:54:45 UTC
Created attachment 875125 [details]
Xorg.0.log
Comment 3 Martin Schmidt 2024-05-27 09:56:20 UTC
Moin, moin,

the hardware is pretty old, but there is no reason to replace it until now. Having many windows and simple graphics - also with good resolution - open at the same time works well. Just a workbench for editing texts, programming and so on.

hwinfo --gfxcard:

12: PCI 100.0: 0300 VGA compatible controller (VGA)            
  [Created at pci.386]
  Unique ID: VCu0.NpzQxy0mUR4
  Parent ID: vSkL.A99aFZ6svBE
  SysFS ID: /devices/pci0000:00/0000:00:01.0/0000:01:00.0
  SysFS BusID: 0000:01:00.0
  Hardware Class: graphics card
  Model: "nVidia GT218 [NVS 300]"
  Vendor: pci 0x10de "nVidia Corporation"
  Device: pci 0x10d8 "GT218 [NVS 300]"
  SubVendor: pci 0x10de "nVidia Corporation"
  SubDevice: pci 0x0862
  Revision: 0xa2
  Driver: "nouveau"
  Driver Modules: "nouveau"
  Memory Range: 0xf6000000-0xf6ffffff (rw,non-prefetchable)
  Memory Range: 0xf0000000-0xf3ffffff (ro,non-prefetchable)
  Memory Range: 0xf4000000-0xf5ffffff (ro,non-prefetchable)
  I/O Ports: 0xe000-0xefff (rw)
  Memory Range: 0x000c0000-0x000dffff (rw,non-prefetchable,disabled)
  IRQ: 28 (1261 events)
  Module Alias: "pci:v000010DEd000010D8sv000010DEsd00000862bc03sc00i00"
  Driver Info #0:
    Driver Status: nouveau is active
    Driver Activation Cmd: "modprobe nouveau"
  Config Status: cfg=no, avail=yes, need=no, active=unknown
  Attached to: #10 (PCI bridge)

Primary display adapter: #12

Also the hardware information of yast recognises only one monitor.

xrandr -q
Screen 0: minimum 1 x 1, current 2560 x 1295, maximum 16384 x 16384
Virtual1 connected primary 2560x1295+0+0 (normal left inverted right x axis y axis) 0mm x 0mm
   2560x1295     60.00*+
   3840x2400     59.97 
....
Virtual2 disconnected (normal left inverted right x axis y axis)
.... the same until Virtual8.

The logfile is uploaded as attachment.
Cheers,
Martin Schmidt
Comment 4 Stefan Dirsch 2024-05-27 11:39:52 UTC
Ok. Your gfx card was released in 2009. The latest proprietary driver still supporting it is 340.108, which is End-Of-Life since a long time now. It's more-or-less impossible to get this driver still building against Kernel 6.4 of Leap 15.6.

I'm afraid you need to rely on nouveau kernel driver.
Comment 5 Stefan Dirsch 2024-05-27 11:41:48 UTC
(In reply to Martin Schmidt from comment #2)
> Created attachment 875125 [details]
> Xorg.0.log

Nothing obivous, but you could try without nouveaou X driver by uninstalling xf86-video-nouveau driver package

  zypper rm xf86-video-nouveau

Then restart X or your complete system. But I'm afraid you found a regression in nouveau kernel driver.
Comment 6 Takashi Iwai 2024-05-27 15:43:39 UTC
First off, please verify that the problem persists with X11 modesetting driver without X11 nouveau driver as suggested in the above.  If yes, install the Leap 15.5 kernel on your Leap 15.6 system with zypper install --oldpackage option, reboot with it and check whether it still works as expected.

If Leap 15.5 kernel is confirmed to work, please boot with drm.debug=0x1e log_buf_len=1M options, get dmesg output after the boot.  Take the logs from both Leap 15.5 and Leap 15.6 kernels for comparison, and upload to Bugzilla.
Comment 7 Martin Schmidt 2024-05-28 06:38:58 UTC
Many thanks. I will do so. But I am not sure, where to define the option

drm.debug=0x1e

best? Which file to edit? /etc/default/grub?
Sorry for asking this stupid question. The man pages are good for those already knowing the essentials.
Cheers
Comment 8 Takashi Iwai 2024-05-28 06:50:56 UTC
(In reply to Martin Schmidt from comment #7)
> Many thanks. I will do so. But I am not sure, where to define the option
> 
> drm.debug=0x1e
> 
> best? Which file to edit? /etc/default/grub?

