From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from [140.186.70.92] (port=42137 helo=eggs.gnu.org)
	by lists.gnu.org with esmtp (Exim 4.43) id 1OySxv-00009v-Se
	for qemu-devel@nongnu.org; Wed, 22 Sep 2010 13:18:45 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.69)
	(envelope-from <tygrys@moo.pl>) id 1OySxu-0001qh-OG
	for qemu-devel@nongnu.org; Wed, 22 Sep 2010 13:18:43 -0400
Received: from moo.pl ([217.149.240.132]:43684)
	by eggs.gnu.org with esmtp (Exim 4.69)
	(envelope-from <tygrys@moo.pl>) id 1OySxu-0001pr-J8
	for qemu-devel@nongnu.org; Wed, 22 Sep 2010 13:18:42 -0400
Received: from localhost (localhost [127.0.0.1])
	by moo.pl (Postfix) with ESMTP id 2BE7018403
	for <qemu-devel@nongnu.org>; Wed, 22 Sep 2010 19:18:39 +0200 (CEST)
Date: Wed, 22 Sep 2010 19:18:32 +0200
From: Leszek Urbanski <tygrys@moo.pl>
Message-ID: <20100922171832.GA28721@moo.pl>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-2
Content-Disposition: inline
Subject: [Qemu-devel] BUG - qdev - partial loss of network connectivity
List-Id: qemu-devel.nongnu.org
List-Unsubscribe: <http://lists.nongnu.org/mailman/listinfo/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: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: qemu-devel@nongnu.org
Cc: tygrys@moo.pl

Hi,

This is a qemu-kvm (i.e. not qemu) bug report. I've been told on IRC (#kvm)
that this bug report should go to this list anyway.


host and guest kernel: 2.6.32.22
arch: amd64
qemu-kvm: 0.12.5 and 0.13.0-rc1

How to reproduce: copy a large (few hundred MB) file to an NFS mount
(guest is the client), using qdev syntax for virtio:

-device virtio-net-pci,netdev=tap1,mac=ab:cd:ef:01:23:45 \
-netdev type=tap,id=tap1,ifname=tap1

all NFS mounts immediately stall and processes go into D state and are stuck
on nfs_wait_bit_uninterruptible. There is no further communication between
the client and the server. It occurs with UDP and TCP mounts.

That same guest, when switched to the old -net/-net syntax:

-net nic,vlan=0,model=virtio,name=eth0,macaddr=ab:cd:ef:01:23:45 \
-net tap,vlan=0,ifname=tap1

works fine. It also works fine on bare metal.

I can reproduce the problem every time within 10 minutes with qdev.


-- 
Leszek "Tygrys" Urbanski, SCSA, SCNA
 "Unix-to-Unix Copy Program;" said PDP-1. "You will never find a more
  wretched hive of bugs and flamers. We must be cautious." -- DECWARS
     http://cygnus.moo.pl/ -- Cygnus High Altitude Balloon