Hardware test for embedded device with multiple NICs

I am in the process of writing software to test Linux based embedded devices after manufacturing. Each of the devices has multiple physical network interfaces, all internally bridged to a Linux bridge br0. **Problem**: - I need to find out if all of the NICs of the DUT are working properly (i.e. nothing went wrong during manufacturing). **The test setup is as follows**: - All Ethernet NICs of the DUT (MAC addresses are known, STP deactivated, IPv4/IPv6 activated) are physically connected to separate ports (let's say 1-3) of a managed switch - The system running the test software (also Linux) is physically connected to the same switch (let's say on port 4). **What I did so far**: - I made sure exactly one of the switch ports (1-3) was enabled (ifAdminStatus=up) via SNMP and then send simply pinged the DUT to see if the connected NIC is working and repeated the process for all switch ports/NICs. This, however, is somewhat slow as the switch needs ~3.5 seconds to (de-)activate a switch port and introduces additional delay until a ping is then successful. This adds up to ~15 seconds which I consider way to long. Ideally, I'd like to activate all switch ports 1-3 simultaneously and do the test in parallel on layer 2 or 3. **Question**: What would be we the best way to design a simple and fast (!) test to find out if all NICs are working and simultanously avoiding all of the inherent problems (broadcast storms due to switching loops, ...)?