From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tejun Heo Subject: Re: [rfc][patch] ide: fix unneeded LBA48 taskfile registers access Date: Thu, 10 Feb 2005 09:01:58 +0900 Message-ID: <420AA476.1040406@gmail.com> References: <4206F2E5.7020501@gmail.com> <200502070959.54973.bzolnier@elka.pw.edu.pl> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Received: from wproxy.gmail.com ([64.233.184.199]:45001 "EHLO wproxy.gmail.com") by vger.kernel.org with ESMTP id S261977AbVBJACF (ORCPT ); Wed, 9 Feb 2005 19:02:05 -0500 Received: by wproxy.gmail.com with SMTP id 67so1400749wri for ; Wed, 09 Feb 2005 16:02:05 -0800 (PST) In-Reply-To: <200502070959.54973.bzolnier@elka.pw.edu.pl> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Bartlomiej Zolnierkiewicz Cc: linux-ide@vger.kernel.org, linux-kernel@vger.kernel.org, Alan Cox , Jeff Garzik , Tejun Heo Hello, Bartlomiej. Happy new lunar year. Bartlomiej Zolnierkiewicz wrote: > > I would prefer to not teach do_rw_taskfile() about ->tf_{in,out}_flags > (and convert all users to use helpers) - it is much simpler this way, > > ->flags field in ide_task_t is needed anyway (32-bit I/O flag). > New lunar year day is one of the biggest holidays here, so I haven't got time to work for a few days. As it's over now, I began to work on ide drivers again. I applied your task->flags patch and am moving my patches over it. One problem is that, with ATA_TFLAG_LBA48, whether to use HOB registers or not cannot be determined separately for writing and reading. So, when initializing flush tasks, if WIN_FLUSH_CACHE_EXT is used, we need to turn on ATA_TFLAG_LBA48 to read error location properly, and we end up unnecessarily writing HOB registers. I think we can... 1. Just leave it as it is. It's not that big a deal. 2. Use another flag(s) to control LBA48 reading/writing separately. 3. do my proposal. :-) I'm currently sticking to #1. Please let me know what you think. Thanks. -- tejun