Bug 23177

Created attachment 10241 [details]
Program version before and after upgrade

Created attachment 10242 [details]
Screenshot error messages from journal (sddm)

How did you exclude the qt5 packages when you updated your cauldron guest?

https://ml.mageia.org/l/arc/dev/2018-06/msg00007.html

(the Qt5 stack update to 5.11.0 is still not finished, qtwebengine5-5.10.1-2.mga7 still needs to be updated)

About not being able to login to a VT, did you try all of the following:

* switch to tty2 after failing to login with SDDM
  
* boot in runlevel 3 
   
* only if you didn't get a login prompt in runlevel 3: 
  switch to a different tty to workaround bug #22620

Keywords: (none) => NEEDINFO
CC: (none) => kde, kernel, marja11

qtwebengine5-5.11.0-1.mga7 should have landed on your mirror.

Is it possible, after updating, to login with SDDM again?

@ Marja,

No, the Qt5 stack and everything else was upgraded to the present state. I had a working login and Plasma desktop before that upgrade. KDE applications were a different matter. Of course, I have a snapshot of that state and can go back at ease.

I cannot imagine that qtwebengine5 is relevant at this early stage.

After the failure I booted in recovery mode. Copy & Paste not working, not able to mount a USB device. So I could only take a screenshot of the error messages as in https://bugs.mageia.org/show_bug.cgi?id=23177#c3.

Thanks for the init 3 hint! That works and I even have a network connection. So I can run further upgrades or even send a file to another client in the local network, e.g. another Qemu/KVM client. But I did not figure out how.

Ulrich

(In reply to Ulrich Beckmann from comment #5)

> 
> Thanks for the init 3 hint! That works and I even have a network connection.
> So I can run further upgrades 

After doing those updates, do you still have the problem with SDDM?

(In reply to Marja Van Waes from comment #6)
> (In reply to Ulrich Beckmann from comment #5)
> 
> > 
> > Thanks for the init 3 hint! That works and I even have a network connection.
> > So I can run further upgrades 
> 
> After doing those updates, do you still have the problem with SDDM?

Yes, I do. I have split now the VM into two, one fully upgraded with the bug, another working one with the x11-* packages on hold.

Jeremiah Summers said in the related https://bugs.mageia.org/show_bug.cgi?id=23163, that a rebuild fixed his problem. So I think that next step should be a rebuild of all 'x11-server|x11-driver' packages.

Ulrich

Afaik all that supports xorg 1.20 are rebuilt

CC: (none) => tmb

(In reply to Ulrich Beckmann from comment #7)
> (In reply to Marja Van Waes from comment #6)

> > 
> > After doing those updates, do you still have the problem with SDDM?
> 
> Yes, I do. I have split now the VM into two, one fully upgraded with the
> bug, another working one with the x11-* packages on hold.
> 

Please attach new_x11.txt that is the result of running as root, after failing to use SDDM and then switching to a VT:

  journalclt -b > new_x11.txt

or, if you need to reboot first to get to a working VT:

  journalclt -b -1 > new_x11.txt

and similarly in the working VM with x11 on hold, after successfully logging in with SDDM, run, again as root:

   journalclt -b > old_x11.txt

and attach old_x11.txt, too.

If one of those files is too large to attach, then please compress it with xz.

First try: a new upgrade of all freshly downloaded x11-packages.
Unfortunately it did not resolve the problem. It led to the same issue.

Ulrich

Created attachment 10253 [details]
Journal of boot with new x11

Error messages refer to ksmserver.

Attachment 10242 is obsolete: 0 => 1

Created attachment 10254 [details]
Journal of boot with old x11

(In reply to Ulrich Beckmann from comment #11)
> Created attachment 10253 [details]
> Journal of boot with new x11
> 
> Error messages refer to ksmserver.

I see three libkscreen segfaults, like this one:

Jun 25 17:14:11 localhost kernel: kscreen_backend[3773]: segfault at 10 ip 00007f6db851db38 sp 00007ffeed9b2060 error 4 in KSC_XRandR.so[7f6db8505000+21000]

and after each of them:

The X11 connection broke (error 1). Did the X11 server die?

Assigning to libkscreen and KDE team.

Ulrich, please give the output of:

   rpm -qa libkscreen

thanks :-)

Source RPM: (none) => libkscreen
Assignee: bugsquad => kde

Ulrich, instead of the above question, can you please run:

  rpm -qa | sort > old_list.txt (in the VM with old X11, of course ;-) )

  rpm -qa | sort > new_list.txt  (in the VM with new X11)

and attach those lists?

Created attachment 10259 [details]
List of installed packages (new X11)

The X11-server did not die. If it died, it was restarted. The login screen always returns after failure.

Created attachment 10260 [details]
List of installed packages (old X11)

After about 2 months of absence I ran a full upgrade now without any versionlocks.  It works now, though I can't say which package is responsible for the fix.

List of installed packages

libkscreen-5.13.4-1.mga7
x11-driver-input-6.0.0-5.mga6
x11-driver-input-libinput-0.28.0-1.mga7
x11-driver-input-wacom-0.36.1-1.mga7
x11-driver-video-7.7-13.mga7
x11-driver-video-amdgpu-18.0.1-1.mga7
x11-driver-video-ati-18.0.1-1.mga7
x11-driver-video-cirrus-1.5.3-8.mga7
x11-driver-video-fbdev-0.5.0-1.mga7
x11-driver-video-glint-1.2.9-2.mga7
x11-driver-video-intel-2.99.917-44.mga7
x11-driver-video-mach64-6.9.6-1.mga7
x11-driver-video-mga-1.6.5-2.mga7
x11-driver-video-neomagic-1.2.9-7.mga7
x11-driver-video-nouveau-1.0.15-3.mga7
x11-driver-video-openchrome-0.6.0-2.mga7
x11-driver-video-qxl-0.1.5-12.mga7
x11-driver-video-r128-6.11.0-1.mga7
x11-driver-video-s3-0.6.5-19.mga7
x11-driver-video-s3virge-1.10.7-9.mga7
x11-driver-video-savage-2.3.9-3.mga7
x11-driver-video-sis-0.10.9-2.mga7
x11-driver-video-sisimedia-0.9.1-9.20091203.28.mga7
x11-driver-video-tdfx-1.4.7-2.mga7
x11-driver-video-trident-1.3.8-2.mga7
x11-driver-video-v4l-0.3.0-1.mga7
x11-driver-video-vesa-2.4.0-2.mga7
x11-driver-video-vmware-13.3.0-3.mga7
x11-server-common-1.20.1-1.mga7
x11-server-xorg-1.20.1-1.mga7
x11-server-xwayland-1.20.1-1.mga7

Best regards,
Ulrich

Status: NEW => RESOLVED
Resolution: (none) => FIXED