[PATCH 0/3] drm/panel: simple: Add mode support to devicetree

Sean Paul seanpaul at chromium.org
Wed Feb 7 15:27:53 UTC 2018


On Wed, Feb 07, 2018 at 09:16:22AM +0000, Eric Anholt wrote:
> Sean Paul <seanpaul at chromium.org> writes:
> 
> > Hey all,
> > Here's a set which allows us to add an "override" mode to the simple
> > panel dt node. The override mode can be used for devices for which the
> > typical display timing is not sufficient, yet the overriding mode should
> > not be applied across the entire platform. 
> >
> > An example of this (and the motivation) is the Chromebook Plus (kevin).
> > If the sharp panel on this laptop is run at the mode advertised in the
> > datasheet (and what is currently in mainline), it creates interference
> > with the touch digitizer. To fix this, we need to run the pixel clock at
> > a slightly higher rate (which we can do by increasing the back porches).
> > This "fix" should not be used on other rockchip devices using this panel
> > since they might not encounter the same interference.
> >
> > If an override mode is present, it will be checked against the panel's
> > display_timing range. When validated, it will be exposed as the
> > preferred mode along with the 'typical' modes generated from the panel's
> > display_timing.
> >
> > This set is based on Linus' master to pick up the edp support in
> > rk3399-gru-kevin.dts.
> 
> Couldn't you just add a different compatible string for the panel
> driver, and use that to have a different mode exposed from the panel?

Yep, there's a couple ways to skin this cat. We could just change the mode to
what the kevin device needs since it's the only one that uses this panel atm
(that's what the original patch in the context link does). We could also expose
multiple modes for the panel and let userspace sort it out.

That said, we already have timing ranges in panel-simple and the goal is to
leverage those such that we don't need additional compatible panels/extra modes.

Sean

-- 
Sean Paul, Software Engineer, Google / Chromium OS


More information about the dri-devel mailing list