I haven’t run this update yet on my main home computer, and I’m glad I didn’t, seeing as how badly this bricked my work computer. Booting the work computer brings up the Grub OS-select screen, and on selecting Manjaro and hitting Enter, I get the usual Linux start-up chatter on TTY1 with the last few lines being:
Started Modem Manager.
Started Network Manager.
Started WPA Supplicant.
Reached target Network.
Started CUPS scheduler.
Finished Permit User Sessions…
Started Command Scheduler.
Starting Hold until boot process finishes up…
Starting Terminate Plymouth Boot Screen…
Starting Manage, Install, and Generate Color Profiles…
The discplay then switches from TTY1 to TTY2 (which is normal). TTY2 starts with a black screen with a single white underscore mark in the upper-left corner, which is normal. But then, nothing. The GUI Linux Login Screen never appears. I’m left staring forever at a blank black screen with a single _ mark. What could be causing this? And how can I fix this?
(Note: This upgrade ran successfully on my Asus notebook, which I’m using right now.)
Addendum: The bricked system CAN operate in console mode on TTY3, so I logged in as root and did “journalctl -n200”. The results were interesting! Program “sddm-greater-qt” gave “segfault”, “Error 6”, and “coredump”. Without SDDM working, I won’t be able to get to the GUI Manjaro Login screen. How can this be fixed?
Further Addendum: on a whim, I typed “sddm ” as root on TTY3. The GUI login screen the popped-up on TTY2 and I was able to log-in as my normal non-root user, HOWEVER, the display is now frozen on my Splash Screen; I can’t get to the desktop. (Maybe because SDDM crashed again?)
I think sddm should run as a service, so the command to start it is systemctl start sddm.service. Using systemctl status sddm.service should also show some more information - perhaps on why it doesn’t run?
Still regression of SDDM and KRunner on Thinkpad T60 with Intel Graphics 945GM / i915:
no text in Login Screen(SDDM) and KDE Launcher KRunner
18/07/2025 21:50 sddm-greeter-qt6 Failed to link shader program: error: Exceeded max nr indirect texture lookups (5/4)
18/07/2025 21:50 sddm-greeter-qt6 Failed to build graphics pipeline state
18/07/2025 21:58 kinfocenter Failed to link shader program: error: Exceeded max ALU instructions (152/64)
18/07/2025 21:58 kinfocenter Failed to build graphics pipeline state
Operating System: Manjaro Linux
KDE Plasma Version: 6.3.6
KDE Frameworks Version: 6.16.0
Qt Version: 6.9.1
Kernel Version: 6.12.38-1-MANJARO (64-bit)
Graphics Platform: X11
Processors: 2 × Intel® Core™2 CPU T5500 @ 1.66GHz
Memory: 1.9 GiB of RAM
Graphics Processor: i915
Manufacturer: LENOVO
Product Name: 1952VA4
System Version: ThinkPad T60
Edit 2025-07-19: similar Regressin in KSystemMonitor. No graphics there.
Ok, as root in TTY3, I typed “systemctl --status sddm.service”, and the result was interesting! The SDDM service is running, even though SDDM itself is not. And I see why.
On running that same command on my notebook PC, I see this line:
SDDM[522] Greeter stopped. SDDM::Auth::HELPER_SUCCESS
followed by several more lines, ending with “SDDM[522] Session started true”
Whereas, on my work desktop PC, I see THIS line instead:
SDDM[652] Greeter stopped. SDDM::Auth::HelperExitStatus(11)
with no more lines, and no mention of an actual SDDM session starting (presumably because of the fact that sddm-helper-qt6 crashed with status “11”).
On the problem PC, typing “journalctl --system -n200” also mentions that “sddm-greeter-qt” sends message “Connect” to “kernel”, producing “segfault”, “Error 6”, and “coredump”.
So, the problem is that a segmentation fault occurs every time “sddm-greeter-qt” sends a “Connect” message to “kernel”. What could be causing this?
