Cpufreq spam in dmesg

We’re recently experience some dmesg spam from cpufreq. We’re using a custom board with jetpack 6.1.

Any idea how we can manage this?

[  689.134328] cpufreq: cpu0,cur:1311000,set:1510400,delta:199400,set ndiv:118
[  689.146403] cpufreq: cpu0,cur:981000,set:806400,delta:174600,set ndiv:63118
[  694.195597] cpufreq: cpu0,cur:881000,set:1510400,delta:629400,set ndiv:118
[  648.639156] cpufreq: cpu0,cur:1317000,set:1510400,delta:193400,set ndiv:118
[  648.675534] cpufreq: cpu0,cur:1319000,set:1510400,delta:191400,set ndiv:118
[  648.677576] cpufreq: cpu4,cur:1238000,set:729600,delta:508400,set ndiv:5788
[  663.837835] cpufreq: cpu0,cur:1509000,set:883200,delta:625800,set ndiv:6918
[  663.838561] cpufreq: cpu0,cur:1322000,set:1510400,delta:188400,set ndiv:118
[  663.851872] cpufreq: cpu0,cur:1299000,set:1510400,delta:211400,set ndiv:118
[  663.852602] cpufreq: cpu0,cur:1001000,set:1510400,delta:509400,set ndiv:118
[  663.853627] cpufreq: cpu4,cur:1230000,set:729600,delta:500400,set ndiv:5788
[  668.922219] cpufreq: cpu0,cur:1508000,set:729600,delta:778400,set ndiv:5718
[  673.970594] cpufreq: cpu0,cur:1109000,set:1510400,delta:401400,set ndiv:118  

Under what circumstances will you hit this issue?
Does it work normally other then these dmesg log?

This “issue” is always there. Everything behaves normally in the OS. Is this normal behavior?

Hi,
it looks like the system keeps hitting thermal condition and throttling. Please run sudo tegrastats to check whether the temperature or current is normal or not.

And please check the section in developer guide:

Jetson Orin Nano Series, Jetson Orin NX Series and Jetson AGX Orin Series — NVIDIA Jetson Linux Developer Guide 1 documentation

We would suggest use power estimator to tune a power mode for the custom board.

With the link provided, I was able to have a fix without knowing the cause. The cpu governor was on schedutil, I changed it to performance.

1 Like