From mboxrd@z Thu Jan 1 00:00:00 1970 From: Sergei Shtylyov Subject: Re: [PATCH 13/15] ide: fix UDMA/MWDMA/SWDMA masks Date: Wed, 31 Jan 2007 23:38:08 +0300 Message-ID: <45C0FE30.4000800@ru.mvista.com> References: <20070119003058.14846.43637.sendpatchset@localhost.localdomain> <20070119003226.14846.87052.sendpatchset@localhost.localdomain> <45B4FFBD.40507@ru.mvista.com> <20070122184620.1c8b87ac@localhost.localdomain> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from gateway-1237.mvista.com ([63.81.120.155]:46964 "EHLO imap.sh.mvista.com" rhost-flags-OK-FAIL-OK-FAIL) by vger.kernel.org with ESMTP id S1030446AbXAaUiN (ORCPT ); Wed, 31 Jan 2007 15:38:13 -0500 In-Reply-To: <20070122184620.1c8b87ac@localhost.localdomain> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Alan Cc: Bartlomiej Zolnierkiewicz , linux-ide@vger.kernel.org, linux-kernel@vger.kernel.org Hello. Alan wrote: >> Ugh, I'm not seeing any *actual* support for MW/SW DMA in this driver... > Thats long been broken. Should be correct in the libata driver I've looked thru the specs and it seemed to me that ULi hardware is much broken PIO wise: their max active time is 8 cycles even on taskfile access which gives 240 ns while standard requeires 290 ns for modes 0 thru 2... I've also noted that the tuneproc() method in both cmd64x.c and alim15x3.c seems to misdo recovery calculation, taking address setup into account -- that should be slightly overclocking PIO modes 0/1 (ULi docs don't shed much light on how it should be calculated)... Well, this seems fixed in libata drivers. > Alan MBR, Sergei