xref: /openbmc/linux/Documentation/gpu/vkms.rst (revision 2208f39c)
1.. _vkms:
2
3==========================================
4 drm/vkms Virtual Kernel Modesetting
5==========================================
6
7.. kernel-doc:: drivers/gpu/drm/vkms/vkms_drv.c
8   :doc: vkms (Virtual Kernel Modesetting)
9
10TODO
11====
12
13If you want to do any of the items listed below, please share your interest
14with VKMS maintainers.
15
16IGT better support
17------------------
18
19- Investigate: (1) test cases on kms_plane that are failing due to timeout on
20  capturing CRC; (2) when running kms_flip test cases in sequence, some
21  successful individual test cases are failing randomly.
22
23- VKMS already has support for vblanks simulated via hrtimers, which can be
24  tested with kms_flip test; in some way, we can say that VKMS already mimics
25  the real hardware vblank. However, we also have virtual hardware that does
26  not support vblank interrupt and completes page_flip events right away; in
27  this case, compositor developers may end up creating a busy loop on virtual
28  hardware. It would be useful to support Virtual Hardware behavior in VKMS
29  because this can help compositor developers to test their features in
30  multiple scenarios.
31
32Add Plane Features
33------------------
34
35There's lots of plane features we could add support for:
36
37- Real overlay planes, not just cursor.
38
39- Full alpha blending on all planes.
40
41- Rotation, scaling.
42
43- Additional buffer formats, especially YUV formats for video like NV12.
44  Low/high bpp RGB formats would also be interesting.
45
46- Async updates (currently only possible on cursor plane using the legacy
47  cursor api).
48
49For all of these, we also want to review the igt test coverage and make sure
50all relevant igt testcases work on vkms.
51
52Prime Buffer Sharing
53--------------------
54
55- Syzbot report - WARNING in vkms_gem_free_object:
56  https://syzkaller.appspot.com/bug?extid=e7ad70d406e74d8fc9d0
57
58Runtime Configuration
59---------------------
60
61We want to be able to reconfigure vkms instance without having to reload the
62module. Use/Test-cases:
63
64- Hotplug/hotremove connectors on the fly (to be able to test DP MST handling
65  of compositors).
66
67- Configure planes/crtcs/connectors (we'd need some code to have more than 1 of
68  them first).
69
70- Change output configuration: Plug/unplug screens, change EDID, allow changing
71  the refresh rate.
72
73The currently proposed solution is to expose vkms configuration through
74configfs.  All existing module options should be supported through configfs
75too.
76
77Writeback support
78-----------------
79
80- The writeback and CRC capture operations share the use of composer_enabled
81  boolean to ensure vblanks. Probably, when these operations work together,
82  composer_enabled needs to refcounting the composer state to proper work.
83
84- Add support for cloned writeback outputs and related test cases using a
85  cloned output in the IGT kms_writeback.
86
87- As a v4l device. This is useful for debugging compositors on special vkms
88  configurations, so that developers see what's really going on.
89
90Output Features
91---------------
92
93- Variable refresh rate/freesync support. This probably needs prime buffer
94  sharing support, so that we can use vgem fences to simulate rendering in
95  testing. Also needs support to specify the EDID.
96
97- Add support for link status, so that compositors can validate their runtime
98  fallbacks when e.g. a Display Port link goes bad.
99
100CRC API Improvements
101--------------------
102
103- Optimize CRC computation ``compute_crc()`` and plane blending ``blend()``
104
105Atomic Check using eBPF
106-----------------------
107
108Atomic drivers have lots of restrictions which are not exposed to userspace in
109any explicit form through e.g. possible property values. Userspace can only
110inquiry about these limits through the atomic IOCTL, possibly using the
111TEST_ONLY flag. Trying to add configurable code for all these limits, to allow
112compositors to be tested against them, would be rather futile exercise. Instead
113we could add support for eBPF to validate any kind of atomic state, and
114implement a library of different restrictions.
115
116This needs a bunch of features (plane compositing, multiple outputs, ...)
117enabled already to make sense.
118