From: "Armando L. Caro, Jr." <[EMAIL PROTECTED]> Date: Thu, 10 Nov 2005 14:17:14 -0500
> Stephen Hemminger wrote: > > It sets cwnd_clamp for the new connection only if the congestion window > > has been set as part of a route entry. A normal connection will store > > the last cwnd in the route metric, but it will not be locked. > ^^^^^^^^^^^^^^^^^^^^^^^^^ > I don't quite follow your response. What do you mean by "it will not be > locked"? I understand that the cwnd stored in the route metric can be > modified in subsequent TCP connections, but for any given TCP > connection, cwnd_clamp enforces an upper limit on cwnd growth. > Therefore, it limits the cwnd that can be cached and potentially hurts > subsequent connections. In other words, the cached cwnd value can only > be decreased, and not increased. Am I missing something? Only when the metric is explicitly loaded into a route by hand by the administrator, can it be locked. So if the metric is merely rememberd by TCP for a routing cache entry, that isn't enough to make it get actually used by future connections. It has to be from a metric loaded explicitly by the administrator into a specific route and locked, in order to be used by TCP. Someone bumps into this ever 3 or 4 months, and gets similarly confused... - To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html