Thanks for the response Mike.  I have tried two different approaches
and in each I have come up with errors I can't get around.

First I tried including the PCI bridge core to the base linux design
that Xilinx provides, I realized that it would be better to see how a
PCI bridge is included by the Base System Builder so I made a dummy
system with PCI capability.  Found out I also need a pci_arbiter and
another DCM module.  The number of connections and ports were a bit
overwhelming to do by hand so I abandoned trying to 'upgrade' the base
system provided.

I instead tried to build a system from scratch that had all the parts
the provided reference system had but with a PCI bridge and arbiter.
Went through and made minor adjusments to some parameters in the
software platform settings and the address mappings and built my BSP.
When I try and generate my bitstream, I get an error in PAR:

ERROR:Place:249 - Automatic clock placement failed.  Please attempt to
analyze
   the global clocking required for this design and either lock the
clock
   placement or area locate the logic driven by the clocks so that that
the
   clocks may be placed in such a way that all logic driven by them may
be
   routed.  The main restriction on clock placement is that only one
clock
   output signal for any Primary / Secondary pair of clocks may enter
any
   region.  For further information see the "Global clocks" section in
the V-II
   Hand-Book (Chapter 2: Design Considerations)

This worries me because I didn't make any changes to what the base
system builder generated that should affect placement (or so I think).
What should I look at to fix this sort of error?  I do see that I have
more DCM modules than the reference system (4 vs 2) due to the
inclusion of PCI.  Any advice on where I should look or an explanation
of this error (did some searching on it and no reference to this error
with 6.3i)?

I can provide more detail if it is helpful...

Thanks,
Joey

Joseph wrote:
> After looking more closely at the block diagram of the ML310:
> http://direct.xilinx.com/products/boards/ml310/current/images/ml310_block.jpg
> It looks like what I really want is to add a pci bridge on the opb bus.

Well, the diagram shows that all of the interesting IO
is on the pci bus. Wouldn't make much of a Linux
machine without that bus.

>  Given a functional bridge, it should recognize NIC on the board as a
> PCI device, right?  

If that's already in the BSP, then yes
else you write PPC405 code.

Can anyone verify that I am on the right track?

You can. Try it and see.

        -- Mike Treseler

After looking more closely at the block diagram of the ML310:
http://direct.xilinx.com/products/boards/ml310/current/images/ml310_block.jpg
It looks like what I really want is to add a pci bridge on the opb bus.
 Given a functional bridge, it should recognize NIC on the board as a
PCI device, right?  Can anyone verify that I am on the right track?

Thanks

Is there a simple tutorial or reference design available that
implements the ethernet IP from Xilinx?  I have stepped through the
"Creating a Linux BSP and System Image for the Base Design" tutorial
provided by Xilinx and would now like to incorporate ethernet support
as a step closer to building something closer to the MontaVista Linux
reference system provided with the board.  Any advice or pointers?

Thanks in advance,
Joey