drm/i915: Fix frontbuffer false positve.

This return 0 without setting atomic bits on fb == crtc->cursor->fb
where causing frontbuffer false positives.

According to Daniel:

The original regression seems to have been introduced in the original
check/commit split:

commit 757f9a3
Author: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Date:   Wed Sep 24 14:20:24 2014 -0300

    drm/i915: move check of intel_crtc_cursor_set_obj() out

Which already cause other trouble, resulting in the check getting moved in

commit e391ea8
Author: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Date:   Wed Sep 24 14:20:25 2014 -0300

    drm/i915: Fix not checking cursor and object sizes

The frontbuffer tracking itself only was broken when we shifted it into
the check/commit logic with:

commit 32b7eee
Author: Matt Roper <matthew.d.roper@intel.com>
Date:   Wed Dec 24 07:59:06 2014 -0800

    drm/i915: Refactor work that can sleep out of commit (v7)

v2: When putting more debug prints I notice the solution was simpler
than I thought. AMS design is solid, just this return was wrong.
Sorry for the noise.

v3: Remove the entire chunck that would probably
    be removed by gcc anyway. (by Daniel)

Cc: Jani Nikula <jani.nikula@intel.com>
Cc: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Cc: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>

drivers/gpu/drm/i915/intel_display.c

@@ -12197,9 +12197,6 @@ intel_check_cursor_plane(struct drm_plane *plane,
 		return -ENOMEM;
 	}
 
-	if (fb == crtc->cursor->fb)
-		return 0;
-
 	/* we only need to pin inside GTT if cursor is non-phy */
 	mutex_lock(&dev->struct_mutex);
 	if (!INTEL_INFO(dev)->cursor_needs_physical && obj->tiling_mode) {