From mboxrd@z Thu Jan  1 00:00:00 1970
From: Arnd Hannemann <hannemann@nets.rwth-aachen.de>
Subject: Re: xen dom0 2.6.32.15 kernel BUG at
	drivers/xen/grant-table.c:583
Date: Mon, 14 Jun 2010 13:09:49 +0200
Message-ID: <4C160DFD.5040707@nets.rwth-aachen.de>
References: <4C15E000.7060509@nets.rwth-aachen.de>
	<alpine.DEB.2.00.1006141156170.3401@kaball-desktop>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7BIT
Return-path: <xen-devel-bounces@lists.xensource.com>
In-reply-to: <alpine.DEB.2.00.1006141156170.3401@kaball-desktop>
List-Unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>,
	<mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
List-Post: <mailto:xen-devel@lists.xensource.com>
List-Help: <mailto:xen-devel-request@lists.xensource.com?subject=help>
List-Subscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>,
	<mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
Sender: xen-devel-bounces@lists.xensource.com
Errors-To: xen-devel-bounces@lists.xensource.com
To: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Cc: "xen-devel@lists.xensource.com" <xen-devel@lists.xensource.com>
List-Id: xen-devel@lists.xenproject.org

Am 14.06.2010 12:57, schrieb Stefano Stabellini:
> On Mon, 14 Jun 2010, Arnd Hannemann wrote:
>> Hi,
>>
>> we have regular but hard to reproduce (wait for a day or two starting domUs) kernel panics (see below) with latest
>> "xen/stable-2.6.32.x" git tree.
>>
>> Any idea, anyone?
>>
> 
> this CS from origin/xen/dom0/gntdev should fix your problem:

Thanks a lot.
We will try this, and post results...

> 
> sstabellini@kaball-desktop:~/xensource/linux-pvops-latest$ git show ad469f0da31bc16b945f9a06710b9d45434d0091
> commit ad469f0da31bc16b945f9a06710b9d45434d0091
> Author: Stefano Stabellini <Stefano.Stabellini@eu.citrix.com>
> Date:   Wed Jun 9 12:34:02 2010 -0700
> 
>     xen/gntdev: use spinlocks rather than rwsem for locking
>     
>     The mmu notifier mechanism calls its callbacks with an rcu lock,
>     which disables preemption.  This means we cannot use any blocking
>     synchronization for locking.
>     
>     Convert all the rwsemas to plain spinlocks.  This requires that
>     the memory allocation and copying to/from userspace be split
>     from the actual datastructure updates since they can't be done
>     under spinlock.
>     
>     Signed-off-by: Stefano Stabellini <Stefano.Stabellini@eu.citrix.com>
>     Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
> 

Best regards,
Arnd