xref: /openbmc/linux/Documentation/gpu/todo.rst (revision 8dfb839c)
1.. _todo:
2
3=========
4TODO list
5=========
6
7This section contains a list of smaller janitorial tasks in the kernel DRM
8graphics subsystem useful as newbie projects. Or for slow rainy days.
9
10Subsystem-wide refactorings
11===========================
12
13De-midlayer drivers
14-------------------
15
16With the recent ``drm_bus`` cleanup patches for 3.17 it is no longer required
17to have a ``drm_bus`` structure set up. Drivers can directly set up the
18``drm_device`` structure instead of relying on bus methods in ``drm_usb.c``
19and ``drm_pci.c``. The goal is to get rid of the driver's ``->load`` /
20``->unload`` callbacks and open-code the load/unload sequence properly, using
21the new two-stage ``drm_device`` setup/teardown.
22
23Once all existing drivers are converted we can also remove those bus support
24files for USB and platform devices.
25
26All you need is a GPU for a non-converted driver (currently almost all of
27them, but also all the virtual ones used by KVM, so everyone qualifies).
28
29Contact: Daniel Vetter, Thierry Reding, respective driver maintainers
30
31Switch from reference/unreference to get/put
32--------------------------------------------
33
34For some reason DRM core uses ``reference``/``unreference`` suffixes for
35refcounting functions, but kernel uses ``get``/``put`` (e.g.
36``kref_get``/``put()``). It would be good to switch over for consistency, and
37it's shorter. Needs to be done in 3 steps for each pair of functions:
38
39* Create new ``get``/``put`` functions, define the old names as compatibility
40  wrappers
41* Switch over each file/driver using a cocci-generated spatch.
42* Once all users of the old names are gone, remove them.
43
44This way drivers/patches in the progress of getting merged won't break.
45
46Contact: Daniel Vetter
47
48Convert existing KMS drivers to atomic modesetting
49--------------------------------------------------
50
513.19 has the atomic modeset interfaces and helpers, so drivers can now be
52converted over. Modern compositors like Wayland or Surfaceflinger on Android
53really want an atomic modeset interface, so this is all about the bright
54future.
55
56There is a conversion guide for atomic and all you need is a GPU for a
57non-converted driver (again virtual HW drivers for KVM are still all
58suitable).
59
60As part of this drivers also need to convert to universal plane (which means
61exposing primary & cursor as proper plane objects). But that's much easier to
62do by directly using the new atomic helper driver callbacks.
63
64Contact: Daniel Vetter, respective driver maintainers
65
66Clean up the clipped coordination confusion around planes
67---------------------------------------------------------
68
69We have a helper to get this right with drm_plane_helper_check_update(), but
70it's not consistently used. This should be fixed, preferrably in the atomic
71helpers (and drivers then moved over to clipped coordinates). Probably the
72helper should also be moved from drm_plane_helper.c to the atomic helpers, to
73avoid confusion - the other helpers in that file are all deprecated legacy
74helpers.
75
76Contact: Ville Syrjälä, Daniel Vetter, driver maintainers
77
78Convert early atomic drivers to async commit helpers
79----------------------------------------------------
80
81For the first year the atomic modeset helpers didn't support asynchronous /
82nonblocking commits, and every driver had to hand-roll them. This is fixed
83now, but there's still a pile of existing drivers that easily could be
84converted over to the new infrastructure.
85
86One issue with the helpers is that they require that drivers handle completion
87events for atomic commits correctly. But fixing these bugs is good anyway.
88
89Contact: Daniel Vetter, respective driver maintainers
90
91Better manual-upload support for atomic
92---------------------------------------
93
94This would be especially useful for tinydrm:
95
96- Add a struct drm_rect dirty_clip to drm_crtc_state. When duplicating the
97  crtc state, clear that to the max values, x/y = 0 and w/h = MAX_INT, in
98  __drm_atomic_helper_crtc_duplicate_state().
99
100- Move tinydrm_merge_clips into drm_framebuffer.c, dropping the tinydrm\_
101  prefix ofc and using drm_fb\_. drm_framebuffer.c makes sense since this
102  is a function useful to implement the fb->dirty function.
103
104- Create a new drm_fb_dirty function which does essentially what e.g.
105  mipi_dbi_fb_dirty does. You can use e.g. drm_atomic_helper_update_plane as the
106  template. But instead of doing a simple full-screen plane update, this new
107  helper also sets crtc_state->dirty_clip to the right coordinates. And of
108  course it needs to check whether the fb is actually active (and maybe where),
109  so there's some book-keeping involved. There's also some good fun involved in
110  scaling things appropriately. For that case we might simply give up and
111  declare the entire area covered by the plane as dirty.
