From mboxrd@z Thu Jan 1 00:00:00 1970 From: =?ISO-8859-1?Q?Rafael_Esp=EDndola?= Subject: Re: unable to set udma >= 3 on ich5 on 2.6.11 Date: Thu, 23 Jun 2005 09:49:51 -0300 Message-ID: <564d96fb050623054973250233@mail.gmail.com> References: <564d96fb05062005064283caaf@mail.gmail.com> <200506212251.27168.rafael.espindola@gmail.com> Reply-To: =?ISO-8859-1?Q?Rafael_Esp=EDndola?= Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7BIT Return-path: Received: from wproxy.gmail.com ([64.233.184.198]:18225 "EHLO wproxy.gmail.com") by vger.kernel.org with ESMTP id S262416AbVFWMty convert rfc822-to-8bit (ORCPT ); Thu, 23 Jun 2005 08:49:54 -0400 Received: by wproxy.gmail.com with SMTP id 68so444707wra for ; Thu, 23 Jun 2005 05:49:52 -0700 (PDT) In-Reply-To: Content-Disposition: inline Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Balazs Ree Cc: linux-ide@vger.kernel.org > To me it seems that there are two implementations of the 80cc ribbon > detection in the kernel - one is the "standard" and the other one is the > one that some drives actually implement? Or how is that? Maybe it would be > good to make the default setting to "true" in the kernel config or rather > just put some more emphasized hint about this in some docs where people > bump into it more easily? It is true that the help text of the config > parameter tells exactly what to do, but otherwise I spent a long time > investigating this problem before finding the "obvious", and I guess I'm > not alone. > No. You aren't alone. If all chipsets supported by the same driver (piix.c for example) use the same algorithm then the there could be two variation of eighty_ninty_three and each driver could use the correct one. If this is not the case, a Documentation/UDMA.txt with a black list would be wonderfull. Rafael