LightDM Login Loop post-upgrade; /dev/dri/card0 is /dev/dri/card1? xrdb cannot open display? No screens found?

Hello,

I have here a complete write-up that I gave to the EndeavorOS Discord on the problem, all the information I have on it, and all the things I’ve tried and any pertinent information.

Last night - haven’t upgraded Endeavor in like 5 months;
Decide to go ahead and update
Boot into it, run the mirror list ranker, then do the update from the welcome screen using eos-update
Everything goes normally other than the lag with the AURs, AND I think I MAY have seen a broken-pipe error, I can’t remember

On next reboot after the upgrade:

  • LightDM greeter appears as normal, Boot up appears to go as normal; No errors in the bootup or when I see the services starting when booting linux before getting to the greeter(Note: My DE is XFCE4)
  • I type my password
  • I hit enter
  • The screen goes black a moment; then it goes right back to the greeter. I can’t launch an X session.

Things that seem to be wrong

  • I’ve noticed that I have an error(I can’t recall if I saw this in Journalctl, or the Xorg logs) that reads:
    open /dev/dri/card0 no such file or directory'
    I went and did ls /dev/dri - and I see that I have three folders in there - There is no card 0, but there is a card 1 folder, and querying its status shows that it is connected.
  • lspci -k shows that my AMD GPU is being detected and is loading the appropriate driver(amdgpu).
  • I cannot remember if it was in journalctl or xorg that I saw these in, but I also recall(and wrote in my notes last night):
environment variable $XAUTHORITY is not set
environment variable $DISPLAY is not set
xrdb cannot open display ''
  • dmesg shows an error that says failed to setup vendor infoframe on connector DP-2 -22

  • There is also this, Fatal server error: (EE) no screens found (EE)

Things I’ve tried (havent tried anything from live usb yet)

  • Re-doing the upgrade via pacman -Syu
  • Reinstalling the drivers
  • Reinstalling linux firmware (note: I seem to only have linux-firmware-amdgpu installed)
  • Reinstalling kernels
  • Logging in via TTY and doing startx from a TTY - X does start, but - either because it’s using the default config, or because something is broken - it simply leads to a black screen with four white terminals
  • Verifying ownership and permissions of the .Xauthority file
  • Making an Xorg config for my card as followed:
Section "Device"
  Identifier "amdgpu"
  Driver "amdgpu"
  BusID # my id (yes i put my id in, i didnt leave it blank >;c)
EndSection```
(simply lead to a new error that i cant remember)
  • I did modprobe -r amdgpu and that made my screen go black(duh), so I waited a second - then typed modprobe amdgpu and that made the display come back again, and it also made /dev/dri/card1 switch to /dev/dri/card0, but it still did not resolve the login loop, and it reverted back to /dev/dri/card1 after rebooting

Please help me ;-;

sudo pacman -R linux-firmware 
sudo pacman -Syu linux-firmware

I already tried this - it changed nothing; Same errors.

In the meantime, I have made a breakthrough in this problem:

  1. I logged into a TTY terminal

  2. I did xinit /usr/bin/xterm – :0, which brought up a black screen with a white terminal

  3. From that white terminal I did startxfce4 - which created an unthemed xfce4

    So… I don’t know what that means :confused:

The core of the issue still remains - I can’t get past the login screen, and using an unthemed XFCE4 launched via xterm isn’t really a good workaround for me lol.

is this AMD GPU?
BTW welcome at the forum :enos:

Yep, RX 570.

And thank you

Seems its not GPU drive rissue as you can start xfce4 graphically

So looks like something with lightdm is failing.

you could try at first reinstralling needed xfce4 packages (including lightdm)
eos-packagelist --install "XFCE4-Desktop"
And reboot
if this is not helping try using a simple Desktop login Manager like ly:
sudo pacman -Sy ly and exchange DMs:
sudo systemctl -f enable ly
and reboot

Will reboot and give that try, thanks :+1:

the error somehow looks like GOU number has changed or you have two gpu installed? ma yits hybrid system with in cpu gpu and decent one? (notebook)

You could show inxi -Gaz

Nope, I only have one GPU installed. No mobo graphics.

The only hardware change I’ve done pertaining to graphics is the removal of one monitor out of three in the past few months, leaving me with just two, but I don’t know why that would be doing this suddenly out of nowhere.

Anyways,

I ran eos-packagelist --install “XFCE4-Desktop”,
The only thing it installed was Discord…

BUT, I also remembered that last night, while troubleshooting things, I actually also did
sudo mv ~/.config/xfce4 ~/.config/xfce4.bak
So I put that back and rebooted, and now I am seeing a new error (at last) and it seems to be a very positive sign:

Whenever I do xterm and/or startxfce4 - I see it complain that Display 0 is already in use. Hell, in order to even get the xterm command to work I had to stop the lightdm service(but something still is using it?)

Whatever the case, it gave me my theme back at least, but I still cannot get lightdm to get me into the session.

Everything else seems to be in order though at least, which is a sigh of relief, and I can at least WORK now, which is very nice.

So even though the problem is still there this is a much better leg-up, thank you lmao.

I guess up next I can try a different greeter. I dunno. I really liked LightDM, I had it setup just the way I like it.

Here is the output of inxi -Gaz (really nice that I can finally post these now)

Graphics:
  Device-1: Advanced Micro Devices [AMD/ATI] Ellesmere [Radeon RX
    470/480/570/570X/580/580X/590] vendor: XFX Pine driver: amdgpu v: kernel
    arch: GCN-4 code: Arctic Islands process: GF 14nm built: 2016-20 pcie:
    gen: 1 speed: 2.5 GT/s lanes: 16 link-max: gen: 3 speed: 8 GT/s ports:
    active: DP-2,HDMI-A-1 empty: DP-1,DP-3,DVI-D-1 bus-ID: 26:00.0
    chip-ID: 1002:67df class-ID: 0300 temp: 56.0 C
  Display: unspecified server: X.Org v: 21.1.18 compositor: xfwm4 v: 4.20.0
    driver: X: loaded: amdgpu unloaded: modesetting,radeon alternate: fbdev,vesa
    dri: radeonsi gpu: amdgpu display-ID: :0.0 screens: 1
  Screen-1: 0 s-res: 3840x1080 s-dpi: 96 s-size: 1017x286mm (40.04x11.26")
    s-diag: 1056mm (41.59")
  Monitor-1: DP-2 mapped: DisplayPort-1 pos: left model: 100002480
    serial: <filter> built: 2019 res: mode: 1920x1080 hz: 60 scale: 100% (1)
    dpi: 103 gamma: 1.2 size: 474x296mm (18.66x11.65") diag: 552mm (21.7")
    ratio: 16:10 modes: max: 1920x1080 min: 720x400
  Monitor-2: HDMI-A-1 mapped: HDMI-A-0 pos: primary,right
    model: LG (GoldStar) FULL HD built: 2016 res: 1920x1080 dpi: 102 gamma: 1.2
    size: 480x270mm (18.9x10.63") diag: 551mm (21.7") ratio: 16:9 modes:
    max: 1920x1080 min: 720x400
  API: EGL v: 1.5 hw: drv: amd radeonsi platforms: device: 1 drv: swrast
    gbm: drv: kms_swrast surfaceless: drv: swrast x11: drv: radeonsi
    inactive: wayland,device-0
  API: OpenGL v: 4.6 compat-v: 4.5 vendor: amd mesa v: 25.2.5-arch1.2
    glx-v: 1.4 direct-render: yes renderer: AMD Radeon RX 570 Series (radeonsi
    polaris10 ACO DRM 3.64 6.17.5-arch1-1) device-ID: 1002:67df
    memory: 7.81 GiB unified: no
  API: Vulkan Message: No Vulkan data available.
  Info: Tools: api: clinfo, eglinfo, glxinfo, vulkaninfo
    de: xfce4-display-settings gpu: amdgpu_top,corectrl
    x11: xdpyinfo, xprop, xrandr

I can see the problem right now —

Display: unspecified

Maybe this guide can help you out.

xfce4 uses display names to set workspaces and wallpapers.. and lightdm too could use display names to adjust the default used display for the password prompt e.t.c

“server unspecified” should be harmless as its running xserver, may running command without a proper Desktop session (as of using xtrem)

you could fully disable it till you find out what is wrong with lightdm: sudo systemctl disable lightdm

Also possible to backup lightdm settings and try to run it without the settings..

sudo mv /etc/lightdm /etc/lightdm.bak
And reinstall EndeavourOS lightdm theme: sudo pacman -S eos-lightdm-slick-theme this will bring back slick theme and setup from defaults..

sudo systemctl start lightdm (or restart if not disabled)

Will give it a try tomorrow; Thanks!

I wound up procrastinating a bit, but eventually I got around to it today.

Tried all of these and unfortunately it still resulted in the same problem, so I decided to just give up on LightDM entirely.

I disabled the service, stopped it, then removed it and installed SDDM and now I can login in with that lol.

So…
Problem solved!!!

(Thank you everybody who helped me try and pin the issue down, I just can’t be bothered anymore lmao… I even ran the commands up above @keescase had in that guide and did not find anything at all that could even remotely be linked beyond imagination to this issue. ;-:wink:

1 Like

dunno how to close the thread for just generally solved so I guess I will just mark that as the solution lol.

I did out of mild curiosity preserve the logs for Xorg and LightDM though so I can look at them later(much later) to see what the hell might have been going on. I must have missed something in one of those. I am pretty sure I scoured them though.

1 Like

This topic was automatically closed 2 days after the last reply. New replies are no longer allowed.