Re: Exynos?

[Joachim Thiemann <joachim.thiemann%gmail.com@localhost>]

On 7 April 2014 14:15, Ottavio Caruso 
<ottavio2006-usenet2012%yahoo.com@localhost> wrote:
> There's no way to '"legally" boot any other kernel than the Chrome OS
> kernel on that Chromebook. I used to have it. If you look on the
> Armedslack mailing list archives you'll find some notes of mine.
>
> You might hack the firmware and unlock it, but I am not sure if the
> Netbsd kernel would compile on it and which open source drivers are
> available in the wild.
>
> I have no idea about other Exynos products. It's a great processor and
> it would be my dream to have Netbsd running on it.
>
> --
> Ottavio


Hi Ottavio,

I had a quick look at the armedslack messages - but I'm not sure these
issues still exist (the "legal" issues) (IANAL etc etc).  I have my
Chromebook in developer mode and it quite happily has u-boot loading
an unsigned kernel (possibly through some chainloading thing).  I'm
not sure I grok the full boot sequence of the arch sd card quite yet
(esp. regarding the u-boot script) but I do know there is a partition
formatted as ext2 with only one file (vmlinux.uimg) that (AFAIK) comes
straight from the arch repositories.  My naive hope is that I could
just plop a u-bootable netbsd kernel in there.

As for peripherals, I would expect that talking to the Mali (?) GPU
with full efficiency would be (ahem) unrealistic to expect - but the
Exynos 5250 datasheet was actually surprisingly easy to find, for
whatever that may be useful (the publicly available datasheets can be
little more than glorified sales brochures).  So talking to the EMMC
and to USB peripherals should not be unreasonable, I hope.

My hope was simply that, assuming u-boot provides an appropriate
environment, I could at least have it load a netbsd kernel and spit
out basic messages on the screen - it would at least give me some hope
:-)

Cheers,
Joachim
(And ARM chromebooks can be found surprisingly cheap second hand
because they are not what some people expect they would be)
-- 
Joachim Thiemann :: http://jthiem.bitbucket.org ::
http://signalsprocessed.blogspot.com