[coreboot] [announce] coreboot for the 21st century

[Aaron Durbin]

On Thu, Mar 27, 2014 at 6:49 AM, Patrick Georgi <patrick at georgi-clan.de> wrote:
> Am 2014-03-27 14:41, schrieb Aaron Durbin:
>
>> As for the Vortex86, I'm sure we can come up with something. Off the
>> top of my head I'd suggest having raminit in bootblock then there
>> isn't a possibility of running into issues.
>
> Long term I'd like to keep it outside the bootblock, but once it runs there
> it's 5 minutes of Makefile changes to move the code to romstage, so yes -
> works for me.

The issues I was referring to is the use of romstage-y += throughout
all the Makefiles. Without clear semantics on what each stage supports
it's very difficult to ensure things don't break for any particular
board/chipset. Either the classes stay the same and we change out the
implementation in the source files by way of macro guards or we draw a
line in the sand and say romstage has ram, e.g. I think having cleaner
sources is the less of 2 evils.

I still think we need to work out clear semantics of the environment
for each stage.

-Aaron