Open rafaeldtinoco opened 1 year ago
But if this was happening in a low throughput environment then it wouldn't be an expected issue right? So just cataloguing this as not an error doesn't seem entirely correct to me, it's circumstantial if this is an error or a warning.
Those outputs are handled by general purpose functions. To tackle this in a proper way, we would need circumstantial information coming from the bpf land.
Description
These errors:
Should be informational I believe. If the perfbuffer gets full, then some packet headers, being sent from SKB progs to userland, might get corrupted indeed. This is not an error. Considering I was sustaining a 5Gbit/sec throughput, the amount of errors is very very low, but still.. its an informative error.
Output of
tracee -v
:Output of
uname -a
:Additional details