From owner-freebsd-ports@FreeBSD.ORG Tue Feb 17 01:27:07 2009 Return-Path: Delivered-To: ports@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 1D8231065672 for ; Tue, 17 Feb 2009 01:27:07 +0000 (UTC) (envelope-from jonathan@kc8onw.net) Received: from vps.kc8onw.net (jonathanstewart-1-pt.tunnel.tserv4.nyc4.ipv6.he.net [IPv6:2001:470:1f06:71d::2]) by mx1.freebsd.org (Postfix) with ESMTP id EAA808FC19 for ; Tue, 17 Feb 2009 01:27:06 +0000 (UTC) (envelope-from jonathan@kc8onw.net) Received: from [10.70.8.101] (nr2-66-117-249-253.fuse.net [66.117.249.253]) by vps.kc8onw.net (Postfix) with ESMTPSA id 7A41C17035; Mon, 16 Feb 2009 15:09:01 -0500 (EST) Message-ID: <499A1261.9000900@kc8onw.net> Date: Mon, 16 Feb 2009 20:26:57 -0500 From: Jonathan User-Agent: Thunderbird 2.0.0.19 (Windows/20081209) MIME-Version: 1.0 To: Chuck Robey References: <200902131919.37778.makc@issp.ac.ru> <49972998.9000504@telenix.org> <49982839.803@kc8onw.net> <4999ADF0.2040405@telenix.org> In-Reply-To: <4999ADF0.2040405@telenix.org> X-Enigmail-Version: 0.95.7 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: ports@freebsd.org Subject: Re: FreeBSD Port: ktorrent-3.1.6_1 X-BeenThere: freebsd-ports@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Porting software to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 17 Feb 2009 01:27:07 -0000 Chuck Robey wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Jonathan wrote: >> Your problem is not related to the one I and the others have. Your >> problem is caused by your upstream being so saturated with data packets >> that the acknowledge packets for the downloads are being delayed or >> dropped. A much more detailed description and more general solution can >> be found here http://www.benzedrine.cx/ackpri.html > > You may be right, I said I didn't understand, but if my upload was supposedly > satured, it makes less sense to me that it never showed as using more that about > 10K (5K for the average, really) and my limit (for both upload & download) was > set to -1 (infinite). I didn't see why that would cause saturation, although > the other results (having the download rate go from very limited to a max value) > do kind of support such an idea. Why would my setting the rates both to > infinite cause saturation? > > Or is maybe the upload rate that's being set being only affecting one use of > upload, but not all uses of upload? That could be twisted in that direction, I > guess, choking off the ability to use uploads for acks, because it's all being > reserved for some other use? Boy, that surprises me, but it's it's what's > meant, it could explain things. If I understand this paragraph correctly, yes, that's exactly what happens. If you check the link I sent earlier it has a detailed explanation with graphs. Jonathan