NetBSD-Bugs archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: kern/45915: panic: pmap_enter_ptpage: can't get KPT page

The following reply was made to PR kern/45915; it has been noted by GNATS.

From: "Michael L. Hitch" <>
Subject: Re: kern/45915: panic: pmap_enter_ptpage: can't get KPT page
Date: Sat, 4 Feb 2012 11:19:23 -0700 (MST)

 On Sat, 4 Feb 2012, Izumi Tsutsui wrote:
 > >  panic(30a22f,0,2704,fec45000,ffffd000) + c
 > >  trap(3e7dc8,8,505,4) + 134
 > >  pmap_enter_ptpage(3386d0,1c48000,0,1c48000,1c48000) + 488
 > >  pmap_kenter_pa(1c48000,ffffa000,3,0,1447d50) + 50
 > >  uvm_km_alloc(342fb8,1f000,0,11,c) + 11a
 > >  pmap_init(1448000,ff800000,3e7f74,3e7f70,353080) + 9e
 > >  uvm_init(c,ffffffff,ffffffff,ffeffffc,fe803882) + 5e
 > This means:
 > - m68k pmap_init() calls uvm_km_alloc() (even before kmem changes)
 > - it seems now uvm_km_alloc() depends on structures which are initialized
 >   in pmap_init() and uvm_km_alloc() is called before that initialization
 > Previously it worked (for more than a decade), but I'm not sure if
 > it was a vaild op or not.
    I was/is a valid operation.  The m68k ports intialize the mmu tables 
 with enough space so memory can be allocated prior to the pmap 
 intialization.  What's happening now is that the kmem changes will 
 allocation substantially more memory before pmap_init() is called, and the 
 initial mmu setup is no longer sufficient.  On my 128MB amiga, the 8 
 kernel PT pages was not sufficient, but increasing that to 12 was.  The 
 kmem changes allocates space base on the physical memory size, and Michael 
 van Elst had some interesting allocations with his 2GB system.
    I suspect that most (or all) m68k ports are going to need to increase 
 the intitial kernel PT page allocations during setup.
 Michael L. Hitch             
 Computer Consultant
 Information Technology Center
 Montana State University       Bozeman, MT     USA

Home | Main Index | Thread Index | Old Index