From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alan Cox Subject: Re: libata PATA support - work items? Date: Thu, 06 Jan 2005 21:29:42 +0000 Message-ID: <1105044491.24187.268.camel@localhost.localdomain> References: <006301c4ee5c$49e6a230$95714109@tw.ibm.com> <41DB2781.7000203@pobox.com> <005701c4f3ce$37d56140$fb7f4109@tw.ibm.com> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: Received: from clock-tower.bc.nu ([81.2.110.250]:39100 "EHLO localhost.localdomain") by vger.kernel.org with ESMTP id S263157AbVAFWeP (ORCPT ); Thu, 6 Jan 2005 17:34:15 -0500 In-Reply-To: <005701c4f3ce$37d56140$fb7f4109@tw.ibm.com> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Albert Lee Cc: Jeff Garzik , "\"Bartlomiej Zolnierkiewicz\"" , "\"IDE Linux\"" , "\"Doug Maxey\"" On Iau, 2005-01-06 at 08:51, Albert Lee wrote: > For the 1st item, > - add PATA device errata/blacklist info > Is there any detail or draft design idea available? There is a fair bit of code in drivers/ide. Most of it is handling drive problems that occurred with early DMA drives and it uses the model/firmware info in the ident string to match drives that claim to do DMA but do not and drives that claim not to but can. The logic is fairly simple and the database is easily transported or shared. It is fairly important because some of these old drives corrupted data if used DMA. (see drivers/ide/ide-dma.c) For PIO problem drives where there is a need to keep PIO modes chosen carefully then ide-lib has a list (ide_pio_blacklist) and the ide-io code knows about pio modes and older style EIDE cycle timing reports (ide_get_best_pio_mode) IDE CD has some DMA blacklists for ATAPI but they should be viewed with caution as many entries were due to a bug in our atapi implementation that was fixed long ago. > For > - C/H/S addressing. > I've an old Segate drive which support CHS-only at hand. > Maybe I can work on it first. That seems the logical start - many of the worst drive errata only apply to old CHS drives anyway.