On Tue, 04 Oct 2022, Ville Syrjala <ville.syrjala@xxxxxxxxxxxxxxx> wrote: > From: Ville Syrjälä <ville.syrjala@xxxxxxxxxxxxxxx> > > I had to stare at the pps stuff again while dealing with the -stable > pps regression. While doing that I figured I'd try to make it so we > don't accidentally use the wrong pps during the panle probe when we > don't yet have the VBT parsed. So here's an attempt at deducing the > correct pps from the hardware state. The approach makes sense. Ack. s8 and /* FIXME not the best place */ connector->panel.vbt.backlight.controller = -1; not the prettiest things though. BR, Jani. > > Cc: Animesh Manna <animesh.manna@xxxxxxxxx> > > Ville Syrjälä (2): > drm/i915: Generalize the PPS vlv_pipe_check() stuff > drm/i915: Try to use the correct power sequencer initially on bxt/glk > > .../gpu/drm/i915/display/intel_connector.c | 3 + > .../drm/i915/display/intel_display_types.h | 8 +- > drivers/gpu/drm/i915/display/intel_pps.c | 105 ++++++++++++++---- > 3 files changed, 93 insertions(+), 23 deletions(-) -- Jani Nikula, Intel Open Source Graphics Center
