preparing pcf50633 for upstream

Andy Green andy at
Fri Oct 3 00:02:21 CEST 2008

Hash: SHA1

Mike (mwester) wrote:

>> Sure, ideally the OM distro should be the reference implementation for
>> these and it wouldn't hurt to document how they work.
> Examples are excellent, but asking someone to decipher the enormous body
> of code in the very dynamic Om distro in to figure something out is not
> the complete solution -- they'll probably not find it unless they
> already know what they're looking for.

I did say, "...and it wouldn't hurt to document it".  Kernel side of
some of it is documented in the page I started for the /sys stuff

...otherwise most of the other points are out of scope for me.

> It would also be nice to outline why that was chosen, to preempt
> re-discussion of the same things that might otherwise occur.  Sure,
> someone would certainly be able to search the mailing list for
> historical discussions, but again, they're not likely to find the right
> topic unless they already know what to look for.
> Despite our protests to the contrary, source code != documentation!

AIUI Mickey Lauer has his eye on defining a bunch of this and related
stuff in a more formal way, I think it can answer this reasonable
request for some docs.

> I have nothing against the use of the defconfigs in the git repo, except:
> $ grep APM_POWER def*
> defconfig-2.6.24:CONFIG_APM_POWER=y
> defconfig-2.6.24-maxmodules:# CONFIG_APM_POWER is not set
> defconfig-gta01:# CONFIG_APM_POWER is not set
> defconfig-gta02:CONFIG_APM_POWER=y
> $
> So which is the right way?    :-(

We seem to be about to throw out APM Emulation, so I guess not having it
will be the right way shortly and solve this inconsistency.  I also
imagine GTA01 getting along fine without it until now means it ain't
that needed...

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


More information about the openmoko-kernel mailing list