[amdgpu] Screens go into flicker cycle after going blank with external monitor(s)

Thanks for the bug report. These log messages look relevant:

[118317.782940] jarek-Inspiron gnome-shell[3471]: Window manager warning: Configuring CRTC 79 with mode 1881 (1920 x 1200 @ 59,950172) at position 0, 0 and transform 0 failed
[118319.278873] jarek-Inspiron gnome-shell[3471]: meta_display_get_monitor_in_fullscreen: assertion 'monitor >= 0 && monitor < n_logical_monitors' failed

and

[94391.010815] jarek-Inspiron gnome-shell[3471]: Window manager warning: Configuring CRTC 79 with mode 1881 (1920 x 1200 @ 59,950172) at position 0, 0 and transform 0 failed
[94391.490646] jarek-Inspiron gnome-shell[3471]: meta_display_get_monitor_in_fullscreen: assertion 'monitor >= 0 && monitor < n_logical_monitors' failed

Can you please reproduce the problem again and then run 'journalctl -b0' to see if the same error messages occur during the flicker?

And next time the flicker occurs, please run:

  journalctl -b0 > journal.txt

and attach the resulting text file here.

I wonder if it matters that some of your configs in ~/.config/monitors.xml don't intersect with the origin 0,0. If you rename that file and then log in again, does the bug stop happening?

As requested attaching a result of a journalctl -b0 command.

This happened after renaming monitor.xml and monitor.xml~ so unfortunaltely that does not help.

Thanks. Can you repeat that in a Wayland session?

I'll try to find some time during the weekend to reproduce it with Wayland. I am using this PC for work during the week and Wayland keeps crashing with entire session whenever I am in a videocall. :)

That's also interesting to me. Please report that crash when you can...

https://wiki.ubuntu.com/Bugs/Responses#Missing_a_crash_report_or_having_a_.crash_attachment

I don't know what changed but I switched back to Wayland a few days ago and cannot repoduce any glitches I suffered. Is this the week I stop regretting buying a laptop on AMD chipset?

It looks like a kernel update was released a few days ago so maybe that fixed it?

My optimism was to early. Journal dump attached. Took it after unplugging one of the monitors and waiting few flickers for desktop to stabilize.

Btw, were should I report that Slack crashes when I right click inside? :D

There's a hardware connection problem which seems to come first in some cases:

cze 23 15:35:36 jarek-Inspiron kernel: [drm] perform_link_training_with_retries: Link(2) training attempt 1 of 4 failed @ rate(10) x lane(2) : fail reason:(1)
...
cze 23 15:35:36 jarek-Inspiron kernel: [drm] perform_link_training_with_retries: Link(2) training attempt 2 of 4 failed @ rate(10) x lane(2) : fail reason:(1)
cze 23 15:35:37 jarek-Inspiron kernel: [drm] perform_link_training_with_retries: Link(2) training attempt 3 of 4 failed @ rate(10) x lane(2) : fail reason:(1)
...
cze 23 15:35:37 jarek-Inspiron kernel: [drm] enabling link 2 failed: 15
cze 23 15:35:38 jarek-Inspiron gnome-shell[3512]: meta_monitor_manager_get_logical_monitor_from_number: assertion '(unsigned int) number < g_list_length (manager->logical_monitors)' failed
cze 23 15:35:38 jarek-Inspiron gnome-shell[3512]: meta_workspace_get_work_area_for_monitor: assertion 'logical_monitor != NULL' failed
cze 23 15:35:38 jarek-Inspiron gnome-shell[3512]: meta_monitor_manager_get_logical_monitor_from_number: assertion '(unsigned int) number < g_list_length (manager->logical_monitors)' failed
cze 23 15:35:38 jarek-Inspiron gnome-shell[3512]: meta_workspace_get_work_area_for_monitor: assertion 'logical_monitor != NULL' failed
cze 23 15:35:38 jarek-Inspiron gnome-shell[3512]: meta_display_get_monitor_in_fullscreen: assertion 'monitor >= 0 && monitor < n_logical_monitors' failed
cze 23 15:35:39 jarek-Inspiron gnome-shell[3512]: 0x3800080: Frame has assigned frame counter but no frame drawn time
cze 23 15:35:39 jarek-Inspiron gnome-shell[3512]: 0x3800080: Frame has assigned frame counter but no frame drawn time
cze 23 15:35:39 jarek-Inspiron gnome-shell[3512]: 0x3800080: Frame has assigned frame counter but no frame drawn time

I suspected this should go to AMD Kernel team. Lots of people complaining about changes AMD has been making for some time.

Can you redirect this error, or should I report it somewhere else?

Already assigned to the kernel yesterday :)

This bug is missing log files that will aid in diagnosing the problem. While running an Ubuntu kernel (not a mainline or third-party kernel) please enter the following command in a terminal window:

apport-collect 2023647

and then change the status of the bug to 'Confirmed'.

If, due to the nature of the issue you have encountered, you are unable to run this command, please add a comment stating that fact and change the bug status to 'Confirmed'.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

apport information

I switched back to Wayland and 2 external monitors today and the flicker seems to be gone (Ubuntu 23.04 with current updates).