From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:59615)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <benh@kernel.crashing.org>) id 1UwK7g-0001f3-Jd
	for qemu-devel@nongnu.org; Mon, 08 Jul 2013 18:41:33 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <benh@kernel.crashing.org>) id 1UwK7f-0007GK-3F
	for qemu-devel@nongnu.org; Mon, 08 Jul 2013 18:41:32 -0400
Message-ID: <1373323278.4446.61.camel@pasglop>
From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Date: Tue, 09 Jul 2013 08:41:18 +1000
In-Reply-To: <87sizosnoj.fsf@codemonkey.ws>
References: <1372315560-5478-1-git-send-email-aik@ozlabs.ru>
	<1372315560-5478-9-git-send-email-aik@ozlabs.ru>
	<87sizozyjl.fsf@codemonkey.ws> <1373319954.4446.52.camel@pasglop>
	<87sizosnoj.fsf@codemonkey.ws>
Content-Type: text/plain; charset="UTF-8"
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [PATCH 08/17] pseries: savevm support for PAPR TCE
 tables
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Anthony Liguori <aliguori@us.ibm.com>
Cc: Alexey Kardashevskiy <aik@ozlabs.ru>, Alexander Graf <agraf@suse.de>, qemu-devel@nongnu.org, qemu-ppc@nongnu.org, Paolo Bonzini <pbonzini@redhat.com>, Paul Mackerras <paulus@samba.org>, David Gibson <david@gibson.dropbear.id.au>

On Mon, 2013-07-08 at 17:15 -0500, Anthony Liguori wrote:

> Hypercall arguments are passed in CPU endianness so what's being stored
> in the tce table is CPU endianness.
> 
> Since VBUFFER just does a blind write() of the full array of uint64s,
> what goes on the wire will be CPU endianness.
> 
> So if you do a savevm on a little endian host and loadvm on a big endian
> host, badness ensues.
> 
> The proper thing to do is use a VARRAY instead of a VBUFFER.  VARRAY
> will handle endian because it treats the data as an array, not as an
> opaque buffer.

Ok, so that's indeed an issue for emulated TCEs because what qemu stores
is not the real (BE) TCE table but a "host native" version of it. I see.

Cheers,
Ben.