Re: "Hard" crash with -current XEN3_DOM0 kernel

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Thu, Jun 19, 2008 at 12:34:27AM +0200, Jean-Yves Migeon wrote:
> Jean-Yves Migeon wrote:
> >Manuel Bouyer wrote:
> >>No problems on a Dell 755, but I'm using serial console, it doesn't have
> >>a real pckbc port (it's a usb-only box). A pckbc is attached though
> >>(I guess the Intel chipset still provides a pckbc, even if it's not 
> >>wired
> >>to an external connector):
> >>pckbc0 at isa0 port 0x60-0x64
> >>pckbdprobe: reset error 5
> >>pmsprobe: reset error 5
> >>  
> >I identified the issue, this patch causes some trouble on my laptop:
> >
> >http://cvsweb.netbsd.org/bsdweb.cgi/src/sys/dev/ic/pckbc.c.diff?r1=1.44&r2=1.45
> > 
> >
> >
> >I will give tech-kern a try.
> Hmm, slight correction.
> 
> Relating to this patch, increasing <s> value (from 100 to 1000 for 
> example) and decreasing the delay() value accordingly makes the dom0 
> boot again.
> 
> Any hint what could cause such an issue? Looks like Xen is at fault 
> there, booting with GENERIC is fine.

The dom0 kernel talks directly to the hardware at this point. The xen kernel
isn't involved, exept it's already handling interrupts at this point.
Maybe it's causing an interrupt storm, which exaust the hypervisor's
stack ?

-- 
Manuel Bouyer, LIP6, Universite Paris VI.           
Manuel.Bouyer%lip6.fr@localhost
     NetBSD: 26 ans d'experience feront toujours la difference
--