Why does the network card on a Linux system boot on its own?

When I was using Telnet to remotely access my target computer on Windows and conducting latency testing with cyclictest, after several hours of execution, I suddenly found that the telnet connection was interrupted:

C:Usersguoya>telnet 192.168.5.1
...
遗失对主机的连接。(Lost connection to host)

At this time, I saw the following message appearing on my target computer, as if the network card eth2 had been restarted:

[18825.382832] asix 1-6:1.0 eth2: Link is Up - 100Mbps/Full - flow control rx/tx

However, at this time, no one or any other process was operating the target device. Of course, there is no process operating eth2 at this moment.

My target computer is using a system that I compiled kernel 6.4.0 to generate bzImage and compiled busybox to generate the root file system,

# taskset -c 2 cyclictest -p 80 -t 1 -n -i 1000 -l 100000000000
/dev/cpu_dma_latency set to 0us
policy: fifo: loadavg: 1.36 1.22 1.14 1/127 1617

T: 0 ( 1553) P:80 I:1000 C:11258227 Min:      2 Act:    3 Avg:    3 Max:      21
Asked By: justSoSo

||

This feels more like a hardware issue, but it’s not really possible to tell with only one information line.

Are there one or more routers (or switches) between your client and target machines? Any of them could have restarted.

Have you checked all the intervening network cables? One of them might be loose.

Answered By: Chris Davies
Categories: Answers Tags: , ,
Answers are sorted by their score. The answer accepted by the question owner as the best is marked with
at the top-right corner.