[KScreen] [Bug 368910] Monitor change not detected on docking

[Matthias Blaicher via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=368910

Matthias Blaicher  changed:

   What|Removed |Added

 Resolution|--- |INVALID
 Status|CONFIRMED   |RESOLVED

--- Comment #9 from Matthias Blaicher  ---
Hi Sebastian,

it turns out, this issue was caused by a broken DVI cable... I guess it somehow
impacted monitor hot-plug detection, but xrandr had an extra look on the I2C
when called, and hence it was detected.

Sorry for the noise!

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 368910] Monitor change not detected on docking

[Matthias Blaicher via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=368910

--- Comment #8 from Matthias Blaicher  ---
(In reply to Sebastian Kügler from comment #4)
> For kicks, could you check if this works in GNOME or Unity (maybe they've
> found such a workaround)?

I just checked - GNOME also does not work So I guess this bug is in the
wrong tracker... Any option on how to proceed from here?

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 368910] Monitor change not detected on docking

[Matthias Blaicher via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=368910

--- Comment #7 from Matthias Blaicher  ---
Hi linuxfluesterer,

I am sorry, but this is a totally unrelated bug, as you acknowledge: 

Neither is it docking related, nor does it have problems enabling your monitor
(looks like a plasmashell crashes to me). Furthermore, you do not execute
"xrandr", nor do you run Plasma 5.8 beta, which this bug report is explicitly
about and which carries multiple multi-screen bugfixes.

If you feel like your bug has not been reported yet, feel free to open a new
one specific to your issues, or wait for 5.8 Plasma to be released and check if
the bug remains.

Very Best,
Matthias

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 368910] Monitor change not detected on docking

[Matthias Blaicher via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=368910

--- Comment #3 from Matthias Blaicher  ---
Hi Sebastian,
thank you for your work and looking into it. As this used to work for KDE4, I
suppose it is not an hardware issue, but the whole software stack and kernel
changed since then...

Regarding xrandr, this is an up-to-date Arch Linux with testing enabled, so it
is the libxrandr 1.5.0-1 package atm.

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 368910] Monitor change not detected on docking

[Matthias Blaicher via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=368910

--- Comment #1 from Matthias Blaicher  ---
Created attachment 101128
  --> https://bugs.kde.org/attachment.cgi?id=101128=edit
Annotated kscreen log

Annotated kscreen log with relevant input from journald of the folliowing
sequence:

1. Boot into KDE undocked
2. Dock laptop
3. Execute xrandr
4. Close lid

While there is nothing happening in the kscreen logs, there is one entry after
docking (point 2) I'd like to point out:

Sep 16 20:51:32 meinpc kwin_x11[6183]: QXcbConnection: XCB error: 3
(BadWindow), sequence: 10182, resource id: 94371848, major code: 18
(ChangeProperty), minor code: 0
Sep 16 20:51:32 meinpc kdeinit5[6147]: QObject::connect: invalid null parameter
Sep 16 20:51:32 meinpc kdeinit5[6147]: QObject::connect: invalid null parameter
Sep 16 20:51:32 meinpc kdeinit5[6147]: QObject::connect: invalid null parameter
Sep 16 20:51:32 meinpc kcminit[6975]: Initializing  "kcm_input" : 
"kcminit_mouse"
Sep 16 20:51:32 meinpc kwin_x11[6183]: QXcbConnection: XCB error: 3
(BadWindow), sequence: 10561, resource id: 92274696, major code: 18
(ChangeProperty), minor code: 0 

Could it be, the ChangeProperty is the event we need, but it is dropped
somehow?

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 368910] New: Monitor change not detected on docking

[Matthias Blaicher via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=368910

Bug ID: 368910
   Summary: Monitor change not detected on docking
   Product: KScreen
   Version: 5.7.95
  Platform: Archlinux Packages
OS: Linux
Status: UNCONFIRMED
  Severity: major
  Priority: NOR
 Component: common
  Assignee: se...@kde.org
  Reporter: matth...@blaicher.com

On a Lenovo Thinkpad L430, running Plasma 5.8 Beta with Arch Linux testing
packages, monitor changes due to docking/undocking are not detected. It also
had the same beviour back with Plasma 5.7. However, a simple execution of
"xrandr" makes kscreen detect the monitor change. Afterwards plasmashell need
to be killed and restarted to have a proper desktop layout.

This happens with the intel as well as the mode setting driver. The dock has
two DVI outputs, but the problem behaves the same on both ports.

Installed packages are:

kde-unstable/plasma-workspace 5.7.95-1 
kde-unstable/kscreen 5.7.95-1
kde-unstable/libkscreen 5.7.95-1

00:02.0 VGA compatible controller: Intel Corporation 3rd Gen Core processor
Graphics Controller (rev 09)

Tomorrow, I will take the time to add propper logs to this bug report. If you
require specific tests, feel free to tell me. I would also be willing to try
kscreen patches.

