I was looking at this with someone else a few weeks ago.
we're faulting on a wired pmap entry and the fault handler enters
a non-wired version of the same mapping.  the only way I can think
that this could happen is that the TLB and the PTEs are getting out of sync.

that kind of bug can be very serious in general, but there are
no other symptoms, so if that is the problem it must only happen in
some very limited circumstance.

-Chuck


On Thu, Jun 02, 2005 at 08:50:25PM +0100, Patrick Welche wrote:
> I'm getting these (new i386 box, -current of today)
> 
> pmap_unwire: wiring for pmap 0xccde9190 va 0xbdbc4000 didn't change!
> pmap_unwire: wiring for pmap 0xccde90c8 va 0xbdbc3000 didn't change!
> pmap_unwire: wiring for pmap 0xcdfd6130 va 0xbda85000 didn't change!
> pmap_unwire: wiring for pmap 0xcdfd64b4 va 0xbda85000 didn't change!
> pmap_unwire: wiring for pmap 0xcdfd6c84 va 0xbda85000 didn't change!
> 
> Should I worry?
> 
> Cheers,
> 
> Patrick