1

So I need to create a simple process to test a large amount of gigabit network ports. Since I don't want to setup two machines and run software on both the test connection my first thought was to create a loopback cable. ANd just see if I can send data and receive what I send. So I looked online and this seemed possible. So I build my own loopback cable with the following pinout:

enter image description here

Now when I plug this cable into a network port I do see the lights coming on. And in Linux I sporadically see a link being established at 10mbit for a short amount of time.

I configured the network interface as follows:

  • Ip address: 192.168.1.50
  • Netmask: 255.255.255.0
  • Gateway: empty
  • Link negotiation: Tried both auto and manual which didn't seem to make a difference

I can sporadically ping myself:

λ ping 192.168.1.50 PING 192.168.1.50 (192.168.1.50) 56(84) bytes of data. From 143.178.160.1 icmp_seq=1 Packet filtered From 143.178.160.1 icmp_seq=2 Packet filtered From 143.178.160.1 icmp_seq=3 Packet filtered From 143.178.160.1 icmp_seq=8 Packet filtered 64 bytes from 192.168.1.50: icmp_seq=9 ttl=64 time=0.048 ms 64 bytes from 192.168.1.50: icmp_seq=10 ttl=64 time=0.032 ms 64 bytes from 192.168.1.50: icmp_seq=11 ttl=64 time=0.038 ms 64 bytes from 192.168.1.50: icmp_seq=12 ttl=64 time=0.043 ms 64 bytes from 192.168.1.50: icmp_seq=13 ttl=64 time=0.034 ms 64 bytes from 192.168.1.50: icmp_seq=14 ttl=64 time=0.033 ms 64 bytes from 192.168.1.50: icmp_seq=15 ttl=64 time=0.074 ms From 143.178.160.1 icmp_seq=16 Packet filtered From 143.178.160.1 icmp_seq=18 Packet filtered From 143.178.160.1 icmp_seq=20 Packet filtered From 143.178.160.1 icmp_seq=21 Packet filtered From 143.178.160.1 icmp_seq=22 Packet filtered From 143.178.160.1 icmp_seq=23 Packet filtered From 143.178.160.1 icmp_seq=24 Packet filtered From 143.178.160.1 icmp_seq=25 Packet filtered From 143.178.160.1 icmp_seq=26 Packet filtered From 143.178.160.1 icmp_seq=27 Packet filtered 64 bytes from 192.168.1.50: icmp_seq=28 ttl=64 time=0.019 ms ...

My other network interface do not seem to interfere:

λ ifconfig enp0s13f0u1u2u1: flags=4099<UP,BROADCAST,MULTICAST> mtu 1500 inet 192.168.1.50 netmask 255.255.255.0 broadcast 192.168.1.255 inet6 fe80::2e0:4cff:fe6c:77a9 prefixlen 64 scopeid 0x20<link> ether 00:e0:4c:6c:77:a9 txqueuelen 1000 (Ethernet) RX packets 0 bytes 0 (0.0 B) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 248 bytes 15360 (15.0 KiB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0 lo: flags=73<UP,LOOPBACK,RUNNING> mtu 65536 inet 127.0.0.1 netmask 255.0.0.0 inet6 ::1 prefixlen 128 scopeid 0x10<host> loop txqueuelen 1000 (Local Loopback) RX packets 1135 bytes 127855 (124.8 KiB) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 1135 bytes 127855 (124.8 KiB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0 vboxnet0: flags=4099<UP,BROADCAST,MULTICAST> mtu 1500 inet 192.168.56.1 netmask 255.255.255.0 broadcast 0.0.0.0 ether 0a:00:27:00:00:00 txqueuelen 1000 (Ethernet) RX packets 0 bytes 0 (0.0 B) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 0 bytes 0 (0.0 B) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0 wlp113s0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500 inet 10.0.24.78 netmask 255.255.255.0 broadcast 10.0.24.255 inet6 fe80::9eb6:d0ff:fe3e:5655 prefixlen 64 scopeid 0x20<link> inet6 fe80::b5d5:5582:6a6a:7f04 prefixlen 64 scopeid 0x20<link> ether 9c:b6:d0:3e:56:55 txqueuelen 4096 (Ethernet) RX packets 404929 bytes 345953679 (329.9 MiB) RX errors 0 dropped 2 overruns 0 frame 0 TX packets 176917 bytes 31814869 (30.3 MiB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

So my question is how can I make this work? I simply want to start testing the interface. Send some udp packets at maximum throughput. First I thought my cable was simply badly made but I tried 3 times making new cables and all of them have the same result so I don't think it's a badly made cable.

Any help is appreciated!

Improve this question
1
  • 2
    1000Base-T, unlike 10Base-T or 100Base-TX, does not use separate transmit and receive pairs that you could loop back from transmit to receive. All pairs are used to both send and receive, and it uses echo-cancelling for this, which kills signals sent from what is received. What you have will not work for 1000Base-T. Commented Jul 19, 2022 at 13:53

1 Answer 1

Reset to default
2

That won't work.

You cannot use a simple loopback for 1000BASE-T due to its use of hybrids and echo compensation. Simply put, the link signals that are sent on the port itself are canceled out and there's nothing to be received. Also, auto negotiation (that is mandatory for 1000BASE-T) won't be able to negotiate the required master/slave relationship.[*1]

I'm afraid there's no way around using a simple active device with a 1000BASE-T port - a small node (RPi-ish) or a basic switch.

[*1] The link you're seeing is due to the feature on some NICs (mostly Broadcom, "Ethernet@Wirespeed") of falling back to 100BASE-TX or 10BASE-T when 1000BASE-T PCS linking fails even when auto negotiation initially determined 1000BASE-T as mutually supported.

Improve this answer
3
  • Can a switch be configured to loopback the data? If so then I would just buy a switch. Commented Jul 19, 2022 at 14:07
  • Some switches can be coaxed into doing that, using a network loop and abusing (unidirectional) port mirroring or source-port filtering to avoid the bridge loop. That's not a supported mode though and you'd need to test it yourself. Commented Jul 19, 2022 at 14:14
  • If you don't mind the looped back frames to carry some VLAN tag, you could translate the VLAN on the loop and tag the translated VLAN on the test port (works with most switches, not all). Commented Jul 19, 2022 at 14:26

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.