Ronald G Minnich wrote:
>
> On Thu, 1 Feb 2001 [EMAIL PROTECTED] wrote:
>
> > - Re-organize hardwaremain.c so that it calls the mainboard_fixup() and
> > mainboard_fixup calls utility functions in hardware main for most of
> > the code that is currently in intel_main(). The current setup is
> > awkward because it has mainboard_fixup() and then
> > final_mainboard_fixup() and yet still isn't very flexible. "Inverting"
> > the function calls like this will prevent endless additions of #ifdef
> > so that it can be customized to the needs of any board.
>
> sounds good, can you send an example.
Will do. I'll leave intel_main() the way it is except for #ifndef'ing
out most of the content based on an option. The call to
mainboard_fixup() will be about the only thing not commented out. This
way it won't break existing code for now. I will then munge a copy of
the content of intel_main() into a set of utility functions that I can
call from mainboard_fixup(). This should obsolete the need for
final_mainboard_fixup().
When this is working and people like it we can phase out support for all
the stuff left in intel_main().
> > - Do the same for linuxbiosmain.c. In particular I can see the need to
> > be able to choose a different kernel and/or different kernel command
> > line based on some tests.
>
> We can try this.
I'll do about the same thing here as for hardwaremain.c. This will also
help support things like loading an initrd image.
> > - Do the fixed mtrr tables in a manner similar to the irq routing tables
> > so that we can have mtrr tables for each board type.
>
> good idea!
Cool. This should be an easy one.
> ron
Cheers!
Ty
--
Tyson D Sawyer iRobot Corporation
Senior Systems Engineer Real World Interface Div.
[EMAIL PROTECTED] Robots for the Real World
603-532-6900 ext 206 http://www.irobot.com
