Bug 505812

Summary:	Wayland: Severe flickering / screen blanking with screenshot region selection on NVIDIA
Product:	[Plasma] kwin	Reporter:	amatira <nospampls.essential677>
Component:	wayland-generic	Assignee:	KWin default assignee <kwin-bugs-null>
Status:	RESOLVED WORKSFORME
Severity:	normal	CC:	kdedev, nate, xaver.hugl
Priority:	NOR
Version First Reported In:	6.3.5
Target Milestone:	---
Platform:	Arch Linux
OS:	Linux
Latest Commit:		Version Fixed/Implemented In:
Sentry Crash Report:

Description amatira 2025-06-20 00:29:14 UTC

On a KDE Wayland session with NVIDIA open drivers (5070ti), using any screenshot tool  (like Spectacle or Flameshot) causes the screen on which it is active (multi-monitor setup) to blank out (go black) for the duration of the selection. Can press ESC to restore the screen then. It seems to work sometimes after a reboot, but uncertain about the reasons why and when. This was happening with Plasma 6.3 as well as with 6.4. Over the range of multiple nvidia drivers as well.

This is a sample captured using the command: strace -e trace=ioctl spectacle 2>&1 | grep -v EAGAIN:

ctl(24, _IOC(_IOC_READ|_IOC_WRITE, 0x46, 0x4f, 0x20), 0x7ffe4b18f4f0) = 0
ioctl(24, _IOC(_IOC_READ|_IOC_WRITE, 0x46, 0x4f, 0x20), 0x7ffe4b18f4f0) = 0
ioctl(24, _IOC(_IOC_READ|_IOC_WRITE, 0x46, 0x4f, 0x20), 0x7ffe4b18f4f0) = 0
ioctl(24, _IOC(_IOC_READ|_IOC_WRITE, 0x46, 0x4f, 0x20), 0x7ffe4b18f4f0) = 0
ioctl(25, _IOC(_IOC_READ|_IOC_WRITE, 0x6d, 0, 0x10), 0x7ffe4b18f5e0) = 0
ioctl(24, _IOC(_IOC_READ|_IOC_WRITE, 0x46, 0x29, 0x10), 0x7ffe4b18f5d0) = 0
ioctl(24, _IOC(_IOC_READ|_IOC_WRITE, 0x46, 0x29, 0x10), 0x7ffe4b18f630) = 0
ioctl(15, _IOC(_IOC_READ|_IOC_WRITE, 0x46, 0x41, 0x38), 0x7ffe4b18f4e0) = 0
ioctl(15, _IOC(_IOC_READ|_IOC_WRITE, 0x46, 0x2a, 0x20), 0x7ffe4b18eb40) = 0
ioctl(15, _IOC(_IOC_READ|_IOC_WRITE, 0x46, 0x29, 0x10), 0x7ffe4b18f520) = 0
ioctl(15, _IOC(_IOC_READ|_IOC_WRITE, 0x46, 0x29, 0x10), 0x7ffe4b18f520) = 0
ioctl(15, _IOC(_IOC_READ|_IOC_WRITE, 0x46, 0x29, 0x10), 0x7ffe4b18f600) = 0
ioctl(15, _IOC(_IOC_READ|_IOC_WRITE, 0x46, 0x29, 0x10), 0x7ffe4b18f550) = 0
ioctl(15, _IOC(_IOC_READ|_IOC_WRITE, 0x46, 0x29, 0x10), 0x7ffe4b18f570) = 0
ioctl(15, _IOC(_IOC_READ|_IOC_WRITE, 0x46, 0x2a, 0x20), 0x7ffe4b18eb60) = 0
ioctl(15, _IOC(_IOC_READ|_IOC_WRITE, 0x46, 0x29, 0x10), 0x7ffe4b18f630) = 0

And after filtering out QT_LOGGING_RULES="*.debug=true" spectacle - there are multiple instances of:

qt.qpa.wayland.backingstore: requestUpdate
qt.qpa.wayland.backingstore: requestUpdate called twice without committing anything

So it could be related.
I would've attached some screenshots, but oh well.. (:

Operating System: CachyOS Linux 
KDE Plasma Version: 6.4.0
KDE Frameworks Version: 6.15.0
Qt Version: 6.9.1
Kernel Version: 6.15.2-3-cachyos (64-bit)
Graphics Platform: Wayland
Processors: 16 × AMD Ryzen 7 5800X 8-Core Processor
Memory: 64 GiB of RAM (62.7 GiB usable)
Graphics Processor: NVIDIA GeForce RTX 5070 Ti

Comment 1 Nate Graham 2025-06-20 14:45:50 UTC

Does it still happen with only one screen?

Comment 2 amatira 2025-06-20 15:08:01 UTC

I kept my PC running over night, meaning it is still the same session where I had the issues, but had both screens turned off.
Turned them both on again and interestingly enough (and of course) - everything works atm.

As soon as it happens again, I'll play around with it a bit and test it.

Comment 3 Zamundaaa 2025-06-25 15:29:41 UTC

This is most likely a driver bug around adaptive sync handling with low fps content. If it happens again, please check if turning off adaptive sync fixes it

Comment 4 amatira 2025-06-25 18:18:02 UTC

The thing is, I am currently on vanilla arch and could not reproduce it so far.
Then again, it was always something that happened so sporadically and also that I couldn't find anyone with the same problem,
that it must have been either a weird combination of my hardware (monitor.. EDID profiles.. I don't know) and / or something that would only happen on CachyOS due to the settings of it, which are by default:

options nvidia NVreg_UsePageAttributeTable=1 \
    NVreg_InitializeSystemMemoryAllocations=0 \
    NVreg_DynamicPowerManagement=0x02 \
    NVreg_RegistryDwords=RMIntrLockingMode=1

It's a tough one. If this happens again or I'll find a solution to this I'll report. What I can say though is, that switching off VRR did not help in my case. That was something I've tested. Since this would mostly happen to me on the desktop, I'd say that it was never a low fps situation. Well, it shouldn't be at least. I kept VRR always on 'Automatic', but yeah, turning it off was no shot.

Comment 5 Bug Janitor Service 2025-07-10 03:48:04 UTC

🐛🧹 ⚠️ This bug has been in NEEDSINFO status with no change for at least 15 days. Please provide the requested information, then set the bug status to REPORTED. If there is no change for at least 30 days, it will be automatically closed as RESOLVED WORKSFORME.

For more information about our bug triaging procedures, please read https://community.kde.org/Guidelines_and_HOWTOs/Bug_triaging.

Thank you for helping us make KDE software even better for everyone!

Comment 6 Bug Janitor Service 2025-07-25 03:47:04 UTC

🐛🧹 This bug has been in NEEDSINFO status with no change for at least 30 days. Closing as RESOLVED WORKSFORME.