From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Fri, 27 Jul 2001 01:05:47 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Fri, 27 Jul 2001 01:05:36 -0400 Received: from mail15b.boca15-verio.com ([208.55.91.59]:44639 "HELO mail15b.boca15-verio.com") by vger.kernel.org with SMTP id ; Fri, 27 Jul 2001 01:05:29 -0400 Message-ID: <3B60F5FB.AB29B413@sigmastorage.com> Date: Thu, 26 Jul 2001 22:02:51 -0700 From: Matt Ryan X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.4.2-2-june14 i686) X-Accept-Language: en MIME-Version: 1.0 To: linux-kernel@vger.kernel.org Subject: device name switching Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Loop-Detect: 1 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Original-Recipient: rfc822;linux-kernel-outgoing [ please CC responses to me ] hello - if I have two hard drives /dev/hda and /dev/hdc, and I pull out the first drive (root partition is raid1 for redundant boot), on reboot I am left with just /dev/hdc. this would seem like the normal, anticipated behavior. however, if I stick in a couple of drives connected to a Promise Ultra 100 card and try the same sort of thing, the drive name assignments reliably are changed. if I start with /dev/hd[aceg] (e and g on the promise card), and pull out hdc, on reboot I have /dev/hd[ace]. if I start with /dev/hd[aceg] and pull out hda, I also end up with /dev/hd[ace]. the higher drives sort of collapse down to fill the 'hole.' this behavior has been reproducible through a series of 2.4 kernels, the last I've tried is 2.4.5. my question is, is this anticipated behavior, or does this surprise anybody? for all I know the BIOS decides the assignments and the kernel has nothing to do with it (also this is an intel d815eea2 chipset, though I'm pretty sure I've seen the same behavior on a Sis-based chipset as well). another reminiscent problem (but probably totally unrelated), is the name assigned to NICs. in a box with the same intel motherboard, with one onboard nic (eepro100) and one PCI nic (3C905C), I have a 2.4.3 kernel and a 2.4.7 kernel. each kernel has the opposite idea of which card is eth0 and which is eth1. curious... Matt