Storing GL function pointers along with their associated extension
name allows to better track the function pointers declared.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Now that the GL extensions are stored in the gl_extensions bitfield,
there's no need to use booleans anymore.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
This commit ensures GL_OES_EGL_image is available before setting up
dma-buf renderer functions because it's not implied by the presence of
the EGL_EXT_image_dma_buf_import extension.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
GL_OES_texture_3D isn't used because of the complexity implied for
correct OpenGL ES 2 support (see commit 734c2278), so there's no need
to check for it. The check also now avoids checking for
GL_EXT_color_buffer_half_float on OpenGL ES 3.2 which includes support
for 16-bit FP renderbuffers by default.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
The GL renderer supports OpenGL ES from version 2.0 to 3.2. It's meant
to correctly link on systems with only OpenGL ES 2, so OpenGL ES 3
symbols can't be used directly. eglGetProcAddress() is currently
defined to not support the querying of non-extension OpenGL ES
functions. Support for that is exposed by the
KHR_get_all_proc_addresses extension. This commit ensures OpenGL ES 3
function addresses can be retrieved by eglGetProcAddress() by checking
for that extension at run-time.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
A new gl_extensions bitfield is added to the renderer to store the
flags supported by the GL implementation. This allows the GL renderer
to easily check for extension support whenever needed.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Get the EGL_WL_bind_wayland_display extension function addresses after
checking for extension availability.
Let the EGL display setup end before trying to bind the Wayland
display in gl-renderer.c. The EGL features log will now report the
extension as supported even if the bind subsequently fails, in which
case a warning is logged.
This commit also avoids calling query_buffer() if the display isn't
bound in fill_buffer_info(), it would otherwise fail or even segfault
if the EGL_WL_bind_wayland_display extension isn't available.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
This feature flag is for explicit sync support.
We replace the explicit sync warning by logging supported fence syncs
along with the report EGL features report.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Get eglCreateImageKHR() and eglDestroyImageKHR() function addresses
depending on EGL_KHR_image_base availability.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
This second feature flag ensures that either the
EGL_EXT_swap_buffers_with_damage or EGL_EXT_swap_buffers_with_damage
extension is available.
Some function pointer addresses are currently retrieved depending on
the availabilty of their associated extension and some others are
retrieved unconditionally. For consistency reasons, we'll try from now
on in this commit set to first load function pointers depending on the
associated extension availability and then flag features once all the
addresses are retrieved.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
This commit introduces feature flags. While an extension flag only
ensures the availability of an extension at run-time, a feature flag
ensures the availability of a minimal OpenGL ES version and/or
extensions in order to easily check for the availability of a specific
feature at run-time.
This first feature ensures the availability of either the
EGL_KHR_no_config_context or EGL_MESA_configless_context extensions.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
A new GET_PROC_ADDRESS() macro is added to get a function address at
run-time using eglGetProcAddress() and to assert() the address isn't
NULL at once.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Storing EGL function pointers along with their associated extension
name allows to better track the function pointers declared.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Now that the EGL extensions are stored in the egl_*_extensions
bitfields, there's no need to use booleans anymore.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
The egl_display_extensions bitfield is added to store the display
extensions supported by the EGL implementation.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
The egl_device_extensions bitfield is added to store the device
extensions supported by the EGL implementation.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
A new egl_client_extensions bitfield is added to the renderer to store
the client flags supported by the EGL implementation. This allows the
GL renderer to easily check for extension support whenever needed.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
This new utility parses extension strings and fills a bitfield of
matched flags. This will be used to store EGL and GL extension flags.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
As weston_windowed_output_get_api needs ARRAY_LENGTH() move helpers to a
libweston/ so other users can re-use that instead of re-defining these
all over. Easier for other front-ends to make use of them.
With this change windowed-output-api.h also includes the helpers header.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Revert active unit to default value right after use so that other
functions can assume the default state. A best practices section is
added to the internal header for reference.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Provide a fixed allocation to each texture unit in order to prevent
conflicts. This fixes a conflict between colour transforms and the
wireframe debug mode.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
The GL renderer is able to use a gbm_device to allocate gbm_bos, which can be
used as DMABUFs.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Support importing dmabuf buffers as renderbuffers and binding them to
FBOs. These can then be rendered to directly, or they can be blitted
into from the shadow render buffer.
How to best create those dmabuf buffers in the backend is an open
question and may vary depending on what external API the backend is
interfacing with.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
The batches debug mode tints each batch of a repaint pass in a
different color in order to highlight batches.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
The wireframe debug mode needs to clear the current renderbuffer in
order to clean up old wireframes lying around. This commit makes this
system generic for upcoming debug modes with similar needs.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
In order to avoid the complexity of handling multiple debug modes at
the same time, this commit makes wireframe a proper debug mode. It
also uses the new tinting system to make the white wireframe pop over
a darkened damaged region.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Don't color sub-mesh differently in wireframe mode. There's not much
interest in being able to distinguish the sub-meshes in a paint node
and this allows to simplify the logic by removing the color vertex
stream.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Current green tint support is a hard-coded Porter-Duff premultiplied
blending of a green source (slightly skewed on the green channel) over
the updated damaged destination. This commit makes green tinting
generic by letting the shader user provide a custom tint color.
