Previously, inverse_evaluate_lut1d was called 3 * len^3 times. Now it is
called only 3 * len times with the help of pre-computed arrays for red
and green channels. Blue channel does not need an array, because there
were no redundant computations. This is a significant win.
Also, allocate a temporary array rgb_in, so cmsDoTransform() can be
called in batches of len triplets. This seemed to be not that big win. I
tried running cmsDoTransform() over len^2 triplets, but that did not
seem to improve performance.
The test I used creates two 3D LUTs, hence two times for a single run.
My representative timing test results per one 3D LUT:
- before: 16 ms and 19 ms
- after: 7 ms and 10 ms
The measurements were done with this patch:
static bool
xform_to_shaper_plus_3dlut(struct weston_color_transform *xform_base,
uint32_t len_shaper, float *shaper,
uint32_t len_lut3d, float *lut3d)
{
struct cmlcms_color_transform *xform = to_cmlcms_xform(xform_base);
struct weston_compositor *compositor = xform_base->cm->compositor;
bool ret;
+ struct timespec begin, end;
+ unsigned i;
- ret = build_shaper(xform->lcms_ctx, xform->cmap_3dlut,
+ clock_gettime(CLOCK_MONOTONIC, &begin);
+ for (i = 0; i < 100; i++)
+ ret = build_shaper(xform->lcms_ctx, xform->cmap_3dlut,
len_shaper, shaper);
if (!ret)
return false;
- ret = build_3d_lut(compositor, xform->cmap_3dlut,
+ for (i = 0; i < 100; i++)
+ ret = build_3d_lut(compositor, xform->cmap_3dlut,
len_shaper, shaper, len_lut3d, lut3d);
if (!ret)
return false;
+ clock_gettime(CLOCK_MONOTONIC, &end);
+ fprintf(stderr, "%s: %" PRId64 " ms\n", __func__, timespec_sub_to_msec(&end, &begin));
return true;
}
Using this command:
$ ./tests/test-color-icc-output -f 8 opaque_pixel_conversion
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
This removes the API impedance mismatch between
weston_color_curve_sample() and weston_color_curve_to_3x1D_LUT() that
was temporarily introduced in the previous commit. That mismatch is now
limited to gl_color_curve_lut_3x1d().
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Future development will need to evaluate pipelines with curves and
matrices. Such pipelines naturally operate on vec3, as matrices cannot
be operated one channel at a time. Make weston_color_curve_sample()
operate on arrays of vec3.
Its currently only caller, weston_color_curve_to_3x1D_LUT(), is modified
to employ a temporary array for the API impedance mismatch. This
workaround will be removed later as weston_color_curve_to_3x1D_LUT()
itself will be converted to operate on vec3 arrays.
weston_v3f_array_to_planar() documentation was generated with AI.
weston_color_curve_sample() is restructured a little bit, attempting to
make it simpler to read.
color-operations.h gets the #includes needed to make it self-standing.
Assisted-by: Github Copilot (Claude Sonnet 3.5)
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
We can handle the mirroring for negative input without conditional this
way. Maybe it's faster, maybe it's not, but I like this style better.
In color-operations we need only one call to the elementary function
rather than two. This helps a lot with code clarity, when we get a vec3
at a time to handle in the future, doing three trivial calls instead of
six with an if-else hassle.
fragment.glsl sheds one layer of function wrapping, which is nice.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
For some cases such as remote control, need to disable the interaction
between input device and the weston. It will not affect the use of input
device by other modules or applications.
Signed-off-by: Elliot Chen <elliot.chen@nxp.com>
since abfe874a51 we no longer do a full view list rebuild after a
surface_set_size(), but that was ensuring we always had an up to
date transform in the repaint loop.
Add a weston_view_update_transform() here to dodge the assert in
the repaint loop.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
This was missed in e169b77430
when leaving planes in the scene graph meant we no longer needed
to calculate this clip.
It's just a few wasted cycles, so no backport required.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Instead of clipping the visible region to the output, clip the entire
region to the output first and save that, then create the visible region
from that.
Now we have both the clipped and visible view regions that the backends
may want to do plane assignment stored in paint nodes, so we can save
some duplicate math.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Assign planes in the future could benefit from visibility information on
pnodes being up to date.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
This is a step towards having correct visibility information in
assign_planes. Instead of visible_next and visible, we now have
visible_previous and visible, with visible being set up in the
visibility update function instead of being copied across in
the late update.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
In the future I plan to move the paint node visibility calculation, and
in doing so the different callers to paint_node_damage_below() will have
different visibility regions.
Add the region as a parameter.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Now that we're validating that buffer size must be an integer multiple of
scale, add some tests to make sure it's happening.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
We're supposed to generate an INVALID_SIZE error if the buffer size isn't
an integer multiple of the scale.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Now that we use single pixel buffer, we need to be careful to only use
scale of 1, as other scales are supposed to generate an INVALID_SIZE
protocol errors (because the buffer size is not an integer multiple
of the scale).
Right now a bad scale just breaks background repaints, but in the future
we'll properly validate scale, which would cause the shell client to
be disconnected.
Fixes: 6cc8f48cd ("clients: Paint desktop-shell color background with single-pixel-buffer")
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Allow dictating which color format we'd like to use. This introduces the
front-end side and the core parts, leaving the EDID parsing and DRM
connector property for later patches.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
When we're going through assign_planes and repaint, give the backend an
opportunity to see what's changed during this repaint.
