Folks,
I am having a strange problem which I am wondering if someone might shed some light on.
I have some old Ethernet data acquisition devices, the Canberra AIM MCAs. These are not TCP/IP, but rather use the IEEE 802.2 Extended SNAP protocol. I have been talking
to these devices for many years from vxWorks, Windows (using WinPCAP) and Linux (using libnet and libpcap). I recently found that I was not able to talk to them from a number of my Linux systems. It partly works: I can build a list of the devices on the
network, I can take “ownership” of the device, and I can read what auxiliary Instrument Control Bus devices are connected to it (ADCs, amplifiers, HVPS, etc.). However, some messages I send to the device never get answered. When I look with Wireshark and
compare the packet sent with a Windows machine and the non-functioning Linux machine the packets look the same. But the Windows machine gets a reply and Linux does not.
On further investigation I found that the problem only occurs when using 10 Gbit Ethernet adapters. When using a 1 Gbit adapter it works fine. I was able to see this in
a single machine that has both 10 Gbit and 1 Gbit adapters. If I use the 1 Gbit adapter it works fine, if I use the 10 Gbit adapter it fails (but kind of works as described above). On 3 separate systems 1 Gbit works, and 3 other systems 10 Gbit fails.
This does not make sense to me. The Canberra device itself is 10 Mbit Ethernet. It is plugged into a 1 Gbit switch. The packets are all “slowed-down” in the switch to
10 Mbit, so I don’t understand why it matters if the network card is 1 Gbit or 10 Gbit.
I also used ethtool to lower the speed/duplex of the 10000T interface to 1000/Full and then to 100/Full and that did not help. I could still see the module, but much functionality
does not work. So a 1 Gbit adapter works, but a 10 Gbit adapter slowed down to 1 Gbit or even 100 Mbit does not work.
Any ideas what can cause these symptoms?
Thanks,
Mark