Xen devel - Jan 2006 - [PATCH] enable ethtool for xen loopback interface (network in dom0)

Hi,

this patch allows disabling the tx checksums on the xen loopback
interface, which is used for the virtual interfaces of dom0 instead 
of netfront/netback.

I need to issue ethtool -K eth0 tx off to work around a bad checksum
problem, which has been reported for domUs a few months ago. I''m
bridging dom0''s eth0 to a domU where I do NAT to a different virtual
interface. Without this change, I get lots of packets with bad
checksums.

Jan

Signed-off-by: Jan Niehusmann <jan@gondor.com>

--- linux-2.6.12/drivers/xen/netback/loopback.c.orig	2006-01-14
12:46:42.000000000 +0100
+++ linux-2.6.12/drivers/xen/netback/loopback.c	2006-01-14 11:59:49.000000000
+0100
@@ -27,6 +27,7 @@
 #include <linux/inetdevice.h>
 #include <linux/etherdevice.h>
 #include <linux/skbuff.h>
+#include <linux/ethtool.h>
 #include <net/dst.h>
 
 static int nloopbacks = 8;
@@ -122,6 +123,12 @@
 	/*dev->mtu             = 16*1024;*/
 }
 
+static struct ethtool_ops network_ethtool_ops +{
+	.get_tx_csum = ethtool_op_get_tx_csum,
+	.set_tx_csum = ethtool_op_set_tx_csum,
+};
+
 static int __init make_loopback(int i)
 {
 	struct net_device *dev1, *dev2;
@@ -141,6 +148,8 @@
 	dev1->features |= NETIF_F_NO_CSUM;
 	dev2->features |= NETIF_F_IP_CSUM;
 
+	SET_ETHTOOL_OPS(dev2, &network_ethtool_ops);
+
 	/*
 	 * Initialise a dummy MAC address for the ''dummy backend''
interface. We
 	 * choose the numerically largest non-broadcast address to prevent the

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

----- Original Message ----- 
From: "Jan Niehusmann" <jan@gondor.com>
> this patch allows disabling the tx checksums on the xen loopback
> interface, which is used for the virtual interfaces of dom0 instead
> of netfront/netback.
Great patch thanks. I finally got around to upgrading a box to Xen 3.0 
testing (changeset 8269:a91e4dcf629d), applied the patch and it works a 
dream.  I can now bridge the xen0 domain onto a VLAN interface with a 
pro/1000 MT server NIC (82546EB).

I am using Fedora Core 4 (which doesn''t appear to support -K ethtool
options
in the initscripts), and have added a /sbin/ifup-local script [1] to the 
system. This in conjunction with adding ''ETHTOOL_KOPTS="tx
off"'' [2] to my
interface configuration works well (Note: I use the standard initscripts to 
bring up the networking, and not ''network-bridge'').

I''m concerned that there may be some performance hit with software 
checksumming, but I''m happy with working rather than faster. Getting
jumbo
frame support in the domU''s would be probably be a better performance
boost.

Greg :-)

--
[1] /sbin/ifup-local
#!/bin/sh

cd /etc/sysconfig/network-scripts
. network-functions

[ -f ../network ] && . ../network

CONFIG="ifcfg-${1}"
source_config

if [ -n "$ETHTOOL_KOPTS" ] ; then
        REALDEVICE=${DEVICE%%:*}
        /sbin/ethtool -K ${REALDEVICE} $ETHTOOL_KOPTS
fi


[2]  /etc/sysconfig/network-scripts/ifcfg-veth0
DEVICE=veth0
TYPE=Ethernet
ONBOOT=yes
IPADDR=192.168.133.1
NETMASK=255.255.255.0
MACADDR=00:aa:00:00:85:01
NOZEROCONF=yes
GATEWAY=192.168.133.254
ARP=yes
ETHTOOL_KOPTS="tx off"


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel