You are not logged in.
Please accept my apologies, if this has been placed in the wrong subforum...
This morning, LightDM would not load properly anymore. It would be stuck in a loop while trying to automatically login my default user to the XFCE desktop.
More precisely, LightDM/X11 would attemt to start (the screen turning completely black), fail immediately and show the (empty) virtual console of tty7 with a flashing cursor in the top left corner. A few seconds later LightDM would attempt to start again. This cycle would repeat indefinitely.
Using any other virtual console was impossible, since every time LightDM attempted to start, the screen would immediately jump back to tty7.
At the same time, it was also impossible to login via SSH (the user would be logged off immediately), so that I had to revert to an ArchLinux32 live USB image to rescue my system. (Remark: Copying files via scp however worked fine.)
LightDM shows the following error messages:
● lightdm.service - Light Display Manager
Loaded: loaded (/usr/lib/systemd/system/lightdm.service; enabled; vendor preset: disabled)
Active: active (running) since Mon 2021-10-25 08:56:59 CEST; 671ms ago
Docs: man:lightdm(1)
Main PID: 792 (lightdm)
Tasks: 6 (limit: 4915)
Memory: 17.1M
CPU: 618ms
CGroup: /system.slice/lightdm.service
├─792 /usr/bin/lightdm
├─798 /usr/lib/Xorg :0 -seat seat0 -auth /run/lightdm/root/:0 -nolisten tcp vt7 -novtswitch
└─807 lightdm --session-child 12 15
Oct 25 08:56:58 archlinux32 systemd[1]: Starting Light Display Manager...
Oct 25 08:56:59 archlinux32 systemd[1]: Started Light Display Manager.
Oct 25 08:56:59 archlinux32 lightdm[807]: pam_succeed_if(lightdm-autologin:auth): requirement "user ingroup autologin" was met by user "marc"
Oct 25 08:56:59 archlinux32 lightdm[807]: pam_systemd_home(lightdm-autologin:account): systemd-homed is not available: Unit dbus-org.freedesktop.home1.service not found.
Oct 25 08:57:00 archlinux32 lightdm[807]: PAM pam_putenv: no variable indicated
Oct 25 08:57:00 archlinux32 lightdm[807]: PAM pam_putenv: no variable indicated
Oct 25 08:57:00 archlinux32 lightdm[807]: PAM pam_putenv: no variable indicated
Oct 25 08:57:00 archlinux32 systemd[1]: lightdm.service: Main process exited, code=dumped, status=8/FPEAll this happened after my last upgrade, affecting the following packages:
[2021-10-24T15:59:16+0200] [ALPM] transaction started
[2021-10-24T15:59:17+0200] [ALPM] upgraded tzdata (2021c-1.0 -> 2021d-1.0)
[2021-10-24T15:59:17+0200] [ALPM] upgraded libffi (3.3-4.1 -> 3.4.2-4.0)
[2021-10-24T15:59:17+0200] [ALPM] upgraded glib2 (2.70.0-1.0 -> 2.70.0-2.5)
[2021-10-24T15:59:18+0200] [ALPM] upgraded guile (2.2.7-2.1 -> 2.2.7-2.4)
[2021-10-24T15:59:18+0200] [ALPM] upgraded libp11-kit (0.24.0-1.1 -> 0.24.0-2.3)
[2021-10-24T15:59:18+0200] [ALPM] upgraded p11-kit (0.24.0-1.1 -> 0.24.0-2.3)
[2021-10-24T15:59:33+0200] [ALPM] upgraded linux (5.13.13.arch1-1.0 -> 5.14.12.arch1-1.0)
[2021-10-24T15:59:37+0200] [ALPM] upgraded llvm-libs (12.0.1-4.0 -> 12.0.1-5.1)
[2021-10-24T15:59:39+0200] [ALPM] upgraded python (3.9.6-1.3 -> 3.9.7-2.1)
[2021-10-24T15:59:39+0200] [ALPM] upgraded wayland (1.19.0-1.1 -> 1.19.0-2.1)
[2021-10-24T15:59:39+0200] [ALPM] upgraded qt5-svg (5.15.2+kde+r7-1.1 -> 5.15.2+kde+r13-1.0)
[2021-10-24T15:59:39+0200] [ALPM] installed libsoup3 (3.0.1-1.0)
[2021-10-24T15:59:40+0200] [ALPM] upgraded tracker3 (3.2.0-1.0 -> 3.2.0-2.1)
[2021-10-24T15:59:41+0200] [ALPM] transaction completedI had to downgrade linux, wayland, glib2/libffi, p11-kit/libp11-kit and tracker3 (libsoup3 was a dependency of tracker3 and could be removed after downgrading) in order to get LightDM working again.
Can anyone confirm this issue?
