CentOS virt - Mar 2012 - CentOS 6.2 / kvm / Windows XP sp3 guest crashes repeatedly after a few minutes

Hello to all:

Recently upgraded from CentOS 6.0 to 6.2 to get USB passthrough support by 
performing "yum update".  USB passthrough works great, but now my
Windows XP
sp3 guest crashes repeatedly after a few minutes.


[root at shackleton-eth0 log]# rpm -qa|egrep
'(kvm|kernel|libvirt-|qemu)'|sort
abrt-addon-kerneloops-2.0.4-14.el6.centos.x86_64
dracut-kernel-004-256.el6_2.1.noarch
gpxe-roms-qemu-0.9.7-6.9.el6.noarch
kernel-2.6.32-220.7.1.el6.x86_64
kernel-2.6.32-71.el6.x86_64
kernel-devel-2.6.32-220.7.1.el6.x86_64
kernel-devel-2.6.32-71.el6.x86_64
kernel-firmware-2.6.32-220.7.1.el6.noarch
kernel-headers-2.6.32-220.7.1.el6.x86_64
libreport-plugin-kerneloops-2.0.5-20.el6.x86_64
libvirt-0.9.4-23.el6_2.6.x86_64
libvirt-client-0.9.4-23.el6_2.6.x86_64
libvirt-python-0.9.4-23.el6_2.6.x86_64
qemu-img-0.12.1.2-2.209.el6_2.4.x86_64
qemu-kvm-0.12.1.2-2.209.el6_2.4.x86_64


This message in /var/log/messages precedes every crash:


[root at shackleton-eth0 log]# grep kvm /var/log/messages
Mar 19 09:09:25 shackleton-eth0 kernel: kvm: 2889: cpu0 unhandled wrmsr: 0x198
data 0
Mar 19 09:15:48 shackleton-eth0 kernel: kvm: 2889: cpu0 unhandled wrmsr: 0x198
data 0
Mar 19 09:25:51 shackleton-eth0 kernel: kvm: 2889: cpu0 unhandled wrmsr: 0x198
data 0
Mar 19 11:21:27 shackleton-eth0 kernel: kvm: 2889: cpu0 unhandled wrmsr: 0x198
data 0
Mar 19 12:07:16 shackleton-eth0 kernel: kvm: 2889: cpu0 unhandled wrmsr: 0x198
data 0
Mar 19 12:16:14 shackleton-eth0 kernel: kvm: 2889: cpu0 unhandled wrmsr: 0x198
data 0
Mar 19 13:46:27 shackleton-eth0 kernel: kvm: 2889: cpu0 unhandled wrmsr: 0x198
data 0
Mar 19 14:06:49 shackleton-eth0 kernel: kvm: 2889: cpu0 unhandled wrmsr: 0x198
data 0


CentOS 6.0 was stable.  Windows XP sp3 did not crash.

Formerly, I was running virtio-win-0.1-15.iso, but now upgraded to 
virtio-win-0.1-22.iso and Windows XP sp3 remains unstable.

i had the same problem with centos 6.3 on ibm x3650 server. windows xp 
with virtio net just reset with any network load and on host 
/var/log/messages says kernel: kvm: 2889: cpu0 unhandled wrmsr: 0x198 
data 0

found the solution on proxmox forum.

just disable jumbo frames on centos host interface ifcfg and ethernet 
switch.

hope this help someone.