i586 with kernel 5.16 - non-blinking cursor after "loading initial RAM disk"

[mrmazda]

I found a modestly newer than OP's Quadro that works:

# xrandr | egrep 'onnect|creen|\*' | grep -v disconn | sort -r
Screen 0: minimum 320 x 200, current 1680 x 1050, maximum 4096 x 4096
DVI-I-1 connected primary 1680x1050+0+0 (normal left inverted right x axis y axis) 474mm x 296mm
   1680x1050     59.97*+  74.89
# inxi -CMSy
System:
  Host: gx270 Kernel: 5.16.0-1-default i686 bits: 32 Desktop: KDE 3.5.10
    Distro: openSUSE Tumbleweed 20220116
Machine:
  Type: Desktop System: Dell product: OptiPlex GX270 v: N/A serial: DJS3241
  Mobo: Dell model: 0Y1057 serial: ..CN698613AD0BFD. BIOS: Dell v: A07
    date: 06/26/2006
CPU:
  Info: single core model: Intel Pentium 4 bits: 32 type: MT cache: 512 KiB
    note: check
  Speed (MHz): avg: 2793 min/max: N/A cores: 1: 2793 2: 2793
# inxi -Gayz
Graphics:
  Device-1: NVIDIA NV34GL [Quadro NVS 280 PCI] driver: nouveau v: kernel
    bus-ID: 01:07.0 chip-ID: 10de:032a class-ID: 0300
  Display: x11 server: X.Org 1.21.1.3 driver: loaded: nouveau display-ID: :0
    screens: 1
  Screen-1: 0 s-res: 1680x1050 s-dpi: 108 s-size: 395x246mm (15.6x9.7")
    s-diag: 465mm (18.3")
  Monitor-1: DVI-I-1 res: 1680x1050 hz: 60 dpi: 90
    size: 474x296mm (18.7x11.7") diag: 559mm (22")
  OpenGL: renderer: NV34 v: 1.5 Mesa 21.3.4 direct render: Yes
#

It shows DVI, but it's going through a required adapter to the display's VGA input. It also works using modesetting display driver instead of nouveau.

[mrmazda]

The Precision M60 has an NVIDIA Quadro FX 700go graphics (driver nouveau according to inxi), it's the xfce desktop I use, LightDM GTK+ greeter.

Graphics:
  Device-1: NVIDIA NV31GLM [Quadro FX Go700] vendor: Dell driver: nouveau
    v: kernel bus-ID: 01:00.0 chip-ID: 10de:031c class-ID: 0300
  Display: server: X.org 1.21.1.3 compositor: xfwm4 driver:
    loaded: modesetting unloaded: fbdev,vesa alternate: nouveau,nv,nvidia
    tty: 130x48
  Message: Advanced graphics data unavailable in console for root.

I suppose you may have a LightDM problem. I even have an ancient NVidia Vanta running on a Pentium 4 with nouveau DDX display driver.

One characteristic all mine have in common is KDM3 for display manager, and on these old CPUs, X is seriously slow to start.

# systemd-analyze critical-chain
...
graphical.target @1min 56.879s
`-display-manager.service @1min 41.298s +15.578s
  `-systemd-logind.service @1min 33.814s +7.478s
    `-basic.target @1min 33.282s
      `-sockets.target @1min 33.281s
        `-telnet.socket @1min 33.281s
          `-sysinit.target @1min 32.339s
            `-systemd-timesyncd.service @1min 31.653s +684ms
              `-systemd-tmpfiles-setup.service @1min 23.657s +7.888s
                `-systemd-journal-flush.service @13.555s +1min 10.070s
                  `-systemd-journald.service @12.576s +964ms
                    `-haveged.service @1min 3.964s
                      `-systemd-journald.socket
                        `-system.slice
                          `--.slice

How long are you allowing for the greeter to show up?

Have you been able to find Xorg.0.log generated when booting with 5.16.0?

[suse_rasputin]

I tried the log from Xorg, but it's definitely from the last successfull boot of 5.15.8, not the botched 5.16 boot (I can see in the log the shutdown after standing over night...). The boot process of 5.16 is stopping (printout on screen) after some 6 seconds (after deleting kernel-vanilla and kernel-default-base and re-installing 5.16.0-1 in Yast).

In the dmesg of a successfull 5.15.8 boot the next entry after the last one I see with 5.16 is "Loading kernel". So the system is failing long before any graphics problem could stop it imo...

[suse_rasputin]

This is what boot looks like (kernel-default 5.16.0-1.1 ) after a zypper dup this morning

URL: https://paste.opensuse.org/049957eb


In dmesg of usccessful boot (kernel 5.15.8) the next line is

"write protecting kernel text and read-oly data : 18328k"

[mrmazda]

In that image it looks to me like it could be the KMS engagement point, where faulty engagement could simply halt terminal I/O while the rest of the system keeps on normally. Have you re-attempted remote login to it? Is the ethernet connector still not lighting up? Can you try booting the old kernel and switching to a different DM with update-alternatives --config default-displaymanager. Does 5.16 boot to a login prompt if you append nomodeset and 3 to the linu line using the E key at the grub menu?

[suse_rasputin]

Update: kernel 5.16.1-1 did not resolve, same stop in boot after 6 sec. No DHCP request whatsoever after 5 min. The machine is used for network maintenance, as long as I have a booting kernel I'm fine, no way to start debugging for hours...

[suse_rasputin]

Update 2: Installed todays updates and added "nomodeset" on reboot.

Now I have here

https://paste.opensuse.org/6baf83bd

sitting for 10 min without any activity of HDD ore something else (network) going forward. Needless to say that kerne 5.15.8 boots just fine.

Any ideas?

[suse_rasputin]

I added now "nomodeset" to boot options in Yast and rebooted. Then the boot stops with the old message after 6.3 sec. If I enter the grub editor (pressing "e") and look, there is the nomodeset. And after pressing F10 for booting the boot process stops at the new end point, i.e. after 11-12 sec with "checking file system", as in the pic from some minutes ago.

Does this make sense at all?

[suse_rasputin]

5.15.2 no improvement. Without editing grub entry: stop after 6-7 sec. When entering grub "e" and remove "nomodeset" -> very fast messages for 11-12 sec then screen turns black, fan running, no activity (screen, network) throughout the next 20 min or so...

[suse_rasputin]

5.15.4 no joy, same error after 6.77 sec. 5.14.8 still doing fine.