Kde stops in blank black screen

[Paul Floyd]

Wheel and operator but not video. Surely that makes no difference for intermittent hangs? I imagine that this is some kind of data race bug (at least for me, maybe not for the OP). I'm using the 470 drivers so I don't think that the linked post applies.

And I have the answer to "does KRunner respond". No.

I can switch virtual screens with ctrl-alt-F2 etc but the screen is corrupt - random coloured blobs. If I login carefully and run "pkill ksm" then back to the login screen and no problems.

 

[t4z3v4r3d]

On Jan 8, 2024:

Where is this the output coming from?

Given your Nvidia GT 730, it should be supported by x11/nvidia-driver-470 as stated in FreeBSD Display Driver – x64 - 470.161.03 - 2022.11.22 as explicitly mentioned there: "GeForce GT 730" and as is mentioned in message #48.

I take it that you're not running bare metal and are using emulators/virtualbox-ose. I don't know precisely how using an emulator might effect the working of other kernel modules, but it is the first "error" ("an 8 kernel: KLD vboxdrv.ko: depends on kernel - not available or version mismatch") that comes up. It would seem that that creates some kind of environment for the rest of the FreeBSD system, including your Nvidia kernel module as referenced in /etc/rc.conf by nvidia-modeset, be it from x11/nvidia-driver-390 or from x11/nvidia-driver-470.

As your message is from Jan 8, your version of emulators/virtualbox-ose is probably not up to date. Ports reported two more recent versions, most notably mentioning "-kmod":

Therefore, I suggest, verify that you're using version 6.1.50_1 - 20 Jan 2024 00:41:39 and try again using x11/nvidia-driver-470

The error message was for a test and I'm already using 470. My machine is running on bare metal too. The "vbox driver error" is just a because of my simple mistake to copy-paste message.
Thanks for reply.

 

[t4z3v4r3d]

I tried to retart every parts of kde on startup process, but nothing is changed, I have no idea why the kde start process is going well with startx command that is running by "-^" argument.
Thanks.

 

[cy@]

startx (xinit) runs within your own context -- session as created by setsid(2) -- whereas IIRC sddm is not PAM aware. There's likely some requirement within your kde environment that needs some session or process group initialization that is done by PAM that sddm doesn't do.

This is the reason I use xdm instead of CDE's dtlogin. dtlogin has no PAM support for my environment -- which uses Kerberos with LDAP and NIS (my LDAP directory is replacing my NIS server).

If it runs under startx it will probably run under xdm. You will need to copy your .xinitrc used by startx (xinit) to an .xsession file, with minor alterations.

If you don't have one you will probably need to create a pam.d/xdm file from your pam.d/login file.

The other thing to look at is your login class. When you login on the console your login class is set by login(1)'s call to setusercontext(3) from login.conf. sddm may not do this correctly.

Another thing for you to investigate: Could sddm be using Wayland? startx(1) uses Xorg. If your sddm is trying to use Wayland, you may have some missing, old, or misconfigured Wayland package.

The other possibility is your .profile/.bashrc/or whatever shell you're using is setting up the environment that kde can work with while sddm may not be. I'm not sure if sddm will invoke your .profile. You may need to check the man page for that.

 

[grahamperrin]

… Could sddm be using Wayland? …

From my x11/nvidia-driver-470 experience (below), I assume that SDDM intelligently omits the Wayland option if the environment lacks support.

x11/sddm with x11/kde5 and x11/kde6: Plasma (Wayland) missing from the SDDM menu

<https://www.freshports.org/x11/sddm/> With [x11/kde5](https://www.freshports.org/x11/kde5/):  The _Plasma (Wayland)_ menu option disappeared months ago, I took no notice at the time (because I never used Wayland). ```text ...

codeberg.org

 

[pming]

Kde5 stops after booting on a blank black screen and the only thing that is alive is the "mouse". After some tries, I faced this error:

Qt.qpa.xcb: could not connect to display

Where are you seeing this error exactly?

Maybe this could help: https://community.kde.org/FreeBSD/Setup

My issue was solved with this
https://discuss.kde.org/t/plasma-starts-with-black-screen/5136/5

TLDR plasma would not start because it was waiting for autofs NFS mount. If you are using that disable it, try to log in and disable the options mentioned in the link above.

The kde is loaded and runing but for the charlie.

No other account can log in, and the messages log file shows that



console-kit-daemon critical: polkit_authority_check_authorization 'polkit_is_authority (authority)' failed

I get the same error in /var/log/messages after updating polkit to 124 but I will create a separate thread.
Edit: Updating to polkit-124_1 solved it for me...

 

[t4z3v4r3d]

There is no hope to solve the issue because of no error logs or sign of root cause.
I will remove all packages and install again.

 

[t4z3v4r3d]

After a time, I tried to run my BSD Box again, and now I'm writing on it. But there was a bad/missed configuration I had to solve.
If I run "Startx" with my user, the KDE will up and run, and thus the issue will show up: the "Log in" process is not working properly, and I am not able to mount my disks that I was able the.
So, what is my mistake in configuration after the update from 12 to 13 ? Or in another word words, how should I search for the root cause of this issue ?
Thanks.