2016-06-21 14:49:02 +03:00
=========================
Kernel Mode Setting (KMS)
=========================
Drivers must initialize the mode setting core by calling
2020-03-23 17:49:25 +03:00
drmm_mode_config_init() on the DRM device. The function
2016-06-21 14:49:02 +03:00
initializes the :c:type: `struct drm_device <drm_device>`
mode_config field and never fails. Once done, mode configuration must
be setup by initializing the following fields.
- int min_width, min_height; int max_width, max_height;
Minimum and maximum width and height of the frame buffers in pixel
units.
- struct drm_mode_config_funcs \*funcs;
Mode setting functions.
2017-03-02 18:16:35 +03:00
Overview
========
.. kernel-render :: DOT
:alt: KMS Display Pipeline
:caption: KMS Display Pipeline Overview
digraph "KMS" {
node [shape=box]
subgraph cluster_static {
style=dashed
label="Static Objects"
node [bgcolor=grey style=filled]
"drm_plane A" -> "drm_crtc"
"drm_plane B" -> "drm_crtc"
"drm_crtc" -> "drm_encoder A"
"drm_crtc" -> "drm_encoder B"
}
subgraph cluster_user_created {
style=dashed
label="Userspace-Created"
node [shape=oval]
"drm_framebuffer 1" -> "drm_plane A"
"drm_framebuffer 2" -> "drm_plane B"
}
subgraph cluster_connector {
style=dashed
label="Hotpluggable"
"drm_encoder A" -> "drm_connector A"
"drm_encoder B" -> "drm_connector B"
}
}
The basic object structure KMS presents to userspace is fairly simple.
Framebuffers (represented by :c:type: `struct drm_framebuffer <drm_framebuffer>` ,
2018-07-09 11:40:10 +03:00
see `Frame Buffer Abstraction`_ ) feed into planes. Planes are represented by
:c:type: `struct drm_plane <drm_plane>` , see `Plane Abstraction`_ for more
details. One or more (or even no) planes feed their pixel data into a CRTC
(represented by :c:type: `struct drm_crtc <drm_crtc>` , see `CRTC Abstraction`_ )
for blending. The precise blending step is explained in more detail in `Plane
Composition Properties`_ and related chapters.
2017-03-02 18:16:35 +03:00
For the output routing the first step is encoders (represented by
:c:type: `struct drm_encoder <drm_encoder>` , see `Encoder Abstraction`_ ). Those
are really just internal artifacts of the helper libraries used to implement KMS
2023-08-15 00:28:22 +03:00
drivers. Besides that they make it unnecessarily more complicated for userspace
2017-03-02 18:16:35 +03:00
to figure out which connections between a CRTC and a connector are possible, and
what kind of cloning is supported, they serve no purpose in the userspace API.
Unfortunately encoders have been exposed to userspace, hence can't remove them
2023-08-15 00:28:22 +03:00
at this point. Furthermore the exposed restrictions are often wrongly set by
2017-03-02 18:16:35 +03:00
drivers, and in many cases not powerful enough to express the real restrictions.
A CRTC can be connected to multiple encoders, and for an active CRTC there must
be at least one encoder.
The final, and real, endpoint in the display chain is the connector (represented
by :c:type: `struct drm_connector <drm_connector>` , see `Connector
Abstraction`_). Connectors can have different possible encoders, but the kernel
driver selects which encoder to use for each connector. The use case is DVI,
which could switch between an analog and a digital encoder. Encoders can also
drive multiple different connectors. There is exactly one active connector for
every active encoder.
Internally the output pipeline is a bit more complex and matches today's
hardware more closely:
.. kernel-render :: DOT
:alt: KMS Output Pipeline
:caption: KMS Output Pipeline
digraph "Output Pipeline" {
node [shape=box]
subgraph {
"drm_crtc" [bgcolor=grey style=filled]
}
subgraph cluster_internal {
style=dashed
label="Internal Pipeline"
{
node [bgcolor=grey style=filled]
"drm_encoder A";
"drm_encoder B";
"drm_encoder C";
}
{
node [bgcolor=grey style=filled]
"drm_encoder B" -> "drm_bridge B"
"drm_encoder C" -> "drm_bridge C1"
"drm_bridge C1" -> "drm_bridge C2";
}
}
"drm_crtc" -> "drm_encoder A"
"drm_crtc" -> "drm_encoder B"
"drm_crtc" -> "drm_encoder C"
subgraph cluster_output {
style=dashed
label="Outputs"
"drm_encoder A" -> "drm_connector A";
"drm_bridge B" -> "drm_connector B";
"drm_bridge C2" -> "drm_connector C";
"drm_panel"
}
}
Internally two additional helper objects come into play. First, to be able to
share code for encoders (sometimes on the same SoC, sometimes off-chip) one or
more :ref: `drm_bridges` (represented by :c:type:`struct drm_bridge
<drm_bridge> `) can be linked to an encoder. This link is static and cannot be
changed, which means the cross-bar (if there is any) needs to be mapped between
the CRTC and any encoders. Often for drivers with bridges there's no code left
at the encoder level. Atomic drivers can leave out all the encoder callbacks to
essentially only leave a dummy routing object behind, which is needed for
backwards compatibility since encoders are exposed to userspace.
