On Mon, 15 May 2000, Michael 'Maki' Kato wrote:

> Wayne,
> 
>   Count me in.  If noting else, I'd like to test, and if possible I'd like to 
> work on some of the drivers etc.
>   I have a Bull Prostation DPX M-20 which is I think a Rx3230 w/ 64Mb of ram 
> and color framebuffer.

Thanks for showing the interest!

I am aware there are a few compatable machines out there.  Hopefully your
Bull machine is one of them.   Once I have a kernel probing hardware it
would be good for you to run it to see if it is identical to the 3230.

>   I'm currently setting up a cross compile environment with NetBSD/i386.  
> With this I'm hoping to build NetBSD/arc for my Olivetti, and then onto 
> NetBSD/mips.

Excellent starting point.  I'm also disadvantaged by not currently
having a native machine running NetBSD in big-endian mode.   My
development is being run for NetBSD/alpha and NetBSD/sparc machines.

>   What is the reason for booting elf instead of using elf2ecoff and booting 
> ecoff.  (debugging maybe?)

Booting ELF is a long way down the track.   The standalone loader was
attacked first to ensure the PROM Firmware routines were usable and it was
a smaller target to build and test with.   It was enough to confirm the
environment was usable.  I was hoping the standalone loader would give me
a remote debugger (which it didn't), so I went onto building ecoff kernels
directly. 

Last night I discovered that the static data from the .data segment isn't
setup (or loaded?) correctly.   The data is there when viewed with
objdump on the source machine, but not in memory when the image is loaded 
via bootp & tftp.
I could not get elf2ecoff running under NetBSD/alpha, so I am currently
using objcopy to generate an ecoff-bigmips binary.   

Does anyone have any comments about the differences (or problems) with the
following techniques to get ECOFF binaries:
  -  linking ELF32 then using elf2ecoff
  -  linking ELF32 then using objcopy
  -  linking ECOFF by setting the target directly in the linker script

Wayne
--
  _____	   	Wayne Knowles,  Systems Manager
 / o   \/   	National Institute of Water & Atmospheric Research Ltd
 \/  v /\   	P.O. Box 14-901 Kilbirnie, Wellington, NEW ZEALAND
  `---'     	Email:   w.knowles@niwa.cri.nz