From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([140.186.70.92]:55586)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <sw@weilnetz.de>) id 1RTfAh-0007pq-Jx
	for qemu-devel@nongnu.org; Thu, 24 Nov 2011 14:41:24 -0500
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <sw@weilnetz.de>) id 1RTfAg-0003ML-HI
	for qemu-devel@nongnu.org; Thu, 24 Nov 2011 14:41:23 -0500
Received: from v220110690675601.yourvserver.net ([78.47.199.172]:49538)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <sw@weilnetz.de>) id 1RTfAg-0003M8-4g
	for qemu-devel@nongnu.org; Thu, 24 Nov 2011 14:41:22 -0500
Message-ID: <4ECE9DAE.2070102@weilnetz.de>
Date: Thu, 24 Nov 2011 20:40:30 +0100
From: Stefan Weil <sw@weilnetz.de>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-15; format=flowed
Content-Transfer-Encoding: 7bit
Subject: [Qemu-devel] [BUG 1.0] Networking in MIPS Malta emulation is broken
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: Avi Kivity <avi@redhat.com>
Cc: QEMU Developers <qemu-devel@nongnu.org>, Aurelien Jarno <aurelien@aurel32.net>

One of my standard tests (booting a MIPS Malta emulation with a
Debian NFS root) no longer works.

Git bisect shows that this commit caused the regression:

commit 5632ae46d5bda798e971dae48ebb318ac2c3686a
Author: Avi Kivity <avi@redhat.com>
Date:   Mon Sep 12 13:00:05 2011 +0300

     mips_malta: move i8259 initialization after piix4 initialization

     i8259 is an ISA device (or at least, depends on the ISA 
infrastructure to
     register its ioport); and the ISA bus is supplied by piix4.  Later 
patches
     make this dependency explicit.

     Use qemu_irq_proxy() to stop the cycle by adding an extra layer of
     indirection.

     Signed-off-by: Avi Kivity <avi@redhat.com>

The test was run with default networking and with an eepro100
network card. In both cases, Linux shows the ethernet devices,
but fails to read init from NFS. I think there are problems
with interrupt handling.

Is there a simple solution? Otherwise I'll examine this issue.
It would be good to get this fixed before 1.0 is released.

Regards,
Stefan