[PATCHv4 04/13] drm/shmobile: Restrict plane loops to only operate on legacy planes

[Laurent Pinchart]

Hi Daniel,

On Friday 28 March 2014 18:53:40 Daniel Vetter wrote:
> On Fri, Mar 28, 2014 at 06:52:50PM +0100, Daniel Vetter wrote:
> > On Fri, Mar 28, 2014 at 04:50:13PM +0100, Laurent Pinchart wrote:
> > > On Thursday 27 March 2014 17:44:29 Matt Roper wrote:
> > > > Ensure that existing driver loops over all planes do not change
> > > > behavior when we begin adding new types of planes (primary and cursor)
> > > > to the DRM plane list in future patches.
> > > > 
> > > > Cc: Laurent Pinchart <laurent.pinchart+renesas at ideasonboard.com>
> > > > Signed-off-by: Matt Roper <matthew.d.roper at intel.com>
> > > 
> > > Acked-by: Laurent Pinchart <laurent.pinchart at ideasonboard.com>
> > > 
> > > I have a question though. The patch set introduces three plane types,
> > > OVERLAY, PRIMARY and CURSOR. What should a driver that has no concept
> > > of primary plane do ? Expose all planes as OVERLAY planes only ?
> > 
> > It's a matter of backwards compat with old userspace. primary/cursor are
> > simply ways to tell the drm core which planes to use to forward the legacy
> > cursor crtc and which plane will be used for the framebuffer in setCrtc.
> > 
> > So until we have the new atomic interface ready your driver kinda needs to
> > expose at least a primary plane, otherwise there's no way to even switch
> > on the crtc.
> > 
> > But besides this backwards compat issue there's no difference and you can
> > specify whatever plane you want as primary/cursor (or none if you don't
> > care about old userspace).
> 
> Well the NULL primary plane probably needs a bit of work on top of Matt's
> patch series here ...

It's the NULL primary plane I'm interested about, to allow a "root" plane not 
to span the whole display. I have no urgent need for this though, I was just 
curious to see if that feature was planned.

-- 
Regards,

Laurent Pinchart