On Mon, Aug 16, 2021 at 01:09:27PM +0300, Jani Nikula wrote: > On Wed, 21 Jul 2021, José Roberto de Souza <jose.so...@intel.com> wrote: > > VBT has support for up two integrated panels but i915 only supports one. > > > > So here stating to add the basic support for two integrated panels > > and moving the DRRS to ddi_vbt_port_info instead of keeping a global > > one. > > Other VBT blocks will be converted in following patches. > > > > While at is also nucking lvds_dither as it is not used. > > If you were to get a mysterious bisect result on this patch, would you > know what's wrong? > > I think there's too much going on in one patch, and this needs to be > split. Small, incremental changes with functional and non-functional > separated if possible. This is incredibly fragile stuff. > > Also, I think we'll need to be able to figure out the panel type based > on EDID, i.e. we need to parse some stuff first, and the panel stuff > only after we've read EDID. Obviously you don't have to do that here, > but this need to be done so that this doesn't become impossible.
My incredibly incomplete start at that: git://github.com/vsyrjala/linux.git vbt_panel_type_pnp_id Totally missing any and all restructuring of the VBT parsing to happen after we have the EDID. And apart from that we also have this horrible chicken vs. egg issue: 1. need the panel type to get power sequencer delays 2. need to turn on the VDD to read the EDID 3. need the EDID to get the panel type -- Ville Syrjälä Intel