112
113Contact: Noralf Trønnes, Daniel Vetter
114
115Fallout from atomic KMS
116-----------------------
117
118``drm_atomic_helper.c`` provides a batch of functions which implement legacy
119IOCTLs on top of the new atomic driver interface. Which is really nice for
120gradual conversion of drivers, but unfortunately the semantic mismatches are
121a bit too severe. So there's some follow-up work to adjust the function
122interfaces to fix these issues:
123
124* atomic needs the lock acquire context. At the moment that's passed around
125  implicitly with some horrible hacks, and it's also allocate with
126  ``GFP_NOFAIL`` behind the scenes. All legacy paths need to start allocating
127  the acquire context explicitly on stack and then also pass it down into
128  drivers explicitly so that the legacy-on-atomic functions can use them.
129
130  Except for some driver code this is done. This task should be finished by
131  adding WARN_ON(!drm_drv_uses_atomic_modeset) in drm_modeset_lock_all().
132
133* A bunch of the vtable hooks are now in the wrong place: DRM has a split
134  between core vfunc tables (named ``drm_foo_funcs``), which are used to
135  implement the userspace ABI. And then there's the optional hooks for the
136  helper libraries (name ``drm_foo_helper_funcs``), which are purely for
137  internal use. Some of these hooks should be move from ``_funcs`` to
138  ``_helper_funcs`` since they are not part of the core ABI. There's a
139  ``FIXME`` comment in the kerneldoc for each such case in ``drm_crtc.h``.
140
141Contact: Daniel Vetter
142
143Get rid of dev->struct_mutex from GEM drivers
144---------------------------------------------
145
146``dev->struct_mutex`` is the Big DRM Lock from legacy days and infested
147everything. Nowadays in modern drivers the only bit where it's mandatory is
148serializing GEM buffer object destruction. Which unfortunately means drivers
149have to keep track of that lock and either call ``unreference`` or
150``unreference_locked`` depending upon context.
151
152Core GEM doesn't have a need for ``struct_mutex`` any more since kernel 4.8,
153and there's a ``gem_free_object_unlocked`` callback for any drivers which are
154entirely ``struct_mutex`` free.
155
156For drivers that need ``struct_mutex`` it should be replaced with a driver-
157private lock. The tricky part is the BO free functions, since those can't
158reliably take that lock any more. Instead state needs to be protected with
159suitable subordinate locks or some cleanup work pushed to a worker thread. For
160performance-critical drivers it might also be better to go with a more
161fine-grained per-buffer object and per-context lockings scheme. Currently only the
162``msm`` driver still use ``struct_mutex``.
163
164Contact: Daniel Vetter, respective driver maintainers
165
166Convert instances of dev_info/dev_err/dev_warn to their DRM_DEV_* equivalent
167----------------------------------------------------------------------------
168
169For drivers which could have multiple instances, it is necessary to
170differentiate between which is which in the logs. Since DRM_INFO/WARN/ERROR
171don't do this, drivers used dev_info/warn/err to make this differentiation. We
172now have DRM_DEV_* variants of the drm print macros, so we can start to convert
173those drivers back to using drm-formwatted specific log messages.
174
175Before you start this conversion please contact the relevant maintainers to make
176sure your work will be merged - not everyone agrees that the DRM dmesg macros
177are better.
178
179Contact: Sean Paul, Maintainer of the driver you plan to convert
180
181Convert drivers to use simple modeset suspend/resume
182----------------------------------------------------
183
184Most drivers (except i915 and nouveau) that use
185drm_atomic_helper_suspend/resume() can probably be converted to use
186drm_mode_config_helper_suspend/resume(). Also there's still open-coded version
187of the atomic suspend/resume code in older atomic modeset drivers.
188
189Contact: Maintainer of the driver you plan to convert
190
191Convert drivers to use drm_fb_helper_fbdev_setup/teardown()
192-----------------------------------------------------------
193
194Most drivers can use drm_fb_helper_fbdev_setup() except maybe:
195
196- amdgpu which has special logic to decide whether to call
197  drm_helper_disable_unused_functions()
198
199- armada which isn't atomic and doesn't call
200  drm_helper_disable_unused_functions()
201
202- i915 which calls drm_fb_helper_initial_config() in a worker
203
204Drivers that use drm_framebuffer_remove() to clean up the fbdev framebuffer can
205probably use drm_fb_helper_fbdev_teardown().
