"David S. Miller" <davem@davemloft.net> wrote on 21.10.2004 08:51:00: > > Also keep in mind that as the connection progresses > the TSO size used will get larger and larger. > > At the beginning, we do not use TSO for the first > packet, and then it progressively grows larger and > larger as the congestion window opens up. Yes, that behavior is clear to me, but even if I send a 100MB file using scp, I do not see large packets in my driver. I only get a lot of packets with tso_size == MTU size - TCP/IP header size. And there shouldn't be a problem with the connection itself. I'm using a virtual connection on a zSeries mainframe with about 10Gpbs throughput. > That being said, no matter what your proposed patch > makes zero sense. Ok, I got that point. The patch is really useless ;-) I guess I was looking for anything to overcome my problem; sorry ;-) - : send the line "unsubscribe linux-net" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
