As the packets are processed in dpdk as burst , if the packet queued in TX ring is more than 40 , It will go in

TX hang state as the H/W descriptor has MAX 40 TX desc limit  .
When TX ring goes in hang state , The dpdk try to reset it , and NIC  should recovered after reset .

But due to dpdk has multiple threads and if NIC reset happened same time with different threads ,

It trigger race condition and reset has failed , Initiating the reboot .

Unit testing and regression testing is done