Hi there, A question related to TCP-FACK, which is a default in Linux: In FACK, the holes between SACK blocks are considered as loss. To a sender, when SACK comes in, loss_out would be non-zero. According to linux-2.6.17.7/net/ipv4/tcp_input.c, function tcp_time_to_recover(), this non-zero loss_out will send the sender into "Recovery" state, in which, CWND could be reduced. In one word, it seems that, FACK would allow SACK holes to reduce CWND. However, in the paper "Congestion Control in Linux TCP", Section 3, subsection Recovery, it says that Recovery state is triggered by "sufficient amount of successive duplicate ACK", to my understand, that means 3-dup. And in Section 3, subsection Disorder, it is also mentioned, SACK cause the sender move into this state. Therefore, I suppose it means SACK holes won't reduce CWND, they merely trigger fast_retrans and fast_recover. Even in the net/ipv4/tcp_input.c file, in the large block of comment right above the function tcp_time_to_recover(), it is also mentioned, CWND should only changed by "claasic VJ rules", which again means, CWND should only changes to 3-dup (or timeout). So, which is correct? The implementation, or the document? Or most probabally, am I wrong somewhere? Thank you guys! malin - : send the line "unsubscribe linux-net" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
