From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Seewer Philippe" Subject: Re: [PATCH] sym53c8xx_2: Add bios_param to sym_glue.c Date: Fri, 17 Feb 2006 15:41:27 +0100 Message-ID: <43F5E097.4080709@bfh.ch> References: <43F5D963.9080009@bfh.ch> <20060217143718.GS12822@parisc-linux.org> Mime-Version: 1.0 Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: 7bit Return-path: Received: from mr1.bfh.ch ([147.87.250.50]:34737 "EHLO mr1.bfh.ch") by vger.kernel.org with ESMTP id S1751461AbWBQOl3 (ORCPT ); Fri, 17 Feb 2006 09:41:29 -0500 Content-class: urn:content-classes:message In-Reply-To: <20060217143718.GS12822@parisc-linux.org> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: Matthew Wilcox Cc: linux-kernel@vger.kernel.org, linux-scsi@vger.kernel.org Matthew Wilcox wrote: > On Fri, Feb 17, 2006 at 03:10:43PM +0100, Seewer Philippe wrote: > >>This patch adds the scsi common function bios_param to the sym53c8xx >>driver. For simplicity i just copied the code from the sym53c416 driver. > > > If the driver doesn't define bios_param, the scsi core calls > scsicam_bios_param, which seems to do everything this patch does, > and more. > > A quick survey suggests that most drivers should have their bios_param > methods deleted. Was there a particular problem you found with the > default scsicam_bios_param implementation? Yes. Using scsicam_bios_parm and other defaults result in a geometry of 64 heads and 32 sectors even for big disks, which is not what the pc bios "gets" from the controller. That is more along the lines of 255/63. Returning a geometry of x/255/63 seems to be the default for bigger disks withing scsi drivers, so i just copied the code...