Addendum: I also see in the system log that “PAM” is malfunctioning because this file is missing:
/usr/lib/security/pam_umax.so
And the PAM error occurs before the sddm-greeter-qt error; could that be the root cause?
I meant one can’t simply run sddm as root and expect it to work properly. So it was no surprise that one can’t login - the surprise was rather that you saw the login screen at all given it didn’t come up when starting. I would reboot, then switch to tty, login as root and check systemctl status sddm.service. Then you can try to start it properly with systemctl start sddm.service or systemctl restart sddm.service. Should the login screen show up then the expectation would be that one can login.
Regarding PAM, the /usr/lib/security/pam_umask.so library is in the pam 1.7.1-1 package. You wrote pam_umax.so but this must be a typo I think. Isn’t this pam package installed and if so how could one of its libraries be missing?
I’m on Testing branch with pam v1.7.1-1 installed, and I don’t have that library. I think it may have been a typo by @robbie-hatley. Possibly they meant /usr/lib/security/pam_umask.so or /usr/lib/security/pam_unix.so, which are included in the pam package on my system:
ls -l /usr/lib/security/ | grep -F 'pam_u'
-rwxr-xr-x 1 root root 14200 Jun 19 18:48 pam_umask.so
-rwxr-xr-x 1 root root 55232 Jun 19 18:48 pam_unix.so
-rwxr-xr-x 1 root root 14200 Jun 19 18:48 pam_usertype.so
Nonetheless, your post(s) have been moved to a dedicated topic, where others may offer assistance; I’m sure someone will help when they are able.
Please start with providing system information as described (below) along with any other information that may be requested.
This seems to narrow it down to sddm-greeter-qt, at face value, but I note there are other reports similar to what you describe.
Hopefully, others can work with you to identify the actual cause, and have you back up and running before too long.
Regards.
Anecdotal: I have performed the same update that you initially blamed for your woes, but without any such issue(s).
System Information
While information from *-fetch type apps might be fine for someone wishing to buy your computer, for Support purposes it’s better to ask your system directly;
Output of the inxi command with appropriate parameters will achieve this (naturally, formatted according to forum guidelines) and will generally be more useful for those wishing to help:
I’m also getting “segfault”, “Error 6”, and “coredump” on “sddm-greeter-qt6” after that last update. I think it’s because authentication fails, because PAM crashes, because of a missing PAM “.so” file. I see there’s another huge update 2 days later; I’ll see if that fixes the problem when I get back to work tomorrow (Mon Jul 21, 2025).
In my imagination, maybe. But if you’d read my post, you’d have seen that I’m staring at a black screen with a single white non-blinking underscore mark in the upper-left corner. GUI does not exist. Keyboard and mouse do nothing. Perhaps you expect me to use ESP or telekinesis? Well, I do wish I had those skills. But alas, I do not.
Besides, i don’t see what any of that has to do with the existing evidence which I’ve presented, namely that “sddm-greeter-qt6” crashes due to “segfault + ‘Error 6’ + coredump”, after PAM crashes due to a missing “.so” file. If you’d like to explain how “customization” can cause “missing PAM ‘.so’ file”, I’m all ears; I not seeing how, personally.
Furthermore, the computer in question was not “customized” any more than any of my other Manjaro-Plasma computers, and they accepted the update and ran after rebooting with no PAM or SDDM problems.
So what it looks like to ME is, the update process got corrupted. However, I’m not seeing any way to “Repeat update, overwriting old update”. Any attempt to run “pacman -Su” as root on TTY3 simply gives “Nothing to do.”
I’m hoping running the next big update that came out just 2 days later will fix the problem. If not, my next step will be to manually copy the missing PAM “.so” file from my notebook to the damaged work PC and see if I can get PAM and SDDM to work again and stop crashing.
But if you have a better idea, certainly I’ll try that.