The second object is for panels, represented by :c:type:`struct drm_panel
<drm_panel> `, see :ref:` drm_panel_helper`. Panels do not have a fixed binding
point, but are generally linked to the driver private structure that embeds
:c:type: `struct drm_connector <drm_connector>` .
Note that currently the bridge chaining and interactions with connectors and
panels are still in-flux and not really fully sorted out yet.
2016-08-29 11:27:51 +03:00
2016-11-14 14:58:23 +03:00
KMS Core Structures and Functions
=================================
2016-08-29 11:27:51 +03:00
2016-11-14 14:58:23 +03:00
.. kernel-doc :: include/drm/drm_mode_config.h
2016-06-21 14:49:02 +03:00
:internal:
2017-03-02 18:16:36 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_mode_config.c
:export:
2021-08-02 10:28:35 +03:00
.. _kms_base_object_abstraction:
2016-11-14 14:58:23 +03:00
Modeset Base Object Abstraction
===============================
2016-08-12 23:48:41 +03:00
2017-03-02 18:16:37 +03:00
.. kernel-render :: DOT
:alt: Mode Objects and Properties
:caption: Mode Objects and Properties
digraph {
node [shape=box]
"drm_property A" -> "drm_mode_object A"
"drm_property A" -> "drm_mode_object B"
"drm_property B" -> "drm_mode_object A"
}
The base structure for all KMS objects is :c:type:`struct drm_mode_object
<drm_mode_object> `. One of the base services it provides is tracking properties,
which are especially important for the atomic IOCTL (see `Atomic Mode
Setting`_). The somewhat surprising part here is that properties are not
directly instantiated on each object, but free-standing mode objects themselves,
represented by :c:type: `struct drm_property <drm_property>` , which only specify
the type and value range of a property. Any given property can be attached
2019-12-04 13:19:33 +03:00
multiple times to different objects using drm_object_attach_property().
2017-03-02 18:16:37 +03:00
2016-11-14 14:58:23 +03:00
.. kernel-doc :: include/drm/drm_mode_object.h
:internal:
.. kernel-doc :: drivers/gpu/drm/drm_mode_object.c
2016-06-21 14:49:02 +03:00
:export:
2017-03-02 18:16:38 +03:00
Atomic Mode Setting
===================
.. kernel-render :: DOT
:alt: Mode Objects and Properties
:caption: Mode Objects and Properties
digraph {
node [shape=box]
subgraph cluster_state {
style=dashed
label="Free-standing state"
"drm_atomic_state" -> "duplicated drm_plane_state A"
"drm_atomic_state" -> "duplicated drm_plane_state B"
"drm_atomic_state" -> "duplicated drm_crtc_state"
"drm_atomic_state" -> "duplicated drm_connector_state"
"drm_atomic_state" -> "duplicated driver private state"
}
subgraph cluster_current {
style=dashed
label="Current state"
"drm_device" -> "drm_plane A"
"drm_device" -> "drm_plane B"
"drm_device" -> "drm_crtc"
"drm_device" -> "drm_connector"
"drm_device" -> "driver private object"
"drm_plane A" -> "drm_plane_state A"
"drm_plane B" -> "drm_plane_state B"
"drm_crtc" -> "drm_crtc_state"
"drm_connector" -> "drm_connector_state"
"driver private object" -> "driver private state"
}
"drm_atomic_state" -> "drm_device" [label="atomic_commit"]
"duplicated drm_plane_state A" -> "drm_device"[style=invis]
}
Atomic provides transactional modeset (including planes) updates, but a
bit differently from the usual transactional approach of try-commit and
rollback:
- Firstly, no hardware changes are allowed when the commit would fail. This
allows us to implement the DRM_MODE_ATOMIC_TEST_ONLY mode, which allows
userspace to explore whether certain configurations would work or not.
