From owner-cvs-src@FreeBSD.ORG Thu Mar 27 16:24:27 2003 Return-Path: Delivered-To: cvs-src@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id BD6F637B401; Thu, 27 Mar 2003 16:24:27 -0800 (PST) Received: from ambrisko.com (adsl-64-174-51-42.dsl.snfc21.pacbell.net [64.174.51.42]) by mx1.FreeBSD.org (Postfix) with ESMTP id D8C1543FBF; Thu, 27 Mar 2003 16:24:26 -0800 (PST) (envelope-from ambrisko@www.ambrisko.com) Received: from www.ambrisko.com (localhost [127.0.0.1]) by ambrisko.com (8.12.6/8.12.6) with ESMTP id h2S0OGhZ054429; Thu, 27 Mar 2003 16:24:16 -0800 (PST) (envelope-from ambrisko@www.ambrisko.com) Received: (from ambrisko@localhost) by www.ambrisko.com (8.12.6/8.12.6/Submit) id h2S0OGsl054428; Thu, 27 Mar 2003 16:24:16 -0800 (PST) (envelope-from ambrisko) From: Doug Ambrisko Message-Id: <200303280024.h2S0OGsl054428@www.ambrisko.com> In-Reply-To: <03d001c2f4b2$ecf461b0$52557f42@errno.com> To: Sam Leffler Date: Thu, 27 Mar 2003 16:24:16 -0800 (PST) X-Mailer: ELM [version 2.4ME+ PL94b (25)] MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII X-Spam-Status: No, hits=-3.3 required=5.0 tests=IN_REP_TO version=2.50 X-Spam-Level: X-Spam-Checker-Version: SpamAssassin 2.50 (1.173-2003-02-20-exp) cc: src-committers@FreeBSD.org cc: Doug Ambrisko cc: Maxime Henrion cc: cvs-src@FreeBSD.org cc: Mike Silbersack cc: cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sys/conf options src/sys/netinet ip_output.c X-BeenThere: cvs-src@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: CVS commit messages for the src tree List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 28 Mar 2003 00:24:28 -0000 Sam Leffler writes: | As I've said already, in the drivers you want to use the minimum-cost | technique to get the packet on the wire. I think your original | single-cluster version is close to what I would do, but so long as this | stuff only happens as an exception to the normal processing path I really | don't care. Just keep stats so we can see how much it's happening. An example of this happening is doing a cvs pserver update. I was seeing around a max of 10 frags being sent down to the driver on the server side of the connection. I saw a bunch of 8 frags. I watched it by putting a printf in the frag. handling part of the driver to count them. This seems to happen if you have an app. the does a bunch of small writes to a socket. Doing a dd if=/dev/null ibs=1 | rsh "> /dev/null" showed it as well if I recall right. Doug A.