The kernel calculates mean deviation as an approximation for
M | rtt - M rtt |,
where M denotes mean value. In other words it's the mean of
the absolute values of the difference between RTT of a single packet
and the mean RTT. The calculation is coded and supplied by good comments in
tcp_rtt_estimator() in tcp_input.c.
[...]
> Now I' ll only verify that the rto got calculated right from the kernel.
Well, I've done the check of the RTO evaluation. I added a debug output
if the kernel increases RTO in more than two or three times at once
(I don't remember) and didn't find anything unexpected.
The only strange place in the RTO evaluation code
was the initialization of SRTT and MDEV values.
Please report if your verification shows a bug in the code.
Regards
Andrey V.
Savochkin
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/faq.html