From owner-freebsd-drivers@FreeBSD.ORG Thu Sep 11 06:06:17 2008 Return-Path: Delivered-To: freebsd-drivers@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 6CAF91065674 for ; Thu, 11 Sep 2008 06:06:17 +0000 (UTC) (envelope-from raul@b2n.org) Received: from mail1.isdefe.es (mail1.isdefe.es [194.15.213.239]) by mx1.freebsd.org (Postfix) with ESMTP id 275D38FC19 for ; Thu, 11 Sep 2008 06:06:16 +0000 (UTC) (envelope-from raul@b2n.org) Received: from turing.b2n.org (unknown [172.24.1.14]) by mail1.isdefe.es (Postfix) with ESMTP id C6BE95C35; Thu, 11 Sep 2008 08:05:57 +0200 (CEST) Received: from [10.10.10.14] (ws.pinlabs.b2n.org [10.10.10.14]) by turing.b2n.org (Postfix) with ESMTP id 4B06E32AC09; Thu, 11 Sep 2008 08:05:57 +0200 (CEST) From: Raul To: Volker In-Reply-To: <48C59A7A.9090102@vwsoft.com> References: <1220724123.6519.8.camel@ws.pinlabs.b2n.org> <48C2E55F.9000901@vwsoft.com> <1220891191.6842.20.camel@ws.pinlabs.b2n.org> <48C59A7A.9090102@vwsoft.com> Content-Type: text/plain; charset=utf-8 Date: Thu, 11 Sep 2008 08:06:13 +0200 Message-Id: <1221113173.6481.36.camel@ws.pinlabs.b2n.org> Mime-Version: 1.0 X-Mailer: Evolution 2.22.3.1 Content-Transfer-Encoding: 8bit Cc: freebsd-drivers@freebsd.org Subject: Re: Adaptec AAR-1225SA REV A2, solved! X-BeenThere: freebsd-drivers@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Writing device drivers for FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 11 Sep 2008 06:06:17 -0000 El lun, 08-09-2008 a las 23:34 +0200, Volker escribió: > First you may try how your IDE adapter behaves under load. Older SiI > chips used to work fine for light load but refused to work under load > (in terms of write operations). I don't know if this is still true for > these newer chips. Well, I'm not sure how much load do you mean so I'll tell you what I did ;). I've set a striped zvol and push 4 parallel dd reading from /dev/zero and 8 tars decompresing src's and ports tarballs. Just to add some 'noise' to the box, a buildworld was also running although from other disk behind other controller ... after some time, the noise coming from the disks changed, I thought, something has happened ... well, tar finished successfully and only the 4 dd continue writing on the zvol. About 6?GB on disk and several thousand files ... So I stop the dd's and run a scrub on the zvol that was completed with 0 errors. After that, I destroyed the striped zvol and set a mirrored one. From them, samba shares, some netcat, scp's ... 24GB has been writen with a more 'normal' load and still without any apparent problem. I don't like too much this _scientific_ approach, but I feel pretty confident about the controller ;). > For the commit: > > 1) I don't have a commit bit, but we may get that in (by the help of > somebody else) > 2) we're already in a freeze situation - don't expect that to be in > -STABLE before 7.1-RELEASE (also it needs to be in HEAD first and then > may get MFC'd). I see. [....] > PS: Can you please file a PR with my patch attached? I'm missing some > time to do that (this way, the change doesn't get lost). Yes, I'll do that. Thanks a lot Volker. Raul