NVIDIA Open GPU Kernel Modules Version
580.142 (Release Build, built Tue Mar 3 19:14:46 UTC 2026) Installed via RPM Fusion package: nvidia-kmod-580.142-2.fc43.src.rpm Full version string: NVIDIA UNIX Open Kernel Module for x86_64 580.142 Release Build (dvs-builder@U22-I3-AD05-51-5)
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.
Operating System and Version
Fedora Linux 43 (Workstation Edition)
Kernel Release
Linux fdesktop 6.19.12-200.fc43.x86_64 #1 SMP PREEMPT_DYNAMIC Sun Apr 12 15:26:33 UTC 2026 x86_64 GNU/Linux Stock Fedora kernel, not self-built. Issue also reproduced on 6.18.16, 6.19.8, 6.19.9, 6.19.10.
Please confirm you are running a stable release kernel (e.g. not a -rc). We do not accept bug reports for unreleased kernels.
Hardware: GPU
NVIDIA GeForce RTX 4070 Ti
Describe the bug
Recurring Xid 56 (display engine hang) on Wayland with a dual DisplayPort monitor setup. The screen freezes completely — no recovery, no TTY switching possible, hard reboot required. Occurs on every boot, typically within 6–40 minutes of login with no specific user action required to trigger it.
The kernel log shows repeated CMDre timeout entries followed by 0x00000200 display engine reset events:
NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000000 00002c88 000100a1 00000007 00000000
NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000000 00000200 00000001 00000005 0030001a
NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000000 00002888 000100a5 00000007 00000000
On one occasion the following was logged immediately before a Xid 56 burst:
NVRM: nvCheckOkFailedNoLog: Check failed: Out of memory [NV_ERR_NO_MEMORY]
returned from _memdescAllocInternal @ mem_desc.c:1359
The issue has been reproduced consistently since March 4, 2026 across driver versions 580.126.18 and 580.142, and kernel versions 6.18.x through 6.19.x.
To Reproduce
Hardware:
- GPU: NVIDIA GeForce RTX 4070 Ti
- CPU: AMD Ryzen 9 7950X (AM5 platform, ASUS B650 motherboard)
- Displays: 2× 2560×1440 via DisplayPort (both at 120 Hz)
- iGPU disabled in BIOS (amdgpu not loaded)
Steps:
- Boot Fedora 43 with the above hardware and driver 580.142 open kernel module
- Log into a GNOME Wayland session (mutter 49.5)
- Use the system normally with both monitors active
- Within 6–40 minutes, one display freezes completely with no recovery possible
The freeze is not triggered by any specific application. It occurs with only gnome-shell running and also with typical workloads (browser, terminal, Electron apps). It does not require elevated GPU load.
Additional trigger: hotplugging a third display via HDMI and then removing it reliably causes Xid 56 to begin firing within minutes of the next session, likely due to connector re-enumeration leaving stale display engine state.
Bug Incidence
Always
nvidia-bug-report.log.gz
nvidia-bug-report1.log.gz
nvidia-bug-report2.log.gz
More Info
- GSP firmware is active (confirmed: GSP Firmware Version: 580.142 in nvidia-smi -q)
- amdgpu module is not loaded (iGPU disabled in BIOS)
- NVreg_PreserveVideoMemoryAllocations=1 is set in /etc/modprobe.d/nvidia.conf but the parameter does not appear in sysfs (/sys/module/nvidia/parameters/ does not exist) and dmesg shows no acknowledgement — it is unclear whether the open kernel module honours NVreg_* parameters at all
- nvidia-persistenced was tested but worsened freeze frequency and was disabled
- The issue appears to be on the Wayland/KMS path; X11 could not be tested as GDM on Fedora 43 with the open kernel module fails to start an X session (black screen)
- monitors.xml contains a single clean configuration; stale multi-monitor configs were ruled out as a cause
- On one session, OBS was explicitly named in a Xid 56 entry (pid=..., name=obs, CMDre ...) while attempting display capture, suggesting the display engine was already degraded and external DMA triggered the visible hang
Expected: dual-monitor Wayland session operates stably indefinitely
Actual: display engine hangs within minutes to tens of minutes on every boot
nvidia-bug-report.log.gz is attached.
NVIDIA Open GPU Kernel Modules Version
580.142 (Release Build, built Tue Mar 3 19:14:46 UTC 2026) Installed via RPM Fusion package: nvidia-kmod-580.142-2.fc43.src.rpm Full version string: NVIDIA UNIX Open Kernel Module for x86_64 580.142 Release Build (dvs-builder@U22-I3-AD05-51-5)
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.
Operating System and Version
Fedora Linux 43 (Workstation Edition)
Kernel Release
Linux fdesktop 6.19.12-200.fc43.x86_64 #1 SMP PREEMPT_DYNAMIC Sun Apr 12 15:26:33 UTC 2026 x86_64 GNU/Linux Stock Fedora kernel, not self-built. Issue also reproduced on 6.18.16, 6.19.8, 6.19.9, 6.19.10.
Please confirm you are running a stable release kernel (e.g. not a -rc). We do not accept bug reports for unreleased kernels.
Hardware: GPU
NVIDIA GeForce RTX 4070 Ti
Describe the bug
Recurring Xid 56 (display engine hang) on Wayland with a dual DisplayPort monitor setup. The screen freezes completely — no recovery, no TTY switching possible, hard reboot required. Occurs on every boot, typically within 6–40 minutes of login with no specific user action required to trigger it.
The kernel log shows repeated CMDre timeout entries followed by 0x00000200 display engine reset events:
NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000000 00002c88 000100a1 00000007 00000000
NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000000 00000200 00000001 00000005 0030001a
NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000000 00002888 000100a5 00000007 00000000
On one occasion the following was logged immediately before a Xid 56 burst:
NVRM: nvCheckOkFailedNoLog: Check failed: Out of memory [NV_ERR_NO_MEMORY]
returned from _memdescAllocInternal @ mem_desc.c:1359
The issue has been reproduced consistently since March 4, 2026 across driver versions 580.126.18 and 580.142, and kernel versions 6.18.x through 6.19.x.
To Reproduce
Hardware:
Steps:
The freeze is not triggered by any specific application. It occurs with only gnome-shell running and also with typical workloads (browser, terminal, Electron apps). It does not require elevated GPU load.
Additional trigger: hotplugging a third display via HDMI and then removing it reliably causes Xid 56 to begin firing within minutes of the next session, likely due to connector re-enumeration leaving stale display engine state.
Bug Incidence
Always
nvidia-bug-report.log.gz
nvidia-bug-report1.log.gz
nvidia-bug-report2.log.gz
More Info
Expected: dual-monitor Wayland session operates stably indefinitely
Actual: display engine hangs within minutes to tens of minutes on every boot
nvidia-bug-report.log.gz is attached.