From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263292AbTKYWuH (ORCPT ); Tue, 25 Nov 2003 17:50:07 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263298AbTKYWuH (ORCPT ); Tue, 25 Nov 2003 17:50:07 -0500 Received: from main.gmane.org ([80.91.224.249]:47754 "EHLO main.gmane.org") by vger.kernel.org with ESMTP id S263292AbTKYWuE (ORCPT ); Tue, 25 Nov 2003 17:50:04 -0500 X-Injected-Via-Gmane: http://gmane.org/ To: linux-kernel@vger.kernel.org From: thomas weidner <3.14159@gmx.net> Subject: do_kern_mount return value question Date: Tue, 25 Nov 2003 23:36:16 +0100 Message-ID: Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7BIT X-Complaints-To: usenet@sea.gmane.org User-Agent: Pan/0.14.2 (This is not a psychotic episode. It's a cleansing moment of clarity.) Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hi, this post is no bug or patch, but a simple question about the kernel source. Why does do_kern_mount return ENODEV when the given filesystem is unknown? EINVAL seems to be better to me. when i change from ENODEV to EINVAL (fs/super.c:713 in 2.6.0-test10) the following szenario works as i except: kernel options: rootfstype=ext3,ext2 1. load ext2 initrd 2. try to mount as ext3 -> fails (ext3 is a module in my configuration and not in the kernel) 3. try to mount it as ext2 -> fine 4. handle initrd 5. try to mount root as ext3 -> fine (ext3 module now loaded) the vanilla kernel stops after 2. because some late boot code thinks the initrd is an invalid device (ENODEV).