On Mon, Apr 11, 2005 at 09:23:54AM +0300, Martti Kuparinen wrote:
> I renamed my dom0 kernel to /netbsd and tried again. DDB dumps all
> 128 MB according to the screen messages but savecore does not find
> anything.
> 
> (copied manually from the screen)
> 
> stopped at pid 309.1 (python2.3) at netbsd:db_read_bytes+0x3b movl 0 
> (%ecx),%eax
> 
> tr
> 
> db_read_bytes(0,4,cae93cd0,0,8) at netbsd_db_read_bytes+0x3b
> db_get_value(0,4,0,0,c060f9a0) at netbsd:db_get_value+0x18
> db_stack_trace_print(0,0,ffff,c0574d76,c031e740) at 
> netbsd:db_stack_trace_print+0x2e4
> ddb_trap_hook(1,1,cae93ddc,c0310d9e,0) at netbsd:ddb_trap_hook+0x98
> db_trap(6,0,cae93e0c,c02fb967,c0b89000) at netbsd:db_trap+0x18
> kdb_trap(6,0,cae93efc,2106,1) at netbsd:kdb_trap+0x119
> trap() at netbsd:trap+0xab
> --- trap (number 6) ---
> ?(1b,caeb3fa4,c9be9130,0,caeb3fa0) at 0
> do_hypervisor_callback(caeb3fa4,0,85d001f,bfa0001f) at 
> netbsd:do_hypervisor_callback+0xa1

It looks like it calls a null function pointer in do_event()
There is probably a locking issue somewhere with the interrupt callback
register. Does is also happen if you create the domains by hand with
xm create -c ?

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