From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261490AbVFZR6d (ORCPT ); Sun, 26 Jun 2005 13:58:33 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261493AbVFZR6d (ORCPT ); Sun, 26 Jun 2005 13:58:33 -0400 Received: from clock-tower.bc.nu ([81.2.110.250]:55750 "EHLO lxorguk.ukuu.org.uk") by vger.kernel.org with ESMTP id S261490AbVFZR6a (ORCPT ); Sun, 26 Jun 2005 13:58:30 -0400 Subject: Re: [PATCH] Allow number of IDE interfaces to be selected (X86) From: Alan Cox To: Nick Warne Cc: Linux Kernel Mailing List In-Reply-To: <200506251641.40922.nick@linicks.net> References: <200506251641.40922.nick@linicks.net> Content-Type: text/plain Content-Transfer-Encoding: 7bit Message-Id: <1119808554.28649.37.camel@localhost.localdomain> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.6 (1.4.6-2) Date: Sun, 26 Jun 2005 18:55:55 +0100 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Sad, 2005-06-25 at 16:41, Nick Warne wrote: > A small patch I done to allow X86 users to select the max number of IDE > interfaces they have - this eliminates the need for passing idex=noprobe on > the command line and/or stops the needless probes at boot on non-existant IDE > interfaces. The needless probe cases patch went into -mm already. Also your change doesn't eliminate the need for noprobe in cases where its not the first or second interface in order. The assumption most x86 users will only have two IDE interfaces is also generally wrong today, although they may fall under the SATA driver on newer boards. I don't think this patch should go in - the one real reason for having it in embedded boxes was saving memory. The right fix for that is already something Bartlomiej has talked about fixing - the static allocation of the ide_hwifs array itself. Alan