206
207Contact: Maintainer of the driver you plan to convert
208
209Clean up mmap forwarding
210------------------------
211
212A lot of drivers forward gem mmap calls to dma-buf mmap for imported buffers.
213And also a lot of them forward dma-buf mmap to the gem mmap implementations.
214Would be great to refactor this all into a set of small common helpers.
215
216Contact: Daniel Vetter
217
218Put a reservation_object into drm_gem_object
219--------------------------------------------
220
221This would remove the need for the ->gem_prime_res_obj callback. It would also
222allow us to implement generic helpers for waiting for a bo, allowing for quite a
223bit of refactoring in the various wait ioctl implementations.
224
225Contact: Daniel Vetter
226
227idr_init_base()
228---------------
229
230DRM core&drivers uses a lot of idr (integer lookup directories) for mapping
231userspace IDs to internal objects, and in most places ID=0 means NULL and hence
232is never used. Switching to idr_init_base() for these would make the idr more
233efficient.
234
235Contact: Daniel Vetter
236
237Core refactorings
238=================
239
240Clean up the DRM header mess
241----------------------------
242
243The DRM subsystem originally had only one huge global header, ``drmP.h``. This
244is now split up, but many source files still include it. The remaining part of
245the cleanup work here is to replace any ``#include <drm/drmP.h>`` by only the
246headers needed (and fixing up any missing pre-declarations in the headers).
247
248In the end no .c file should need to include ``drmP.h`` anymore.
249
250Contact: Daniel Vetter
251
252Add missing kerneldoc for exported functions
253--------------------------------------------
254
255The DRM reference documentation is still lacking kerneldoc in a few areas. The
256task would be to clean up interfaces like moving functions around between
257files to better group them and improving the interfaces like dropping return
258values for functions that never fail. Then write kerneldoc for all exported
259functions and an overview section and integrate it all into the drm book.
260
261See https://dri.freedesktop.org/docs/drm/ for what's there already.
262
263Contact: Daniel Vetter
264
265Make panic handling work
266------------------------
267
268This is a really varied tasks with lots of little bits and pieces:
269
270* The panic path can't be tested currently, leading to constant breaking. The
271  main issue here is that panics can be triggered from hardirq contexts and
272  hence all panic related callback can run in hardirq context. It would be
273  awesome if we could test at least the fbdev helper code and driver code by
274  e.g. trigger calls through drm debugfs files. hardirq context could be
275  achieved by using an IPI to the local processor.
276
277* There's a massive confusion of different panic handlers. DRM fbdev emulation
278  helpers have one, but on top of that the fbcon code itself also has one. We
279  need to make sure that they stop fighting over each another.
280
281* ``drm_can_sleep()`` is a mess. It hides real bugs in normal operations and
282  isn't a full solution for panic paths. We need to make sure that it only
283  returns true if there's a panic going on for real, and fix up all the
284  fallout.
285
286* The panic handler must never sleep, which also means it can't ever
287  ``mutex_lock()``. Also it can't grab any other lock unconditionally, not
288  even spinlocks (because NMI and hardirq can panic too). We need to either
289  make sure to not call such paths, or trylock everything. Really tricky.
290
291* For the above locking troubles reasons it's pretty much impossible to
292  attempt a synchronous modeset from panic handlers. The only thing we could
293  try to achive is an atomic ``set_base`` of the primary plane, and hope that
294  it shows up. Everything else probably needs to be delayed to some worker or
295  something else which happens later on. Otherwise it just kills the box
296  harder, prevent the panic from going out on e.g. netconsole.
297
298* There's also proposal for a simplied DRM console instead of the full-blown
299  fbcon and DRM fbdev emulation. Any kind of panic handling tricks should
300  obviously work for both console, in case we ever get kmslog merged.
301
302Contact: Daniel Vetter
303
304Clean up the debugfs support
305----------------------------
306
307There's a bunch of issues with it:
308
309- The drm_info_list ->show() function doesn't even bother to cast to the drm
310  structure for you. This is lazy.
311
312- We probably want to have some support for debugfs files on crtc/connectors and
313  maybe other kms objects directly in core. There's even drm_print support in
314  the funcs for these objects to dump kms state, so it's all there. And then the
315  ->show() functions should obviously give you a pointer to the right object.
316
317- The drm_info_list stuff is centered on drm_minor instead of drm_device. For
318  anything we want to print drm_device (or maybe drm_file) is the right thing.
319
320- The drm_driver->debugfs_init hooks we have is just an artifact of the old
321  midlayered load sequence. DRM debugfs should work more like sysfs, where you
322  can create properties/files for an object anytime you want, and the core
323  takes care of publishing/unpuplishing all the files at register/unregister
324  time. Drivers shouldn't need to worry about these technicalities, and fixing
325  this (together with the drm_minor->drm_device move) would allow us to remove
326  debugfs_init.
