Re: [coreboot] [poll] device_t

[Stefan Reinauer]

* ron minnich  [150507 21:35]:
> one counter-question: is romcc ever going away, or at least is the usage
> gong to change such that no code would ever use the uint32 version of
> device_t? 
>
>
> If it's *never* going away then I think the change makes no sense. If romcc is
> going away, then I would favor the change.

With our current bootblock concept, it is never going away on x86 (for
bootblock usage)

Stefan

[poll] device_t

[Patrick Georgi]

2015-05-07 22:00 GMT+02:00 Stefan Reinauer :
> With our current bootblock concept, it is never going away on x86 (for
> bootblock usage)

Which isn't that much of a problem once we provide separate headers
for x86 bootblock code. There's really very tiny overlap.
That could then be reused to deal with raminit on romcc-boards, too:
from coreboot's point of view, raminit is just an overly large
piece of cache-as-ram code, followed by a raminit noop.
This is simplified by the lack of the need for development tools (eg
printk) to develop new non-car x86 raminits.

One of these days...
Patrick