For a temporary command line, you can add it at GRUB.  Press 'e' on the boot entry, and edit there, press ctrl-X to boot.
Comment 9 Martin Schmidt 2024-05-28 12:37:18 UTC
I had a little accident. Nothing serious, but I cannot continue now. I will upload the logs in the beginning of the next week.
Comment 10 Felix Miata 2024-05-29 07:53:26 UTC
Similar NVidia cards, but with discrete output ports for 2 or 3 displays, plus two connected displays each, are working fine here:
# inxi -CMSz --vs --hostname
inxi 3.3.34-00 (2024-04-13)
System:
  Host: hp945 Kernel: 6.4.0-150600.21-default arch: x86_64 bits: 64
  Desktop: KDE v: 3.5.10 Distro: openSUSE Leap 15.6
Machine:
  Type: Desktop System: HP-Pavilion product: RX900AA-ABA a6010n v: N/A
    serial: <filter>
  Mobo: ASUSTek model: LEONITE v: 5.00 serial: <filter> BIOS: Phoenix
    v: 5.10 date: 01/30/2007
CPU:
  Info: dual core model: Intel Core2 6600 bits: 64 type: MCP cache: L2: 4 MiB
  Speed (MHz): avg: 1600 min/max: 1600/2400 cores: 1: 1600 2: 1600
# inxi -Gaz
Graphics:
  Device-1: NVIDIA GT218 [GeForce 210] vendor: eVga.com. driver: nouveau
    v: kernel non-free: series: 340.xx status: legacy (EOL~2019-12-xx) last:
    release: 340.108 kernel: 5.4 xorg: 1.20 arch: Tesla process: 40-80nm
    built: 2006-2013 pcie: gen: 1 speed: 2.5 GT/s lanes: 16 ports:
    active: DVI-I-1,HDMI-A-1 empty: VGA-1 bus-ID: 01:00.0 chip-ID: 10de:0a65
    class-ID: 0300 temp: 56.0 C
  Display: x11 server: X.Org v: 1.21.1.11 compositor: kwin driver: X:
    loaded: modesetting unloaded: fbdev,vesa alternate: nouveau,nv,nvidia
    dri: nouveau gpu: nouveau display-ID: :0 screens: 1
  Screen-1: 0 s-res: 3600x1200 s-dpi: 120 s-size: 762x254mm (30.00x10.00")
    s-diag: 803mm (31.62")
  Monitor-1: DVI-I-1 pos: right model: Dell P2213 serial: <filter>
    built: 2012 res: 1680x1050 hz: 60 dpi: 90 gamma: 1.2
    size: 473x296mm (18.62x11.65") diag: 558mm (22") ratio: 16:10 modes:
    max: 1680x1050 min: 720x400
  Monitor-2: HDMI-A-1 mapped: HDMI-1 pos: primary,left model: NEC EA243WM
    serial: <filter> built: 2011 res: 1920x1200 hz: 60 dpi: 94 gamma: 1.2
    size: 519x324mm (20.43x12.76") diag: 612mm (24.1") ratio: 16:10 modes:
    max: 1920x1200 min: 640x480
  API: EGL v: 1.5 hw: drv: nvidia nouveau platforms: device: 0 egl: 1.4
    drv: nouveau device: 1 drv: swrast gbm: egl: 1.4 drv: nouveau surfaceless:
    egl: 1.4 drv: nouveau x11: egl: 1.4 drv: nouveau inactive: wayland
  API: OpenGL v: 4.5 compat-v: 3.3 vendor: mesa v: 23.3.4 glx-v: 1.4
    direct-render: yes renderer: NVA8 device-ID: 10de:0a65 memory: 983.4 MiB
    unified: no
#
# inxi -CMSz --vs --hostname
inxi 3.3.34-00 (2024-04-13)
System:
  Host: p5bse Kernel: 6.4.0-150600.21-default arch: x86_64 bits: 64
  Desktop: KDE v: 3.5.10 Distro: openSUSE Leap 15.6
Machine:
  Type: Desktop Mobo: ASUSTeK model: P5B SE v: Rev 1.xx serial: <filter>
    BIOS: American Megatrends v: 1103 date: 06/04/2009
CPU:
  Info: dual core model: Intel Core2 Duo E7500 bits: 64 type: MCP cache:
    L2: 3 MiB
  Speed (MHz): avg: 1596 min/max: 1596/2926 cores: 1: 1596 2: 1596
# inxi -Gaz
Graphics:
  Device-1: NVIDIA GF119 [NVS 310] vendor: Hewlett-Packard driver: nouveau
    v: kernel non-free: series: 390.xx+ status: legacy (EOL~2022-11-22) last:
    release: 390.157 kernel: 6.0 xorg: 1.21 arch: Fermi code: GF1xx
    process: 40/28nm built: 2010-2016 pcie: gen: 1 speed: 2.5 GT/s lanes: 16
    ports: active: DP-1,DP-2 empty: none bus-ID: 01:00.0 chip-ID: 10de:107d
    class-ID: 0300 temp: 50.0 C
  Display: x11 server: X.Org v: 1.21.1.11 compositor: kwin driver: X:
    loaded: modesetting unloaded: fbdev,vesa alternate: nouveau,nv,nvidia
    dri: nouveau gpu: nouveau display-ID: :0 screens: 1
  Screen-1: 0 s-res: 2560x2490 s-dpi: 120 s-size: 541x527mm (21.30x20.75")
    s-diag: 755mm (29.73")
  Monitor-1: DP-1 pos: primary,bottom model: Acer K272HUL serial: <filter>
    built: 2018 res: 2560x1440 hz: 60 dpi: 109 gamma: 1.2
    size: 598x336mm (23.54x13.23") diag: 686mm (27") ratio: 16:9 modes:
    max: 2560x1440 min: 720x400
  Monitor-2: DP-2 pos: top model: Dell P2213 serial: <filter> built: 2012
    res: 1680x1050 hz: 60 dpi: 90 gamma: 1.2 size: 473x296mm (18.62x11.65")
    diag: 558mm (22") ratio: 16:10 modes: max: 1680x1050 min: 720x400
  API: EGL v: 1.5 hw: drv: nvidia nouveau platforms: device: 0 drv: nouveau
    device: 1 drv: swrast gbm: drv: nouveau surfaceless: drv: nouveau x11:
    drv: nouveau inactive: wayland
  API: OpenGL v: 4.5 compat-v: 4.3 vendor: mesa v: 23.3.4 glx-v: 1.4
    direct-render: yes renderer: NVD9 device-ID: 10de:107d memory: 474.6 MiB
    unified: no
#
Comment 11 Stefan Dirsch 2024-05-29 09:58:37 UTC
@ Felix seems you have two gfx cards with 10de:107d and 10de:0a65. @Martin has 10de:10d8

