[Intel-gfx] [PATCH 2/7] drm/i915/fbc: replace a loop with drm_atomic_get_existing_crtc_state()

[Ville Syrjälä]

On Fri, Nov 11, 2016 at 05:46:16PM +0000, Chris Wilson wrote:
> On Fri, Nov 11, 2016 at 02:57:36PM -0200, Paulo Zanoni wrote:
> > Much simpler. Thanks to Ville for pointing this.
> > 
> > Cc: Ville Syrjälä <ville.syrjala at linux.intel.com>
> > Reported-by: Ville Syrjälä <ville.syrjala at linux.intel.com>
> > Signed-off-by: Paulo Zanoni <paulo.r.zanoni at intel.com>
> > ---
> >  drivers/gpu/drm/i915/intel_fbc.c | 17 +++++++----------
> >  1 file changed, 7 insertions(+), 10 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/intel_fbc.c b/drivers/gpu/drm/i915/intel_fbc.c
> > index ded77bd..b53b884 100644
> > --- a/drivers/gpu/drm/i915/intel_fbc.c
> > +++ b/drivers/gpu/drm/i915/intel_fbc.c
> > @@ -1071,7 +1071,7 @@ void intel_fbc_choose_crtc(struct drm_i915_private *dev_priv,
> >  	struct drm_plane *plane;
> >  	struct drm_plane_state *plane_state;
> >  	bool fbc_crtc_present = false;
> > -	int i, j;
> > +	int i;
> >  
> >  	mutex_lock(&fbc->lock);
> >  
> > @@ -1092,6 +1092,7 @@ void intel_fbc_choose_crtc(struct drm_i915_private *dev_priv,
> >  	for_each_plane_in_state(state, plane, plane_state, i) {
> >  		struct intel_plane_state *intel_plane_state =
> >  			to_intel_plane_state(plane_state);
> > +		struct intel_crtc_state *intel_crtc_state;
> >  
> >  		if (!intel_plane_state->base.visible)
> >  			continue;
> > @@ -1099,16 +1100,12 @@ void intel_fbc_choose_crtc(struct drm_i915_private *dev_priv,
> >  		if (!intel_fbc_can_choose(to_intel_crtc(plane_state->crtc)))
> >  			continue;
> >  
> > -		for_each_crtc_in_state(state, crtc, crtc_state, j) {
> > -			struct intel_crtc_state *intel_crtc_state =
> > -				to_intel_crtc_state(crtc_state);
> > -
> > -			if (plane_state->crtc != crtc)
> > -				continue;
> > +		intel_crtc_state = to_intel_crtc_state(
> > +			drm_atomic_get_existing_crtc_state(state,
> > +							   plane_state->crtc));
> 
> My knowledge of atomic api is not good enough to say whether this can
> return NULL at this point.

If the plane is part of the state, then plane_state->crtc will be part
of the state as well.

Reviewed-by: Ville Syrjälä <ville.syrjala at linux.intel.com>

Just to reaffirm I wanted to point out what
drm_atomic_set_crtc_for_plane() does, but looks like that one was
written with the assumption that it could fail, which is not correct
since drm_atomic_get_plane_state() will always add the current crtc
to the state, and drm_atomic_set_crtc_for_plane() will always add the
new crtc to the state.

drm_atomic_set_crtc_for_connector() is written in the way I was
expecting, so I think I'll fire off a patch for
drm_atomic_set_crtc_for_plane() to follow the same pattern.

-- 
Ville Syrjälä
Intel OTC