Hi list, I developed a basic client/server program, with bidirectional communication. Basically, server accept connection from a client and they send packets to them for a few time every 20milliseconds. I printed out tfrc_rx_info RTT values in a file in order to get a graph. What I get is a graph who starts from 0 and converge to the same RTT value. So this seems pretty bad to me, and I wanted to share that with you, dccp developers. I guess you are probably aware of that, I just trying to figure out what's happening in the receive struct. An other point is, if I only stop sending without closing socket, (that means my application blocks on "read" or something like this), when I restart the sending, RTT is completely false. It could be 23000 microseconds or 40000 microseconds instead of 300us or 400us, and it converges again to the same value. Is it something I would have to expect since dccp is in development? Just to be clear I'm talking about RTT in tfrc_rx_info structure, not in tfrc_tx_info which looks fine to me. I'm using the latest dave Miller's tree 2.6.21-rc7 on Debian testing. Best regards, Patrick. - To unsubscribe from this list: send the line "unsubscribe dccp" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