Last edited by levi (2021-10-28 14:51:02)
Offline
I also get it with 'slim':
Stack trace of thread 1841:
#0 0x00000000b791a585 round_and_return (libc.so.6 + 0x3f585)
#1 0x00000000b791c574 ____strtod_l_internal (libc.so.6 + 0x41574)
#2 0x00000000b7916b60 strtod (libc.so.6 + 0x3bb60)
#3 0x00000000b7673709 n/a (libfontconfig.so.1 + 0x28709)
#4 0x00000000b7677f0f n/a (libfontconfig.so.1 + 0x2cf0f)
#5 0x00000000b753c274 n/a (libexpat.so.1 + 0xb274)
#6 0x00000000b753e43e n/a (libexpat.so.1 + 0xd43e)
#7 0x00000000b753f851 n/a (libexpat.so.1 + 0xe851)
#8 0x00000000b7537a89 XML_ParseBuffer (libexpat.so.1 + 0x6a89)
#9 0x00000000b76768e8 n/a (libfontconfig.so.1 + 0x2b8e8)
#10 0x00000000b7677135 n/a (libfontconfig.so.1 + 0x2c135)
#11 0x00000000b7677443 n/a (libfontconfig.so.1 + 0x2c443)
#12 0x00000000b7677a1d n/a (libfontconfig.so.1 + 0x2ca1d)
#13 0x00000000b753c274 n/a (libexpat.so.1 + 0xb274)
#14 0x00000000b753e43e n/a (libexpat.so.1 + 0xd43e)
#15 0x00000000b753f851 n/a (libexpat.so.1 + 0xe851)
#16 0x00000000b7537a89 XML_ParseBuffer (libexpat.so.1 + 0x6a89)
#17 0x00000000b76768e8 n/a (libfontconfig.so.1 + 0x2b8e8)
#18 0x00000000b7677135 n/a (libfontconfig.so.1 + 0x2c135)
#19 0x00000000b765f6d4 n/a (libfontconfig.so.1 + 0x146d4)
#20 0x00000000b765a9a3 n/a (libfontconfig.so.1 + 0xf9a3)
#21 0x00000000b765f9a8 FcInit (libfontconfig.so.1 + 0x149a8)
#22 0x00000000b771ea83 XftInit (libXft.so.2 + 0xda83)
#23 0x00000000b771b2a9 XftFontMatch (libXft.so.2 + 0xa2a9)
#24 0x00000000b771b5e9 XftFontOpenName (libXft.so.2 + 0xa5e9)
#25 0x00000000b7ce07e3 _ZN5PanelC2EP9_XDisplayimP3CfgRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEENS_9PanelTypeE (libslim.so.1.3.6 + 0x167e3)
#26 0x00000000004499eb _ZN3App3RunEv (slim + 0xc9eb)
#27 0x00000000004426f2 main (slim + 0x56f2)
#28 0x00000000b78f998d __libc_start_main (libc.so.6 + 0x1e98d)
#29 0x0000000000442865 _start (slim + 0x5865)
This is rounding in the middle of the C library reading fonts. hard to belileve this has anything to do with libffi.
Offline
startx says:
The XKEYBOARD keymap compiler (xkbcomp) reports:
> Warning: Could not resolve keysym XF86BrightnessAuto
> Warning: Could not resolve keysym XF86DisplayOff
> Warning: Could not resolve keysym XF86Info(EE) 15: /usr/lib/Xorg (_start+0x35) [0x526c55]
(EE)
(EE) Floating point exception at address 0xb666278e
(EE)
Fatal server error:
(EE) Caught signal 8 (Floating point exception). Server aborting
(EE)
(EE)
So, this is X being completely nuked on pentium4, I fail to see why..
Offline
I am running X using the pentium4 builds here now, updated last night. I did experience something strange when I plugged in my phone earlier, the X session terminated but it didn't drop me back into bash, but I still got kernel messages about the USB port resolving, but this machine does sometimes do strange stuff when you plug or unplug stuff, so it might be on it's last legs. But provided I keep that to a minimum it seems fairly stable still, and X doesn't segfault regularly for me.
I am using the testing repo though. I'm not sure which of my updated packages came from the testing repos, but if you've any suggestion I could check.
Last edited by levi (2021-10-25 19:26:17)
Architecture: pentium4, Testing repos: Yes, Hardware: EeePC 901+2GB RAM+OS half on the SD card.
Offline
Offline
Both mesa and all the font packages I have installed come not from the testing repos, so they should be the same for all of us.
Architecture: pentium4, Testing repos: Yes, Hardware: EeePC 901+2GB RAM+OS half on the SD card.
Offline
My architectrue is pentium4.
I've just successfully upgraded to the latest xorg-server and xorg-server-common. The situation now is as follows:
Only upgrading wayland causes X11 to crash.
Only upgrading linux from kernel version 5.13.13 to 5.14.12 also causes X11 to crash.
I haven't tested any of the other packages described in my initial post.
Is the announcement concerning libffi33 relevant for this issue?
Last edited by mhdry (2021-10-27 17:01:53)
Offline
Nope. I don't think so (at least installing/uninstalling libffi33 doesn't make a difference).
I'm currently diffing all packages between pentium4 testing and pentium4 stable, and I
also start to see no relevant differences anymore but the Linux kernel.
Offline
Yep, it's then kernel causing a floating point exception while X is starting. This is really, really weird behaviour.
Fixed for me it seems with kernel 5.14.14, also lightdms show up again..
Offline
Yes, I'm also on 5.14.14 and have experienced very little unexpected behaviour on this old machine. Maybe I dodged a bullet here as I went from 5.14.9 right to 5.14.14.
Architecture: pentium4, Testing repos: Yes, Hardware: EeePC 901+2GB RAM+OS half on the SD card.
Offline
Thank you for your quick response. Yesterday evening all updates went through; LightDM is working fine with kernel 5.14.14. I think this thread can be marked as solved...
BTW: I have not seen the USB issues described by @levi yet. However I'm not running an EeePC, but a Thinkpad X60.
Offline
Yeah, I think that's just this machine showing its age.
Architecture: pentium4, Testing repos: Yes, Hardware: EeePC 901+2GB RAM+OS half on the SD card.
Offline