On Wed, May 25, 2022 at 09:46:24PM +0200, Javier Martinez Canillas wrote: > On 3/10/22 14:11, Dominik Kierner wrote: ... > > # DRM Mode Configuration via Device Tree > > > > In the old fbdev driver, the display modes are hard-coded, which means > > for every new display configuration, a new patch needs to be mainlined, > > which slows down official Kernel support and > > puts burden on the maintainers. > > Additionally, with the DRM-subsystem supporting height and length > > information, for scaling, this opens up a lot of new combinations. > > The SSD1306 for example, is available in multiple resolutions like > > 128x64 and 96x16 and comes in different sizes per resolution as well. > > Just to name a few: > > * 128x64 0.96" (22x11mm) > > * 128x64 1.3" (30x15mm) > > * 96x16 0.69" (18x3mm) > >> Instead of hard-coding, I would suggest something along the lines of > > of_get_drm_display_mode(). > > The displays won't need to support multiple modes at the same time, > > let alone support for switching between them, > > so the one-time invocation of this expensive function might be worth it. > > maybe a new and simpler function that could be named: > > of_get_drm_display_mode_simple() > > This makes sense to me as well. What about non-OF platforms? Please, do not spread OF-only interfaces, and use fwnode instead. > > Providing a mode could later prove useful for a conversion to > > drm_panel, if that is feasible. > > > > But for a function like this, I have to chicken out. -- With Best Regards, Andy Shevchenko
