On Mon, Feb 08, 2016 at 03:58:18PM +0000, Andre Przywara wrote:
> Hi,
> 
> On 08/02/16 15:54, Rob Herring wrote:
> > On Thu, Feb 04, 2016 at 05:51:51PM +0100, Maxime Ripard wrote:
> >> Hi Andre,
> >>
> >> On Tue, Feb 02, 2016 at 04:53:58PM +0000, Andre Przywara wrote:
> >>>> So, droping it in the filenames, why not. But I'd really like to keep
> >>>> the same compatible scheme.
> >>>
> >>> And I still don't get this: in the DT compatible scheme we always have a
> >>> vendor prefix, so allwinner,a64 is surely not a mysterious ARM Ltd. core
> >>> or a new Apple SoC. Instead it is the A64 from Allwinner, full stop. So
> >>> why should we add an arbitrary and confusing sun50i naming to it (when
> >>> it actually should be more like "sun8i-a64").
> >>
> >> I don't decide on their marketing names. And I know you want to start
> >> anew with the arm64 SoCs, but the truth is, you don't. Most of the
> >> compatibles in the DTSI are from earlier SoCs, and we have to keep
> >> that legacy and remain consistent with it. With all the good and bad
> >> things a legacy imply.
> > 
> > I have to agree. Unless there is some agreement to move to another 
> > naming scheme, then just follow the same pattern. If sunXi is just a 
> > made up name outside of Allwinner to provide some logical grouping of 
> > SoCs, then yes, that probably should not have been done.
> 
> So I still don't like it, but will not waste my time or energy on that
> front.
> 
> Maxime, do you want "allwinner,sun50i-a64" or would
> "allwinner,sunxi-a64" be OK as well?

The former will be fine.

Thanks!
Maxime

-- 
Maxime Ripard, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com