Errors in dmesg on AGX Orin

Autonomous Machines Jetson & Embedded Systems Jetson AGX Orin
1

I see the following errors many times in dmesg when workloads are running. What do they mean?

[81390.503065] nvgpu: 17000000.ga10b nvgpu_ioctl_channel_reg_ops:1001 [ERR] dbg regops failed
[81390.511819] nvgpu: 17000000.ga10b validate_reg_op_offset:399 [ERR] invalid regop offset: 0x50ba54
[81390.521939] nvgpu: 17000000.ga10b nvgpu_regops_exec:235 [ERR] invalid op(s)
[81390.530627] nvgpu: 17000000.ga10b nvgpu_ioctl_channel_reg_ops:1001 [ERR] dbg regops failed

2

What kind of application you ran?

4

A deep learning training workload. What do these errors indicate?

5

Hi,
We would need to replicate the error and check further. Please share the steps for reproducing it on Orin developer kit.

6

It doesn’t seem to happen every time I run the workload, but only sometimes. Let me check if I can reproduce it and then share steps.

Any general idea on what the errors are related to?