- This would still allow setting and rollback of just the software state,
simplifying conversion of existing drivers. But auditing drivers for
correctness of the atomic_check code becomes really hard with that: Rolling
back changes in data structures all over the place is hard to get right.
- Lastly, for backwards compatibility and to support all use-cases, atomic
updates need to be incremental and be able to execute in parallel. Hardware
doesn't always allow it, but where possible plane updates on different CRTCs
should not interfere, and not get stalled due to output routing changing on
different CRTCs.
Taken all together there's two consequences for the atomic design:
- The overall state is split up into per-object state structures:
:c:type: `struct drm_plane_state <drm_plane_state>` for planes, :c:type:`struct
drm_crtc_state <drm_crtc_state>` for CRTCs and :c:type:` struct
drm_connector_state <drm_connector_state>` for connectors. These are the only
objects with userspace-visible and settable state. For internal state drivers
2023-08-15 00:28:22 +03:00
can subclass these structures through embedding, or add entirely new state
2019-12-04 13:00:11 +03:00
structures for their globally shared hardware functions, see :c:type:`struct
drm_private_state<drm_private_state>`.
2017-03-02 18:16:38 +03:00
- 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>`
2017-12-14 23:30:54 +03:00
container. Driver private state structures are also tracked in the same
structure; see the next chapter. Only when a state is committed is it applied
to the driver and modeset objects. This way rolling back an update boils down
to releasing memory and unreferencing objects like framebuffers.
2017-03-02 18:16:38 +03:00
2019-12-04 13:00:11 +03:00
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.
2019-12-04 13:19:33 +03:00
drm_atomic_get_crtc_state(). Locking only protects the software data
2019-12-04 13:00:11 +03:00
structure, ordering of committing state changes to hardware is sequenced using
:c:type: `struct drm_crtc_commit <drm_crtc_commit>` .
2017-03-02 18:16:38 +03:00
Read on in this chapter, and also in :ref: `drm_atomic_helper` for more detailed
coverage of specific topics.
2017-12-14 23:30:53 +03:00
Handling Driver Private State
-----------------------------
.. kernel-doc :: drivers/gpu/drm/drm_atomic.c
:doc: handling driver private state
2016-06-21 14:49:02 +03:00
Atomic Mode Setting Function Reference
2017-03-02 18:16:38 +03:00
--------------------------------------
2016-06-21 14:49:02 +03:00
2016-08-12 23:48:46 +03:00
.. kernel-doc :: include/drm/drm_atomic.h
2016-06-21 14:49:02 +03:00
:internal:
2017-03-02 18:16:36 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_atomic.c
:export:
2018-09-05 16:57:11 +03:00
Atomic Mode Setting IOCTL and UAPI Functions
--------------------------------------------
.. kernel-doc :: drivers/gpu/drm/drm_atomic_uapi.c
:doc: overview
.. kernel-doc :: drivers/gpu/drm/drm_atomic_uapi.c
:export:
2016-11-14 14:58:23 +03:00
CRTC Abstraction
================
.. kernel-doc :: drivers/gpu/drm/drm_crtc.c
2017-01-25 09:26:57 +03:00
:doc: overview
CRTC Functions Reference
--------------------------------
2016-11-14 14:58:23 +03:00
.. kernel-doc :: include/drm/drm_crtc.h
:internal:
2017-01-25 09:26:57 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_crtc.c
:export:
2020-12-16 23:22:18 +03:00
Color Management Functions Reference
------------------------------------
.. kernel-doc :: drivers/gpu/drm/drm_color_mgmt.c
:export:
.. kernel-doc :: include/drm/drm_color_mgmt.h
:internal:
2016-06-21 14:49:02 +03:00
Frame Buffer Abstraction
2016-08-12 23:48:41 +03:00
========================
2016-06-21 14:49:02 +03:00
2016-08-12 23:48:48 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_framebuffer.c
:doc: overview
2016-06-21 14:49:02 +03:00
2016-08-15 17:07:02 +03:00
Frame Buffer Functions Reference
--------------------------------
.. kernel-doc :: include/drm/drm_framebuffer.h
:internal:
2017-03-02 18:16:36 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_framebuffer.c
:export:
2016-06-21 14:49:02 +03:00
