Bug 466208

Just gave this a try with a nearly-identical setup:
- Laptop
- USB-C dock with HDMI screen plugged into it
- Laptop internal screen disabled in KScreen KCM
-Reboot

After logging in, the laptop screen is correctly turned off and the external screen is the only one that's turned on.

To debug this further, can you please provide the following information?
1. What GPU does the laptop use?
2. Before reboot, can you verify that the internal screen is marked as disabled in System Settings' Display and Monitor page? Or is it enabled?
3. After reboot, can you verify that the internal screen is marked as enabled in System Settings' Display and Monitor page? Or is it disabled?
3. Can you attach the output of `kscreen-doctor -o` in a terminal window before reboot, and after reboot?

I know what the issue was.
I was not using the settings center. Instead I always used this new widget in the tray that allows to selected display layout (switch to external display)

When I do everything in the configuration center all works as expected.

Is that the idea that the widget is not persistent?

The widget is provided by NVIDIA and does not manipulate the screen arrangement in a way that's compatible with the rest of the system. I strongly recommend never using it to change the screen arrangement.

Created attachment 156660 [details]
displaywidget

I don't think it is an NVIDIA, looks like pure Plasma tray widget.
See screenshot attached

Aha, indeed that's not NVIDIA.

(In reply to Nate Graham from comment #1)
> To debug this further, can you please provide the following information?
> 1. What GPU does the laptop use?
> 2. Before reboot, can you verify that the internal screen is marked as
> disabled in System Settings' Display and Monitor page? Or is it enabled?
> 3. After reboot, can you verify that the internal screen is marked as
> enabled in System Settings' Display and Monitor page? Or is it disabled?
> 3. Can you attach the output of `kscreen-doctor -o` in a terminal window
> before reboot, and after reboot?
Can you please answer these questions?

Created attachment 156668 [details]
widget-before_reboot

Created attachment 156669 [details]
widget-after-reboot

Created attachment 156670 [details]
controlcenter-before-reboot

Created attachment 156671 [details]
controlcenter-after-reboot

(In reply to Nate Graham from comment #6)
> (In reply to Nate Graham from comment #1)
> > To debug this further, can you please provide the following information?
> > 1. What GPU does the laptop use?
> > 2. Before reboot, can you verify that the internal screen is marked as
> > disabled in System Settings' Display and Monitor page? Or is it enabled?
> > 3. After reboot, can you verify that the internal screen is marked as
> > enabled in System Settings' Display and Monitor page? Or is it disabled?
> > 3. Can you attach the output of `kscreen-doctor -o` in a terminal window
> > before reboot, and after reboot?
> Can you please answer these questions?

1. Dual gpu with Nvidia+build-in, prime-select not configured for any driver but nvidia loaded anyway
2. If I switch with screen of using the widget it is also marked as off in control center
3. It is marked as enabled
4. Attaching files

Aha, thanks. This is Bug 460341, then.

*** This bug has been marked as a duplicate of bug 460341 ***

Apparently not fixed with the fix for Bug 460341; re-opening.

I have this exact same issue that presents after sleep/resume, where my second monitor, which I usually keep turned off, is re-activated. Of course the same happens after rebooting as well.

It's not limited to proprietary drivers, I'm using an AMD card.

Stephane, let's use this bug report for Sebastian's specific issue. Can you submit a new one for your issue? Thanks!

*** Bug 467365 has been marked as a duplicate of this bug. ***

(In reply to Nate Graham from comment #15)
> Stephane, let's use this bug report for Sebastian's specific issue. Can you
> submit a new one for your issue? Thanks!

I'm not sure mine is worthy of another bug report just to be closed as DUPLICATE, as my issue is the same.

If I clear `.local/share/kscreen`, everything works as intended, KScreen creates one config, and everything works fine. When I resume from sleep or reboot, the screens keep the last configuration.

But after a while, not sure what's the trigger, I see 3 different configs stored under `.local/share/kscreen` and none of my configurations are retained. It seems that the logic to determine what is a new Display Arrangement or a pre-existing one in KScreen to be completely borked, and it might be because I'm making use of the Extend Screen feature like OP did.

Can reproduce the issue when using the Extend button in the widget. The KCM works fine. The code paths must be subtly different here.

I have started using a custom script to toggle my second monitor with kscreen-doctor and this issue does not appear. It presents only after using the Extend widget. 

So for me the only workaround was to delete all kscreen configuration files to start again fresh, and just use kscreen-doctor, avoiding the widget at all. 

On Tue, 25 Jul 2023, at 22:22, Nate Graham wrote:
> https://bugs.kde.org/show_bug.cgi?id=466208
>
> Nate Graham <nate@kde.org> changed:
>
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>            Keywords|                            |multiscreen
>              Status|REPORTED                    |CONFIRMED
>             Summary|Display arrangement is lost |Preference to disable
>                    |after reboot when using     |laptop screen while
>                    |"Extend" button in KScreen  |external screen is
>                    |widget                      |connected is lost after
>                    |                            |reboot when using "Extend"
>                    |                            |button in KScreen widget
>      Ever confirmed|0                           |1
>
> --- Comment #18 from Nate Graham <nate@kde.org> ---
> Can reproduce the issue when using the Extend button in the widget. The KCM
> works fine. The code paths must be subtly different here.
>
> -- 
> You are receiving this mail because:
> You are on the CC list for the bug.

Can still reproduce in Plasma 6.

Git commit ae84480fbfdc684b8ee4b0207d3ce679f6fb4cd7 by Xaver Hugl.
Committed on 05/10/2023 at 20:29.
Pushed by zamundaaa into branch 'master'.

outputconfigurationstore: add new config system

Instead of an external service (like KScreen) storing and restoring output configurations,
with this commit KWin takes over that responsibility. This allows it to, among other things,
generate appropriate configs for new sets of outputs immediately, and take KWin-internal information
about outputs into account when generating them.
Related: bug 455082, bug 457430, bug 474021, bug 469653, bug 466342, bug 470863, bug 466556

M  +1    -0    CMakeLists.txt
M  +1    -0    autotests/integration/kwin_wayland_test.cpp
M  +1    -0    src/CMakeLists.txt
M  +5    -0    src/backends/drm/drm_output.cpp
M  +9    -0    src/core/output.cpp
M  +11   -0    src/core/output.h
M  +1    -0    src/core/outputconfiguration.h
M  +7    -0    src/main.cpp
M  +1    -0    src/main.h
M  +0    -4    src/main_wayland.cpp
M  +709  -20   src/outputconfigurationstore.cpp
M  +68   -4    src/outputconfigurationstore.h
M  +1    -0    src/utils/CMakeLists.txt
M  +8    -0    src/utils/edid.cpp
M  +8    -0    src/utils/edid.h
A  +54   -0    src/utils/orientationsensor.cpp     [License: GPL(v2.0+)]
A  +38   -0    src/utils/orientationsensor.h     [License: GPL(v2.0+)]
M  +34   -1    src/wayland/outputdevice_v2.cpp
M  +1    -0    src/wayland/outputdevice_v2.h
M  +12   -1    src/wayland/outputmanagement_v2.cpp
M  +49   -6    src/workspace.cpp
M  +2    -0    src/workspace.h

https://invent.kde.org/plasma/kwin/-/commit/ae84480fbfdc684b8ee4b0207d3ce679f6fb4cd7

*** Bug 472732 has been marked as a duplicate of this bug. ***