Le mar. 6 avr. 2021 à 20:03, Heinrich Schuchardt <xypron.glpk@gmx.de> a écrit :
On 4/6/21 4:34 PM, François Ozog wrote:
> Hi
>
> following a mail discussion on trusted-firmware A we will have a discussion
> on HOBs during this week's Trusted Substrate
> <http://trusted-substrate.org>architecture
> council call <https://linaro-org.zoom.us/j/94563644892>.
>
> In addition to mail thread, here is a deck
> <https://docs.google.com/presentation/d/1KyqkyqoGXHUynZSI2hRFBoO_P7sFwgUJWhjW-Vm2SNk/edit?usp=sharing>
> I'll use to start things up.
>
> Cheers
>
> FF
>

Dear François,

as long as TF-A uses HOB structures only internally I am not worried.

But when it comes to the interface to other parts of the firmware we
should avoid creating multiple parallel interfaces.

The interface between secure firmware and U-Boot is not ARM specific but
also exists on RISC-V. Both on ARM and RISC-V we currently use
device-trees to transfer system information.
as far as I can see this is a possibility but not a generalized method across all platforms. I wish we could find something standard.
I am missing this on your
slide 2 "Overall picture for the discussion".

I cannot see anything stopping TF-A from packaging the information that
you want to transfer into the device-tree instead of a
EFI_HOB_RESOURCE_DESCRIPTOR.
That is certainly a good method. That DT can even be embedded into a GUID hob.

 From the U-Boot side I consider this as the preferred solution.
That is a very valid one. But need consensus. Also the scope and bindings need to be clearly defined 


Best regards

Heinrich
_______________________________________________
boot-architecture mailing list
boot-architecture@lists.linaro.org
https://lists.linaro.org/mailman/listinfo/boot-architecture
--
François-Frédéric Ozog | Director Linaro Edge & Fog Computing Group
T: +33.67221.6485
francois.ozog@linaro.org | Skype: ffozog