DRM Format Handling
2016-08-12 23:48:41 +03:00
===================
2016-06-21 14:49:02 +03:00
2018-08-21 19:16:11 +03:00
.. kernel-doc :: include/uapi/drm/drm_fourcc.h
:doc: overview
Format Functions Reference
--------------------------
2016-10-18 01:41:09 +03:00
.. kernel-doc :: include/drm/drm_fourcc.h
:internal:
2016-06-21 14:49:02 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_fourcc.c
:export:
2023-08-21 16:15:56 +03:00
.. _kms_dumb_buffer_objects:
2016-06-21 14:49:02 +03:00
Dumb Buffer Objects
2016-08-12 23:48:41 +03:00
===================
2016-06-21 14:49:02 +03:00
2016-11-14 14:58:21 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_dumb_buffers.c
:doc: overview
2016-06-21 14:49:02 +03:00
2016-09-21 11:59:24 +03:00
Plane Abstraction
=================
2016-09-21 11:59:25 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_plane.c
:doc: overview
2016-09-21 11:59:24 +03:00
Plane Functions Reference
-------------------------
.. kernel-doc :: include/drm/drm_plane.h
:internal:
.. kernel-doc :: drivers/gpu/drm/drm_plane.c
:export:
2020-12-16 23:22:16 +03:00
Plane Composition Functions Reference
-------------------------------------
.. kernel-doc :: drivers/gpu/drm/drm_blend.c
:export:
2020-12-16 23:22:17 +03:00
Plane Damage Tracking Functions Reference
-----------------------------------------
.. kernel-doc :: drivers/gpu/drm/drm_damage_helper.c
:export:
.. kernel-doc :: include/drm/drm_damage_helper.h
:internal:
2024-04-09 19:30:41 +03:00
Plane Panic Feature
-------------------
.. kernel-doc :: drivers/gpu/drm/drm_panic.c
:doc: overview
Plane Panic Functions Reference
-------------------------------
.. kernel-doc :: include/drm/drm_panic.h
:internal:
.. kernel-doc :: drivers/gpu/drm/drm_panic.c
:export:
2016-08-12 23:48:41 +03:00
Display Modes Function Reference
================================
2016-06-21 14:49:02 +03:00
2016-08-12 23:48:41 +03:00
.. kernel-doc :: include/drm/drm_modes.h
:internal:
.. kernel-doc :: drivers/gpu/drm/drm_modes.c
:export:
2016-06-21 14:49:02 +03:00
2016-08-12 23:48:53 +03:00
Connector Abstraction
=====================
.. kernel-doc :: drivers/gpu/drm/drm_connector.c
:doc: overview
Connector Functions Reference
-----------------------------
2016-08-12 23:48:50 +03:00
.. kernel-doc :: include/drm/drm_connector.h
:internal:
.. kernel-doc :: drivers/gpu/drm/drm_connector.c
:export:
2017-03-29 19:42:32 +03:00
Writeback Connectors
--------------------
.. kernel-doc :: drivers/gpu/drm/drm_writeback.c
:doc: overview
2022-01-11 23:27:14 +03:00
.. kernel-doc :: include/drm/drm_writeback.h
:internal:
2017-03-29 19:42:32 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_writeback.c
:export:
2016-08-29 11:27:49 +03:00
Encoder Abstraction
===================
2016-08-29 11:27:50 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_encoder.c
:doc: overview
Encoder Functions Reference
---------------------------
2016-08-29 11:27:49 +03:00
.. kernel-doc :: include/drm/drm_encoder.h
:internal:
.. kernel-doc :: drivers/gpu/drm/drm_encoder.c
:export:
2016-06-21 14:49:02 +03:00
KMS Locking
2016-08-12 23:48:41 +03:00
===========
2016-06-21 14:49:02 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_modeset_lock.c
:doc: kms locking
.. kernel-doc :: include/drm/drm_modeset_lock.h
:internal:
.. kernel-doc :: drivers/gpu/drm/drm_modeset_lock.c
:export:
KMS Properties
==============
2020-12-17 14:32:12 +03:00
This section of the documentation is primarily aimed at user-space developers.
For the driver APIs, see the other sections.
2021-07-20 17:35:44 +03:00
Requirements
------------
KMS drivers might need to add extra properties to support new features. Each
new property introduced in a driver needs to meet a few requirements, in
addition to the one mentioned above:
* It must be standardized, documenting:
* The full, exact, name string;
* If the property is an enum, all the valid value name strings;
* What values are accepted, and what these values mean;
* What the property does and how it can be used;
* How the property might interact with other, existing properties.
