On Wed, Feb 08, 2006 at 07:45:00AM +0000, lha@NetBSD.org wrote:
> >Environment:
> System: NetBSD xen2-netbsd4-arla.lab.it.su.se 3.99.15 NetBSD 3.99.15 (XENU) #0: Thu Feb 2 13:53:02 CET 2006 root@xen2-netbsd-netbsd.it.su.se:/usr/src/sys/arch/i386/compile/XENU i386
> Architecture: i386
> Machine: i386
> >Description:
> 
> 	the linux dom0 xenhost that my arla lab machine
> 	seems to have something weired going on in the network
> 	land, iether in the physical end, or in the xen/dom0 end.
> 
> 	anyway, when running my nightly tests for arla my netbsd domU
> 	fell over with this backtrace
> 
> db> tr
> xen_network_handler(c0855000,c9b77e14,c07c3480,c9b77dbc,1) at netbsd:xen_network_handler+0x18c
> evtchn_do_event(4,c9b77e14,0,4c,2) at netbsd:evtchn_do_event+0xee
> do_hypervisor_callback(c9b77e14,0,11,c0470031,c07c0011) at netbsd:do_hypervisor_callback+0x13f
> hypervisor_callback(4,c0395520,c0855000,5,c085501c) at netbsd:hypervisor_callback+0x64
> xennet_interface_status_change(c045d3c4,0,c9b82b60,0,10) at netbsd:xennet_interface_status_change+0x1db

Could you instrument xennet_interface_status_change() to see which
status_change messages are received before this happens ?

-- 
Manuel Bouyer <bouyer@antioche.eu.org>
     NetBSD: 26 ans d'experience feront toujours la difference
--