[Arm-netbook] device tree not the answer in the ARM world

[Ken Phillis Jr]

On Mon, May 6, 2013 at 1:55 PM, Scott Sullivan <scott at ss.org> wrote:
> On 05/06/2013 02:35 PM, Ken Phillis Jr wrote:
>> On Mon, May 6, 2013 at 1:19 PM, luke.leighton <luke.leighton at gmail.com> wrote:
>>> On Mon, May 6, 2013 at 7:04 PM, Ken Phillis Jr <kphillisjr at gmail.com> wrote:
>>>
>>>> As for already supported platforms on UEFI, I know that the Beagle
>>>> Board [2][3] and the Samsung Origen [4] both offer support for UEFI
>>>> Directly.
>>>
>>>   and the iMX6, which supports reading of UEFI partitions.  that just
>>> leaves every other SoC out there, of which there are hundreds, which
>>> don't support it.  and it's only the boot phase.
>>>
>>>   l.
>>>
>>
>> There is no requirement that every SOC handle UEFI partitions.  The
>> Beagle board uses a TI OMAP3530, and the Origen Board uses a Samsung
>> Exynos 4 Quad. To my knowledge UEFI Partition support is not a feature
>> found on either SoC.Instead what happens is the EFI/UEFI loader takes
>> the place of Daz U-Boot and functions similarly to the DUET package
>> that was created for x86/x86_64 based systems in order to allow these
>> systems to develop EFI/UEFI enabled operating systems.
>
> Correct, there is no requirement (business or technical), so we're left
> with the mess we have. Sure, us techies can figure out, patch, replace,
> or workaround the defective by design products, but that is not a solution.
>
> Even in this case your citing the 'optional' third party firmware as
> 'see it can be done'. That's been known along time now, but until it
> ships in products in mass volume, and is profitable, the burden of
> making this crap bearable will always been on the shoulder of the open
> source community. We will continue to be ever in 'catch up' mode.
>
> It's a burden the distro's are having to take on and we're getting a
> duplication of effort. Just take a look at the supported hardware lists
> for some of the ARM linux distro's [1][2]. This is only a small
> smattering of the available hardware, and it's all hard work. There just
> isn't the man power to go through making 'easy' to use scripts to
> literally high-jack the original boot environments just to use the
> computers as the general purpose devices they are.
>
> [1]: http://archlinuxarm.org/platforms
> [2]: https://fedoraproject.org/wiki/Architectures/ARM
>
> The distro's should not be responsible for this, it should be the OEMs
> just like in the x86 world.
>
>
> Luke has had the right idea for a while now, you have to change the
> system from the top down and develop products bottom up that provide
> incrementally steps towards our goals. Then reinvesting the profits to
> fund further increments.
>


I understand this and ARM is actually recommending that All SoC
implementations support UEFI on ARM64 based platforms in place of Daz
U-Boot. This is in the UEFI Learning Center[1] and can be found in
multiple articles released by Andrew Sloss ( From ARM, Inc) and
articles released by Prax Fang ( from American Megatrends). I believe
this is possibly a Top down approach by ARM since they are at the top
telling all the Implementers at the bottom that they should use UEFI.
Either way, There is already support for Cortex A8 ( and newer)
devices that exist in the TianoCore source repository, and this means
that UEFI as a viable option for the EOMA-68 A10 powered devices. The
only part I can think of that Luke may not like is the fact that
TianoCore is under a BSD Style license, and Daz-Uboot is under GPL.
Either way I doubt this is a problem, because as long as vendors
handle the boot loader properly ( and follow the license agreement)
Things should be good. I know that there is many examples of devices
using U-Boot ( along with linux) and failing to produce a single line
of code to the Open Source community. Anyways UEFI can be used with
some careful planning and documentation to help make things like the
Device Trees work as intended.

[1] http://www.uefi.org/learning_center/

>
> Luke,
>
> Question for you. Assuming down the road after the first (or even
> second) EMOA-68 card is being sold, would you dedicate resources to
> tackling a UEFI boot environment for future cards? The iMX6 care sounds
> like a candidate if it already has it as you say.
>
> --
> Scott Sullivan
>
> _______________________________________________
> arm-netbook mailing list arm-netbook at lists.phcomp.co.uk
> http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
> Send large attachments to arm-netbook at files.phcomp.co.uk