@Felix you happen to have nvidia G04/390.157 driver installed, which apparently still supports you cards. But it no longer supports Martin's card I checked.
Comment 12 Felix Miata 2024-05-29 13:36:45 UTC
I have three more old NVidia cards connected to dual displays via multiple GPU connectors, but only two so far with 15.6 - both working as expected - and none matching OP's PCI-ID. I've never installed proprietary graphics drivers for Linux on any computer I own.

# inxi -CMSz --vs --hostname
inxi 3.3.34-00 (2024-04-13)
System:
  Host: big41 Kernel: 6.4.0-150600.21-default arch: x86_64 bits: 64
  Desktop: KDE v: 3.5.10 Distro: openSUSE Leap 15.6
Machine:
  Type: Desktop Mobo: BIOSTAR model: T41 HD serial: N/A
    BIOS: American Megatrends v: 080015 date: 09/22/2009
CPU:
  Info: dual core model: Intel Core2 Duo E7600 bits: 64 type: MCP cache:
    L2: 3 MiB
  Speed (MHz): avg: 1603 min/max: 1603/3066 cores: 1: 1603 2: 1603
# inxi -Gaz
Graphics:
  Device-1: NVIDIA G84 [GeForce 8600 GT] vendor: XFX Pine driver: nouveau
    v: kernel non-free: series: 340.xx status: legacy (EOL~2019-12-xx) last:
    release: 340.108 kernel: 5.4 xorg: 1.20 arch: Tesla process: 40-80nm
    built: 2006-2013 pcie: gen: 1 speed: 2.5 GT/s lanes: 16 ports:
    active: DVI-I-1,DVI-I-2 empty: none bus-ID: 01:00.0 chip-ID: 10de:0402
    class-ID: 0300 temp: 67.0 C
  Display: x11 server: X.Org v: 1.21.1.11 compositor: kwin driver: X:
    loaded: modesetting unloaded: fbdev,vesa alternate: nouveau,nv,nvidia
    dri: swrast gpu: nouveau display-ID: :0 screens: 1
  Screen-1: 0 s-res: 3600x1200 s-dpi: 120 s-size: 762x254mm (30.00x10.00")
    s-diag: 803mm (31.62")
  Monitor-1: DVI-I-1 pos: primary,left model: NEC EA243WM serial: <filter>
    built: 2011 res: 1920x1200 hz: 60 dpi: 94 gamma: 1.2
    size: 519x324mm (20.43x12.76") diag: 612mm (24.1") ratio: 16:10 modes:
    max: 1920x1200 min: 640x480
  Monitor-2: DVI-I-2 pos: right model: Dell P2213 serial: <filter>
    built: 2012 res: 1680x1050 hz: 60 dpi: 90 gamma: 1.2
    size: 473x296mm (18.62x11.65") diag: 558mm (22") ratio: 16:10 modes:
    max: 1680x1050 min: 720x400
  API: EGL v: 1.5 platforms: device: 1 drv: swrast gbm: drv: kms_swrast
    surfaceless: drv: swrast x11: drv: swrast inactive: wayland,device-0
  API: OpenGL v: 4.5 vendor: mesa v: 23.3.4 glx-v: 1.4 direct-render: yes
    renderer: llvmpipe (LLVM 17.0.6 128 bits) device-ID: ffffffff:ffffffff
    memory: 3.74 GiB unified: yes