327
328Contact: Daniel Vetter
329
330KMS cleanups
331------------
332
333Some of these date from the very introduction of KMS in 2008 ...
334
335- drm_mode_config.crtc_idr is misnamed, since it contains all KMS object. Should
336  be renamed to drm_mode_config.object_idr.
337
338- drm_display_mode doesn't need to be derived from drm_mode_object. That's
339  leftovers from older (never merged into upstream) KMS designs where modes
340  where set using their ID, including support to add/remove modes.
341
342Better Testing
343==============
344
345Enable trinity for DRM
346----------------------
347
348And fix up the fallout. Should be really interesting ...
349
350Make KMS tests in i-g-t generic
351-------------------------------
352
353The i915 driver team maintains an extensive testsuite for the i915 DRM driver,
354including tons of testcases for corner-cases in the modesetting API. It would
355be awesome if those tests (at least the ones not relying on Intel-specific GEM
356features) could be made to run on any KMS driver.
357
358Basic work to run i-g-t tests on non-i915 is done, what's now missing is mass-
359converting things over. For modeset tests we also first need a bit of
360infrastructure to use dumb buffers for untiled buffers, to be able to run all
361the non-i915 specific modeset tests.
362
363Extend virtual test driver (VKMS)
364---------------------------------
365
366See the documentation of :ref:`VKMS <vkms>` for more details. This is an ideal
367internship task, since it only requires a virtual machine and can be sized to
368fit the available time.
369
370Contact: Daniel Vetter
371
372Driver Specific
373===============
374
375tinydrm
376-------
377
378Tinydrm is the helper driver for really simple fb drivers. The goal is to make
379those drivers as simple as possible, so lots of room for refactoring:
380
381- backlight helpers, probably best to put them into a new drm_backlight.c.
382  This is because drivers/video is de-facto unmaintained. We could also
383  move drivers/video/backlight to drivers/gpu/backlight and take it all
384  over within drm-misc, but that's more work. Backlight helpers require a fair
385  bit of reworking and refactoring. A simple example is the enabling of a backlight.
386  Tinydrm has helpers for this. It would be good if other drivers can also use the
387  helper. However, there are various cases we need to consider i.e different
388  drivers seem to have different ways of enabling/disabling a backlight.
389  We also need to consider the backlight drivers (like gpio_backlight). The situation
390  is further complicated by the fact that the backlight is tied to fbdev
391  via fb_notifier_callback() which has complicated logic. For further details, refer
392  to the following discussion thread:
393  https://groups.google.com/forum/#!topic/outreachy-kernel/8rBe30lwtdA
394
395- spi helpers, probably best put into spi core/helper code. Thierry said
396  the spi maintainer is fast&reactive, so shouldn't be a big issue.
397
398- extract the mipi-dbi helper (well, the non-tinydrm specific parts at
399  least) into a separate helper, like we have for mipi-dsi already. Or follow
400  one of the ideas for having a shared dsi/dbi helper, abstracting away the
401  transport details more.
402
403- tinydrm_gem_cma_prime_import_sg_table should probably go into the cma
404  helpers, as a _vmapped variant (since not every driver needs the vmap).
405  And tinydrm_gem_cma_free_object could the be merged into
406  drm_gem_cma_free_object().
407
408- tinydrm_fb_create we could move into drm_simple_pipe, only need to add
409  the fb_create hook to drm_simple_pipe_funcs, which would again simplify a
410  bunch of things (since it gives you a one-stop vfunc for simple drivers).
411
412- Quick aside: The unregister devm stuff is kinda getting the lifetimes of
413  a drm_device wrong. Doesn't matter, since everyone else gets it wrong
414  too :-)
415
416- also rework the drm_framebuffer_funcs->dirty hook wire-up, see above.
417
418Contact: Noralf Trønnes, Daniel Vetter
419
420AMD DC Display Driver
421---------------------
422
423AMD DC is the display driver for AMD devices starting with Vega. There has been
424a bunch of progress cleaning it up but there's still plenty of work to be done.
425
426See drivers/gpu/drm/amd/display/TODO for tasks.
427
428Contact: Harry Wentland, Alex Deucher
429
430i915
431----
432
433- Our early/late pm callbacks could be removed in favour of using
434  device_link_add to model the dependency between i915 and snd_had. See
435  https://dri.freedesktop.org/docs/drm/driver-api/device_link.html
436
437Outside DRM
438===========
439