The goal is to reuse that system for the upcoming debug modes.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Set up debug mode initial infrastructure using a dedicated key binding
and make shaders debug a debug mode.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Render wireframe within paint nodes instead of drawing lines in a
second pass. The wireframe is blended over the node in a single draw
call. This slightly simplifies the logic by removing the computation
of a second set of indices and enables wireframe anti-aliasing using
Celes and Abraham's "Fast and versatile texture-based wireframe
rendering" paper from 2011.
Celes and Abraham use a one-dimensional set of texture coords for each
triangle edge, 1.0 for the 2 vertices defining the edge and 0.0 for
the other vertex, which basically define barycentric coords. Texture
mapping and the mip chain is then exploited to give a constant-width
edge. The main drawback of the technique is that contour edges of
node's damage mesh are drawn half as thick as interior lines since
each triangle draws half of each line's thickness.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
This makes the code slightly easier to read and prevents using
incorrect locations now that shader permutations can provide different
vertex streams.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Index vertices from the damage mesh as lines and emit a single draw
call in fan debug mode. A new shader path and an additional vertex
stream are added in order to filter the color of the solid shader
variant per sub-mesh.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
A paint node with 'n' rects damaged by 'm' quads emits 'n*m' OpenGL
draw calls. This commit batches the 'n*m' clipped polygons into an
indexed triangle strip damage mesh using degenerate triangles. A
single draw call per paint node is emitted to reduce API overhead.
Fan debug mode is disabled for now and will be added back using
batching in the next commits.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Until now, all the curves would be represented with 3x1D LUT's. Now we
support LINPOW and POWLIN curves (arbitrary names that we've picked).
We can use these curves to represent LittleCMS curves type 1, 4 and
their inverses -1, -4. The reason why we want that is because we gain
precision using the parametric curves (compared to the LUT's);
Surprisingly we had to increase the tolerance of the sRGB->adobeRGB MAT
test. Our analysis is that the inverse EOTF power-law curve with
exponent 1.0 / 2.2 amplifies errors more than the LUT, specially for
input (optical) values closer to zero.
That makes sense, because this curve is more sensible to input values
closer to zero (i.e. little input variation results in lots of output
variation). And this model makes sense, as humans are more capable of
perceiving changes of light intensity in the dark.
But the downside of all that is that for input values closer to zero, a
little bit of noise increases significantly the error.
Signed-off-by: Leandro Ribeiro <leandro.ribeiro@collabora.com>
In the next commit we'll add support for more color curves. So move
lut_3x1d to an union.
Signed-off-by: Leandro Ribeiro <leandro.ribeiro@collabora.com>
Plug async read back support to OpenGL ES 2 implementations using
GL_NV_pixel_buffer_object, GL_OES_mapbuffer extensions and
GL_EXT_map_buffer_range.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
ReadPixels() implies a synchronous read back of the render buffer to
return pixel data. OpenGL ES 3 adds asynchronous read back support by
writing the pixel data into a dedicated buffer object. This commit
adds asynchronous read back support to the output capture code. It
spawns a read back request and schedules a timeout a few frames later
in order to store the pixels into the client SHM buffer.
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Let the graphics hardware handle the transformation from surface
position to texture coordinates. Paint nodes now have a single vertex
position attribute from which texture coordinates are derived. A new
vertex shader variant handles the transformation.
Signed-off-by: Loïc Molinari <loic.molinari@gmail.com>
Add support for creating surfaceless outputs and rendering to FBOs.
The backend has to create FBOs with the create_fbo API and pass the
resulting weston_renderbuffer handles to repaint_output.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Output repaint uses a pair of fence syncs to profile GPU execution by
retrieving their timestamps once signalled. While the end timestamp
can be rather inaccurate in some cases (drivers reusing sync objects
from previous command buffers), the begin timestamp is never correct
because fence syncs are signalled on command buffer completion.
Get rid of the begin fence sync and use the EXT_disjoint_timer_query
extension to measure the actual repaint duration and extrapolate the
begin timestamp from the end one.
Fixes#342
Signed-off-by: Loïc Molinari <loic.molinari@gmail.com>
Use struct pixel_format_info pointers instead of uint32_t drm fourcc
values at the API surface for output and image creation.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
This services output capture tasks for the 'framebuffer' and 'full
framebuffer' pixel sources.
Both pixel sources come from the same source: the EGLSurface. The only
difference is the area. The EGLSurface contains the borders used for
output decorations, hence 'full framebuffer' is possible to capture.
We use GL_ANGLE_pack_reverse_row_order extension to make glReadPixels
return the image data in the layout we need for wl_shm buffers directly.
Without the extension we have to flip manually.
Another extension to the same effect is MESA_pack_invert, but this is
not specified for GL ES. It also uses a different token value, so it
cannot be directly substituted even if supported.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Add matrix in color.h
Matrix is used as an optimized method for
color mapping vs 3DLUT.
Nothing sets color mapping to matrix yet.
Signed-off-by: Vitaly Prosyak <vitaly.prosyak@amd.com>
