From: Blaisorblade <blaisorblade_spam@yahoo.it>
To: user-mode-linux-user@lists.sourceforge.net
Subject: Re: [uml-user] Harddisk Shutdown while UML Guest Shutdown
Date: Wed, 3 Nov 2004 22:09:52 +0100
User-Agent: KMail/1.7.1
Cc: Roland Kaeser <roli8200@yahoo.de>, LKML <linux-kernel@vger.kernel.org>
References: <20041103192155.86313.qmail@web26104.mail.ukl.yahoo.com>
In-Reply-To: <20041103192155.86313.qmail@web26104.mail.ukl.yahoo.com>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200411032209.52449.blaisorblade_spam@yahoo.it>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 3788
Lines: 93

On Wednesday 03 November 2004 20:21, Roland Kaeser wrote:
> Hello
>
> Thank You for Your answer. But my host kernel is a 2.6.7 with the skas
> patch. The guest is a 2.6.9. Just an another info: The behaviour does not
> depend on the hostfs configuration. It also happens when I use a ubd
> device.

> And no, the HOST!! freezes after exit of the guest kernel.

> And i get a 
> Kernel panic from the HOST!! kernel, this in case the host (ide) harddisk
> drive spins down (but not spins up anymore).
Yes, I already understood this.

> My idea is that some routines 
> to spin down the harddisk are been routed outside the uml guest kernel or
> not been sucessfully removed for the uml architecture.

> Is it possible that 
> the /sbin/halt binary can have made something with the hosts harddisk?
I don't think that /sbin/halt itself can do anything. I would be suspicious of 
hdparm, instead. It would be a UML bug anyway, but it's the more likely thing 
it's possible to cause this.

> How can i get the kernel panic message from the host?

If you are on a virtual console, you see them on your screen. Otherwise, how 
did you guess you have a kernel panic? Messages on /var/log/messages (I guess 
no)? Or what?

> Roland

>  --- Blaisorblade <blaisorblade_spam@yahoo.it> schrieb:
> > On Wednesday 03 November 2004 08:53, Roland Kaeser wrote:
> > > Hello

> > > I starting the UML Kernel with the following command:

> > > /install/uml/kernel/vmlinux mem=128M devfs=nomount con=xterm
> > > eth0=tuntap,,,1.1.1.1 eth1=tuntap,,,2.2.2.2 root=/dev/root
> > > rootflags=/install/uml/instances/hostfc2install rootfstype=hostfs

> > > The system starts up successfully, and opens all the xterm for the
> > > consoles. I can login and work with the guest system.
> > > Then I shut down the guest system from inside the guest system using
> > > the command init 0 or shutdown.

> > > The guest shuts down normally.
> > > But in the moment when the uml kernel exits in the xterm where I
> > > started it, I CAN HEAR the harddisk of the host goes down.

> > Ok, with this explaination, I can believe that this happens. Sorry for my
> > first answer, I first took the most likely idea, i.e. the hard disk spins
> > down because it's no more used, but the host kernel is still alive and
> > kicking, and the hard drive will spin up when needed. Anyway, this
> > description, w

> > You are using a vanilla host kernel (2.6.9), so you have no SKAS patch
> > running, i.e. this is a host kernel bug, actually, and it is pretty
> > severe. Attach your .config in next email.

> > Can you please report that (including the kernel panic message) to LKML,
> > CC:ing the -devel list? Also, could you try that with different host
> > kernels? Different setups? Are you running your UML as root or not? If
> > you are running it as root, then try running it with an unprivileged
> > userid.

> > > After this, nothing happens more on the host, even the mouse freezes. I
> > > cannot start an other program or even access the harddisk or login on a
> > > other tty.

> > > It freezes the whole guest system.

> > You mean obiously the "host" system (i.e. the hardware one; "guest"
> > refers to the UML instance).

> > > About a minute later I get a
> > > Kernel Panic from the host kernel

> > Try to copy and send the text of the panic, (most of the numbers can be
> > omitted, even because they are often hard to decode; the list of function
> > calls is more important than anything else).

-- 
Paolo Giarrusso, aka Blaisorblade
Linux registered user n. 292729

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/