Subject: Re: panic in new ieee1394 code
To: KIYOHARA Takashi <kiyohara@kk.iij4u.or.jp>
From: Michael van Elst <mlelstv@serpens.de>
List: current-users
Date: 07/21/2005 20:12:54
On Fri, Jul 22, 2005 at 01:59:00AM +0900, KIYOHARA Takashi wrote:
> Hi, mlelstv!
> 
> 
> From: mlelstv@serpens.de (Michael van Elst)
> Date: Thu, 21 Jul 2005 05:54:38 +0000 (UTC)
> 
> > There seems to be some confusion in the driver.
> > 
> > attach disk 1 -> sd0 appears
> > detach disk 1 -> no reaction
> > attach disk 1 -> sd0 is still there and can be used (!)
> > detach disk 1 -> no reaction
> > attach disk 2 -> sd1 appears and can be used
> > detach disk 2 -> sd0 detaches and the kernel panics with a 0xdeadbeef dereference
> 
> How did you do 'detach disk' sd0?

I pulled the firewire cable from the disk 2 enclosure.


> From: Dave Huang <khym@azeotrope.org>
> Date: Wed, 20 Jul 2005 13:55:19 -0500
> 
> > I had noticed that too, but then saw this in the sbp(4) manpage:
> > 
> >      Some users familiar with umass(4) might wonder why the device is
> >      not detached at the scsi(4) layer when the device is unplugged.
> >      It is detached only if the device has not been plugged again
> >      during several bus resets. This is for preventing to detach an
> >      active file system even when the device cannot be probed
> >      correctly for some reason after a bus reset or when the device is
> >      temporary disconnected because the user changes the bus topology.
> >      If you want to force to detach the device, run fwctl -r several
> >      times.
> 
> Please log.  ;-)

Hmm. I guess I have to get used to that scheme :-/

Greetings,
-- 
                                Michael van Elst
Internet: mlelstv@serpens.de
                                "A potential Snark may lurk in every tree."