From owner-freebsd-hackers@FreeBSD.ORG Tue Aug 23 04:48:54 2005 Return-Path: X-Original-To: freebsd-hackers@freebsd.org Delivered-To: freebsd-hackers@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 12ECC16A41F for ; Tue, 23 Aug 2005 04:48:54 +0000 (GMT) (envelope-from oceanare@pacific.net.sg) Received: from smtpgate2.pacific.net.sg (smtpgate2.pacific.net.sg [203.120.90.28]) by mx1.FreeBSD.org (Postfix) with SMTP id 1C3DC43D45 for ; Tue, 23 Aug 2005 04:48:52 +0000 (GMT) (envelope-from oceanare@pacific.net.sg) Received: (qmail 13967 invoked from network); 23 Aug 2005 04:48:51 -0000 Received: from maxwell2.pacific.net.sg (203.120.90.192) by smtpgate2.pacific.net.sg with SMTP; 23 Aug 2005 04:48:51 -0000 Received: from [192.168.0.107] ([210.24.246.165]) by maxwell2.pacific.net.sg with ESMTP id <20050823044850.DYRQ28012.maxwell2.pacific.net.sg@[192.168.0.107]>; Tue, 23 Aug 2005 12:48:50 +0800 Message-ID: <430AAA72.9090603@pacific.net.sg> Date: Tue, 23 Aug 2005 12:47:46 +0800 From: Erich Dollansky Organization: oceanare pte ltd User-Agent: Mozilla Thunderbird 1.0.6 (X11/20050802) X-Accept-Language: en-us, en MIME-Version: 1.0 To: m.ehinger@ltur.de References: In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: freebsd-hackers@freebsd.org Subject: Re: IBM Active Protection System Approach X-BeenThere: freebsd-hackers@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Technical Discussions relating to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 23 Aug 2005 04:48:54 -0000 Hi, m.ehinger@ltur.de wrote: > > what would be the best approach to implement aps on FreeBSD? > Go as deep as possible into the disk driver. The reason is pretty simple. Just stop any write access to the disk as soon as the machine starts moving above a given limit. If the movement is above a higher level, start also to move the heads out. As the heads are moved out when a fall starts, the chance that the heads hit the surface is minimised. The driver can then move the heads back after a certain amount of time passed by without any event and continue operation. > Would an daemon be sufficient for that? Reaction time? What about an kernel thread? > Getting the heads out of the danger zone is one thing but the most crucial thing from my point of view is the write access. If the heads are moved away while writing the current part of the track is damaged. I have seen a Thinkpad of the first lot being so nervous moving the heads avway at the slidest movement of the machine to slow work down to a crawl being seated in a bus or plane. > Other solutions? I did some work like this a very long time ago on RSX. Stopping a write brought for us the cases of damaged file systems down to some 10% of the original value. Erich