From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:60087)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <kraxel@redhat.com>) id 1ZAv9A-0006S1-Tn
	for qemu-devel@nongnu.org; Fri, 03 Jul 2015 03:12:29 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <kraxel@redhat.com>) id 1ZAv97-0000ux-P8
	for qemu-devel@nongnu.org; Fri, 03 Jul 2015 03:12:28 -0400
Received: from mx1.redhat.com ([209.132.183.28]:51923)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <kraxel@redhat.com>) id 1ZAv97-0000ul-Kp
	for qemu-devel@nongnu.org; Fri, 03 Jul 2015 03:12:25 -0400
Message-ID: <1435907542.26896.8.camel@redhat.com>
From: Gerd Hoffmann <kraxel@redhat.com>
Date: Fri, 03 Jul 2015 09:12:22 +0200
In-Reply-To: <55962F84.1090902@redhat.com>
References: <1435568020-8669-1-git-send-email-kraxel@redhat.com>
	<1435568020-8669-3-git-send-email-kraxel@redhat.com>
	<55962F84.1090902@redhat.com>
Content-Type: text/plain; charset="UTF-8"
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [SeaBIOS] [PATCH 02/18] virtio: run drivers in
	32bit mode
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: Paolo Bonzini <pbonzini@redhat.com>
Cc: seabios@seabios.org, qemu-devel@nongnu.org, "Michael S. Tsirkin" <mst@redhat.com>

On Fr, 2015-07-03 at 08:45 +0200, Paolo Bonzini wrote:
> 
> On 29/06/2015 10:53, Gerd Hoffmann wrote:
> > virtio version 1.0 registers can (and actually do in the qemu
> > implementation) live in mmio space.  So we must run the blk and
> > scsi virtio drivers in 32bit mode, otherwise we can't access them.
> > 
> > This also allows to drop a bunch of GET_LOWFLAT calls from the virtio
> > code in the following patches.
> 
> This isn't really necessary though.  The config access capability exists
> for this reason.  At least as a follow up it would be nice to add a switch.

Just saw mst's patch implementing it in qemu.  I'll have a look. 

I think I want keep the virtio drivers as 32bit drivers.  The modern
virtio bar is a 64bit though, so accessing via config access capability
might be needed even for the 32bit drivers in case it is mapped above
4G ...

cheers,
  Gerd