On Wed, 27 Mar 2013 01:13:48 +0100 Daniel Vetter <daniel at ffwll.ch> wrote: > On Tue, Mar 26, 2013 at 04:33:07PM -0700, Jesse Barnes wrote: > > Read the current hardware state to retrieve the active mode and populate > > our CRTC config if that mode matches our presumptions. > > > > v2: check that get_hw_state gave us a valid pipe (Imre) > > add clock_get for ILK+ (Jesse) > > > > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk> > > Please preheat your wrath-dispenser ... > > Atm the mode retrieval logic is smashed into setup_hw_state. Imo this > needs to be part of the general hw state readout, and for paranoia needs > to be of the usual cross-checking after each modeset. I was thinking about this last night too; I don't like reading the state in the fb layer either, it really belongs in intel_display somewhere. > Some later patches from my pipe_config series (after the pieces just > resend) add some basic infrastructure for this, including lax matching > ruels (e.g. for the clock cross-checking after a modeset, since we don't > yet put the _real_ hw dotclock into adjusted_mode->clock). I'll check it out. Thanks, -- Jesse Barnes, Intel Open Source Technology Center
