Skip to content
  1. Jun 07, 2016
  3. Jun 04, 2016
  5. Jun 03, 2016
    • Chris Wilson's avatar
      drm: Store the plane's index · 490d3d1b
      Chris Wilson authored 
      

Currently the plane's index is determined by walking the list of all
planes in the mode and finding the position of that plane in the list. A
linear walk, especially a linear walk within a linear walk as frequently
conceived by i915.ko [O(N^2)] quickly comes to dominate profiles.

The plane's index is constant for as long as no earlier planes are
removed from the list. For all drivers, planes are static, determined
at boot and then untouched until shutdown. In fact, there is no locking
provided to allow for dynamic removal of planes/encoders/crtcs.

v2: Convert drm_crtc_index() and drm_encoder_index() as well.
v3: Stop adjusting the indices upon removal; consider the list
construct-only.

Signed-off-by: default avatarChris Wilson <chris@chris-wilson.co.uk>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Matt Roper <matthew.d.roper@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: default avatarMatt Roper <matthew.d.roper@intel.com>
[danvet: Fixup typo in kerneldoc that Matt spotted.]
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1464375900-2542-1-git-send-email-chris@chris-wilson.co.uk
      490d3d1b
    • Gustavo Padovan's avatar
      drm/fence: add fence to drm_pending_event · 1b47aaf9
      Gustavo Padovan authored 
      

Now a drm_pending_event can either send a real drm_event or signal a
fence, or both. It allow us to signal via fences when the buffer is
displayed on the screen. Which in turn means that the previous buffer
is not in use anymore and can be freed or sent back to another driver
for processing.

v2: Comments from Daniel Vetter
	- call fence_signal in drm_send_event_locked()
	- remove unneeded !e->event check

v3: Remove drm_pending_event->destroy to fix a leak when e->file_priv
is not set.

Reviewed-by: default avatarSean Paul <seanpaul@chromium.org>
Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> (v2)
[danvet: fix one e->destroy in arcpgu due to rebasing.]
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1464818821-5736-13-git-send-email-daniel.vetter@ffwll.ch
      1b47aaf9
  7. Jun 02, 2016
    • Daniel Vetter's avatar
      drm/atomic-docs: Spelling fixups · 60c9e190
      Daniel Vetter authored 
      Eric nicely pointed these out, but I failed at git add and lost them.
This fixes up

commit 2f196b7c


Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Thu Jun 2 16:21:44 2016 +0200

    drm/atomic: Add drm_atomic_crtc_state_for_each_plane_state

to actually do what it says on the tin^Wcommit message.

Signed-off-by: default avatarDaniel Vetter <daniel.vetter@intel.com>
      60c9e190
    • Daniel Vetter's avatar
      drm: Consolidate crtc arrays in drm_atomic_state · 5d943aa6
      Daniel Vetter authored 
      

It's silly to have 2 mallocs when we could tie these two together.

Also, Gustavo adds another one in his per-crtc out-fence patches. And
I want to add more stuff here for nonblocking commit helpers.

In the future we can use this to store a pointer to the preceeding
state, making an atomic update entirely free-standing. This will be
needed to be able to queue them up with a depth > 1.

Cc: Gustavo Padovan <gustavo@padovan.org>
Reviewed-by: default avatarMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1464818821-5736-12-git-send-email-daniel.vetter@ffwll.ch
      5d943aa6
    • Daniel Vetter's avatar
      drm: Consolidate plane arrays in drm_atomic_state · b8b5342b
      Daniel Vetter authored 
      

It's kinda pointless to have 2 separate mallocs for these. And when we
add more per-plane state in the future it's even more pointless.

Right now there's no such thing planned, but both Gustavo's per-crtc
fence patches, and some nonblocking commit helpers I'm playing around
with will add more per-crtc stuff. It makes sense to also consolidate
planes, just for consistency.

In the future we can use this to store a pointer to the preceeding
state, making an atomic update entirely free-standing. This will be
needed to be able to queue them up with a depth > 1.

Cc: Gustavo Padovan <gustavo@padovan.org>
Reviewed-by: default avatarMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1464818821-5736-11-git-send-email-daniel.vetter@ffwll.ch
      b8b5342b
    • Daniel Vetter's avatar
      drm: Consolidate connector arrays in drm_atomic_state · 63e83c1d
      Daniel Vetter authored 
      

It's kinda pointless to have 2 separate mallocs for these. And when we
add more per-connector state in the future it's even more pointless.

Right now there's no such thing planned, but both Gustavo's per-crtc
fence patches, and some nonblocking commit helpers I'm playing around
with will add more per-crtc stuff. It makes sense to also consolidate
connectors, just for consistency.

In the future we can use this to store a pointer to the preceeding
state, making an atomic update entirely free-standing. This will be
needed to be able to queue them up with a depth > 1.

Cc: Gustavo Padovan <gustavo@padovan.org>
Reviewed-by: default avatarMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1464818821-5736-10-git-send-email-daniel.vetter@ffwll.ch
      63e83c1d
    • Daniel Vetter's avatar
      drm/atomic: Add drm_atomic_crtc_state_for_each_plane_state · 2f196b7c
      Daniel Vetter authored 
      

... and use it in msm&vc4. Again just want to encapsulate
drm_atomic_state internals a bit.

