Walking the format/modifier list to try to find out if our FB is
compatible with the plane is surprisingly expensive. Since the plane's
capabilities are static over the lifetime of the KMS device, cache the
set of planes for which the FB is theoretically
format/modifier-compatible when it's created, and use that to do an
early cull of the set of acceptable planes.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Currently each drm_fb takes a reference on a client buffer it wraps.
This prevents us from being able to reuse a drm_fb in multiple places
(e.g. two views of the same client buffer) simultaneously, or even back
to back.
Move the buffer reference to the plane state, as preparation for
allowing drm_fb to be cached inside the weston_buffer.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Currently we take a reference on the underlying client buffer every time
we materialise a drm_fb from a view, and release it when the drm_fb is
destroyed. This means that we need to create and destroy a drm_fb every
time we want to use it, which is pathologically unperformant on some
platforms.
To start working towards being able to cache drm_fb, only take the
reference when we apply it to a plane state.
Signed-off-by: Daniel Stone <daniels@collabora.com>
No sense walking the plane list every frame if we can't use it because
it's neither a SHM buffer nor a client buffer we can directly import as
a framebuffer.
Signed-off-by: Daniel Stone <daniels@collabora.com>
In commit "libweston: add initial dma-buf feedback implementation" we've
added initial support to dma-buf feedback, but it was not using the
feedback from DRM-backend to add useful per-surface feedback.
In this patch we add this. The scanout tranche of the per-surface
feedback is based on the union of formats/modifiers of the primary and
overlay planes available. These are intersected with the
formats/modifiers supported by the renderer device.
Also, it's important to mention that the scene can change a lot and we
can't predict much. So this patch also adds hysteresis to the dma-buf
feedback. We wait a few seconds to be sure that we reached stability
before adding or removing the scanout tranche from dma-buf feedback and
resending them. This help us to avoid spamming clients and leading to
unnecessary buffer reallocations on their end.
Here's an example of what we want to avoid:
1. We detect that a view was not placed in a plane only because its
format is not supported by the plane, so we add the scanout tranche
to the feedback and send the events.
2. A few milliseconds after, the view gets occluded. So now the view
can't be placed in a plane anymore. We need to remove the scanout
tranche and resend the feedback with formats/modifiers optimal for
the renderer device. The client will then reallocate its buffers.
3. A few milliseconds after, the view that was causing the occlusion
gets minimized. So we got back to the first situation, in which the
format of the view is not compatible with the plane. Then we need to
add a scanout tranche and resend the feedback...
This patch is based on previous work of Scott Anderson (@ascent).
Signed-off-by: Leandro Ribeiro <leandro.ribeiro@collabora.com>
Signed-off-by: Scott Anderson <scott.anderson@collabora.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Add enum try_view_on_plane_failure_reasons to help us to keep track of
the reason why promoting view to a plane failed. We also add a variable
to struct weston_paint_node so that we can update this information in
each output repaint.
This will be used in the next commits, in which we add proper surface
dma-buf feedback support.
Signed-off-by: Leandro Ribeiro <leandro.ribeiro@collabora.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
The DRM backend uses changes in the cursor view memory address and
surface damage to detect when it needs to re-upload to a cursor plane
framebuffer.
However, when a cursor view is destroyed and then recreated, e.g., when
the pointer cursor surface is updated, the newly created view may have
the same memory address as the just destroyed one. If no new cursor
buffer is provided (because it was attached, committed and used
previously) when this address reuse occurs, then there also isn't any
updated surface damage and the backend doesn't update the cursor plane
framebuffer at all.
To fix this issue utilize the destroy signal to track when the cursor
view is destroyed, and clear the cached cursor_view value in drm_output.
After clearing the cached value, the next cursor view is always
considered new and thus uploaded to the plane properly.
Signed-off-by: Alexandros Frantzis <alexandros.frantzis@collabora.com>
See: https://gitlab.freedesktop.org/wayland/weston/-/issues/467#note_814985
This starts building the framework required for implementing color
management.
