Jetson Orin Nano Devkit / JetPack 7.2 (R39.2.0) — nvarguscamerasrc fails with NvRmGpuLibOpen error=4, DCE RPC control call failures at boot

Robotics & Edge Computing Jetson Systems Jetson Orin Nano
1

Environment

Hardware: Jetson Orin Nano Super Developer Kit (P3767-0005)
L4T: R39 (release) REVISION 2.0, DATE Mon Jun 1 09:28:48 UTC 2026
JetPack: 7.2
Kernel: 6.8.12-1021-tegra (PREEMPT, built Jun 1 2026)
OS: Ubuntu 24.04
Packages: nvidia-l4t-core / -init / -initrd all 39.2.0-20260601141651
Cameras: 2x Arducam B0181 IMX219 autofocus via CSI (J20 + J219)
User: member of video and render groups

Symptom

Every gscam / nvarguscamerasrc client on this system segfaults during
CameraProvider initialization. Reproducer:

gst-launch-1.0 nvarguscamerasrc sensor-id=0 !
‘video/x-raw(memory:NVMM), width=1280, height=720, format=NV12,
framerate=30/1’ ! nvvidconv ! fakesink

Client-side output:
libnvrm_gpu.so: NvRmGpuLibOpen failed, error=4
(then SIGSEGV, exit code -11)

nvargus-daemon log at the moment of failure:
nvargus-daemon: (Argus) Error 0x00000006:
(propagating from src/api/GlobalProcessState.cpp,
function createCameraProvider(), line …)
nvargus-daemon: === gscam_node[PID]: CameraProvider failed to initialize
nvargus-daemon: WARNING: CameraProvider was not destroyed before client
connection terminated. The client may have abnormally terminated.

Root cause appears to be earlier, in the NVRM kernel module load. dmesg at boot:
[11.920981] NVRM: loading NVIDIA UNIX Open Kernel Module for aarch64
TempVersion Release Build (r595_00) … Jun 1 2026
[12.212696] NVRM: GPU0 rpcRmApiControl_dce: NVRM_RPC_DCE:
Failed RM ctrl call cmd:0x731341 result 0xffff:
Failure: Generic Error [NV_ERR_GENERIC]
[12.213962] NVRM: GPU0 rpcRmApiControl_dce: NVRM_RPC_DCE:
Failed RM ctrl call cmd:0x730282 result 0x1f:
Invalid argument to call [NV_ERR_INVALID_ARGUMENT]

The two RPC failures to DCE occur ~300 ms after NVRM module load and every
subsequent GPU userspace call inherits the broken state, including
NvRmGpuLibOpen from Argus. Note the driver banner reads “TempVersion
Release Build (r595_00)” which is unusual — production banners don’t
normally say TempVersion.

What I’ve already ruled out

  • nvargus-daemon is active/running (systemctl status green)
  • User has video + render group membership
  • CMA is not exhausted: CmaTotal 262144 kB, CmaFree 252652 kB
  • No cma_alloc failures in dmesg
  • No pending nvidia-* package upgrades in apt (apt list --upgradable is empty)
  • sudo systemctl restart nvargus-daemon then relaunch: same failure
  • Failure is deterministic across many reboots and every launch attempt
  • Both sensors on separate CSI ports fail identically, so not a sensor
    hardware issue; both IMX219 modules were confirmed working under an
    earlier L4T on this same board

Ask

Is the “TempVersion” driver banner intentional in the shipped 7.2 image?
Is a patched nvidia-l4t-* package expected to land in apt, or is a full
JetPack reinstall the recommended path? The DCE RPC failure at kernel
module load is the first thing I’d want investigated — everything else
downstream (Argus / gscam / any CUDA client) inherits the broken GPU state.

2

Hello @znhartman!

Based on the title and content of your topic, it looks like it may receive better visibility and feedback in a different category. We took the liberty of moving it for you.

If this was an incorrect assessment, please send me a direct message.

Disclaimer: this moderation suggestion and message were generated with AI assistance.

4

hello znhartman,

may I know your reproduce steps? is it native JP-7.2 public release?

the NVRM/DCE GPU initialization failure does not looks like a camera sensor or Argus plugin defect.
let’s try with Applications Using V4L2 IOCTL Directly to use V4L2 IOCTL to verify basic functionality.
for instance,
$ v4l2-ctl -d /dev/video0 --set-fmt-video=width=1920,height=1080,pixelformat=RG10 --set-ctrl bypass_mode=0 --stream-mmap --stream-count=100

you may also gather the complete logs for reference. such as.. $ dmesg > klogs.txt