Reproducible: Always

Steps to Reproduce:
1. Plug/Unplug laptop to dock

Actual Results:  
Nothing happens. Only detected when xrandr is executed.

Expected Results:  
Direct detection of monitor changes, proper Plasma desktop applied after
screens are adjusted.

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 363238] Docked laptop lid close does not enable external monitor

[Matthias Blaicher via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=363238

--- Comment #9 from Matthias Blaicher  ---
(In reply to Sebastian Kügler from comment #8)
> You can tell it not to suspend in Systemsettings | Power Management | Button
> Events handling. There's a checkbox that allows to not suspend when the lid
> is closed and an external monitor is connected.

Yes, but in this case, the external monitor is initially disabled and I expect
it to switch to it once the lid is closed. In this case, suspend was not
suppressed. I had to disable logind suspend events to make it work...

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 363238] Docked laptop lid close does not enable external monitor

[Matthias Blaicher via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=363238

Matthias Blaicher  changed:

   What|Removed |Added

 Status|NEEDSINFO   |RESOLVED
 Resolution|WAITINGFORINFO  |FIXED

--- Comment #7 from Matthias Blaicher  ---
I cannot reproduce this bug anymore with latest git, since the Laptop now goes
to sleep - even though its on AC power and docked, while the second monitor is
connected. Stupid, but hey

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 363238] Docked laptop lid close does not enable external monitor

[Matthias Blaicher via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=363238

--- Comment #6 from Matthias Blaicher  ---
Yes, indeed. I have seen your recent activity and that laptop is already
running kscreen git as of 4-5 days ago. Unfortunately, the laptop still has
issues detecting changes in monitors while docking (only discovered upon the
first xrandr call on command line). This occurs both with the intel as well as
with the modesetting driver. It might very well be a kernel driver issue?

Since it is not my personal laptop I haven't investigated further, I'll try to
have a look next weekend again.

-- 
You are receiving this mail because:
You are watching all bug changes.

[gwenview] [Bug 335498] Gwenview delaying kde startup

[Matthias Blaicher via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=335498

--- Comment #2 from Matthias Blaicher  ---
Interestingly, I have just run into the same issue again - just to discover
that I have already commented on this bug. So this issue still exists on
current KDE (I'm using Arch testing.)

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 351025] When external monitor is disconnected, the laptop monitor doesn't become active again

[Matthias Blaicher via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=351025

--- Comment #29 from Matthias Blaicher  ---
As a final data point, I've applied https://git.reviewboard.kde.org/r/125451/
on top of Plasma Desktop 5.6.4, and I have to say, this is godsend! No issues
anymore regarding removing screens, suspend in dock, docking, and all
combinations of it. Hell yeah!

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 345225] External monitors not activated after resuming from sleep

[Matthias Blaicher via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=345225

Matthias Blaicher  changed:

   What|Removed |Added

 CC||matth...@blaicher.com

--- Comment #9 from Matthias Blaicher  ---
While I have not yet fully understood the full standby logic of kscreen. It
seems to me that two things might change how the logic behaves:
 - The time it takes to suspend (kscreen uses a timer internally)
 - The state of the laptop lid

Hence, does it make a difference to you if the lid is opened or closed?

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 351025] When external monitor is disconnected, the laptop monitor doesn't become active again

[Matthias Blaicher via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=351025

Matthias Blaicher  changed:

   What|Removed |Added

 CC||matth...@blaicher.com

--- Comment #28 from Matthias Blaicher  ---
Given my huge issues with multi-monitors in general on multiple laptops. I've
decided to try reproducing other people's issues first. 

I am running Arch Linux with enabled testing repositories, which is
Qt: 5.6.0
Plasma Desktop: 5.6.4
KDE Applications: 16.04.1 
KScreen and LibKscreen are from git master (c949f3d, d80fede)
I have a intel Haswell card.

I've followed the steps described in the original post with an external VGA
monitor and could not reproduce the described state where all monitors are
disabled. 

However, the plasmashell  barely became active on the new monitor, with panels
missing and wrong background.  After some time (maybe after pressing Alt+F2),
the panels popped up again. 

