drm/todo: Add entry to rename drm_atomic_state
authorMaxime Ripard <mripard@kernel.org>
Thu, 14 Dec 2023 10:09:16 +0000 (11:09 +0100)
committerMaxime Ripard <mripard@kernel.org>
Mon, 18 Dec 2023 10:45:40 +0000 (11:45 +0100)
The name of the structure drm_atomic_state is confusing. Let's add an
entry to our todo list to rename it.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://lore.kernel.org/r/20231214100917.277842-5-mripard@kernel.org
Signed-off-by: Maxime Ripard <mripard@kernel.org>
Documentation/gpu/todo.rst

index 41a264bf84ce02127ee00ac646b216ed532a9963..fb9ad120b141478a657012f752a5207e98984c73 100644 (file)
@@ -120,6 +120,29 @@ Contact: Daniel Vetter, respective driver maintainers
 
 Level: Advanced
 
+Rename drm_atomic_state
+-----------------------
+
+The KMS framework uses two slightly different definitions for the ``state``
+concept. For a given object (plane, CRTC, encoder, etc., so
+``drm_$OBJECT_state``), the state is the entire state of that object. However,
+at the device level, ``drm_atomic_state`` refers to a state update for a
+limited number of objects.
+
+The state isn't the entire device state, but only the full state of some
+objects in that device. This is confusing to newcomers, and
+``drm_atomic_state`` should be renamed to something clearer like
+``drm_atomic_commit``.
+
+In addition to renaming the structure itself, it would also imply renaming some
+related functions (``drm_atomic_state_alloc``, ``drm_atomic_state_get``,
+``drm_atomic_state_put``, ``drm_atomic_state_init``,
+``__drm_atomic_state_free``, etc.).
+
+Contact: Maxime Ripard <mripard@kernel.org>
+
+Level: Advanced
+
 Fallout from atomic KMS
 -----------------------