NVIDIA Open GPU Kernel Modules Version
610.57.04
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
CachyOS
Kernel Release
7.1.6-1-cachyos
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
GeForce RTX 4060
Describe the bug
I'm running into repeated GPU VA space mapping failures with Cyberpunk 2077 through Proton-CachyOS (Steam) on Linux, RTX 4060, driver 610.57.04 (nvidia-open-dkms). It only seems to happen when DLSS upscaling is turned on, and it lines up with visible screen corruption / static-like artifacting laid over the frame.
The kernel log shows the same handful of virtual addresses getting hit over and over throughout the session (e.g. 0x636c0000, 0x155600000, 0x2d600000, each one recurring multiple times over about 30 minutes).
Relevant kernel log lines:
NVRM: nvCheckFailedNoLog: Check failed: 0 == (pMapNode->gpuMask & gpuMask) @ gpu_vaspace.c:4552
NVRM: nvCheckOkFailedNoLog: Check failed: Invalid argument to call [NV_ERR_INVALID_ARGUMENT] (0x0000001F) returned from _gvaspaceMappingInsert(pGVAS, pGpu, pVASBlock, vaLo, vaHi, flags) @ gpu_vaspace.c:2022
NVRM: nvCheckFailedNoLog: Check failed: NV_OK == status @ virt_mem_allocator_gm107.c:2560
NVRM: dmaAllocMapping_GM107: can't update VA space for mapping @vaddr=0x636c0000
I also tested the exact same scenario on the proprietary 610.57.04 package. This specific VA space error does not occur there, but the session still crashes, this time with an Xid 109 (CTX SWITCH TIMEOUT):
NVRM: GPU at PCI:0000:01:00: GPU-4a505203-96c5-e80c-29d9-e49e7a6dca0a
NVRM: Xid (PCI:0000:01:00): 109, pid=8223, name=GameThread, channel 0x0000000a, errorString CTX SWITCH TIMEOUT, Info 0x14c010
Later in the same proprietary session, a CPU watchdog stall showed up, followed by dozens of repeated NVKMS/GEM memory allocation failures:
clocksource: Watchdog remote CPU 4 read timed out
[drm] [nvidia-drm] [GPU ID 0x00000100] Failed to allocate NVKMS video memory for GEM object, trying to fall back to sysmem
[drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to allocate NVKMS memory for GEM object
For comparison, I also tested with the older 580.x proprietary driver, and the game runs fine with no crashes or corruption at all. This suggests a regression was introduced somewhere between the 580 and 610 branches, showing up as this VA space mapping bug on the open kernel module, and as a different but related crash pattern on the 610 proprietary module, while being completely absent on 580.
No overclock/undervolt was in use (stock clocks and voltages), and GPU temperatures stayed normal (under 80°C) throughout testing, so hardware and thermal causes were ruled out.
This looks like it might be related to (issue #1191 that disappeared from here), which reports a similar NVKMS/GEM allocation failure pattern on a 5090 D running Ghostty/GTK4, also on the 610.x branch. The NVKMS/GEM failures I hit on proprietary 610 look like the same family of issue, just triggered by a different GPU and workload.
To Reproduce
- GPU: RTX 4060
- Driver: 610.57.04
- Launch Cyberpunk 2077 via Steam using Proton-CachyOS (I tried other protons and the same happened with others, but I used Proton-CachyOS the most.)
- Enable DLSS upscaling in Settings → Graphics
- Play for approximately 10-30 minutes of normal gameplay
- Observe sudden screen corruption (static/noise artifacts overlaying parts of the frame) correlating with VA space mapping errors in dmesg just before crashing
Bug Incidence
Always
nvidia-bug-report.log.gz
nvidia-bug-report.log.gz
More Info
Attaching a screenshot showing the visual corruption captured during an active bug event that correlates with the dmesg output submitted separately.
The VA space error specifically is confirmed absent on proprietary 610.57.04, so the checkbox above holds for the bug as reported. That said, the proprietary package is not clean under the same test conditions either, it fails differently (Xid 109, then watchdog stall, then NVKMS/GEM allocation failures), while 580.x proprietary has none of these issues. Flagging this as extra context in case it points to a broader regression across the 610.x branch, even though the specific bug reported here is scoped correctly to the open kernel module.
NVIDIA Open GPU Kernel Modules Version
610.57.04
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
CachyOS
Kernel Release
7.1.6-1-cachyos
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
GeForce RTX 4060
Describe the bug
I'm running into repeated GPU VA space mapping failures with Cyberpunk 2077 through Proton-CachyOS (Steam) on Linux, RTX 4060, driver 610.57.04 (nvidia-open-dkms). It only seems to happen when DLSS upscaling is turned on, and it lines up with visible screen corruption / static-like artifacting laid over the frame.
The kernel log shows the same handful of virtual addresses getting hit over and over throughout the session (e.g. 0x636c0000, 0x155600000, 0x2d600000, each one recurring multiple times over about 30 minutes).
Relevant kernel log lines:
I also tested the exact same scenario on the proprietary 610.57.04 package. This specific VA space error does not occur there, but the session still crashes, this time with an Xid 109 (CTX SWITCH TIMEOUT):
Later in the same proprietary session, a CPU watchdog stall showed up, followed by dozens of repeated NVKMS/GEM memory allocation failures:
For comparison, I also tested with the older 580.x proprietary driver, and the game runs fine with no crashes or corruption at all. This suggests a regression was introduced somewhere between the 580 and 610 branches, showing up as this VA space mapping bug on the open kernel module, and as a different but related crash pattern on the 610 proprietary module, while being completely absent on 580.
No overclock/undervolt was in use (stock clocks and voltages), and GPU temperatures stayed normal (under 80°C) throughout testing, so hardware and thermal causes were ruled out.
This looks like it might be related to (issue #1191 that disappeared from here), which reports a similar NVKMS/GEM allocation failure pattern on a 5090 D running Ghostty/GTK4, also on the 610.x branch. The NVKMS/GEM failures I hit on proprietary 610 look like the same family of issue, just triggered by a different GPU and workload.
To Reproduce
Bug Incidence
Always
nvidia-bug-report.log.gz
nvidia-bug-report.log.gz
More Info
Attaching a screenshot showing the visual corruption captured during an active bug event that correlates with the dmesg output submitted separately.
The VA space error specifically is confirmed absent on proprietary 610.57.04, so the checkbox above holds for the bug as reported. That said, the proprietary package is not clean under the same test conditions either, it fails differently (Xid 109, then watchdog stall, then NVKMS/GEM allocation failures), while 580.x proprietary has none of these issues. Flagging this as extra context in case it points to a broader regression across the 610.x branch, even though the specific bug reported here is scoped correctly to the open kernel module.
