TRex Link Down on XL710 Interfaces Despite Driver Binding and ARP Unresolved

[surajsureshkumar]

So I have been trying out trex for a week and facing this link down issue on both my interfaces.

System Environment:

1. Using TRex v3.04 with two Ethernet Controller XL710 for 40GbE QSFP+ interfaces (PCI IDs: 0000:02:00.0 and 0000:02:00.1).
2. The interfaces are bound to the i40e DPDK driver.
3. The system operates on a Linux environment with hugepages configured correctly and DPDK-bound ports.

Problem: When starting TRex using sudo ./t-rex-64 -i, both interfaces (ports) are consistently showing "link down" status in the TRex console. This persists even after:

1. Running trex> portattr --link up command, which returns "SUCCESS" but the link status remains "DOWN."
2. The destination MAC address is unresolved for the default gateways configured (1.1.1.1 and 2.2.2.2), which also fails ARP resolution.

Troubleshooting Attempted:

1. Unbinding and Rebinding interfaces through TRex and DPDK devbind tools.
   2.Verified ports were correctly bound to i40e.
2. Cables and network equipment have been checked and confirmed to be functional, yet the link remains down.
3. The system was able to ping 1.1.1.1 but not 2.2.2.2 when bound to the kernel drivers, indicating that network settings (routes or gateways) may be an issue.
4. Looked into the github issues and google groups to find a possible solution, but got nothing.

Request for Assistance: Looking for guidance on:

1. Resolving the "link down" issue despite the ports being bound successfully and reporting the link as "up" in the software.
2. Any suggestions on how to debug unresolved ARP issues or troubleshoot further at the driver or physical link level.

Can someone please help me with this