Artem Prilutskiy

Yes, this patch works very well. And by using it we reached performance of DPDK‘s ixgbe twice with less CPU resources.

After applying the fix i rewrote my code. Now transmits in multiple threads. 2-4 queues per thread. It’s production system, which can use different methods to accelerate transmission/reception of UDP....

Your bug in two words: wrong interpretation of array access. Just imagine, which address has, for example, `rx_queue[1]` since the size of `struct ixgbe_rx_queue` doesn't include actual length of `virtual_addresses[]`...

I've fixed my comment. Correct explanation is there. 01:30 AM here, I have to be in bed ;)

It seems someone mixed things. Guys added value 4 to the host settings that allows to remove OVCM bit, I guess

And OVCM attribute has bit 2 (0x04) and not 6 (0x40). Please read ts_10236102v020401p.pdf page 79. Bit 6 (0x40) indicates encrypted transmission.

All good with the sequence numbers. From time to time Reflector creates at least the second connection context for the same srcip:srcport. Probably when reflector is behind the NAT.

Uhhh :) This is wrong address for this claim ))) Please draw call flow and give it to Yentel, then we check configuration of master and research the issue.

Most possible you have configured TA transmission for Hytera repeaters on your master. Just ask Rudy or Yentel to turn it off.

Tested. It seems like the main branch has no such problem. I can guess it was fixed in 4.3.x. Unfortunately upcoming version of Debian 12 has only 4.1.6. I created...