The main new interface is struct weston_color_manager. This commit also
adds a no-op color manager implementation, which is used if no other
color manager is loaded. This no-op color manager simply provides
identity color transforms for everything, so that Weston keeps running
exactly like before.
weston_color_manager interface is incomplete and will be extended later.
Colorspace objects are not introduced in this commit. However, when
client content colorspace and output colorspace definitions are
combined, they will produce color transformations from client content to
output blending space and from output blending space to output space.
This commit introduces a placeholder struct for color transforms,
weston_color_transform. Objects of this type are expected to be heavy to
create and store, which is why they are designed to be shared as much as
possible, ideally making their instances unique. As color transform
description is intended to be generic in libweston core, renderers and
backends are expected to derive their own state for each transform
object as necessary. Creating and storing the derived state maybe be
expensive as well, more the reason to re-use these objects as much as
possible. E.g. GL-renderer might upload a 3D LUT into a texture and keep
the texture around. DRM-backend might create a KMS blob for a LUT and
keep that around.
As a color transform depends on both the surface and the output, a
transform object may need to be created for each unique pair of them.
Therefore color transforms are referenced from weston_paint_node. As
paint nodes exist for not just surface+output but surface+view+output
triplets, the code ensures that all paint nodes (having different view)
for the same surface+output have the same color transform state.
As a special case, if weston_color_transform is NULL, it means identity
transform. This short-circuits some checks and memory allocations, but
it does mean we use a separate member on weston_paint_node to know if
the color transform has been initialized or not.
Color transformations are pre-created at the weston_output
paint_node_z_order_list creation step. Currently the z order lists
contain all views globally, which means we populate color transforms we
may never need, e.g. a view is never shown on a particular output.
This problem should get fixed naturally when z order lists are
constructed "pruned" in the future: to contain only those paint nodes
that actually contribute to the output's image.
As nothing actually supports color transforms yet, both renderers and
the DRM-backend assert that they only get identity transforms. This
check has the side-effect that all surface-output pairs actually get a
weston_surface_color_transform_ref even though it points to NULL
weston_color_transform.
This design is inspired by Sebastian Wick's Weston color management
work.
Co-authored-by: Sebastian Wick <sebastian@sebastianwick.net>
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Iterate paint nodes instead of the global view list. Right now this does
not change behavior.
This is a step towards using per-output view lists that can then be
optimized for the output in libweston core.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
If sprites_are_broken, then we will only ever arrive in renderer_only
mode, so this case will be caught by the checks above.
Signed-off-by: Daniel Stone <daniels@collabora.com>
fb->format is *pixel_format_info, whereas fb->format->format is the
actual DRM/wl_shm format code we want to see here. Fix the drm_debug()
call accordingly.
Signed-off-by: Bastian Krause <bst@pengutronix.de>
In commit "backend-drm: do not import dmabuf buffers with no modifiers
to KMS" we've stopped to import dmabuf with no modifiers to KMS.
In this patch we document why we can still import wl_drm buffers with no
modifiers to KMS.
Signed-off-by: Leandro Ribeiro <leandro.ribeiro@collabora.com>
In commit "libweston: add struct weston_drm_format" struct
weston_drm_format and its helper functions were added to libweston.
Also, unit tests for this API have been added in commit "tests: add unit
tests for struct weston_drm_format".
Start to use this API in the DRM-backend, as it enhances the code by
avoiding repetition and ensuring correctness.
Signed-off-by: Scott Anderson <scott.anderson@collabora.com>
Co-authored-by: Leandro Ribeiro <leandro.ribeiro@collabora.com>
Signed-off-by: Leandro Ribeiro <leandro.ribeiro@collabora.com>
In drm_output_prepare_plane_view() we iterate the list of planes and add
them as candidates to promote the view to one of them.
Cursor planes do not support buffers other than wl_shm (at least for
now). So when we have a dmabuf or an EGL buffer in the view, the
function drm_output_plane_cursor_has_valid_format() returns false and
the cursor plane is not added to the candidate list.
