From owner-cvs-all@FreeBSD.ORG Thu Mar 27 19:13:34 2003 Return-Path: Delivered-To: cvs-all@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 0893A37B405 for ; Thu, 27 Mar 2003 19:13:34 -0800 (PST) Received: from relay.pair.com (relay.pair.com [209.68.1.20]) by mx1.FreeBSD.org (Postfix) with SMTP id 821BB43FBF for ; Thu, 27 Mar 2003 19:13:31 -0800 (PST) (envelope-from silby@silby.com) Received: (qmail 3825 invoked from network); 28 Mar 2003 03:13:30 -0000 Received: from niwun.pair.com (HELO localhost) (209.68.2.70) by relay.pair.com with SMTP; 28 Mar 2003 03:13:30 -0000 X-pair-Authenticated: 209.68.2.70 Date: Thu, 27 Mar 2003 21:10:05 -0600 (CST) From: Mike Silbersack To: Luigi Rizzo In-Reply-To: <20030327153352.A66323@xorpc.icir.org> Message-ID: <20030327210503.G601@odysseus.silby.com> References: <200303260452.h2Q4quap015364@www.ambrisko.com> <20030326183351.GJ57674@elvis.mu.org> <20030327013224.P7674@odysseus.silby.com> <20030327153352.A66323@xorpc.icir.org> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Spam-Status: No, hits=-25.4 required=5.0 tests=AWL,EMAIL_ATTRIBUTION,IN_REP_TO,QUOTED_EMAIL_TEXT, REFERENCES,REPLY_WITH_QUOTES autolearn=ham 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: cvs-src@FreeBSD.org cc: Maxime Henrion cc: Sam Leffler cc: cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sys/conf options src/sys/netinet ip_output.c X-BeenThere: cvs-all@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: CVS commit messages for the entire tree List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 28 Mar 2003 03:13:36 -0000 On Thu, 27 Mar 2003, Luigi Rizzo wrote: > so individual writes of 513+ bytes will result in wasting up to 75% > of the socket buffer space. At the very least, i would drop the > 'm->m_len <= MCLBYTES / 4' check to reduce the waste. > > cheers > luigi Ooooh, that gives me a good idea. While we're walking the queues in tcp_drain trying to free mbufs, it might be a good idea to run m_defrag. /me scribbles on TODO list. (Of course, if you have 0 mbufs + clusters free, that would totally blow up in your face, so it'll have to be more subtle.) Mike "Silby" Silbersack