[coreboot] first output: dbe62 and v3
Carl-Daniel Hailfinger
c-d.hailfinger.devel.2006 at gmx.net
Wed Mar 5 17:06:58 CET 2008
On 05.03.2008 16:58, ron minnich wrote:
> On Wed, Mar 5, 2008 at 5:49 AM, Carl-Daniel Hailfinger
> <c-d.hailfinger.devel.2006 at gmx.net> wrote:
>
>> That depends on your goals. We once discussed how to fail over to a
>> initram/stage2/payload downloaded via serial. That can take a lot of
>> time. I seem to remember that some power button control on some Geode
>> boards has to be done in the first 4 seconds after poweron. That timing
>> constraint will bite us when downloading over serial. Of course, if I'm
>> wrong (or we can perform failsafe powerbutton control in stage 1),
>> discard my objection.
>>
>
> right, but the enable ENABLES the timeout. Safest thing to do is wait
> until stage 2 at least to enable such things. I.e. later.
>
OK, then I'm totally fine doing all the powerbutton handling in stage2.
> Although I see your point now. Sigh. mainboard_pre_payload?
>
I'd simply decide on a phase in stage2 where all this happens (except
time-critical stuff).
> We're finding that the device tree needs to span stage1 to
> mainboard_pre_payload, now. We had not planned for this. We're going
> to have to rethink it a bit.
>
See my other mail from 5 minutes ago for a possible solution for that
problem.
> If we do this it will clean a lot of things up.
>
And the bootblock will grow a lot. I'd still like to avoid device tree
handling in stage 1 and initram.
> Good work for denver.
>
Speaking of Denver, how many people have confirmed to come?
Regards,
Carl-Daniel
--
http://www.hailfinger.org/
More information about the coreboot
mailing list