In this patch we move the responsibility of doing this from
drm_output_plane_cursor_has_valid_format() to
drm_output_prepare_plane_view() itself, as the incompatibility between
other types of buffers and cursor planes is different from the
incompatibility between formats. This makes the code easier to read
and also documents the current incompatibility between cursor planes
and buffers that are not created through wl_shm.
Signed-off-by: Leandro Ribeiro <leandro.ribeiro@collabora.com>
In some situations, like positioning a sub-surface that exceeds the
output's dimensions we would adjust the plane state dimensions to some
lower values to that of the buffer. That would ultimately trip the cursor
update function because the buffer itself actually exceeds the maximum
size/dimension of the cursor.
The plane state destination co-ordinates is the area of the view which
is visible on the output, which in some situations, would actually be
smaller than the original buffer dimensions (making it so that it will
pass the cropping/scaling check), but depending on of
how large is the surface buffer, it would tripping the assert wrt to
cursor width/height dimensions.
This hasn't been seen so far due to the fact that until recently we had
a cursor surface that always reached the cursor plane and that was
already being set-up by default (with desktop-shell, which is no longer
the case), and also because kiosk-shell, which doesn't set-up a cursor
surface, was not available.
This adds a check to skip placing the view in the cursor plane if the
buffer dimensions exceed the cursor permitted width/height.
(Suggested-by Daniel Stone).
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
There are some places where we can make some cosmetic changes
to make code simpler and easier to read. Make these cosmetic
changes. Note that they do not change the code behavior.
Signed-off-by: Leandro Ribeiro <leandro.ribeiro@collabora.com>
Try to make drm_output_state_propose a little more clear by reducing
divergence between plane and renderer modes towards the end, removing
a possibly-surprising conditional continue.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Reduce the scope of surface_overlap to where it's actually used, which
is only in the per-view loop, where it gets initialised and destroyed
every time.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Previously we assumed that cursor planes occluded nothing and would
always be blended, but overlay and scanout planes would always occlude
what's behind them. This is not actually true, as we can support alpha
blending on any kind of plane type now.
Remove the special case, which might hopefully fix some weird display
issues along the way. (Noticed by inspection.)
Signed-off-by: Daniel Stone <daniels@collabora.com>
We used to use planes_region for the output regions which were being
displayed on hardware planes; before we grew zpos awareness, we couldn't
have any planes overlapping with each other, since the ordering would be
undefined.
Since the zpos awareness though, this region is unused, so we can just
remove it.
Signed-off-by: Daniel Stone <daniels@collabora.com>
If a surface is not visible, then is does not matter if the view is on multiple
outputs. It will be skipped anyways when the output is rendered.
So check first if the surface is acually visible on the output before doing any
checks that might force rendering. This avoids unnecessary rendering.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
Mode change from mixed-mode to renderer-only means we should no longer
try to place views in HW planes (as we composite everything into the
primary plane) thus we should avoid that whenever that happens.
In the same time we need to be able to place in mixed-mode/renderer-only
mode the cursor view into the cursor plane (if one is available).
This patch adds a further check to skip plane assignment when disabling
overlay support (when we switch to renderer-only mode), when drivers do
not have atomic-modeset or it has been disabled intentionally.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Adds a further assert() to make sure we're not checking against invalid
values. This was seen in the wild when the kernel rejects the commit for
overlay resulting in a check for invalid zpos values.
Fixes: #304
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
when switching to mixed-mode of compositing
This way we avoid an (incorrect) duplicate check of zpos values. Also,
this would be needed because the renderer needs have the lowest zpos value
available as we don't (yet) properly support underlays, the primary
plane serves as our renderer.
Adds also a check to see if we try to assign a view to a plane with
a lower zpos value than the one assigned to the primary when switching
to mixed-mode of compositing.
Fixes: #304
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
It makes much more sense to be there. It adds some additional drm_debug()
statements to provide reason for failing to place the view in the HW
plane. Makes the reason for failing more accurate.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Suggested-by: Daniel Stone <daniel.stone@collabora.com>
Avoids the need to retrieve the DRM framebuffer in each function and
re-uses the one got before constructing the zpos candidate list.
