Re: [PATCH 01/10] dt-bindings: display: renesas: Add R-Car LVDS encoder DT bindings

[Laurent Pinchart <laurent.pinchart@ideasonboard.com>]

Hi Geert,

On Monday, 15 January 2018 10:05:59 EET Geert Uytterhoeven wrote:
> On Mon, Jan 15, 2018 at 7:59 AM, Laurent Pinchart wrote:
> > On Monday, 15 January 2018 08:55:29 EET Simon Horman wrote:
> >> On Fri, Jan 12, 2018 at 03:29:48PM +0200, Laurent Pinchart wrote:
> >>> On Friday, 12 January 2018 12:13:18 EET Geert Uytterhoeven wrote:
> >>>> As this is a new binding, please use "renesas,<soc>-lvds".
> >>> 
> >>> I've recently been thinking that we made the wrong choice, <ip>-<soc>
> >>> would be better in my opinion as it aligns with <ip>-<version>, but
> >>> it's too late to change that, so I'll change the order here.
> >> 
> >> My recollection is that in the beginning we had a bit of a mixture but
> >> leaned towards <ip>-<soc>, which made sense in my opinion. However, after
> >> some discussion it was agreed that the best-practice for upstream was to
> >> use <soc>-<ip>. Unless that situation has changed lets stock with using
> >> <soc>-<ip> for new bindings.
> > 
> > Sure, that was my plan, and it seems I failed to explain it clearly. I too
> > believe that <ip>-<soc> would be better, but as we have standardized on
> > <soc>- <ip> and as there's no strong reason to reconsider that decision
> > at the moment, the next version of this patch will use <soc>-<ip>. It was
> > a mistake in v1, not an attempt to change what we had agreed on.
> 
> Note that I believe you have to consider the full tuple
> "<vendor>,<soc>-<ip>" to see the light: <soc> is more closely tied to
> <vendor>, than <ip> is.

I suppose there are pros and cons for both options :-) I see <soc> more as a 
version qualifier.

-- 
Regards,

Laurent Pinchart