  API: Vulkan v: 1.3.275 layers: 1 device: 0 type: cpu name: llvmpipe (LLVM
    17.0.6 128 bits) driver: N/A device-ID: 10005:0000 surfaces: xcb,xlib
#
# inxi -CMSz --vs --hostname
inxi 3.3.34-00 (2024-04-13)
System:
  Host: gb970 Kernel: 6.4.0-150600.21-default arch: x86_64 bits: 64
  Desktop: TDE (Trinity) v: R14.1.2 Distro: openSUSE Leap 15.6
Machine:
  Type: Desktop Mobo: Gigabyte model: GA-970A-D3 serial: N/A BIOS: Award
    v: F10 date: 05/30/2012
CPU:
  Info: quad core model: AMD Phenom II X4 965 bits: 64 type: MCP cache:
    L2: 2 MiB
  Speed (MHz): avg: 3423 min/max: N/A cores: 1: 3423 2: 3423 3: 3423 4: 3423
# inxi -Gaz
Graphics:
  Device-1: NVIDIA GF108 [GeForce GT 630] vendor: Gigabyte driver: nouveau
    v: kernel non-free: series: 390.xx+ status: legacy (EOL~2022-11-22) last:
    release: 390.157 kernel: 6.0 xorg: 1.21 arch: Fermi code: GF1xx
    process: 40/28nm built: 2010-2016 pcie: gen: 1 speed: 2.5 GT/s lanes: 16
    ports: active: DVI-I-1,HDMI-A-1 empty: VGA-1 bus-ID: 01:00.0
    chip-ID: 10de:0f00 class-ID: 0300 temp: 48.0 C
  Display: x11 server: X.Org v: 1.21.1.11 compositor: Twin v: 3.0 driver: X:
    loaded: modesetting unloaded: fbdev,vesa alternate: nouveau,nv,nvidia
    dri: nouveau gpu: nouveau display-ID: :0 screens: 1
  Screen-1: 0 s-res: 3600x1200 s-dpi: 120 s-size: 762x254mm (30.00x10.00")
    s-diag: 803mm (31.62")
  Monitor-1: DVI-I-1 pos: right model: Dell P2213 serial: <filter>
    built: 2012 res: 1680x1050 hz: 60 dpi: 90 gamma: 1.2
    size: 473x296mm (18.62x11.65") diag: 558mm (22") ratio: 16:10 modes:
    max: 1680x1050 min: 720x400
  Monitor-2: HDMI-A-1 mapped: HDMI-1 pos: primary,left model: NEC EA243WM
    serial: <filter> built: 2011 res: 1920x1200 hz: 60 dpi: 94 gamma: 1.2
    size: 519x324mm (20.43x12.76") diag: 612mm (24.1") ratio: 16:10 modes:
    max: 1920x1200 min: 640x480
  API: EGL v: 1.5 hw: drv: nvidia nouveau platforms: device: 0 drv: nouveau
    device: 1 drv: swrast gbm: drv: nouveau surfaceless: drv: nouveau x11:
    drv: nouveau inactive: wayland
  API: OpenGL v: 4.5 compat-v: 4.3 vendor: mesa v: 23.3.4 glx-v: 1.4
    direct-render: yes renderer: NVC1 device-ID: 10de:0f00 memory: 982.4 MiB
    unified: no
  API: Vulkan v: 1.3.275 layers: 3 device: 0 type: cpu name: llvmpipe (LLVM
    17.0.6 128 bits) driver: N/A device-ID: 10005:0000 surfaces: xcb,xlib
#
Comment 13 Martin Schmidt 2024-06-06 12:45:10 UTC
Dear all, back to some minor extend ...

