noctalia-shell crashes when disconnecting HDMI on ASUS E510 (Intel Celeron N4020)

[Angel173hu]

When disconnecting an HDMI monitor, Niri logs the following error:

ERROR niri::backend::tty: missing surface in vblank callback for crtc crtc::Handle(117)

and the process noctalia-shell immediately crashes.
This happens consistently on my ASUS E510 laptop with Intel Celeron N4020, but does not happen on another laptop with the same Niri configuration (noctalia-shell stays alive after HDMI disconnect).
It seems to be a bug specific to the combination of Niri + the graphics driver on this hardware.

Steps to reproduce

Start a Niri session with noctalia-shell (spawn-sh-at-startup).
Connect an HDMI monitor (detected correctly as HDMI-A-1).
Disconnect the HDMI monitor.
Niri shows the “missing surface in vblank callback” error and noctalia-shell crashes.
Relevant logs

DEBUG niri::backend::tty: disconnecting connector: "HDMI-A-1"
feb 10 09:50:08 enyel-cachyos-x8664 niri[1503]: 2026-02-10T15:50:08.494385Z DEBUG niri::niri: putting output eDP-1 at x=0 y=0
feb 10 09:50:08 enyel-cachyos-x8664 niri[1503]: 2026-02-10T15:50:08.565347Z ERROR niri::backend::tty: missing surface in vblank callback for crtc crtc::Handle(117)

System Information

• niri version: 25.11 (b35bcae)
• quickshell version: 0.2.1 (distributed by Arch Linux)
• Distro: CachyOS (Arch-based, selected Niri as desktop environment, noctalia came preinstalled)
• GPU: Intel UHD Graphics (integrated in Celeron N4020)
• CPU: Intel Celeron N4020 (ASUS E510 laptop)

[Angel173hu]

Hi, thanks a lot for looking into this so quickly!
Yes, a PR would definitely be helpful.

For context:

• I’m running CachyOS (Arch-based) and selected Niri as my desktop environment.
• Noctalia came preinstalled with this setup.
• My hardware is an ASUS E510 laptop with Intel Celeron N4020 and integrated Intel UHD Graphics.
• Niri version: 25.11 (b35bcae)
• quickshell version: 0.2.1 (distributed by Arch Linux)

On this machine, disconnecting HDMI consistently causes noctalia-shell to crash with the “missing surface in vblank callback” error. On another laptop with the same config, the crash does not happen, so it seems hardware/driver-specific.

Thanks again for offering to help!

[YaLTeR]

The comment by @AiGentsy reads like a bot, and also contains no information whatsoever

[Angel173hu]

Oh D: El El mar, 10 de feb de 2026 a la(s) 10:45 a.m., Ivan Molodetskikh < ***@***.***> escribió:

…

*YaLTeR* left a comment (niri-wm/niri#3426) <#3426 (comment)> The comment by @AiGentsy <https://github.com/AiGentsy> reads like a bot, and also contains no information whatsoever — Reply to this email directly, view it on GitHub <#3426 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AZJ7DB5NGBG6J6PWDCIF2UL4LIDJPAVCNFSM6AAAAACUUKZBE2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTQNZZGI2DMMZWHA> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[Sempyos]

Can you provide the output of journalctl --user-unit niri -xe when it happens

[fishman]

I think this is a qt bug triggered in quickshell. I had reported this to the noctalia repo, but it seems like it's quickshell that crashes.

[Sempyos]

Can you link to that..?

[fishman]

noctalia-dev/noctalia-shell#1141 (comment)

quickshell-mirror/quickshell#477

https://github.com/qt/qtbase/blob/8a27a432a095f962e5f27cdfa76646519f568c8e/src/plugins/platforms/wayland/qwaylanddisplay.cpp#L412

[fishman]

This ones links might be more useful noctalia-dev/noctalia-shell#1822

[bryanparadis]

This ones links might be more useful noctalia-dev/noctalia-shell#1822

I put a lot of context in there.

I think that basically what is happening is that WlOutput is being used by Qt when it's already been destroyed or the hotplugs somehow sent a bad one along. I am guessing that Niri still is showing it as valid when it shouldn't be.

You can see it trying to do vblank on a missing surface. Some kind of object lifetime issue. I am not at all familiar with rust and this is a pretty gnarly thing to look at.

You can reproduce by replugging quickly 5-10 times hdmi in your monitor very consistently. I see quickshell, qt and steam stuff crash when doing this. Obviously things could handle it better but I think something is going wrong smithay/niri.

[bryanparadis]

quickshell-mirror/quickshell#578
https://github.com/qt/qtbase/pull/136

[cmeissl]

@bryanparadis can you test if #3502 fixes the issue?