Re: [PATCH v2 03/19] arm64: renesas: Update 'vsps' property

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

Hi Jacopo,

On Thu, Aug 22, 2019 at 12:00:51PM +0200, Jacopo Mondi wrote:
> On Wed, Aug 21, 2019 at 02:16:02PM +0200, Geert Uytterhoeven wrote:
> > On Sat, Jul 6, 2019 at 4:07 PM Jacopo Mondi <jacopo+renesas@jmondi.org> wrote:
> > > Update the 'vsps' property in the R-Car Gen3 SoC device tree files to
> > > match what's in in the documentation example.
> >
> > double in (no worries, I'll fix that up myself)
> >
> > > Signed-off-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
> >
> > Thanks!
> >
> > Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
> > i.e. will queue in renesas-devel for v5.4.
> >
> > BTW, any plans to add channel indices to the vsps properties in the
> > remaining DTS files?
> 
> According to what I read in the parsing 'vsps' proeprties parsing
> routine rcar_du_vsps_init(), if the channel index is not specified, it
> is defaulted to 0.
> 
>            /* Store the VSP pointer and pipe index in the CRTC. */
>            rcdu->crtcs[i].vsp = &rcdu->vsps[j];
>            rcdu->crtcs[i].vsp_pipe = cells >= 1 ? args.args[0] : 0;
> 
> The DU bindings documentation does not state that the channel index is
> optional, so yes, it might be worth changing this and simplify the
> parsing routing to always assume the channel index is there.
> 
> What does Laurent think? Is this worth the effort?

I think it's worth the effort to keep the upstream DT sources in sync
with the latest bindings, but We can't change the code as we need to
ensure backward compatibility. So, yes to the DT update, but the driver
should stay unmodified (or should receive a new comment explaining the
required backward compatibility if there's not one already).

-- 
Regards,

Laurent Pinchart