with xf86-video-nouveau installed:

- booting (advanced options for 15.6) with kernel 5.14 works well on the first screen and ends up in the desired 2-screen environment.
- booting (advanced options for 15.6) with kernel 6.4 switches to the second screen after "loading intial ramdisk". The first screen goes to sleep forever. The login screen is blured, but password input works anyway - just guessing where to type. Then a normal 1-screen kde comes up.

I had deinstalled xf86-video-nouveau. Doing so I did not get a login screen at all, when booting kernel 6.4. The system was accessable remotely via ssh, so I reinstalled this module with zypper this way. 

Deinstalling also Mesa-dri-nouveau makes the system almost irresponsive, some fallback using a small graphics device in the memory seems to do a very slow job. I reinstalled this module.  

modeprobe -D nouveau

shows different results for both kernels.
insmod /lib/modules/5.14.21-150500.55.59-default/kernel/drivers/video/fbdev/core/fb_sys_fops.ko.zst 
insmod /lib/modules/5.14.21-150500.55.59-default/kernel/drivers/video/fbdev/core/sysimgblt.ko.zst 
insmod /lib/modules/5.14.21-150500.55.59-default/kernel/drivers/video/fbdev/core/sysfillrect.ko.zst 
insmod /lib/modules/5.14.21-150500.55.59-default/kernel/drivers/video/fbdev/core/syscopyarea.ko.zst 

is missing with 6.4. Also
insmod /lib/modules/5.14.21-150500.55.59-default/kernel/drivers/gpu/drm/drm.ko.zst 

does not have a 6.4-counterpart.
I am producing the dmesg - logs now and will upload them soon.
Cheers
Comment 14 Martin Schmidt 2024-06-06 14:26:02 UTC
Created attachment 875358 [details]
dmesg for kernel 5.14
Comment 15 Martin Schmidt 2024-06-06 14:26:43 UTC
Created attachment 875359 [details]
dmesg for kernel 6.4
Comment 16 Felix Miata 2024-06-06 16:13:12 UTC
(In reply to Stefan Dirsch from comment #11)
> @ Felix seems you have two gfx cards with 10de:107d and 10de:0a65. @Martin
> has 10de:10d8

My 107d and Martin's are both Teslas. According to Wikipedia, his is a Quadro, while my 107d is a GeForce. I have also two other Teslas, both GeForce, one 10de:06e4, for which I haven't installed 15.6 yet, the other 10de:0402, which does not reproduce this:
# inxi -GSxxz
System:
  Kernel: 6.4.0-150600.21-default arch: x86_64 bits: 64 compiler: gcc v: 7.5.0
  Desktop: KDE v: 3.5.10 tk: Qt v: 3.3.8c wm: kwin dm: 1: KDM 2: XDM
    Distro: openSUSE Leap 15.6
Graphics:
  Device-1: NVIDIA G84 [GeForce 8600 GT] vendor: XFX Pine driver: nouveau
    v: kernel arch: Tesla pcie: speed: 2.5 GT/s lanes: 16 ports:
    active: DVI-I-1,DVI-I-2 empty: none bus-ID: 01:00.0 chip-ID: 10de:0402
    temp: 72.0 C
  Display: x11 server: X.Org v: 1.21.1.11 compositor: kwin driver: X:
    loaded: modesetting unloaded: fbdev,vesa alternate: nouveau,nv,nvidia
    dri: swrast gpu: nouveau display-ID: :0 screens: 1
  Screen-1: 0 s-res: 3600x1200 s-dpi: 120
  Monitor-1: DVI-I-1 pos: primary,left model: NEC EA243WM res: 1920x1200
    dpi: 94 diag: 612mm (24.1")
  Monitor-2: DVI-I-2 pos: right model: Dell P2213 res: 1680x1050 dpi: 90
    diag: 558mm (22")
  API: EGL v: 1.5 platforms: device: 1 drv: swrast gbm: drv: kms_swrast
    surfaceless: drv: swrast x11: drv: swrast inactive: wayland,device-0
  API: OpenGL v: 4.5 vendor: mesa v: 23.3.4 glx-v: 1.4 direct-render: yes
    renderer: llvmpipe (LLVM 17.0.6 128 bits) device-ID: ffffffff:ffffffff
  API: Vulkan v: 1.3.275 surfaces: xcb,xlib device: 0 type: cpu driver: N/A
    device-ID: 10005:0000
#
None of mine have DMS-59 output though.
 
