[U-Boot] [ARM] RFC: Add board support for Colorado Engineering TK1-SOM

[Stephen Warren]

On 08/23/2016 04:47 PM, Peter Chubb wrote:
>>>>>> "Stephen" == Stephen Warren <swarren at wwwdotorg.org> writes:
>
> Stephen> On 08/22/2016 04:29 PM, Peter Chubb wrote:

>>> diff --git a/board/nvidia/tk1-som/pinmux-config-tk1-som.h
>>> b/board/nvidia/tk1-som/pinmux-config-tk1-som.h
>
>>> +/* + * Copyright (c) 2016, NVIDIA CORPORATION. All rights
>>> reserved.  + * + * SPDX-License-Identifier: GPL-2.0+ + */ + +/* + *
>>> THIS FILE IS AUTO-GENERATED - DO NOT EDIT!  + * + * To generate
>>> this file, use the tegra-pinmux-scripts tool available from + *
>>> https://github.com/NVIDIA/tegra-pinmux-scripts + * Run
>>> "board-to-uboot.py tk1-som".  + */
>
> Stephen> Did you generate this yourself, or extract it from some CEI
> Stephen> U-Boot source tree? I ask because the file is obviously based
> Stephen> on an old version of the pinmux spreadsheet and pinmux
> Stephen> scripts (e.g. the unsupported OWR pin is configured, there's
> Stephen> no tk1_som_mipipadctrlgrps variable, etc. If you generated
> Stephen> this yourself, if you could update it based on the latest
> Stephen> spreadsheet that'd be great. If you're just taking this from
> Stephen> some CEI U-Boot source tree, then the commit description
> Stephen> should mention this, and you likely need to preserve the
> Stephen> original CEI authorship and Signed-off-by lines.
>
> I generated this myself from a spreadsheet provided by Colorado
> Engineering and the current tegra-pinmux-scripts.   Is there any
> easy way to see the differences in versions between the one this was
> based on (v09 of the upstream NVIDIA sheet) and the current v11 one to
> be able to transform the spreadsheet?

Probably the easiest way is to save the v9 and v11 versions of the 
Jetson TK1 spreadsheet as CSV, and then diff that. There are tons of 
columns in the spreadsheet that aren't terribly relevant though, so the 
lines will be long and hard to interpret. It might be better to go 
through the whole process of importing v9 and v11 into 
tegra-pinmux-scripts, which will extract just the useful information, 
and then diff the data files there. That should give a list of changes 
you can use to bring the v9-based TK1 SOM spreadsheet up to match the 
v11 spreadsheet.

BTW, I meant to mention: can you also send a patch to 
tegra-pinmux-scripts to add support for this board there too?

>>> +CONFIG_USE_PRIVATE_LIBGCC=y +CONFIG_CPU_V7_HAS_NONSEC=y
>>> +CONFIG_CPU_V7_HAS_VIRT=y +CONFIG_ARMV7_NONSEC=y
>>> +CONFIG_ARMV7_VIRT=y +CONFIG_SUPPORT_SPL=y +CONFIG_SPL=y
>>> +CONFIG_SPL_BUILD=y +CONFIG_ARMV7_PSCI=y
>
> Stephen> None of those are in jetson-tk1_defconfig. Was this addition
> Stephen> deliberate?
>
> Yes, except the USE_PRIVATE_LIBGCC which I think is an error.  The
> main use-case I have for the board is for virtualisation; enabling
> booting in nonsec mode is essential for this.  And it's harmless, as
> the feature has to be enabled via an environment variable.

Yes, looks like that's in the original Jetson TK1 spreadsheet too. I 
assume the mistake you're referring to is that I mentioned that config 
option in my mail as being added by you? Having that option enabled is 
necessary for the U-Boot SPL to work, unless your toolchain's libc is 
built for a very old instruction set, and we don't want to require 
people to use only such toolchains.