11-24-2019 06:14 PM
I am trying to test the Xapp 1305 on my zcu102 board. Since I don't have 2 boards available, I am using the ML4026 SFP loopback adapter. I already did a ping on the zcu102 board (running Linux) by pinging the i.p address of the 10G interface enumerated as eth1. I then checked the /proc/net/dev file and found that the same packets sent on the eth1 interface were received on the same interface without errors. I am not sure if this is how it works. Is this the way to do a loopback test on the SFP loopback module? Please see the attached image.
My conclusion was that if there are 866bytes of data sent and received on the interface, then it's working. But I am not really sure about that since (as you can see) a dmesg of the interface brings up the "couldn't find phy i/f ". I tried to look at the datasheet of the module and found no way to test it. Probably this is a dumb question with a very simple solution, but I am a bit lost here.
11-29-2019 01:54 AM
Hi @alexkeys ,
Which SFP ports are you using to do loopback on a board?
Are these SFP ports enabled and pin constrained in hardware design?
11-29-2019 08:25 PM
Hi @pvenugo ,
Thanks for your reply.
I am using only one of the four ports available - the SFP0 to be specific (top right).
To answer your second question:
In both of these, the results were the same whenever I tried to connect to the 10G interface, eth1 (the error I am asking about)4. A senior co-worker told me there is no way to do a hardware loopback test yet. So now I am in a bit of a dilemma
12-02-2019 07:37 AM
I assume you're talking about the 10G XAPP1305 design, correct?
If so, that "couldn't find phy i/f" message is expected, since the 10G core does not have an MDIO interface to tie in to the phydev framework. You can safely ignore this message.