The const threading is a bit awkward in vc4 since C sucks, but I still
think it's worth to enforce this. Eventually I want to make all the
obj->state pointers const too, but that's a lot more work ...

v2: Provide safe macro to wrap up the unsafe helper better, suggested
by Maarten.

v3: Fixup subject (Maarten) and spelling fixes (Eric Engestrom).

Cc: Eric Anholt <eric@anholt.net>
Cc: Rob Clark <robdclark@gmail.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: default avatarMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1464877304-4213-1-git-send-email-daniel.vetter@ffwll.ch
      2f196b7c
    • Daniel Vetter's avatar
      drm/doc: Appease sphinx · 2e7a5701
      Daniel Vetter authored 
      

Mostly this is unexpected indents. But really it's just a
demonstration for my patch, all these issues have been found&fixed
using the correct source file and line number support I just added.
All line numbers have been perfectly accurate.

One issue looked a bit fishy in intel_lrc.c, where I don't quite grok
what sphinx is unhappy about. But since that file looks like it has
never seen a proper kernel-doc parser I figured better to fix in a
separate path.

v2: Use fancy new &drm_device->struct_mutex linking (Jani).

Cc: Jani Nikula <jani.nikula@intel.com>
Cc: linux-doc@vger.kernel.org
Cc: Jonathan Corbet <corbet@lwn.net>
Acked-by: default avatarJani Nikula <jani.nikula@intel.com>
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@intel.com>
      2e7a5701
    • Daniel Vetter's avatar
      drm/exynos: Use for_each_crtc_in_state · 8e01550e
      Daniel Vetter authored 
      

We want to hide drm_atomic_state internals better.

v2: Use drm_crtc_mask (Maarten).

Cc: Inki Dae <inki.dae@samsung.com>
Acked-by: default avatarInki Dae <inki.dae@samsung.com>
Reviewed-by: default avatarMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1464818821-5736-8-git-send-email-daniel.vetter@ffwll.ch
      8e01550e
    • Daniel Vetter's avatar
      drm/omap: Use for_each_plane_in_state · 82072573
      Daniel Vetter authored 
      

We want to hide drm_atomic_stat internals a bit better.

v2: Use drm_crtc_mask (Maarten).

Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: default avatarMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1464818821-5736-7-git-send-email-daniel.vetter@ffwll.ch
      82072573
    • Daniel Vetter's avatar
      drm/vc4: Use for_each_plane_in_state · 833cd78a
      Daniel Vetter authored 
      

We want to hide drm_atomic_stat internals a bit better.

Cc: Eric Anholt <eric@anholt.net>
Reviewed-by: default avatarMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1464818821-5736-6-git-send-email-daniel.vetter@ffwll.ch
      833cd78a
    • Daniel Vetter's avatar
      drm/rcar-du: Use for_each_*_in_state · 09ad807b
      Daniel Vetter authored 
      

We want to hide drm_atomic_state internals better.

v2: Use drm_crtc_mask (Maarten).

Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: default avatarMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1464818821-5736-5-git-send-email-daniel.vetter@ffwll.ch
      09ad807b
    • Daniel Vetter's avatar
      drm/msm: Use for_each_*_in_state · 8d76b79f
      Daniel Vetter authored 
      

We want to hide drm_atomic_state internals

v2: Review from Maarten:
- remove whitespace change in rockchip driver that slipped in.
- use drm_crtc_mask insted of open-coding it.

Cc: Rob Clark <robdclark@gmail.com>
Reviewed-by: default avatarMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1464818821-5736-4-git-send-email-daniel.vetter@ffwll.ch
      8d76b79f
    • Daniel Vetter's avatar
      drm/i915: Use drm_atomic_get_existing_plane_state · 831655e5
      Daniel Vetter authored 
      

We want to encapsulate the drm_atomic_state internals.

v2: Use intel_atomic_get_existing_plane_state (Maarten).

Reviewed-by: default avatarMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1464818821-5736-3-git-send-email-daniel.vetter@ffwll.ch
      831655e5
    • Daniel Vetter's avatar
      drm/hdlcd: Clean up crtc hooks · 9fb3fc70
      Daniel Vetter authored 
      

Those are all no longer needed for a pure atomic driver.

Cc: Liviu Dudau <Liviu.Dudau@arm.com>
Tested-by: default avatarLiviu Dudau <Liviu.Dudau@arm.com>
Acked-by: default avatarLiviu Dudau <Liviu.Dudau@arm.com>
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1464818821-5736-17-git-send-email-daniel.vetter@ffwll.ch
      9fb3fc70
    • Daniel Vetter's avatar
      drm/atomic-helper: use for_each_*_in_state more · be9174a4
      Daniel Vetter authored 
      

This avois leaking drm_atomic_state internals into the helpers. The
only place where this still happens after this patch is drm_atomic_helper_swap_state().
It's unavoidable there, and maybe a good indicator we should actually
move that function into drm_atomic.c.

