[Intel-gfx] [PATCH] drm: Global atomic state handling

[Daniel Vetter]

On Fri, Feb 06, 2015 at 11:58:58AM +0200, Jani Nikula wrote:
> On Wed, 05 Nov 2014, Daniel Vetter <daniel.vetter at ffwll.ch> wrote:
> 
> [heavily edited to just show the debugs]
> 
> > +struct drm_atomic_state *
> > +drm_atomic_state_alloc(struct drm_device *dev)
> > +{
> > +	DRM_DEBUG_KMS("Allocate atomic state %p\n", state);
> > +}
> 
> > +void drm_atomic_state_clear(struct drm_atomic_state *state)
> > +{
> > +	DRM_DEBUG_KMS("Clearing atomic state %p\n", state);
> > +}
> 
> > +void drm_atomic_state_free(struct drm_atomic_state *state)
> > +{
> > +	DRM_DEBUG_KMS("Freeing atomic state %p\n", state);
> > +}
> 
> > +struct drm_crtc_state *
> > +drm_atomic_get_crtc_state(struct drm_atomic_state *state,
> > +			  struct drm_crtc *crtc)
> > +{
> > +	DRM_DEBUG_KMS("Added [CRTC:%d] %p state to %p\n",
> > +		      crtc->base.id, crtc_state, state);
> > +}
> 
> > +struct drm_plane_state *
> > +drm_atomic_get_plane_state(struct drm_atomic_state *state,
> > +			  struct drm_plane *plane)
> > +{
> > +	DRM_DEBUG_KMS("Added [PLANE:%d] %p state to %p\n",
> > +		      plane->base.id, plane_state, state);
> > +}
> 
> > +struct drm_connector_state *
> > +drm_atomic_get_connector_state(struct drm_atomic_state *state,
> > +			  struct drm_connector *connector)
> > +{
> > +	DRM_DEBUG_KMS("Added [CONNECTOR:%d] %p state to %p\n",
> > +		      connector->base.id, connector_state, state);
> > +}
> 
> > +int drm_atomic_check_only(struct drm_atomic_state *state)
> > +{
> > +	DRM_DEBUG_KMS("checking %p\n", state);
> > +}
> 
> > +int drm_atomic_commit(struct drm_atomic_state *state)
> > +{
> > +	DRM_DEBUG_KMS("commiting %p\n", state);
> > +}
> 
> These seem to be rather verbose [1], is this normal or indicative of
> something going awry? If normal, it does seem quite noisy in the logs.

For debugging screw-ups in the state-preparetion, checks and commit
they're fairly useful imo. So definitely valuable for developing atomic
drivers.

Should we instead have a new DRM_DEBUG_ATOMIC to be able to filter these
out as needed?
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch