On Fri, Aug 24, 2007 at 08:37:57PM +0200, Quentin Garnier wrote:
> On Fri, Aug 24, 2007 at 08:12:20PM +0200, Manuel Bouyer wrote:
> > On Fri, Aug 24, 2007 at 02:19:48PM +0200, Quentin Garnier wrote:
> > > On Fri, Aug 24, 2007 at 12:24:14PM +0100, Patrick Welche wrote:
> > > > Have you possibly just solved port-i386/36611 ? (The computer in question is
> > > > no longer with us.)
> > > 
> > > Actually, no, it's a different code path.  36611 is something I remeber
> > > Manuel Bouyer having with one of his Dells.
> > > 
> > > The reason I didn't commit my patch yesterday is because I wanted to
> > > make the sanitisation look exactly like the one in the code path that
> > > troubled 36611 and Manuel, but I couldn't find it!
> > > 
> > > Manuel, was it ever committed?
> > 
> > It was: revision 1.47 of sys/arch/x86/x86/mpacpi.c
> 
> Ah, right, in the end we only checked fot device status, not the
> validity of the device number.

... and I remember having seen 1.47, so I already had that code in the
kernel which panicked in 36611 (1.47 is from April, panic was in July)
so there must be yet something else. Shame I no longer have that computer.

Cheers,

Patrick