> Well we can move things on RiscPC first and then do the SHARK and CATS 
> afterwards. Everything is already setup to allow for different mappings on 
> each machine. Before doing this we need a complete map of the memory as it is 
> layed out now along with extents for each of the current sections. Then we 
> need to figure out how the current mappings work so that they can be moved.
> Time to figure out what the hell is going on in initarm().
> 
> Mike
> 

If we are moving things around, I'd vote that we move the VM_MAX_USER 
address down to somewhere similar to the value on i386 (if x86 users don't 
need all that extra user VM, I'm sure arm users can manage :-).  Someone 
mentioned that btriscbsd (btnetbsd) couldn't work out the load value from 
the object file, but objdump seems to manage it OK, so why can't btnetbsd?

R.

shark1:gcc [590] $ objdump -h /netbsd 

/netbsd:     file format a.out-arm-netbsd

Sections:
Idx Name          Size      VMA       LMA       File off  Algn
  0 .text         00149fe0  f0000020  f0000020  00000020  2**2
                  CONTENTS, ALLOC, LOAD, CODE
  1 .data         0000e000  f014a000  f014a000  0014a000  2**2
                  CONTENTS, ALLOC, LOAD, DATA
  2 .bss          000288cc  f0158000  f0158000  00000000  2**2
                  ALLOC