The best_encoder field of connector_state wasn't properly set when a
connector was being disabled, leading to an incosistent atomic state.

For now, this doesn't cause anything to blow up, because everywhere
we're using connector_state->best_encoder there is a check for
connector_state->crtc which is properly initialized. I reached the issue
while testing some patches I haven't sent out yet, that remove the usage
of intel_connector->new_encoder from check_digital_port_conflicts(). In
that case, it would be possible to trigger the converted version of the
WARN in that function.
Signed-off-by: NAnder Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
[danvet: Add commit message augmentation Ander supplied.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>