From mboxrd@z Thu Jan 1 00:00:00 1970 From: Sergei Shtylyov Subject: Re: [PATCH] ata: Add Intel SCH PATA support (revised) Date: Wed, 30 Apr 2008 15:17:18 +0400 Message-ID: <4818553E.6040606@ru.mvista.com> References: <20080426140007.63cfeede@dxy.sh.intel.com> <20080428094123.59d055c7@core> <20080428170751.097bdbe6@dxy.sh.intel.com> <200804282101.33012.bzolnier@gmail.com> <20080429113905.757703fd@dxy.sh.intel.com> <48174EF9.8020208@ru.mvista.com> <20080430104931.3aaab288@dxy.sh.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from homer.mvista.com ([63.81.120.155]:35892 "EHLO imap.sh.mvista.com" rhost-flags-OK-FAIL-OK-FAIL) by vger.kernel.org with ESMTP id S1757344AbYD3LR7 (ORCPT ); Wed, 30 Apr 2008 07:17:59 -0400 In-Reply-To: <20080430104931.3aaab288@dxy.sh.intel.com> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Alek Du Cc: Bartlomiej Zolnierkiewicz , Alan Cox , jgarzik@pobox.com, linux-ide@vger.kernel.org, rob.rhoads@intel.com Hello. Alek Du wrote: >> NAK this patch completely -- both libata and IDE parts. In the future it would be worth to post 2 separate patches, one for IDE and one for libata -- since they should go thru the different maintainer trees. >> Luckily, I have just had a quick look at the datasheet to which you've >>linked and that was enough to conclude that SCH is *not* compatible to >>ICH/PIIX family, so absolutely needs a separate driver! > Most registers are same Well, if you mean the standard BMDMA and standard PCI config. space registers (0x00 thru 0x3f), then those are mostly the same for the majority of the PCI IDE controllers. Why not stick your code into one of the other driver instead? ;-) > except for port enable bit, IOCFG and DMA timing. Now And also PIO and UDMA timings/enable. Actually, all the vendor-specific PCI config. registers are different from PIIX/ICH -- and that's what makes the PATA drivers different from each other. > I know why my patch just works -- the BIOS set the timing correctly. Yeah, and we have the special drivers which rely on the BIOS doing all the mode programming... > Thanks for point this out. I will post a separate driver patch later. In the meantwhile the only viable solution would be to stick your code into drivers/ide/pci/generic.c and drivers/ata/ata_generic.c. No more hacks, please. :-) > Thanks, > Alek MBR, Sergei