[U-Boot-Users] pci.c in MPC8349ADS

[U-Boot-Users] pci.c in MPC8349ADS

[David Updegraff]

Hi.

I wonder why PCI support was added to the MPC83xx in board/mpc8349ads as 
a board-specific thing, instead of a cpu-architecture-general thing in 
cpu/mpc83xx ?

Aside from the I2C setup of the IO-expander, it seems that the remainder 
of PCI initialization is 83xx generic..

-dbu.

[U-Boot-Users] pci.c in MPC8349ADS

[Kumar Gala]

On Nov 15, 2005, at 9:22 AM, David Updegraff wrote:

> Hi.
>
> I wonder why PCI support was added to the MPC83xx in board/ 
> mpc8349ads as a board-specific thing, instead of a cpu-architecture- 
> general thing in cpu/mpc83xx ?
>
> Aside from the I2C setup of the IO-expander, it seems that the  
> remainder of PCI initialization is 83xx generic..
>
> -dbu.

Eventually it should be.  However someone needs to get full support  
working for PCI2 and such before we go down that path and looking at  
what should be done in the board code and what should be done in cpu/ 
mpc83xx.

- kumar

[U-Boot-Users] pci.c in MPC8349ADS

[Wolfgang Denk]

In message <4379FD30.9040105@cray.com> you wrote:
> 
> I wonder why PCI support was added to the MPC83xx in board/mpc8349ads as 
> a board-specific thing, instead of a cpu-architecture-general thing in 
> cpu/mpc83xx ?

At the time, when we added PCI support for  the  TQM834x  boards,  it
looked  as  if  it was board specific. When new 83xx boards get added
and it turns out to be more generic then the code should be moved  to
cpu/mpc83xx;  the  question is how many #ifdef's will be needed - the
TQM834x will probably require some.

> Aside from the I2C setup of the IO-expander, it seems that the remainder 
> of PCI initialization is 83xx generic..

Yes, in theory. When you have to deal with real  hardware  (and  it's
buggy early incarnations) things may look different.

Best regards,

Wolfgang Denk

-- 
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Life would be so much easier if we could  just  look  at  the  source
code.                                                   -- Dave Olson