You are not logged in.
In the middle of an usual system updating (pacman -Syu), I was forced to login again through Lightdm.
One of the packages that was being updated was lightdm (as well as lightdm-gtk-greeter), so I suspect of them.
Even doing a downgrade does not fixed the problem.
Now, every time I log in (through lightdm), just some seconds after reaching the desktop, I am redirected to a new login (showing briefly on the console: Start: Accounts.service, between logout and login again).
After this new login it doesn't logout by itself any more.
When I try to execute some command using sudo, or a root terminal (but it seems only if some writing is involved, as when I redirect stdout to a file on my user home directory), it logouts and go to lightdm login again.
The same occurs when yay will install a package (probably using sudo).
It seems to be something related to lightdm and to root permissions, but I could not find anything that seems relevant on dmesg or journalctl.
Any help will be highly appreciated.
lightdm 1:1.32.0-4
lightdm-gtk-greeter 1:2.0.8-3
lightdm-gtk-greeter-settings 1.2.2-7
linux 5.19.9.arch1-1
Armando
Offline
This sounds more like your session crashes, maybe the display server.
I could not find anything that seems relevant on dmesg or journalctl
Let's see. After being logged out at least once, run from a console login
sudo journalctl -b | curl -F 'file=@-' 0x0.stand post the url you get imn return.
Offline
Here is the url: http://0x0.st/oV6k.txt
I've forced the logout/login issuing the comand:
ls -lh > teste.txton a root terminal on my user home directory.
That was about set 21 18:41h
Last edited by amb (2022-09-21 21:47:05)
Armando
Offline
set 21 18:41:11 AMB systemd-coredump[28746]: Process 28244 (Xorg) of user 0 dumped core.
…
Stack trace of thread 28244:
#0 0x00007f621dae349c n/a (libc.so.6 + 0x8849c)
#1 0x00007f621da93958 raise (libc.so.6 + 0x38958)
#2 0x00007f621da7d53d abort (libc.so.6 + 0x2253d)
#3 0x0000563c4da6e8d0 OsAbort (Xorg + 0x1578d0)
#4 0x0000563c4da70273 FatalError (Xorg + 0x159273)
#5 0x0000563c4da75c0c n/a (Xorg + 0x15ec0c)
#6 0x00007f621da93a00 n/a (libc.so.6 + 0x38a00)
#7 0x00007f621dae349c n/a (libc.so.6 + 0x8849c)
#8 0x00007f621da93958 raise (libc.so.6 + 0x38958)
#9 0x00007f621da7d53d abort (libc.so.6 + 0x2253d)
#10 0x00007f621da7d45c n/a (libc.so.6 + 0x2245c)
#11 0x00007f621da8c486 __assert_fail (libc.so.6 + 0x31486)
#12 0x0000563c4daf27f5 n/a (Xorg + 0x1db7f5)
#13 0x0000563c4daf3ec5 DRI2Authenticate (Xorg + 0x1dcec5)
#14 0x0000563c4daf9379 n/a (Xorg + 0x1e2379)
#15 0x0000563c4d956057 n/a (Xorg + 0x3f057)
#16 0x00007f621da7e290 n/a (libc.so.6 + 0x23290)
#17 0x00007f621da7e34a __libc_start_main (libc.so.6 + 0x2334a)
#18 0x0000563c4d956475 _start (Xorg + 0x3f475)
Stack trace of thread 28248:
#0 0x00007f621db62d36 epoll_wait (libc.so.6 + 0x107d36)
#1 0x0000563c4da6e4cb n/a (Xorg + 0x1574cb)
#2 0x0000563c4da6a469 n/a (Xorg + 0x153469)
#3 0x00007f621dae174d n/a (libc.so.6 + 0x8674d)
#4 0x00007f621db63700 n/a (libc.so.6 + 0x108700)
ELF object binary architecture: AMD x86-64You're using nouveau and
nouveau.config=NvBios=PRAMIN nouveau.noaccel=1why do you disable the acceleration?
Consider using
https://aur.archlinux.org/packages/nvidia-390xx-dkms
https://aur.archlinux.org/packages/nvidia-390xx-utils
Or at least globally (/etc/profile, this needs to affect the X11 server itself) export LIBGL_ALWAYS_SOFTWARE=1
You may also get away w/ "LIBGL_DRI2_DISABLE=1" and "LIBGL_DRI3_DISABLE=1", but I'd not rely on the HW at all w/ that configuration.
But the nvidia driver will give you VASTLY greater performance and very likely be much more stable as well.
Offline
When I inspected journalctl log I filtered it so I missed the coredump (noticed it soon after sending the url here).
I've disabled hardware acceleration because it rendered nouveau a little unstable, and changed from nvidia drivers to nouveau because at some point they became much more unstable then nouveau with acceleration.
I'll try your recommendations about turning on acceleration on nouveau again, specially that you stated that that would be more stable (I thought that the inverse would be true), and report here again.
In fact video performance is not a problem to me, I use this computer mainly for programming (backend).
Armando
Offline
try your recommendations about turning on acceleration on nouveau again, specially that you stated that that would be more stable
nananananana. nah.
The nvidia binary driver will turn out to be far more reliable, not sure about progress w/ nouveau itr.
Offline