Re: kernel diagnostic assertion in evtchn.c

[Sarton O'Brien <bsd-xen%roguewrt.org@localhost>]

Sarton O'Brien wrote:
> Manuel Bouyer wrote:
> > On Sat, Jun 07, 2008 at 09:37:21PM +0200, Manuel Bouyer wrote:
> > > [...]
> > > > evtchn_do_event: ih 0xc1584660 ih_level 0 != i 8
> > > > ih handlers:
> > > >  0xc1584660 fn 0xc06b7e00
> > > >  0xc1584ea0 fn 0xc04fa890
> > > A handler with ih_level isn't somthing expected.
> > > Can you look in your kernel binary what functions
> > > 0xc06b7e00 and 0xc04fa890 are ?
> > > nm /netbsd should be enough to find it.
> >
> > I think I found it (ih_level wasn't initialized for xenevt_processevt()),
> > should be fixed with xenevt.c 1.28
> >
> > Running a HVM guest did increase the probability of seeing it, but 
> > even without
> > HVM guests I would have expected to see it from time to time ...
>
> I don't use hvm at all but have had consistent crashes when using smb in 
> general for high volume transfers. That seems to be about the only thing 
> in common.
>
> I still need to find this usb serial port belkin thing i have but will 
> test properly when I do.

The belkin dock/usb-to-serial adapter works (after needing the vista 
drivers on xp :S ) I haven't had the same problem since trying to 
replicate but now I'm stuck on another issue which is preventing me from 
testing the system properly (see thread unresponsive after mount msdos 
and cd /mnt within amd64 dom0).

I _believe_ the original problem has gone ... but can't see the forest 
for the trees at the moment.

Sarton