[maemo-developers] [maemo-developers] OS 2007 / 770 hacker edition

[From: Frantisek Dufka]

Hello Carlos and Markku

this is pretty cool :-)

> 
> Have a look at Markku’s notes (http://maemo.org/maemowiki/Os2007On770) 
> if you are interested in the gory details. He’ll be updating these notes 
> as he continues to work on this.

Quite interesting reading. Found some of those bits about initfs and 
kernel too.

> 
> An image will be available soon for download and flashing, to 770 users 
> only.

I guess including some sources - kernel at least (I am interested and 
maybe could even help) or all modified packages.


> 
> After that, Maemo developers with a 770 and a bit of time on their hands 
> are encouraged to check it out.

I will and definitely try to help if possible. Few questions/ideas:

Did you try to leave old kernel from 2.2006 and backport changes needed 
(if any) from rx-34 kernel? Seems easier than using rx-34 kernel 
directly (arch/arm/omap1/ is outdated). What newer kernel gives us?

Same for initfs, what does break, dsme,bme? Thanks for the toolchain 
name, I had to build static version of evkey for bootmenu since I 
couldn't find correct uclibc toolchain. there is 'We used rx-34-initfs 
from apt-repository as a base' where is this, something internal?


> We will work on this still for some time though it is hard to say how 
> long. It will also depend on how this is received. It might be that we 
> can get more of the DSP and Multimedia to work, or maybe not.

Could leaving old kernel, dspgateway, dsp_dld and all n770 codecs help 
with this? Updated gstreamer could be nice, though. Maybe some fixes in 
newer gst dsp plugins to make it work with older DSP stuff can make this 
possible?

> 
> We can’t justify doing much work on this hybrid, our development focus 
> is really on the N800. However, if we manage to make further progress we 
> will release updated images. We’ll look at any patches that people in 
> the Maemo community might turn up and consider applying them and 
> releasing new images. We can’t promise we will do a good job there but 
> let’s see what we can manage.

Well I think community can (in theory) keep working on this and continue 
  as long as is needed. What is not clear is how far we can go without 
those few but important closed bits (initfs - dsme/bme, config partition 
format, cal-tool etc., dsp stuff) and implications they have on the rest.

> 
> This is again, something new for us, and we hope to learn something in 
> the process.

Yes, at least it can be another push to opensource something needed for 
this or reveal how far we can go without it. I am pleasantly surprised 
by the progress so far.

Best regards,
Frantisek