On Sun, 31 Dec 2006, John Klos wrote:

> In NetBSD 4, in tree for the last few weeks at least up until two days ago's 
> tree, I've been seeing panics:
>
> panic: enter: out of address space
> pmap_enter_ptpage(a88bcf0,e000000)+3d6
> pmap_enter
> uvm_fault_internal
> trap(8,481,e000060)+5d2
> fault stkadj()+0
>
> (copied sloppily by hand via paper, so errors may exist). This is on a system 
> with 132 megs of memory and 384 megs of swap, and it can pretty consistently 
> be reproduced by stopping and restarting milter-greylist.
>
> Anyone have any ideas? Might this be related to the recent pmap changes? I 
> will try to get this machine onto a serial console soon if anyone wants to 
> poke it.

   This is a known limitation of the m68k pmap on 68040 and 68060 
processors.  The user program is attempting to map 256M or more of user 
memory.  Many of the m68k ports using 4K pagesize will have this problem 
as long as they allow 256M or more of user address space.  A workaround
would be to set USRSTACK to a lower limit (I think 0x0e000000), which 
should prevent a user process from hitting this limitation.

--
Michael L. Hitch			mhitch@montana.edu
Computer Consultant
Information Technology Center
Montana State University	Bozeman, MT	USA