> @Felix you happen to have nvidia G04/390.157 driver installed, which
> apparently still supports you cards. But it no longer supports Martin's card
> I checked.

I don't understand this. My own NVidias have never while in my possession been used with proprietary drivers. Mine normally run on nouveau kernel & drm, modesetting DIX, as inxi suggests.

Does removing the DMS-59 and reinstalling, or a different Y-cable change anything?

Is this related?:
https://gitlab.freedesktop.org/drm/nouveau/-/issues/255
Comment 17 Stefan Dirsch 2024-06-06 18:23:27 UTC
(In reply to Felix Miata from comment #16)
> > @Felix you happen to have nvidia G04/390.157 driver installed, which
> > apparently still supports you cards. But it no longer supports Martin's card
> > I checked.
> 
> I don't understand this. My own NVidias have never while in my possession
> been used with proprietary drivers. Mine normally run on nouveau kernel &
> drm, modesetting DIX, as inxi suggests.

Because of this in comment#12

[...]
390.xx+ status: legacy (EOL~2022-11-22) last: release: 390.157
[....]

This is clearly G04 proprietary driver still found on your system somehow by inxi.
Comment 18 Felix Miata 2024-06-06 19:07:06 UTC
(In reply to Stefan Dirsch from comment #17)
> (In reply to Felix Miata from comment #16)
> > I don't understand this. My own NVidias have never while in my possession
> > been used with proprietary drivers. Mine normally run on nouveau kernel &
> > drm, modesetting DIX, as inxi suggests.
 
> Because of this in comment#12
 
> [...]
> 390.xx+ status: legacy (EOL~2022-11-22) last: release: 390.157
> [....]

Had I understood that OP's device was a Tesla, I would not have included the "Host: gb970" inxi output, as that and its 390.157 are Fermi, not Tesla.
 
> This is clearly G04 proprietary driver still found on your system somehow by
> inxi.

What it is reporting there is the proprietary driver version applicable to that device ID. It is part of the extended information option "a" of inxi I use when desiring to show maximum detail of the Display, Screen & Monitor information that follows "G"'s device information.

"driver: nouveau" on the "Device-1" line shows the loaded kernel module. In comment #16 I used inxi -GSxxz instead of -GSaz, where you can note that additional information was excluded.

# inxi -G | wc -lm
      9     475
# inxi -Gx | wc -lm
     10     576
# inxi -Gxx | wc -lm
     17     881
# inxi -Gxxx | wc -lm
     19    1076
# inxi -Ga | wc -lm
     21    1300
Comment 19 Stefan Dirsch 2024-06-06 21:13:26 UTC
Ok. So inxi has some driver database, which even has information about legacy proprietary drivers from nvidia. I wasn't aware of this. Whatever, this is completely off-topic. Meanwhile Martin provided the kernel logs.
Comment 20 Martin Schmidt 2024-06-06 21:42:20 UTC
Some additional information:
meanwhile in the main repository repo-oss there is a new kernel version. I have also installed this. Now I have:
6.4.0-150600.17.2
6.4.0-150600.21.2

17.2 comes from @System, Vendor Suse LLC.

Booting from the 21.2-kernel ends up with single screen system, now on the first screen, but with only 640x480 resolution. 