Signed-off-by: Daniel Stone <daniels@collabora.com>
If a surface changes size or opaque region, or a view changes position,
we don't need to rebuild the entire view list from scratch: we can just
rebuild what's changed and nothing else.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Nothing uses this, and it was also being set incorrectly, as it would
return true when there was no damage to actually be flushed.
Signed-off-by: Daniel Stone <daniels@collabora.com>
In order to streamline our checks and to make it easier to ensure
that we replace paint nodes with solid placeholders correctly.
Fix the failure reason in one case while on it.
Signed-off-by: Robert Mader <robert.mader@collabora.com>
In order to be more in line with weston_view_is_opaque() - which we
use in most cases to set the value - and ensure the expectations of
existing places in the code are honored, see usages in both
draw_paint_node() implementations.
An exception here are holes - these always need to get painted as fully
opaque in the renderer path.
Signed-off-by: Robert Mader <robert.mader@collabora.com>
I heard rumoours this is mostly unnecessary to check nowdays since it is
the default.
The real reason is that it crashes sometimes:
$ ci-fairy check-merge-request --require-allow-collaboration --junit-xml=results.xml
Traceback (most recent call last):
File "/usr/lib/python3.12/site-packages/gitlab/exceptions.py", line 344, in wrapped_f
return f(*args, **kwargs)
^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3.12/site-packages/gitlab/mixins.py", line 125, in get
server_data = self.gitlab.http_get(path, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3.12/site-packages/gitlab/client.py", line 813, in http_get
result = self.http_request(
^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3.12/site-packages/gitlab/client.py", line 779, in http_request
raise gitlab.exceptions.GitlabHttpError(
gitlab.exceptions.GitlabHttpError: 403: 403 Forbidden
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/usr/bin/ci-fairy", line 8, in <module>
sys.exit(ci_fairy())
^^^^^^^^^^
File "/usr/lib/python3.12/site-packages/click/core.py", line 1161, in __call__
return self.main(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3.12/site-packages/click/core.py", line 1082, in main
rv = self.invoke(ctx)
^^^^^^^^^^^^^^^^
File "/usr/lib/python3.12/site-packages/click/core.py", line 1697, in invoke
return _process_result(sub_ctx.command.invoke(sub_ctx))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3.12/site-packages/click/core.py", line 1443, in invoke
return ctx.invoke(self.callback, **ctx.params)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3.12/site-packages/click/core.py", line 788, in invoke
return __callback(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3.12/site-packages/click/decorators.py", line 33, in new_func
return f(get_current_context(), *args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3.12/site-packages/ci_fairy.py", line 1795, in check_merge_request
mr = p.mergerequests.get(merge_request_iid)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3.12/site-packages/gitlab/v4/objects/merge_requests.py", line 523, in get
return cast(ProjectMergeRequest, super().get(id=id, lazy=lazy, **kwargs))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3.12/site-packages/gitlab/exceptions.py", line 346, in wrapped_f
raise error(e.error_message, e.response_code, e.response_body) from e
gitlab.exceptions.GitlabGetError: 403: 403 Forbidden
Observed in https://gitlab.freedesktop.org/elliot_chen/weston/-/jobs/86174879
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
This should be checking for a valid transform so it doesn't use corners
of an axis aligned bounding box for free-form transformed views.
It should still check surf_xform for color management, but it should only
do that when surf_xform_valid is true.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
This test is critical, so only do it once, and throw some asserts around
to make sure we don't mess it up.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Video underlay "holes" are fully transparent but must be rendered
fully opaque. However, they appear to be fully transparent to
the current gl-renderer test which is missing a need_hole check.
Add a paint node attribute for this so it can be more easily
checked in assign_planes and renderers.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Remove maybe_replace_paint_node entirely and place the parts of it
properly between early and late updates.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
We'll need this in both the early and late paint node updates soon,
so this saves a couple of lines of code.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
We're going to assert on a bunch of stuff, so let's put it in a function
to remove clutter.
For now it's trivial, but I'll be adding more tests shortly.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
It's the same as buffer->direct_display essentially 100% of the time,
except maybe if someone set weston_direct on a single-pixel-buffer, but
that's madness.
Just drop it entirely, and let the only existing reader of the variable
get it from the buffer directly.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
At the moment this is only replacing two clear calls with one. However,
when client_buffer starts using properly-bracketed CPU access, this will
become much more tedious. Introduce a helper now to make it easier.
Signed-off-by: Daniel Stone <daniels@collabora.com>
We always have the name of a reference image to compare to; comparison
is kind of impossible without one.
Signed-off-by: Daniel Stone <daniels@collabora.com>
It's pretty trivial to make verify_screen_content() and
capture_screenshot_from_output() support decorations, so we can reuse
those in output-decorations tests rather than open-coding.
Signed-off-by: Daniel Stone <daniels@collabora.com>
get_middle_row() is a pure CPU accessor which only needs to operate on a
pixman_image_t. Pass this directly instead of struct buffer.
Signed-off-by: Daniel Stone <daniels@collabora.com>
A very marginal benefit, but why not. Any tests using screen capture are
now using a refresh rate of 0 (redraw immediately on capture request),
whilst others are using HIGHEST_OUTPUT_REFRESH.
Signed-off-by: Daniel Stone <daniels@collabora.com>
When we fail out because we couldn't even start a container or bring the
runner environment up, just retry the job.
Signed-off-by: Daniel Stone <daniels@collabora.com>
