summaryrefslogtreecommitdiff
path: root/Documentation/gpu
diff options
context:
space:
mode:
authorDaniel Vetter <daniel.vetter@ffwll.ch>2019-12-04 11:00:11 +0100
committerDaniel Vetter <daniel.vetter@ffwll.ch>2019-12-10 21:44:02 +0100
commit0380c6846a4be705b2d5c964b01ba1e5aaa3f5df (patch)
tree13cb9637c78c33a4802da82ac8afb5d5de2946d0 /Documentation/gpu
parent3cc1430cdbca12044571e3849cd57ac2e0d695c1 (diff)
drm/atomic: Update docs around locking and commit sequencing
Both locking and especially sequencing of nonblocking commits have evolved a lot. The details are all there, but I noticed that the big picture and connections have fallen behind a bit. Apply polish. Motivated by some review discussions with Thierry. v2: Review from Thierry Reviewed-by: Thierry Reding <treding@nvidia.com> Cc: Thierry Reding <treding@nvidia.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20191204100011.859468-1-daniel.vetter@ffwll.ch
Diffstat (limited to 'Documentation/gpu')
-rw-r--r--Documentation/gpu/drm-kms.rst11
1 files changed, 10 insertions, 1 deletions
diff --git a/Documentation/gpu/drm-kms.rst b/Documentation/gpu/drm-kms.rst
index c68588ce4090..b9330343d1bc 100644
--- a/Documentation/gpu/drm-kms.rst
+++ b/Documentation/gpu/drm-kms.rst
@@ -260,7 +260,8 @@ Taken all together there's two consequences for the atomic design:
drm_connector_state <drm_connector_state>` for connectors. These are the only
objects with userspace-visible and settable state. For internal state drivers
can subclass these structures through embeddeding, or add entirely new state
- structures for their globally shared hardware functions.
+ structures for their globally shared hardware functions, see :c:type:`struct
+ drm_private_state<drm_private_state>`.
- An atomic update is assembled and validated as an entirely free-standing pile
of structures within the :c:type:`drm_atomic_state <drm_atomic_state>`
@@ -269,6 +270,14 @@ Taken all together there's two consequences for the atomic design:
to the driver and modeset objects. This way rolling back an update boils down
to releasing memory and unreferencing objects like framebuffers.
+Locking of atomic state structures is internally using :c:type:`struct
+drm_modeset_lock <drm_modeset_lock>`. As a general rule the locking shouldn't be
+exposed to drivers, instead the right locks should be automatically acquired by
+any function that duplicates or peeks into a state, like e.g.
+:c:func:`drm_atomic_get_crtc_state()`. Locking only protects the software data
+structure, ordering of committing state changes to hardware is sequenced using
+:c:type:`struct drm_crtc_commit <drm_crtc_commit>`.
+
Read on in this chapter, and also in :ref:`drm_atomic_helper` for more detailed
coverage of specific topics.