Signed-off-by: default avatarDaniel Vetter <daniel.vetter@intel.com>
Reviewed-by: default avatarMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1464818821-5736-2-git-send-email-daniel.vetter@ffwll.ch
      be9174a4
    • Chris Wilson's avatar
      drm: Only create a cmdline mode if no probed modes match · 5f0c3f99
      Chris Wilson authored 
      The intention of using video=<connector>:<mode> is primarily to select
the user's preferred resolution at startup. Currently we always create a
new mode irrespective of whether the monitor has a native mode at the
desired resolution. This has the issue that we may then select the fake
mode rather the native mode during fb_helper->inital_config() and so
if the fake mode is invalid we then end up with a loss of signal. Oops.
This invalid fake mode would also be exported to userspace, who
potentially may make the same mistake.

To avoid this issue, we filter out the added command line mode if we
detect the desired resolution (and clock if specified) amongst the
probed modes. This fixes the immediate problem of adding a duplicate
mode, but perhaps more generically we should avoid adding a GTF mode if
the monitor has an EDID that is not GTF-compatible, or similarly for
CVT.

Was meant to fix a regression from

commit eaf99c74


Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Wed Aug 6 10:08:32 2014 +0200

    drm: Perform cmdline mode parsing during connector initialisation

but Radek explained that the original bug is no longer reproducible on
latest kernels.

v2: Explicitly delete our earlier cmdline mode
v3: Mode pruning should now be sufficient to delete stale cmdline modes
v4: Compute the vrefresh for the probed mode

Reported-by: default avatarRadek Dostál <rd@radekdostal.com>
Signed-off-by: default avatarChris Wilson <chris@chris-wilson.co.uk>
Cc: Radek Dostál <rd@radekdostal.com>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Cc: Julia Lemire <jlemire@matrox.com>
Cc: Dave Airlie <airlied@redhat.com>
Reviewed-by: default avatarAlex Deucher <alexander.deucher@amd.com>
Reviewed-by: default avatarVille Syrjälä <ville.syrjala@linux.intel.com>
[danvet: Drop cc: stable since no longer a pressing bugfix, just
nice-to-have.]
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1464774651-20376-1-git-send-email-chris@chris-wilson.co.uk
      5f0c3f99
    • Christian König's avatar
      dma-buf/fence: add signal_on_any to the fence array v2 · f7104568
      Christian König authored 
      

If @signal_on_any is true the fence array signals if any fence in the array
signals, otherwise it signals when all fences in the array signal.

v2: fix signaled test and add comment suggested by Chris Wilson.

Signed-off-by: default avatarChristian König <christian.koenig@amd.com>
Reviewed-by: default avatarGustavo Padovan <gustavo.padovan@collabora.co.uk>
Acked-by: default avatarSumit Semwal <sumit.semwal@linaro.org>
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1464786612-5010-4-git-send-email-deathsimple@vodafone.de
      f7104568
    • Gustavo Padovan's avatar
      dma-buf/fence: add fence_array fences v6 · b3dfbdf2
      Gustavo Padovan authored 
      

struct fence_array inherits from struct fence and carries a
collection of fences that needs to be waited together.

It is useful to translate a sync_file to a fence to remove the complexity
of dealing with sync_files on DRM drivers. So even if there are many
fences in the sync_file that needs to waited for a commit to happen,
they all get added to the fence_collection and passed for DRM use as
a standard struct fence.

That means that no changes needed to any driver besides supporting fences.

To avoid fence_array's fence allocates a new timeline if needed (when
combining fences from different timelines).

v2: Comments by Daniel Vetter:
	- merge fence_collection_init() and fence_collection_add()
	- only add callbacks at ->enable_signalling()
	- remove fence_collection_put()
	- check for type on to_fence_collection()
	- adjust fence_is_later() and fence_later() to WARN_ON() if they
	are used with collection fences.

v3: - Initialize fence_cb.node at fence init.

    Comments by Chris Wilson:
	- return "unbound" on fence_collection_get_timeline_name()
	- don't stop adding callbacks if one fails
	- remove redundant !! on fence_collection_enable_signaling()
	- remove redundant () on fence_collection_signaled
	- use fence_default_wait() instead

v4 (chk): Rework, simplification and cleanup:
	- Drop FENCE_NO_CONTEXT handling, always allocate a context.
	- Rename to fence_array.
	- Return fixed driver name.
	- Register only one callback at a time.
	- Document that create function takes ownership of array.

v5 (chk): More work and fixes:
	- Avoid deadlocks by adding all callbacks at once again.
	- Stop trying to remove the callbacks.
	- Provide context and sequence number for the array fence.

v6 (chk): Fixes found during testing
	- Fix stupid typo in _enable_signaling().

Signed-off-by: default avatarGustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: default avatarChristian König <christian.koenig@amd.com>
Reviewed-by: default avatarGustavo Padovan <gustavo.padovan@collabora.co.uk>
Acked-by: default avatarSumit Semwal <sumit.semwal@linaro.org>
[danvet: Improve commit message as suggested by Gustavo.]
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1464786612-5010-3-git-send-email-deathsimple@vodafone.de
      b3dfbdf2