next-generation (now) MPU discussion / U-Boot vs

Andy Green andy at
Thu Apr 24 09:55:48 CEST 2008

Hash: SHA1

Somebody in the thread at some point said:

| Now for real: we got a lot of additional sw-fiddling with low level
things on
| the small chip when comparing to the 12,95$ one, and we even aren't
100% sure
| where we will hit the limits the small RAM and missing hw-serial-I/O is
| imposing (are we?).

No we are not sure, but it is in the right ballpark for being doable.

Also, "additional" is not correct, the fiddling is going to get done
somewhere, previously U-Boot soaked up a lot of this crap and that had
downsides for us.

| Andy: can you give to us a short memo summarizing the benefits of a
MPU we
| wouldn't get any other way, briefly explaining background and way it
| work for each of those points? So everybody involved in the decision
on MPU
| is clear on what we are talking about, enabling us to make a more
| decision.

Yeah I will put something together later trying to summarize the
elements of the discussion and contrasting against GTA02 methodology.

It sounds like in this meeting we still first have to decide if we have
an MPU at all before which one :-)  Not having one reduces the cost
impact of the MPU to "optimal levels" :-)  Rather than close my ears to
this I think we should examine the two concepts in parallel and have a
beauty contest tomorrow by studying the ramifications either way.

- -Andy
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora -


More information about the openmoko-kernel mailing list