The use of stage spans for 3dluts was missed when the new cached
mappings were introduced, meaning that it would try to map and unmap
memory that might already have a cached mapping.
Vulkan does not support mapping memory multiple times, so make sure the
3dlut code also uses the cached mapping to avoid segfaults after unmap.
It's cumbersome for compositors to guard every
wlr_color_transform_ref() or wlr_color_transform_unref() call
behind a #if WLR_HAS_COLOR_MANAGEMENT. Moreover, none of the LCMS2
types are used in our public API.
Instead, always install the color.h header, and add a stub for
wlr_color_transform_init_linear_to_icc().
e08d52bb introduced a bug when rotating by 90 or 270 degrees and doing
non-uniform scaling (different scale factor on X and Y axes). The scale
factor was calculated as the ratio between the src box and the rotated
dst box. But scaling is applied after rotation, so the scale factor
should instead be the ratio between rotated src box and dst box.
Reverts 77006e5565. It turns out this
change really breaks the labwc drop-shadow implementation (which relies
on scaling up single-pixel-wide buffers).
After reverting this there's still something subtley wrong with scaling
pixel locations (shown by `weston-scaler -b`) but I can't see an obvious
way to make both weston-scaler and labwc drop-shadows work in the same
way with pixman and gles2. I'll have a harder look at this but in the
mean time reverting this patch seems to make things less broken on
average.
UNDEFINED when used as source layout means that the contents of
the underlying memory becomes undefined. This isn't what we want
here: we don't want to mutate the imported pixel data.
The Vulkan spec isn't really clear what the proper value should be
here, but after discussing with driver developers [1] it seems like
UNDEFINED isn't the right one. The recommendation is to use GENERAL
instead.
[1]: https://github.com/ValveSoftware/gamescope/issues/356
The old code to render transformed textures with pixman would run
composite over the whole output regardless of the texture size. When
rendering something small this caused a huge performance hit.
Rewrite the transform branch of render_pass_add_texture to:
- Only composite over the rectangle we're drawing to
- Generally try to make things a lot clearer and some comments
Fixes: https://gitlab.freedesktop.org/wlroots/wlroots/-/issues/3832
This change makes it possible to support both the direct srgb-format
pipeline and indirect (color-managed, or non-srgb-format) pipeline
for the same render buffer.
Fix source crop in the pixman render backend. It was being applied by
using a source offset as arguments to pixman_image_composite32(). But
this is wrong because the source crop should get applied before all the
other transforms, not after them (or at least this is how it works in
the other wlroots render backends). Instead, apply the source crop as
yet another matrix transform when we're doing transforms (Or keep it the
same as previously if there's no other transforming going on).
Closes: https://gitlab.freedesktop.org/wlroots/wlroots/-/issues/3830
The translations for flipped offsets were set the same as non-flipped
ones which was totally wrong and meant that any textures with
flipped-transforms rendered entirely outside the viewport and were
basically invisible.
Instead of having separate getters for shm formats and DMA-BUF
formats, use the same pattern as wlr_output_impl.get_primary_formats
with a single function which takes buffer caps as input.
Pre-multipled sRGB values need to be un-multiplied before conversion
to linear and then re-multiplied after. Compare shaders/texture.frag.
This fixes an issue in labwc where titlebar corners (rendered as
ARGB textures) did not match the rest of the titlebar (rendered as
a solid wlr_scene_rect).
Note: 0.17 has other instances in render/vulkan/renderer.c that need
the same fix.
The spec for GL_EXT_disjoint_timer_query says
> The GetInteger64vEXT command is required only if OpenGL ES 3.0 or later
> is not supported.
Some GLES 3.2 implementations like the proprietary mali driver on the
rk3566 based OrangePi advertise GL_EXT_disjoint_timer_query but lack
glGetInteger64vEXT. Use glGetInteger64v instead.
Saving the old context and immediately making our own context
current is a common pattern. Let's make it easier to do.
No functional change, just refactoring.
In query_modifier_support, the calloc for either or both of render_mods
and texture_mods may fail, in which case both are freed for convenience.
However, if one is non-NULL, vulkan_format_props_finish will try to free
it again.
NULL them to avoid double-free.
If drm_format_set_extend fails, we need to make sure each wlr_drm_format
is cleaned up together with the formats array. Finish the set to take
care of it.
Stop trying to maintain a per-file _POSIX_C_SOURCE. Instead,
require POSIX.1-2008 globally. A lot of core source files depend
on that already.
Some care must be taken on a few select files where we need a bit
more than POSIX. Some files need XSI extensions (_XOPEN_SOURCE) and
some files need BSD extensions (_DEFAULT_SOURCE). In both cases,
these feature test macros imply _POSIX_C_SOURCE. Make sure to not
define both these macros and _POSIX_C_SOURCE explicitly to avoid
POSIX requirement conflicts (e.g. _POSIX_C_SOURCE says POSIX.1-2001
but _XOPEN_SOURCE says POSIX.1-2008).
Additionally, there is one special case in render/vulkan/vulkan.c.
That file needs major()/minor(), and these are system-specific.
On FreeBSD, _POSIX_C_SOURCE hides system-specific symbols so we need
to make sure it's not defined for this file. On Linux, we can
explicitly include <sys/sysmacros.h> and ensure that apart from
symbols defined there the file only uses POSIX toys.
- Add POSIX 1993.09 compliance macro in source files that use
"struct timespec";
- Add POSIX 2001.12 compliance macro in source files that use
"struct sigaction" and the SA_SIGINFO macro, or the fchmod()
function;
- Add POSIX 2008.09 compliance macro in source files that use the
getline() function.
These compliance macros are enough for wlroots to compile with the
git-master version of uClibc-ng.
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Since we only use the backend capabilities here we can simply pass
them in directly. This allows other locations to create an allocator
even if they don't have a backend. They can simply specify the caps
they want instead.
Some opaque pixel formats (nv12, p010) require per-plane bytes_per_block
info. However, it doesn't make sense to store them in
wlr_pixel_format_info, as they will never be useful (currently, this
info is used for shm, which doesn't have a concept of multi-planar
buffers.)
Let's define a separate array and function for determining whether a
pixel format has alpha.
wl_drm is a legacy interface superseded by the linux-dmabuf
protocol. All clients should migrate.
As a first step, stop creating the wl_drm global by default.
This should let us discover any remaining issues in clients.
Compositors can still manually create the global if they want to.
As a second step, we can completely drop our implementation.
Kenny Levinsen
Simon Ser
David Turner
Kirill Primak
Manuel Stoeckl
John Lindgren
Biswapriyo Nath
Guido Günther
Paul Cercueil
Austin Shafer
Leo Li
Félix Poisot
Roman Gilg