From owner-cvs-all@FreeBSD.ORG Sun Jul 3 01:26:07 2005 Return-Path: X-Original-To: cvs-all@freebsd.org 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 CB67A16B292; Sun, 3 Jul 2005 01:01:25 +0000 (GMT) (envelope-from ps@mu.org) Received: from elvis.mu.org (elvis.mu.org [192.203.228.196]) by mx1.FreeBSD.org (Postfix) with ESMTP id A71704522B; Sun, 3 Jul 2005 00:50:12 +0000 (GMT) (envelope-from ps@mu.org) Received: by elvis.mu.org (Postfix, from userid 1000) id 30B256EA19; Sat, 2 Jul 2005 17:40:36 -0700 (PDT) X-Original-To: ps@mu.org Delivered-To: ps@mu.org Received: from mx2.freebsd.org (mx2.freebsd.org [216.136.204.119]) by elvis.mu.org (Postfix) with ESMTP id 72F3B5C9BF for ; Mon, 21 Feb 2005 14:10:40 -0800 (PST) Received: from hub.freebsd.org (hub.freebsd.org [216.136.204.18]) by mx2.freebsd.org (Postfix) with ESMTP id E91A2561E8 for ; Mon, 21 Feb 2005 22:10:38 +0000 (GMT) (envelope-from owner-src-committers@FreeBSD.org) Received: by hub.freebsd.org (Postfix) id 5661B16A4DE; Mon, 21 Feb 2005 22:10:27 +0000 (GMT) Delivered-To: ps@freebsd.org Received: by hub.freebsd.org (Postfix, from userid 538) id 8AFC116A4D2; Mon, 21 Feb 2005 22:10:05 +0000 (GMT) Delivered-To: src-committers@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 56AAD16A4D0; Mon, 21 Feb 2005 22:10:04 +0000 (GMT) Received: from ylpvm29.prodigy.net (ylpvm29-ext.prodigy.net [207.115.57.60]) by mx1.FreeBSD.org (Postfix) with ESMTP id 75BE943D5D; Mon, 21 Feb 2005 22:10:03 +0000 (GMT) (envelope-from kbyanc@posi.net) Received: from gateway.posi.net (adsl-63-201-89-53.dsl.snfc21.pacbell.net [63.201.89.53]) by ylpvm29.prodigy.net (8.12.10 outbound/8.12.10) with ESMTP id j1LM9hh2019956; Mon, 21 Feb 2005 17:09:43 -0500 Received: from localhost (localhost [127.0.0.1]) by gateway.posi.net (Postfix) with ESMTP id C8D7675E077; Mon, 21 Feb 2005 15:12:54 -0800 (PST) From: Kelly Yancey To: Pawel Worach In-Reply-To: <4216C84D.2010903@telia.com> Message-ID: <20050221150749.L1802@gateway.posi.net> References: <200502160022.j1G0MK3P059683@repoman.freebsd.org> <42166EAA.2060001@telia.com> <42167133.9030207@freebsd.org> <42167415.6060608@telia.com> <20050218190716.S35160@gateway.posi.net> <4216C84D.2010903@telia.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-src-committers@FreeBSD.org Precedence: bulk X-Loop: FreeBSD.ORG X-Spam-Checker-Version: SpamAssassin 3.0.2 (2004-11-16) on elvis.mu.org X-Spam-Status: No, score=-1.5 required=5.0 tests=AWL, BAYES_00, J_CHICKENPOX_31, RCVD_IN_SORBS autolearn=no version=3.0.2 X-Spam-Level: Cc: cvs-src@freebsd.org, src-committers@freebsd.org, cvs-all@freebsd.org, Colin Percival , Kelly Yancey Subject: Re: cvs commit: src/lib/libfetch http.c X-BeenThere: cvs-all@freebsd.org X-Mailman-Version: 2.1.5 List-Id: CVS commit messages for the entire tree List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Date: Sun, 03 Jul 2005 01:26:07 -0000 X-Original-Date: Mon, 21 Feb 2005 15:12:54 -0800 (PST) X-List-Received-Date: Sun, 03 Jul 2005 01:26:07 -0000 On Sat, 19 Feb 2005, Pawel Worach wrote: > Kelly Yancey wrote: > > On Sat, 19 Feb 2005, Pawel Worach wrote: > > > [ tcpdump output snipped ] > > > > > >>Solaris box is: SunOS proxy 5.9 Generic_112233-12 sun4u sparc SUNW,Ultra-4 > >>FreeBSD box is: FreeBSD client 6.0-CURRENT FreeBSD 6.0-CURRENT #0: Wed Feb 16 > >>14:39:08 CET 2005 > >>Squid is: Squid Cache: Version 2.5.STABLE6-20041001 > >> > >>Both machines are connected to the same switch and are on the same subnet. > >> > > > > > > I don't suppose you have half_closed_clients set to off in your squid > > config? > > Yes I do. This proxy has 4500-5000 unique clients and now only fetch(1) has a > problem with it :) > The issue then isn't the TCP_NOPUSH, but rather the call to shutdown(2). There is no reason to disallow half-closed TCP connections from talking to your proxy, in fact squid defaults to allowing them. Nonetheless, I've been testing an updated diff which replaces the call to shutdown(2) with setting the TCP_NODELAY socket option and clearing the TCP_NOPUSH option instead. This appears to be sufficient to force the pending request to be sent without closing the client half of the TCP connection. I'll post the diff to -net shortly. Kelly -- Kelly Yancey - kbyanc@{posi.net,FreeBSD.org} - kelly@nttmcl.com