From owner-freebsd-net@FreeBSD.ORG Sun Mar 19 23:20:56 2006 Return-Path: X-Original-To: freebsd-net@freebsd.org Delivered-To: freebsd-net@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 202DA16A401 for ; Sun, 19 Mar 2006 23:20:56 +0000 (UTC) (envelope-from sam@errno.com) Received: from ebb.errno.com (ebb.errno.com [69.12.149.25]) by mx1.FreeBSD.org (Postfix) with ESMTP id C252E43D46 for ; Sun, 19 Mar 2006 23:20:55 +0000 (GMT) (envelope-from sam@errno.com) Received: from [10.0.0.199] ([10.0.0.199]) (authenticated bits=0) by ebb.errno.com (8.12.9/8.12.6) with ESMTP id k2JNKqo7045139 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Sun, 19 Mar 2006 15:20:53 -0800 (PST) (envelope-from sam@errno.com) Message-ID: <441DE754.6020004@errno.com> Date: Sun, 19 Mar 2006 15:20:52 -0800 From: Sam Leffler Organization: Errno Consulting User-Agent: Thunderbird 1.5 (Macintosh/20051201) MIME-Version: 1.0 To: Darren Pilgrim References: <200603191738.19442.max@love2party.net> <441DE5B2.1060202@bitfreak.org> In-Reply-To: <441DE5B2.1060202@bitfreak.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Max Laier , freebsd-net@freebsd.org Subject: Re: New version of iwi(4) - Call for testers X-BeenThere: freebsd-net@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Networking and TCP/IP with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 19 Mar 2006 23:20:56 -0000 Darren Pilgrim wrote: > Max Laier wrote: >> On Sunday 19 March 2006 16:47, dima wrote: >>>> >>>> the new version at: >>>> http://people.freebsd.org/~mlaier/new_iwi/20060315.both.tgz > >> >>> The new driver didn't pass cvsup test at my laptop :( >>> It fails large file upload either. >>> >>> It's definitely a flow control problem. Is taskqueue designed to address >>> this? >> >> We found that this problem is completely unrelated to iwi, but a >> general problem with software encryption in net80211. This should be >> fixed with ieee80211_output.c rev. 1.40 and will be MFCed shortly. >> Note that this uses m_unshare, which also is not yet in RELENG_6. > > Are you referring to the problem in cvsup tests where it will suddenly > stop with a "Network write failure" error? Yes. The issue was that when crypto was done in the host it was sometimes being done in-place on mbufs still owned by the socket (exactly when depended on a lot of things but turning down the mtu increased the likelihood). If this happened and tcp retransmitted the previously encrypted data then it would send garbage. As Max said this was in the net80211 layer and affected all drivers depending on the host to do crypto. It'll get mfc'd this week (re permitting). Sam