[PATCH 0/3] sunxi: switch three SoC families to OF_UPSTREAM

[Sumit Garg]

On Wed, 22 Jan 2025 at 16:49, Andre Przywara <andre.przywara at arm.com> wrote:
>
> On Wed, 22 Jan 2025 12:03:24 +0530
> Sumit Garg <sumit.garg at linaro.org> wrote:
>
> Hi Sumit,
>
> > On Sun, 19 Jan 2025 at 22:13, Andre Przywara <andre.przywara at arm.com> wrote:
> > >
> > > This series converts boards with the Allwinner F1C100s/F1C200s, A10, A10s
> > > and A13s SoCs over to OF_UPSTREAM. The .dts and .dtsi files were
> > > identical between the kernel and U-Boot repository, so this is a
> > > no-brainer.
> >
> > Glad to see this series especially the diffstat.
>
> Yeah, love that too, though the mailing list server blocked 2/3 and 3/3
> initially because of their size ;-)
>
> > > However there are four boards (tablets, I guess) that do not have a .dts
> >
> > Can you share the board names you refer to here?
>
> Ah, sorry, sure, it's:
> iNet_3F_defconfig       sun4i-a10-inet-3f.dts
> iNet_3W_defconfig       sun4i-a10-inet-3w.dts
> Ampe_A76_defconfig      sun5i-a13-ampe-a76.dts
> iNet_86VS_defconfig     sun5i-a13-inet-86vs.dts
>
> As you can see, they are extremely minimal, and also identical
> between each other, except for the model names. So there is cleanup
> potential regardless, but we have bigger fish to fry.
>
> > > file in the kernel repository. For now I kept them using the "old
> > > style", so they continue working. This requires to keep a copy of the
> > > SoC .dtsi file,
> >
> > Is the SoC .dtsi file present in dts/upstream? If yes then it is
> > better to not maintain a local copy for them as the local .dts board
> > files are able to include SoC .dtsi file from dts/upstream directory
> > too.
>
> Oh, now I feel stupid for having not even tried that!
> That indeed works, so I will remove the .dtsi, but keep the .dts files in,
> at least for now. Jernej has a point that those boards could also be
> removed, but I didn't want to pull the plug too early and unnecessarily.
>
> > > I would be grateful for any advice as how to handle this
> > > situation best. Ideally we would submit the DTs to the kernel, but I
> > > don't know how feasible this is, given their age, the missing device
> > > availability and their basic DT nature.
> >
> > Lets have only local board .dts files for them including other
> > dependencies from dts/upstream tree. This will at least give folks who
> > care about these boards a chance to uprev the corresponding DTs.
>
> Yes, I agree, I didn't know that this was even possible. But works for me,
> certainly.
>
> If people don't mind, I will apply those patches now.

That's fine by me. Feel free to add:

Acked-by: Sumit Garg <sumit.garg at linaro.org>

-Sumit

> In my test builds
> the .dtb files came out identical, so there is really no risk in this
> anyway.
>
> Cheers,
> Andre