On the contrary, it seems you don’t understand the metaphor “bricked”. It means “to become as useful as a brick”. For a work PC, which exists expressly for the purpose of running GUI programs (including CAD, spreadsheet, and word-processor programs) to be unable to do so, is the epitome of “brickness”, yes. It’s a brick. I’ll paint it dark red with rough texture and send you a photo, if you want.
Mod Edit - let’s use Hidden sections for side arguments and focus on the task in hand…
On the contrary, I do. While your interpretation may be convenient to enhance your displeasure, the machine is still functional; albeit temporarily incapacitated by what appears to be a software issue.
The cause seems unclear, though several other reports indicate similar issues, which I linked here in another topic.
While your frustration is understandable, I’ll remind you that most here are Manjaro users, like you, and attempt to help where they can; as best they can. Noting your previous commentary, some consideration for that fact certainly will not go astray.
Regards.
Keep it. I’ve encountered many retardant bricks in my travels.
Let me see, from my brain I extract the definition that Bricked means Total, unrecoverable hardware failure. You can manage this by screwing your BIOS up, so that the computer won’t initiate.
However, your system boots successfully through a significant portion of the Linux boot process, as evidenced by the TTY output and starting services like CUPS… it also can reach TTY3 and log in and execute commands… it appears to be running some form of Linux I have yet to see a brick do this… though I’m interested to do experiments to see if I can find a block of concrete to run ‘journalctl -n200’ as a comparison.
Severely inconvenient
More than a little frustrating
Bricked
Have we tried pacman -S sddm sddm-kcm qt5-quickcontrols2 qt5-graphicaleffects to try to fix any corrupted files…
Let’s try resetting SDDM config:
mv /etc/sddm.conf /etc/sddm.conf.brick # see what I did there ;)
sddm --example-config > /etc/sddm.conf # A fresh default
Greeter themes can be dodgy; I messed up mine a few years back:
Also, as a quick reference - use backticks `and/or tilde ~ to enclose your code - and if it’s LONG, you can use the icon and ‘Hide Details’, go up to the ‘Summary’ and type your heading, then arrow down, enter your ~~~, paste and add another ~.
Like this...
[details="Like This"]
~~~
Pasted some ■■■■ in here...
~~~
[/details]
You strike me as an academic person; I don’t want to challenge your knowledge nor do I want to fight you; the nature of asking in a forum is not to cast blame but to find a way to repair your installation; I am merely trying to provide input to the thought process by asking questions, by suggesting what may cause a given incident; I am never trying to win an argument; such a fight is a child’s game - I won’t participate in that.
I did read - but without any means of reproducing - and with personal systems which barely causes issues - one look back at experience, and experience tells me that Nvidia GPU and/or a customised greeter theme could be the issue.
Since you did not mention the GPU my initial hunch is with customisation - but it can have other root as well.
When you are missing important system files - you have definitely customised your system - albeit not the customisation I had in mind
For obvious reasons - I cannot say why a certain system file is missing - suffice to say, they rarely go missing by themselves - you lean towards a failure in the process - could be - but without physical access to the patient your guess is as good as mine.
When one understand how pacman -Syu work
download new metadata
lookup installed packages in metadata
if newer package exist - add it to the queue
download the queue with signatures (/var/carche/pacman/pkg)
verify the each package in the queue using the corresponding signature
if one package does not match the process is aborted
a package update is technically a transaction
pre-removal hook
complete removal of package related files
post-remove hook
pre-install hook
unpacking of the content to the file system
post-install hook
One will realise the only way pacman can fail, is if the PKGBUILD maintainer made an error; packages are signed after they have been created; and a damaged package (not matching signature) will stop the process.
If the system encounters a fatal power outage in the middle of a transaction - all kinds of bugs will emerge.
Since you already stated that one or more systems did not complain one has to look into the specific system’s configuration and hardware layout and what could have gone wrong.
I have yet to see a brick do this… though I’m interested to do experiments to see if I can find a block of concrete to run ‘journalctl -n200’ as a comparison.
icon and ‘Hide Details’, go up to the ‘Summary’ and type your heading, then arrow down, enter your ~~~, paste and add another ~.