[Arm-netbook] firefly 3399 all source software disclosed?

[Luke Kenneth Casson Leighton]

---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68


On Mon, May 29, 2017 at 10:12 PM, David Niklas <doark at mail.com> wrote:
> On Mon, 8 May 2017 05:45:36 +0100
> Luke Kenneth Casson Leighton <lkcl at lkcl.net> wrote:
>> ---
>> crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
>>
>>
>> On Sun, May 7, 2017 at 10:10 PM, zap <zapper at openmailbox.org> wrote:
>> >
>> >
>> > On 05/07/2017 04:29 PM, ronwirring at Safe-mail.net wrote:
>> >> All software for the mali-t860 is open source?
>>
>>  none.  MALI is proprietary.
>
> I'm confused.
> Luke, if you plan on making an RK3399 into an eoma project how can you
> get RYF status if the mali GPU is closed source?

 this was discussed a year ago or so.  same process as for the EOMA68-A20

> For that matter, how can you get RYF cert. for your current eoma68
> project?

 by leaving out the proprietary crap, simple as that.  see below.

> Unless I'm mistaken and it uses a different GPU?

 it is actually a different GPU but that does not change the
assessment process carried out by the FSF.

> Or you just leave the HW crippled?

 if the FSF considered the device to be "crippled" by it not having
the 3D engine running, such that there was a genuine risk that people
would actively seek out the installation of proprietary software.

 in the case of e.g. a proprietary on-board WIFI device that *would*
constitute a genuine risk of people *actively* seeking out proprietary
firmware, and consequently the FSF quite naturally refuses to certify
devices that contain non-removable proprietary on-board WIFI chips.

 however in this case it actually turns out that if you use the
proprietary 3D GPU for the tasks that i suspect you *believe* will
quotes accelerate quotes certain operations (such as X11), the MALI
embedded GPU (or its associated proprietary software - we can't
actually tell which because we DON'T HAVE THE DAMN SOURCE) is so
piss-poor at its job that it actually SLOWS DOWN CERTAIN OPERATIONS of
X11.

given that 2D acceleration is already covered by fbturbo, and works
really well *and is entirely libre software*, the *need* for the 3D
engine just for basic Small-Office / Home-Office and day-to-day usage
is NOT A CONCERN.

 so does that make it clear that the evaluation process (which was
described a year ago) is not just a hard-and-fixed process?

now, if on the other hand this was a dedicated Games Console product,
*that would be an entirely different matter*.  applying for RYF
Certification on a 3D Games Console product which has a 3D GPU which
*only works with proprietary software* would probably constitute too
much of a risk that buyers *WOULD* in fact go out of their way to
download the proprietary drivers.

but this design *isn't being sold as a 3D Games Console*, is it?

does that help clarify?

l.