From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263118AbTJPTQr (ORCPT ); Thu, 16 Oct 2003 15:16:47 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263119AbTJPTQq (ORCPT ); Thu, 16 Oct 2003 15:16:46 -0400 Received: from tmr-02.dsl.thebiz.net ([216.238.38.204]:29701 "EHLO gatekeeper.tmr.com") by vger.kernel.org with ESMTP id S263118AbTJPTQo (ORCPT ); Thu, 16 Oct 2003 15:16:44 -0400 To: linux-kernel@vger.kernel.org Path: gatekeeper.tmr.com!davidsen From: davidsen@tmr.com (bill davidsen) Newsgroups: mail.linux-kernel Subject: Re: usb-storage kills lilo (2.6-test[67]) Date: 16 Oct 2003 19:06:50 GMT Organization: TMR Associates, Schenectady NY Message-ID: References: <20031012051107.GA1881@defiant> X-Trace: gatekeeper.tmr.com 1066331210 18384 192.168.12.62 (16 Oct 2003 19:06:50 GMT) X-Complaints-To: abuse@tmr.com Originator: davidsen@gatekeeper.tmr.com Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org In article <20031012051107.GA1881@defiant>, Norbert Kiesel wrote: | (sorry for the flashy subject, but I could not come up with another one | :-) | | Hi, I got problems with usb-storage and linux-2.6-test[67]. AFAICS, | test5 works fine (still have to retest to make sure). | | Problem is that inserting my USB flash memory disk makes /dev/sda (and | /dev/sda1) appear in /proc/partitions, but removing it does not remove | the entries from /proc/partitions. lilo is reading /proc/partitions | (verified trough strace) and dies because /dev/sda is gone. | | Possibly related to that is the next time I insert the flash disk, is | shows up as /dev/sdb1 (and adds this to /proc/partiotions, too). I have seen similar behaviour with flash readers. I have CF and memstick readers, and after a photo session I'm likely to have a bunch of each to unload. If I do the CF first, the device is sda. Even changing media, still sda. But when I unplug and connect the memstick reader, that becomes sdb, and sda is shown as missing in action. I didn't report it because I lack time to characterize it properly and I can live with it. But since you report the problem, I'll throw this info out in case it helps someone else understand exactly what's happening. I *believe* I first saw it in test4, I haven't been running 2.6 kernels lately on that machine, so I can't say if it's still an issue. -- bill davidsen CTO, TMR Associates, Inc Doing interesting things with little computers since 1979.