If we ever hit bugs in our readout code we may have set plane_state->visible without setting a fb/crtc. Our code implicitly assumes that plane_state->visible is only true when both are set, so we will either null deref in intel_fbc_choose_crtc() or do something bad during the actual commit which cares even more. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@xxxxxxxxxxxxxxx> --- drivers/gpu/drm/i915/intel_atomic_plane.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/intel_atomic_plane.c b/drivers/gpu/drm/i915/intel_atomic_plane.c index e067fb531a29..d754a8ba9ffb 100644 --- a/drivers/gpu/drm/i915/intel_atomic_plane.c +++ b/drivers/gpu/drm/i915/intel_atomic_plane.c @@ -117,10 +117,10 @@ int intel_plane_atomic_check_with_state(const struct intel_crtc_state *old_crtc_ struct intel_plane *intel_plane = to_intel_plane(plane); int ret; + intel_state->base.visible = false; if (!intel_state->base.crtc && !old_plane_state->base.crtc) return 0; - intel_state->base.visible = false; ret = intel_plane->check_plane(crtc_state, intel_state); if (ret) return ret; @@ -152,6 +152,7 @@ static int intel_plane_atomic_check(struct drm_plane *plane, const struct drm_crtc_state *old_crtc_state; struct drm_crtc_state *new_crtc_state; + new_plane_state->visible = false; if (!crtc) return 0; -- 2.18.0 _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