Using yast to scan through all installed packages found with keyword "kernel" reveals more related parts (kernel-option etc.) that have a different version number in the last digit, when insatlled from the repository repo-oss. I upgraded all installed packages to the latest version, always from the repo-oss repository. Rebooting again leads back to the old behaviour as reported - single but highly resolving display 2. 

For me it is confusing that booting with identical driver packages but different kernels leads to varying display configurations. 

I just do not know how kernel and drivers interact. Could it be that the problem is a new kernel that cannot make proper use from the driver? May be the choice of repositories that I've got is not appropriate? It is confusing for me that modprebe -D nouveau gives very different result with kernal 5.15 and 6.4. 
Cheers
Comment 21 Felix Miata 2024-06-06 23:53:46 UTC
(In reply to Martin Schmidt from comment #20)

> Booting from the 21.2-kernel ends up with single screen system, now on the
> first screen, but with only 640x480 resolution. 

These two are behavior common to nouveau kernel module needed but not loaded.

@Martin, are you using only standard repos?:
# | Alias        | Enabled | GPG Check | URI
--+--------------+---------+-----------+------------------------------------------------------------------
1 | NonOSS       | Yes     | (r ) Yes  | http://download.opensuse.org/distribution/leap/15.6/repo/non-oss/
2 | OSS          | Yes     | (r ) Yes  | http://download.opensuse.org/distribution/leap/15.6/repo/oss/
3 | Update       | Yes     | (r ) Yes  | http://download.opensuse.org/update/leap/15.6/oss/
4 | UpdateBP     | Yes     | (r ) Yes  | http://download.opensuse.org/update/leap/15.6/backports/
5 | UpdateNonOSS | Yes     | (r ) Yes  | http://download.opensuse.org/update/leap/15.6/non-oss/
6 | UpdateSLE    | Yes     | (r ) Yes  | http://download.opensuse.org/update/leap/15.6/sle/
Comment 22 Martin Schmidt 2024-06-07 06:24:13 UTC
Good morning. I have:
---+---------------------------------------+---------+-----------+--------
 1 | NVIDIA:repo-non-free                  | Yes     | (r ) Yes  | Yes
 2 | ftp.gwdg.de-openSUSE_Leap_$releasever | No      | ----	 | ----
 3 | openSUSE:repo-non-oss                 | Yes     | (r ) Yes  | Yes
 4 | openSUSE:repo-non-oss-debug           | No      | ----	 | ----
 5 | openSUSE:repo-openh264                | Yes     | (r ) Yes  | Yes
 6 | openSUSE:repo-oss                     | Yes     | (r ) Yes  | Yes
 7 | openSUSE:repo-oss-debug               | No      | ----	 | ----
 8 | openSUSE:repo-oss-source              | No      | ----	 | ----
 9 | openSUSE:update-backports             | Yes     | (r ) Yes  | Yes
10 | openSUSE:update-backports-debug       | No      | ----	 | ----
11 | openSUSE:update-non-oss               | Yes     | (r ) Yes  | Yes
12 | openSUSE:update-non-oss-debug         | No      | ----	 | ----
13 | openSUSE:update-oss                   | Yes     | (r ) Yes  | Yes
14 | openSUSE:update-oss-debug             | No      | ----	 | ----
15 | openSUSE:update-sle                   | Yes     | (r ) Yes  | Yes
16 | openSUSE:update-sle-debug             | No      | ----	 | ----
17 | repo-backports-debug-update           | No      | ----	 | ----
18 | repo-backports-update                 | Yes     | (r ) Yes  | Yes
19 | repo-sle-debug-update                 | No      | ----	 | ----
20 | repo-sle-update                       | Yes     | (r ) Yes  | Yes
21 | rpm                                   | Yes     | ( p) Yes  | No
22 | x86_64                                | No      | ----	 | ----

I have disabled the Nvidia repository several times, but got the message that is is most probably enabled again by the "Nvidia Service" - which reveals as true. I could not find such a service. Anyway, there are no Nvidia drivers installed, as possibly seen from the uploaded logs. I never added the other repositories by hand. They where proposed during the upgrade step in earlier system versions (14.x). rpm is empty. 

