NetBSD -current with Xen 3.0.3: kernel panic while starting domU

[Rainer Brinkmoeller <rainer.brinkmoeller%web.de@localhost>]

Hi,

at a machine which i updated from NetBSD 3 with Xen 2.0.7 to NetBSD 
-current with Xen 3.0.3 i've got a kernel panic while starting the third 
domU:

panic: kernel diagnostic assertion "pcg->pcg_objects[idx].pcgo_va == NULL"
failed: file "/usr/src/sys/kern/subr_pool.c", line 2008
Stopped in pid 104.3 (python2.4) at    netbsd: cpu_Debugger+0x4: popl
%
ebp
db>

Then enter "bt" at this prompt. Next you see the list of function i've got.

On Wed, Nov 08, 2006 at 08:59:53PM +0100, Hubert Feyrer wrote:

> > > >cpu_Debugger +0x4
> > > >panic +0x155
> > > >__assert + 0x2e
> > > >pool_cache_get_paddr
> > > >bt_free +0x1f
> > > >vmem_xfree +0x190
> > > >xen_shm_unmap +0xa9
> > > >xbdback_iodome +0x12f
> > > >biodome +0x111
> > > >wddone +0x82
> > > >wdc_ata_bio_done +0x88
> > > >wdc_ata_bio_intr +0x1ba
> > > >wdcintr +0xf3
> > > >pciide_pci_intr +0x49
> > > >evtchn_do_event +0xbd
> > > >do_hypervisor_callback +0xfa
> > > >hypervisor_callback +0x64
> > > >bt_alloc +0x24
> > > >vmem_xalloc +0xf1
> > > >vmem_alloc +0xd2
> > > >kmem_alloc +0x45
> > > >kmem_zalloc +0x1e
> > > >sa_upcall_userret +0x23e
> > > >upcallret +0x5f
> > > >sa_yield +0xe9
> > > >sys_sa_yield +0x7b
> > > >syscall_plain +0xb3
> > > >--- syscall (number 334) ---
> > > >0xbba0b267
> >    
> >
> >
> > I hope Manuel has a clue here... :->
>  
>
> Not really, as I didn't track the vmem(9) stuff. 
> Rainer, you should post this to the port-xen list; I hope the author of vmem
> can look at this, and especially see if it's OK to call vmem_xfree() while
> we're already in vmem_xalloc(). Maybe there's some spl calls missing here.

done

Meanwhile i did reboot and all (4) domU did boot fine now.
Further information needed?

regards,
Rainer