hyprlock crashes on monitor power cycle (DPMS) since 0.9.4 [NVIDIA]

[umibear]

Regression?

Yes, worked on 0.9.2,

Hyprlock Info and Version

hyprlock version v0.9.4

Hyprlock config

source = ~/.config/omarchy/current/theme/hyprlock.conf

general {
ignore_empty_input = true
}

background {
monitor =
color = $color
path = ~/.config/omarchy/current/background
blur_passes = 3
}

animations {
enabled = false
}

input-field {
monitor =
size = 650, 100
position = 0, 0
halign = center
valign = center

inner_color = $inner_color
outer_color = $outer_color
outline_thickness = 4

font_family = CaskaydiaMono Nerd Font
font_color = $font_color

placeholder_text = Enter Password
check_color = $check_color
fail_text = <i>$FAIL ($ATTEMPTS)</i>

rounding = 0
shadow_passes = 0
fade_on_empty = false

}

auth {
fingerprint:enabled = false
}

cat ~/.config/omarchy/current/theme/hyprlock.conf
$color = rgba(28,27,25, 1.0)
$inner_color = rgba(28,27,25, 0.8)
$outer_color = rgba(216,208,192, 1.0)
$font_color = rgba(216,208,192, 1.0)
$check_color = rgba(224,128,96, 1.0)

Compositor Info and Version

System/Version info

Hyprland 0.54.3 built from branch v0.54.3 at commit 521ece463c4a9d3d128670688a34756805a4328f clean (version: bump to 0.54.3).
Date: Fri Mar 27 18:17:50 2026
Tag: v0.54.3, commits: 7014

Libraries:
Hyprgraphics: built against 0.5.0, system has 0.5.1
Hyprutils: built against 0.12.0, system has 0.12.0
Hyprcursor: built against 0.1.13, system has 0.1.13
Hyprlang: built against 0.6.8, system has 0.6.8
Aquamarine: built against 0.10.0, system has 0.10.0

Version ABI string: 521ece463c4a9d3d128670688a34756805a4328f_aq_0.10_hu_0.12_hg_0.5_hc_0.1_hlg_0.6
no flags were set


System Information:
System name: Linux
Node name: omarchy
Release: 6.19.11-arch1-1
Version: #1 SMP PREEMPT_DYNAMIC Thu, 02 Apr 2026 23:33:01 +0000

Libraries:
Hyprgraphics: built against 0.5.0, system has 0.5.1
Hyprutils: built against 0.12.0, system has 0.12.0
Hyprcursor: built against 0.1.13, system has 0.1.13
Hyprlang: built against 0.6.8, system has 0.6.8
Aquamarine: built against 0.10.0, system has 0.10.0



GPU information:
01:00.0 VGA compatible controller [0300]: NVIDIA Corporation GB203 [GeForce RTX 5080] [10de:2c02] (rev a1) (prog-if 00 [VGA controller])
NVRM version: NVIDIA UNIX Open Kernel Module for x86_64  595.58.03  Release Build  (root@omarchy)


os-release: NAME="Arch Linux"
PRETTY_NAME="Arch Linux"
ID=arch
BUILD_ID=rolling
ANSI_COLOR="38;2;23;147;209"
HOME_URL="https://archlinux.org/"
DOCUMENTATION_URL="https://wiki.archlinux.org/"
SUPPORT_URL="https://bbs.archlinux.org/"
BUG_REPORT_URL="https://gitlab.archlinux.org/groups/archlinux/-/issues"
PRIVACY_POLICY_URL="https://terms.archlinux.org/docs/privacy-policy/"
LOGO=archlinux-logo


plugins:

Explicit sync: supported
GL ver: 3.2
Backend: drm

Monitor info:
        Panel DP-1: 3440x1440, DP-1 LG Electronics LG ULTRAGEAR 111NTNH0L615 -> backend drm
                explicit ✔️
                edid:
                        hdr ✔️
                        chroma ✔️
                        bt2020 ✔️
                vrr capable ✔️
                non-desktop ❌

Description

Hyprlock crashes when the monitor is powered off while the lock screen is active, causing Hyprland to display the "oopsie daisy" screen. It worked on 0.9.3-2, on resume i didn't have wallpaper visible, only black screen and password field but it was working without crashes.
Note: nvidia-open-dkms was also upgraded on the same day (to 595.58.03)

How to reproduce

1. Lock screen (hyprlock running)
2. Power off monitor (physical button or hyprctl dispatch dpms off)
3. Power monitor back on
4. Hyprlock is gone — Hyprland shows "oopsie daisy screen"

Crash reports, logs, images, videos

hyprlock-verbose.log

[imanubdesigner]

Yes! Same here with Omarchy 3.5.1

[vaxerski]