Takes another reference for the scanout as to live the state, like
there's one for the overlay bit.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
As we already have a potential plane available to use, pass it
over the _prepare_overlay_view instead of trying to find one
from the backend plane list.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Suggested-by: Daniel Stone <daniel.stone@collabora.com>
In this manner we will allow views to reach the overlay (or underlays)
even if the damage tracking will detect that the new view will
occlude the view underneath it.
Renames occluded_region to planes_region, and uses occluded_region
to represent the region where we add each view's visible-and-opaque region.
Sprinkle some comments about each region.
Re-uses the view's clipped region to determine visible-and-opaque region
which is accumulated (for both renderer and HW planes cases) into
occluded_region. The current view's clipped_region is then checked against
occluded_region.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Suggested-by: Daniel Stone <daniel.stone@collabora.com>
We can determine if the pixel format used by the clients buffer is
scan-out capable much sooner, so do it when constructing the zpos
candidate list. It also removes the checks in their respective
prepare_ functions.
Avoids the situation where we'd need to retrieve the DRM framebuffer each time
when checking the pixel format.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
function
The idea is to place pixel the format checks in a common part and until
then, to make it available as a function so we can re-use easily.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
In order to better optimize view assignment to HW planes, we construct
an intermediary zpos candidate list which is used aggregate all suitable
planes for handling scan-out capable client buffers.
We go over it twice: once to construct it and once to pick-and-choose a
suitable plane based its highest zpos position.
In order to maintain the view order correctly we track current zpos
value being applied to the plane state and use it when trying to place
a view into a plane.
Pass the computed zpos value to be applied to the plane state.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Suggested-by: Daniel Stone <daniel.stone@collabora.com>
Due to an error in driving GitLab, this commit erroneously contained the
entirety of !267 (zpos support in the KMS backend) squashed into one
single commit, pushed into master.
In order to keep the history clean, this is being reverted; a rebased
version of !267 with the clear individual commits which were already
present will be applied in its place.
This reverts commit 95e3b0deae.
Make GBM optional in case GL renderer is disabled. This allows to
build Weston with DRM backend without Mesa dependencies.
Signed-off-by: Stefan Agner <stefan@agner.ch>
Currently drm-layer supports HDCP1.4 using connector property:
Content Protection. This property if available for a platform, can be
read and set for requesting content-protection.
Also, the patch series [1] adds HDCP2.2 support in drm, and patch [2]
adds support to send udev events for change in connector properties,
made by the kernel.
This patch adds these HDCP connector properties in weston, and exposes
the content-protection support to the client for drm-backend.
It adds the enums to represent 'Content Protection' and 'Content Type'
connector properties exposed by drm layer. It adds a member
'protection' in drm_output_state, to store the desired protection
from the weston_output in the drm-backend output-repaint cycle. This
is then used to write the HDCP connector properties for the drm_heads
attached to the drm_output.
The kernel sends uevents to the user-space for any change made by it
in the "Content Protection" connector property. No event is sent in
case of change in the property made by the user-space.
It means, when there is a change of the property value from "DESIRED"
to "ENABLE" i.e. successful authentication by the kernel, a uevent
will be generated, but in case of userspace requesting for disabling
the protection by writing "UNDESIRED" into the property, no uevent
will be generated.
This patch also adds support for handling new udev events for HDCP
connector property changes. Any such change, triggers change in the
weston_head's current_protection.
[1] https://patchwork.freedesktop.org/series/57233/#rev7
[2] https://patchwork.freedesktop.org/patch/303903/?series=57233&rev=7
Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
drm_assign_planes() is called to separate views out and decide what will
be taken out for plane composition and what will be left for the
renderer to compose.
It calls drm_output_propose_state() in order to find a good
configuration, which itself has a number of helpers that it calls. Break
these out into a separate file.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Daniel Stone
Leandro Ribeiro
Emmanuel Gil Peyrot
Alexandros Frantzis
Pekka Paalanen
Bastian Krause
Scott Anderson
Marius Vlad
Maxime Roussin-Bélanger
Michael Olbrich
Marius Vlad
Stefan Agner
Ankit Nautiyal