[cip-dev] [PATCH 01/14] ARM: dts: r8a7743: initial SoC device tree

[ben.hutchings@codethink.co.uk (Ben Hutchings)]

On Wed, 2017-08-23 at 07:36 +0000, Biju Das wrote:
> 
> 
> > -----Original Message-----
> > From: Ben Hutchings [mailto:ben.hutchings at codethink.co.uk]
> > Sent: 22 August 2017 19:08
> > To: Biju Das <biju.das@bp.renesas.com>
> > Cc: Chris Paterson <Chris.Paterson2@renesas.com>; cip-dev at lists.cip-
> > project.org
> > Subject: Re: [PATCH 01/14] ARM: dts: r8a7743: initial SoC device tree
> >
> > On Tue, 2017-08-08 at 10:50 +0100, Biju Das wrote:
> > > The  initial r8A7743 SoC device tree including CPU0, GIC and timer.
> > >
> > > Signed-off-by: Biju Das <biju.das@bp.renesas.com>
> >
> > I compared this to the upstream version (commit
> > 34e8d993a68ae459ad98c27afc07647e439deacc) and I roughly understand why
> > the clocks are described differently, but can you explain why there's no soc
> > node?
> 
> Thanks for the comments.
> 
> RZ/G1M(r8a7743) and R-Car M2(r8a7791) are almost identical. To make consistant
> with other R-Car Gen2 dtsi patches on 4.4 kernel version, I have dropped the soc node.

But in the upstream device tree sources, r8a7743.dtsi has an soc node
and r8a7791.dtsi does not.  If it's OK for the two chips to be described
differently upstream, I don't see why they should be described
consistently here.

> Do you see any issues with this approach?
[...]

I think that consistency with upstream is more important than
consistency within the platform.  (Of course, the kernel and FDT need to
actually *work*, and if introducing the soc node gets in the way of that
then we shouldn't do it.)

Ben.

-- 
Ben Hutchings
Software Developer, Codethink Ltd.