On 11/6/25 12:56 PM, Paolo Abeni wrote: > On 10/30/25 3:34 PM, [email protected] wrote: >> From: Chia-Yu Chang <[email protected]> >> >> Before this patch, num_retrans = 0 for the first SYN/ACK and the first >> retransmitted SYN/ACK; however, an upcoming change will need to >> differentiate between those two conditions. > > AFAICS, send_synack is invoked with a NULL dst only on retransmissions. > Perhaps you could use that info instead? moving forward and backward a > counter is not so nice.
... except you need to propagate the information to nested call. Possibly adding a TCP_SYNACK_RETRANS synack_type would fit? /P