Reattaching the monitor resulted In the expected beaviour, disabling the Laptop
monitor and only enabling the VGA screen. However, plasmashell behaved
strangely again: While the wallpaper was now right, the panels were in the
middle of the screen, probably at the positions from the old laptop screen.
This was followed by a plasmashell crash, involving Kscreen in some kind:

[KCrash Handler]
#5  0x7f894a69a7f0 in Plasma::Applet::actions() const () from
/usr/lib/libKF5Plasma.so.5
#6  0x00446303 in ?? ()
#7  0x0044647f in ?? ()
#8  0x7f894622b71c in QMetaObject::activate(QObject*, int, int, void**) ()
from /usr/lib/libQt5Core.so.5
#9  0x7f894b366d73 in
KScreen::Output::apply(QSharedPointer const&) () from
/usr/lib/libKF5Screen.so.7
#10 0x7f894b3480a4 in
KScreen::Config::apply(QSharedPointer const&) () from
/usr/lib/libKF5Screen.so.7
#11 0x7f894b354586 in
KScreen::ConfigMonitor::Private::updateConfigs(QSharedPointer
const&) () from /usr/lib/libKF5Screen.so.7
#12 0x7f894b356709 in
KScreen::ConfigMonitor::Private::edidReady(QDBusPendingCallWatcher*) () from
/usr/lib/libKF5Screen.so.7
#13 0x7f894622b71c in QMetaObject::activate(QObject*, int, int, void**) ()
from /usr/lib/libQt5Core.so.5
#14 0x7f894704167f in
QDBusPendingCallWatcher::finished(QDBusPendingCallWatcher*) () from
/usr/lib/libQt5DBus.so.5
#15 0x7f8947041778 in ?? () from /usr/lib/libQt5DBus.so.5
#16 0x7f894622c179 in QObject::event(QEvent*) () from
/usr/lib/libQt5Core.so.5
#17 0x7f89473b47bc in QApplicationPrivate::notify_helper(QObject*, QEvent*)
() from /usr/lib/libQt5Widgets.so.5
#18 0x7f89473b995f in QApplication::notify(QObject*, QEvent*) () from
/usr/lib/libQt5Widgets.so.5
#19 0x7f89461fe280 in QCoreApplication::notifyInternal2(QObject*, QEvent*)
() from /usr/lib/libQt5Core.so.5
#20 0x7f89462001fc in QCoreApplicationPrivate::sendPostedEvents(QObject*,
int, QThreadData*) () from /usr/lib/libQt5Core.so.5
#21 0x7f89462542c3 in ?? () from /usr/lib/libQt5Core.so.5
#22 0x7f8941db2dd7 in g_main_context_dispatch () from
/usr/lib/libglib-2.0.so.0
#23 0x7f8941db3040 in ?? () from /usr/lib/libglib-2.0.so.0
#24 0x7f8941db30ec in g_main_context_iteration () from
/usr/lib/libglib-2.0.so.0
#25 0x7f89462546cf in
QEventDispatcherGlib::processEvents(QFlags) ()
from /usr/lib/libQt5Core.so.5
#26 0x7f89461fc46a in
QEventLoop::exec(QFlags) () from
/usr/lib/libQt5Core.so.5
#27 0x7f8946204a0c in QCoreApplication::exec() () from
/usr/lib/libQt5Core.so.5
#28 0x0041c17c in ?? ()
#29 0x7f8945875741 in __libc_start_main () from /usr/lib/libc.so.6
#30 0x0041c2e9 in _start ()

Why does plasma anything to do with libkscreen?

In conclusion, maybe this is about the plasmashell not coming up again instead
of the monitor change not executed? At least #3 described seeing the mouse
cursor.

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 363238] Docked laptop lid close does not enable external monitor

[Matthias Blaicher via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=363238

--- Comment #2 from Matthias Blaicher  ---
Created attachment 99060
  --> https://bugs.kde.org/attachment.cgi?id=99060=edit
kscreen monitor output until crash

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 363238] Docked laptop lid close does not enable external monitor

[Matthias Blaicher via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=363238

--- Comment #1 from Matthias Blaicher  ---
Created attachment 99059
  --> https://bugs.kde.org/attachment.cgi?id=99059=edit
Logs collected by systemd

-- 
You are receiving this mail because:
You are watching all bug changes.

[KScreen] [Bug 363238] New: Docked laptop lid close does not enable external monitor

[Matthias Blaicher via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=363238

Bug ID: 363238
   Summary: Docked laptop lid close does not enable external
monitor
   Product: KScreen
   Version: 5.6.4
  Platform: Archlinux Packages
OS: Linux
Status: UNCONFIRMED
  Severity: crash
  Priority: NOR
 Component: libkscreen
  Assignee: se...@kde.org
  Reporter: matth...@blaicher.com

I have two available displays on a laptop with Intel gfx:
- Enabled Panel (LVDS1), id 67, 1600x900
- Disabled External Display (HDMI2), id 72, 2048x1152

Wen closing the lid, I'd expect the external monitor to get active, instead it
tries to enable a config without any monitor and gives up on that. A
kscreen-console runing in parallel crashes inside libKF5Screen
(KScreen::Output::name()).

I've attached debug logs. Feel free to let me test patches etc.

Reproducible: Always

Steps to Reproduce:
1. Laptop docked, panel enabled
2. Close Lid


Actual Results:  
Nothing happens, libkscreen crashes

Expected Results:  
Move to target monitor.

-- 
You are receiving this mail because:
You are watching all bug changes.