Here is, what I've got for a fresh system using the web installer on another (virtual) machine: 
 1 | openSUSE-Leap-15.6-1        | Yes     | (r ) Yes  | Yes
 2 | repo-backports-debug-update | No	   | ----      | ----
 3 | repo-backports-update       | Yes     | (r ) Yes  | Yes
 4 | repo-debug                  | No	   | ----      | ----
 5 | repo-debug-non-oss          | No	   | ----      | ----
 6 | repo-debug-update           | No	   | ----      | ----
 7 | repo-debug-update-non-oss   | No	   | ----      | ----
 8 | repo-non-oss                | Yes     | (r ) Yes  | Yes
 9 | repo-openh264               | Yes     | (r ) Yes  | Yes
10 | repo-oss                    | Yes     | (r ) Yes  | Yes
11 | repo-sle-debug-update       | No	   | ----      | ----
12 | repo-sle-update             | Yes     | (r ) Yes  | Yes
13 | repo-source                 | No	   | ----      | ----
14 | repo-update                 | Yes     | (r ) Yes  | Yes
15 | repo-update-non-oss         | Yes     | (r ) Yes  | Yes

The missing openSUSE-Leap-15.6-1 in the setup in question above is confusing me. But historically, the system that is not working with the latest kernel was installed from dvd with version 42.1 or so and upgraded many times. May be it is time to start from scratch?
Comment 23 Felix Miata 2024-06-07 08:47:19 UTC
@Martin: repo names mean little. They are free to be modified to the admin's liking, which I do as a matter of course. It's the URL that matters most. To get output including it requires either modifying /etc/zypp/zypper.conf, or including the -u switch when running zypper lr. Otherwise, the repo name is duplicated, while the URL is omitted. To get the relevant output seen in comment #21 I have the option "repoListColumns = au" set in zypper.conf.

To eliminate NVidia repos, openSUSE-repos-Leap and openSUSE-repos-Leap-NVIDIA need to be removed and locked. That removes the repos service. If an admin does that, he must manually ensure all required repos exist and are enabled, which is what comment #21 shows is the state of that installation.

That said, I see nothing among your repos' names that /suggests/ a source that /might/ produce your trouble, with the possible exception of openSUSE-repos-Leap-NVIDIA.
Comment 24 Martin Schmidt 2024-06-07 11:06:47 UTC
I see, I update the list. For clarity only the enabled repositories. 
 repo-non-oss (15.6)                                                                         | Yes     | http://cdn.opensuse.org/distribution/leap/15.6/repo/non-oss
 repo-oss (15.6)                                                                             | Yes     | http://cdn.opensuse.org/distribution/leap/15.6/repo/oss
 update-backports (15.6)                                                                     | Yes     | http://cdn.opensuse.org/update/leap/15.6/backports
 update-non-oss (15.6)                                                                       | Yes     | http://cdn.opensuse.org/update/leap/15.6/non-oss
 update-oss (15.6)                                                                           | Yes     | http://cdn.opensuse.org/update/leap/15.6/oss
 update-sle (15.6)                                                                           | Yes     | http://cdn.opensuse.org/update/leap/15.6/sle
 repo-openh264 (15.6)                                                                        | Yes     | http://codecs.opensuse.org/openh264/openSUSE_Leap
 Update repository of openSUSE Backports                                                     | Yes     | http://download.opensuse.org/update/leap/15.6/backports/
 Update repository with updates from SUSE Linux Enterprise 15                                | Yes     | http://download.opensuse.org/update/leap/15.6/sle/
 repo-non-free (15.6)                                                                        | Yes     | https://download.nvidia.com/opensuse/leap/15.6

Cheers
Comment 25 Martin Schmidt 2024-06-14 17:38:16 UTC
Hi, I tried to continue. A brandnew system setup reveals the same shortcoming - only one display. Now I tried to disable nouveau. I downloaded the nvidia driver installer selected for my graphics card. After booting with nomodeprobe and blacklisting the nouveau-driver (this was done by the nvidia installer) I get a 640x480 display. Now I tried to build the nvidia driver into the kernel. After installing the kernel souces and devel-files, the build fails, since <stdarg.h> is not found. After linking <path>/linux/stdarg.h to <path>/stdarg.h this error is gone. Now <asm/kmap_types.h> is missing. Indeed, this is not among the kernel source files. For Redhat, I find in some forum that this file is removed from the kernel distribution. May be the kernel driver installer does not work for the 6.4-kernel? Also on a 5.14-machine the missing include file is not there. But here a more recent nvidia-driver is installed for the gpus. 

In summery I am tempted to give up and find another graphics card. 

Thanks to all for their input and help. 

Cheers,
Martin