@PointerDilemma maaax

[vaxerski]

please get a debug stacktrace in the meantime, instructions are the same as for hyprland https://wiki.hypr.land/Crashes-and-Bugs/#obtaining-a-debug-stacktrace

[umibear]

I tried to generate a debug stacktrace with the guide but no coredump was generated for hyprlock. With the help of an AI assistant (I don't have enough knowledge to debug on my own), I used strace instead. Sharing findings if it's any helpful

Command used:

strace -f -e trace=all -o /tmp/hyprlock-strace.log ~/hyprlock-debug/build/hyprlock

debug build of hyprlock 0.9.4, compiled with -DCMAKE_BUILD_TYPE=Debug. Triggered crash by powering off monitor

No SIGSEGV or SIGABRT was observed. All threads are killed by SIGINT {si_code=SI_KERNEL}, with all GPU ioctls returning success before death:

8356 futex(0x557e8fd76cac, ...) <unfinished ...>
8360 <... poll resumed>) = 0 (Timeout)
8360 recvmsg(3, {msg_namelen=0}, MSG_DONTWAIT|MSG_CMSG_CLOEXEC) = -1 EAGAIN (Resource temporarily unavailable)
8360 poll([{fd=3, events=POLLIN}], 1, 5000 <unfinished ...>
8356 <... futex resumed>) = ? ERESTART_RESTARTBLOCK (Interrupted by signal)
8356 --- SIGINT {si_signo=SIGINT, si_code=SI_KERNEL} ---
8362 +++ killed by SIGINT +++
8361 +++ killed by SIGINT +++
8360 +++ killed by SIGINT +++
8358 +++ killed by SIGINT +++
8359 +++ killed by SIGINT +++
8356 +++ killed by SIGINT +++

If this isn't useful i'm happy to try again but will need some guidance on how to proceed.

[vaxerski]

I don't think we can send SIGINT to ourselves by accident. Odd.
I suspect ca2aa5e, can you try reverting it?

[PointerDilemma]

Judging by the logs, culprit is probably #988 although I don't see how.
It doesn't crash for some reason.

@umibear

since you have a local build can you git revert ca2aa5e5e3bec40ca79d5343eb54ea0f8c8a9e8c build again and see if that fixes it?

[umibear]

I did the revert, rebuilt and tested. Crash still occurs on monitor power cycle.

[umibear]

I downgraded the packages updated around when I started seeing crashes, but none of them fixed it:

• hyprlock → 0.9.3-2
• nvidia-open-dkms → 590.48.01
• xorg-xwayland → 24.1.9

Still crashing

One observation: the crash only happens if the monitor has been off for ~30+ seconds. Powering it back on quickly (within ~5 seconds) does not trigger the crash. Not sure if that's useful but wanted to mention it.

[umibear]

Seems like i found the culprit, i was wrong to assume it was hyprlock update which just happened to come at the same time. In hyprlock.conf when i comment out path and delete animations block i don't get the crash anymore.
Note: I hadn't made any changes to my hyprlock config when the crashes started — the config with path and animations was working at some point.

Two independent config elements that trigger the crash:

1. path = in the background block (any image)
2. animations {} block

i commented out the path and deleted animations block and don't crash anymore.

working Hyprlock config

source = ~/.config/omarchy/current/theme/hyprlock.conf

general {
    ignore_empty_input = true
}

background {
    monitor =
    #path = ~/.config/omarchy/current/background
    color = $color
}

input-field {
    monitor =
    size = 650, 100
    position = 0, 0
    halign = center
    valign = center
    inner_color = $inner_color
    outer_color = $outer_color
    outline_thickness = 4
    font_family = CaskaydiaMono Nerd Font
    font_color = $font_color
    placeholder_text = Enter Password
    check_color = $check_color
    fail_text = <i>$FAIL ($ATTEMPTS)</i>
    rounding = 0
    shadow_passes = 0
    fade_on_empty = false
}

Tested with all packages on latest versions (hyprlock 0.9.4, nvidia 595.58.03, xwayland 24.1.10).

[PointerDilemma]

Thanks a lot for debugging!!
So I tried to reproduce based on this, but wasn't able to.
But considering all the dmabuf logs, maybe this: #992 fixes it.
Could you someone that can reproduce test that branch?

[umibear]

Tested PR #992 branch (commit 03d1cdc). Re-enabled path = in config and ran the debug build. Powered off monitor, got the crash. Log attached.

hyprlock-pr992.log

[PointerDilemma]

OK thanks how about #994? That one has some additional logging as well, which may be helpful.

[umibear]

#994 fixed the issue. Tested few times. Ran debug build, manual power off, waited 1 min working lockscreen. Second test normal way, let hypridle listener to kick in and dpms off monitor, waited approx 4 minutes and still got working lockscreen.