On Sat, May 18, 2013 at 2:58 PM, Jean-Francois Moine <moinejf@xxxxxxx> wrote: > On Sat, 18 May 2013 14:23:19 -0400 > Rob Clark <robdclark@xxxxxxxxx> wrote: > >> > These parameters should not be there. It seems to me that the DT is the >> > right place. >> >> You might not want to directly have a hard DT dependency in tda998x, >> as the encoder could be used on non-DT platforms. Although a DT to >> encoder-params helper might be a nice idea for platforms which do have >> DT. > > If I correctly understand: > > - Russell does not use any DT, so his drm driver should be declared in > some cubox-setup code in mach-dove/ > > - this code should also declare the tda998x > > - the drm driver contains/passes parameters to the tda998x > > As the connection Dove LCD <-> tda998x is Cubox specific, the question > is: why are'nt the tda998x parameters in the cubox-setup code? ok, maybe I am misunderstanding you. I think the parameters should be filled in by the board file on a non-DT setup. But the part in drivers/gpu/drm/i2c should not pull them directly out of DT, or should have an arrangement like #ifdef CONFIG_OF .. pull params out of DT .. #else .. use params passed in from via params struct, which is populated in board file .. #endif to accommodate non-DT builds. (Although I think just having a helper to populate 'struct tda998x_encoder_params' from DT seems cleaner.) BR, -R > -- > Ken ar c'hentañ | ** Breizh ha Linux atav! ** > Jef | http://moinejf.free.fr/ _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/dri-devel