* It must provide a generic helper in the core code to register that
property on the object it attaches to.
* Its content must be decoded by the core and provided in the object's
associated state structure. That includes anything drivers might want
to precompute, like struct drm_clip_rect for planes.
* Its initial state must match the behavior prior to the property
introduction. This might be a fixed value matching what the hardware
does, or it may be inherited from the state the firmware left the
system in during boot.
* An IGT test must be submitted where reasonable.
2024-04-10 15:20:06 +03:00
For historical reasons, non-standard, driver-specific properties exist. If a KMS
driver wants to add support for one of those properties, the requirements for
new properties apply where possible. Additionally, the documented behavior must
match the de facto semantics of the existing property to ensure compatibility.
Developers of the driver that first added the property should help with those
tasks and must ACK the documented behavior if possible.
2016-08-29 11:27:55 +03:00
Property Types and Blob Property Support
----------------------------------------
2016-08-29 11:27:57 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_property.c
:doc: overview
2016-08-29 11:27:55 +03:00
.. kernel-doc :: include/drm/drm_property.h
:internal:
.. kernel-doc :: drivers/gpu/drm/drm_property.c
:export:
2021-10-05 23:23:13 +03:00
.. _standard_connector_properties:
2016-11-17 11:56:48 +03:00
Standard Connector Properties
-----------------------------
.. kernel-doc :: drivers/gpu/drm/drm_connector.c
:doc: standard connector properties
2018-05-15 16:59:27 +03:00
HDMI Specific Connector Properties
2018-07-02 12:10:23 +03:00
----------------------------------
2018-05-15 16:59:27 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_connector.c
:doc: HDMI connector properties
2022-11-17 12:28:48 +03:00
Analog TV Specific Connector Properties
2022-11-28 11:19:36 +03:00
---------------------------------------
2022-11-17 12:28:48 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_connector.c
:doc: Analog TV Connector Properties
2020-05-21 14:09:31 +03:00
Standard CRTC Properties
------------------------
.. kernel-doc :: drivers/gpu/drm/drm_crtc.c
:doc: standard CRTC properties
2020-12-17 14:32:13 +03:00
Standard Plane Properties
-------------------------
.. kernel-doc :: drivers/gpu/drm/drm_plane.c
:doc: standard plane properties
2022-08-04 18:01:07 +03:00
.. _plane_composition_properties:
2016-09-21 11:59:27 +03:00
Plane Composition Properties
----------------------------
.. kernel-doc :: drivers/gpu/drm/drm_blend.c
:doc: overview
2016-08-12 23:48:51 +03:00
2023-11-24 01:13:00 +03:00
.. _damage_tracking_properties:
2020-12-16 23:22:15 +03:00
Damage Tracking Properties
--------------------------
2018-05-24 05:04:08 +03:00
2021-07-23 11:34:57 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_plane.c
:doc: damage tracking
2018-05-24 05:04:08 +03:00
2016-09-21 11:59:29 +03:00
Color Management Properties
---------------------------
.. kernel-doc :: drivers/gpu/drm/drm_color_mgmt.c
:doc: overview
2016-11-14 14:58:24 +03:00
Tile Group Property
-------------------
.. kernel-doc :: drivers/gpu/drm/drm_connector.c
:doc: Tile group
2016-11-22 03:11:28 +03:00
Explicit Fencing Properties
---------------------------
2018-09-05 16:57:11 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_atomic_uapi.c
2016-11-22 03:11:28 +03:00
:doc: explicit fencing properties
2018-10-04 21:38:42 +03:00
Variable Refresh Properties
---------------------------
.. kernel-doc :: drivers/gpu/drm/drm_connector.c
:doc: Variable refresh properties
2023-10-23 10:46:13 +03:00
Cursor Hotspot Properties
---------------------------
.. kernel-doc :: drivers/gpu/drm/drm_plane.c
:doc: hotspot properties
2016-06-21 14:49:02 +03:00
Existing KMS Properties
-----------------------
2018-02-20 01:53:52 +03:00
The following table gives description of drm properties exposed by various
modules/drivers. Because this table is very unwieldy, do not add any new
properties here. Instead document them in a section above.
