nvidia-drm NVKMS GEM allocation failure followed by Ghostty SIGSEGV in libnvidia-eglcore.so.610.43.02

[derekszen]

NVIDIA Open GPU Kernel Modules Version

nvidia-open-dkms 610.43.02-2 / nvidia-utils 610.43.02-2

Please confirm this issue does not happen with the proprietary driver (of the same version). This issue tracker is only for bugs specific to the open kernel driver.

• I confirm that this does not happen with the proprietary driver package.

I have not tested the proprietary driver package of the same version yet. I am filing this as an open-kernel-module incident because the system is currently using nvidia-open-dkms, and the kernel-side failure came from nvidia_drm immediately before a userspace crash in NVIDIA EGL.

Operating System and Version

Arch Linux / Omarchy 3.8.2

Kernel Release

Linux omarchyNeptune 7.0.11-arch1-1 #1 SMP PREEMPT_DYNAMIC Tue, 02 Jun 2026 18:26:58 +0000 x86_64 GNU/Linux

Please confirm you are running a stable release kernel (e.g. not a -rc). We do not accept bug reports for unreleased kernels.

• I am running on a stable kernel release.

Hardware: GPU

NVIDIA GeForce RTX 5090 D, driver 610.43.02

Describe the bug

I hit a GPU/rendering failure on Arch Linux with the open NVIDIA kernel module stack. I do not have a reliable reproducer yet, but the timeline shows nvidia-drm allocation errors immediately before a Ghostty/GTK4 process crashed in NVIDIA EGL code. The machine then appears to have hung or rebooted abnormally.

Environment:

GPU: NVIDIA GeForce RTX 5090 D
Driver: 610.43.02
nvidia-open-dkms: 610.43.02-2
nvidia-utils: 610.43.02-2
Kernel: 7.0.11-arch1-1
Hyprland: 0.55.2-2
Ghostty: 1.3.1-arch2
GTK4: 4.22.4-1
Omarchy: 3.8.2

Observed timeline:

2026-06-11 15:15:10 CST
kernel: [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to allocate NVKMS memory for GEM object
kernel: [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to allocate NVKMS memory for GEM object
kernel: ghostty[1007163]: segfault at 0 ip 00007fe2a64f4390 sp 00007ffe361311d0 error 6 in libnvidia-eglcore.so.610.43.02[af4390,7fe2a5c19000+c46000] likely on CPU 7 (core 7, socket 0)

Coredump details:

Signal: SIGSEGV
Executable: /usr/bin/ghostty
Command line:
ghostty --class=org.omarchy.screensaver --config-file=~/.local/share/omarchy/default/ghostty/screensaver --font-size=18 -e omarchy-screensaver

The coredump stack starts in libnvidia-eglcore.so.610.43.02, then continues into the GTK4 render stack:

#0  libnvidia-eglcore.so.610.43.02 + 0xaf4390
#1  libnvidia-eglcore.so.610.43.02 + 0xa38c33
#2  libnvidia-eglcore.so.610.43.02 + 0xa43b5c
#3  libnvidia-eglcore.so.610.43.02 + 0x77e3a4
#4  libgtk-4.so.1 + 0x5cf7b8
#7  gsk_renderer_render (libgtk-4.so.1 + 0x59d9a3)
#23 ghostty + 0x135de95

The previous boot journal stopped at 2026-06-11 15:15:24 CST, and the next boot began at 2026-06-11 15:16:35 CST, suggesting the machine either hung or rebooted abnormally after the crash. I do not yet know whether the Ghostty crash caused the reboot/hang, was a symptom of an earlier GPU state problem, or was only temporally correlated.

After reboot, the GPU appeared healthy:

NVIDIA GeForce RTX 5090 D, 610.43.02, 32607 MiB total, 1926 MiB used, 35 C, 5% utilization

To Reproduce

I do not have a reliable reproducer yet. The affected userspace process was Omarchy's Ghostty-based screensaver:

ghostty --class=org.omarchy.screensaver --config-file=~/.local/share/omarchy/default/ghostty/screensaver --font-size=18 -e omarchy-screensaver

Bug Incidence

Once

nvidia-bug-report.log.gz

Not attached yet. I can generate and review nvidia-bug-report.log.gz if needed, but I did not want to upload a full machine report before checking it for local/private details. I understand the issue may be closed without this file.

More Info

Questions:

• Is Failed to allocate NVKMS memory for GEM object expected to be recoverable, or can it leave the EGL/GTK rendering path in a state where clients crash inside libnvidia-eglcore.so?
• Are there specific NVIDIA debug flags, module parameters, or additional logs that would be useful if this happens again?

[derekszen]

Related public reports I found after filing this issue. I did not find an exact all-fields match for this incident, but these look close enough to be useful context:

• NVIDIA forum: 610.43.02 pageflip timeout / black screen / reboot on Arch/Wayland, with libnvidia-eglcore.so.610.43.02 in the stack: https://forums.developer.nvidia.com/t/610-43-02-pageflip-timed-out-black-screen-randomly-while-actively-used-reboot-required-to-fix-arch/372045
• NVIDIA forum: driver 610, kernel 7.0.x, repeated NVKMS GEM allocation errors, display loss, libnvidia-eglcore.so.610.43.02, reboot needed: https://forums.developer.nvidia.com/t/rtx-3070-ti-610-430-02-gpu-mmu-tlb-failure-causes-display-loss-on-wayland/372569
• NVIDIA forum: nvidia-open-dkms / Wayland / XWayland with exact Failed to allocate NVKMS memory for GEM object: https://forums.developer.nvidia.com/t/nvidia-open-dkms-wayland-xwayland-failed-to-allocate-nvkms-memory-for-gem-object/371922
• NVIDIA/open-gpu-kernel-modules Unable to allocate RAM for VRAM when VRAM is full #1078: open-kernel report for VRAM/full allocation failures; comments include the exact NVKMS GEM error and sysmem fallback behavior: Unable to allocate RAM for VRAM when VRAM is full #1078
• NVIDIA/open-gpu-kernel-modules [GTX 1650 4GB] VRAM fragmentation causes nvAssertFailedNoLog in virt_mem_allocator_gm107.c leading to Wayland crash and system deadlock on shutdown #1165: Wayland crash/deadlock after NVKMS allocation failures; mentions opening GPU-accelerated apps such as Ghostty/browser as a trigger in a stressed GPU state: [GTX 1650 4GB] VRAM fragmentation causes nvAssertFailedNoLog in virt_mem_allocator_gm107.c leading to Wayland crash and system deadlock on shutdown #1165
• NVIDIA/egl-wayland Supported devices list request. #185: Ghostty coredumps plus exact NVKMS GEM allocation errors under GPU memory pressure, on a different driver/GPU: System memory is never used for GPU processes in Wayland causing crashes, but it works fine in X11 egl-wayland#185

The exact combination I observed remains: Omarchy screensaver launched via Ghostty, nvidia-open-dkms/nvidia-utils 610.43.02, Failed to allocate NVKMS memory for GEM object, then Ghostty SIGSEGV in libnvidia-eglcore.so.610.43.02.