2016-06-21 14:49:02 +03:00
.. csv-table ::
:header-rows: 1
:file: kms-properties.csv
Vertical Blanking
=================
2017-05-24 17:51:45 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_vblank.c
:doc: vblank handling
2016-06-21 14:49:02 +03:00
Vertical Blanking and Interrupt Handling Functions Reference
------------------------------------------------------------
2017-05-31 12:21:46 +03:00
.. kernel-doc :: include/drm/drm_vblank.h
2016-07-15 22:48:01 +03:00
:internal:
2017-03-02 18:16:36 +03:00
2017-05-31 12:21:46 +03:00
.. kernel-doc :: drivers/gpu/drm/drm_vblank.c
2017-03-02 18:16:36 +03:00
:export:
drm/vblank: Add vblank works
Add some kind of vblank workers. The interface is similar to regular
delayed works, and is mostly based off kthread_work. It allows for
scheduling delayed works that execute once a particular vblank sequence
has passed. It also allows for accurate flushing of scheduled vblank
works - in that flushing waits for both the vblank sequence and job
execution to complete, or for the work to get cancelled - whichever
comes first.
Whatever hardware programming we do in the work must be fast (must at
least complete during the vblank or scanout period, sometimes during the
first few scanlines of the vblank). As such we use a high-priority
per-CRTC thread to accomplish this.
Changes since v7:
* Stuff drm_vblank_internal.h and drm_vblank_work_internal.h contents
into drm_internal.h
* Get rid of unnecessary spinlock in drm_crtc_vblank_on()
* Remove !vblank->worker check
* Grab vbl_lock in drm_vblank_work_schedule()
* Mention self-rearming work items in drm_vblank_work_schedule() kdocs
* Return 1 from drm_vblank_work_schedule() if the work was scheduled
successfully, 0 or error code otherwise
* Use drm_dbg_core() instead of DRM_DEV_ERROR() in
drm_vblank_work_schedule()
* Remove vblank->worker checks in drm_vblank_destroy_worker() and
drm_vblank_flush_worker()
Changes since v6:
* Get rid of ->pending and seqcounts, and implement flushing through
simpler means - danvet
* Get rid of work_lock, just use drm_device->event_lock
* Move drm_vblank_work item cleanup into drm_crtc_vblank_off() so that
we ensure that all vblank work has finished before disabling vblanks
* Add checks into drm_crtc_vblank_reset() so we yell if it gets called
while there's vblank workers active
* Grab event_lock in both drm_crtc_vblank_on()/drm_crtc_vblank_off(),
the main reason for this is so that other threads calling
drm_vblank_work_schedule() are blocked from attempting to schedule
while we're in the middle of enabling/disabling vblanks.
* Move drm_handle_vblank_works() call below drm_handle_vblank_events()
* Simplify drm_vblank_work_cancel_sync()
* Fix drm_vblank_work_cancel_sync() documentation
* Move wake_up_all() calls out of spinlock where we can. The only one I
left was the call to wake_up_all() in drm_vblank_handle_works() as
this seemed like it made more sense just living in that function
(which is all technically under lock)
* Move drm_vblank_work related functions into their own source files
* Add drm_vblank_internal.h so we can export some functions we don't
want drivers using, but that we do need to use in drm_vblank_work.c
* Add a bunch of documentation
Changes since v4:
* Get rid of kthread interfaces we tried adding and move all of the
locking into drm_vblank.c. For implementing drm_vblank_work_flush(),
we now use a wait_queue and sequence counters in order to
differentiate between multiple work item executions.
* Get rid of drm_vblank_work_cancel() - this would have been pretty
difficult to actually reimplement and it occurred to me that neither
nouveau or i915 are even planning to use this function. Since there's
also no async cancel function for most of the work interfaces in the
kernel, it seems a bit unnecessary anyway.
* Get rid of to_drm_vblank_work() since we now are also able to just
pass the struct drm_vblank_work to work item callbacks anyway
Changes since v3:
* Use our own spinlocks, don't integrate so tightly with kthread_works
Changes since v2:
* Use kthread_workers instead of reinventing the wheel.
Cc: Tejun Heo <tj@kernel.org>
Cc: dri-devel@lists.freedesktop.org
Cc: nouveau@lists.freedesktop.org
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Co-developed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Lyude Paul <lyude@redhat.com>
Acked-by: Dave Airlie <airlied@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200627194657.156514-4-lyude@redhat.com
2020-04-17 22:33:13 +03:00
Vertical Blank Work
===================
.. kernel-doc :: drivers/gpu/drm/drm_vblank_work.c
:doc: vblank works
Vertical Blank Work Functions Reference
---------------------------------------
.. kernel-doc :: include/drm/drm_vblank_work.h
:internal:
.. kernel-doc :: drivers/gpu/drm/drm_vblank_work.c
:export: