Right now the Vulkan renderer blocks until the frame is complete
after rendering. This is necessary because Vulkan doesn't
interoperate well with implicit sync we use everywhere else.
Use the new kernel API to import a sync_file into a DMA-BUF to
avoid blocking.
We need to wait for the pending command buffer to complete before
re-using stage buffers. Otherwise we'll overwrite the stage buffer
with new contents before the texture is fully uploaded.
We need to wait for any pending command buffer to complete before
we're able to fully destroy a struct wlr_vk_texture: the Vulkan
spec requires the VkDescriptorSet to be kept alive.
So far we've done this in vulkan_end(), after blocking until the
command buffer completes. We'll soon stop blocking, so move this
logic in get_command_buffer(), where we check which commands buffers
have completed in a non-blocking fashion.
vkCmdCopyBufferToImage requires that the buffer offset be a multiple
of the texel block size, which for single plane uncompressed formats
is the same as the number of bytes per pixel. This commit adds an
alignment parameter to vulkan_get_stage_span which ensures that the
provided span (and the sequence of image copy operations derived which
use it) have this alignment.
Let's just forward-declare struct wlr_backend instead.
We need to fixup the Vulkan renderer: it needs makedev(), which
got included by chance via <wlr/backend.h> → <wlr/backend/session.h>
→ <libudev.h>.
We'll use this function from wlr_shm too.
Add some assertions, use int32_t (since the wire protocol uses that,
and we don't want to use 16-bit integers on exotic systems) and
switch the stride check to be overflow-safe.
Call glGetGraphicsResetStatusKHR in wlr_renderer_begin to figure
out when a GPU reset occurs. Destroy the renderer when this
happens (the OpenGL context is defunct).
Allow to get whether has alpha channel of the VkImage, it can help an
optimization to disable blending when the texture doesn't have alpha.
Because the VkFormat isn't enough because it's always set to
VK_FORMAT_B8G8R8A8_SRGB or VK_FORMAT_R8G8B8A8_SRGB.
`modifiers` and `external_only` are never initialized, and free'd later. This
commit explicitly initializes them to NULL to prevent segfaults on `free()`
It's not safe to destroy any resources which might still be in-use
by the GPU. Wait for any asynchronous tasks to complete before
destroying everything.
Before re-using a VkCommandBuffer, we need to wait for its
operations to complete. Right now we unconditionally wait for
rendering to complete in vulkan_end(), however we have plans to
fix this [1]. To fully avoid blocking, we need to handle multiple
command buffers in flight at the same time (e.g. for multi-output,
or for rendering followed by texture uploads).
Implement a pool of command buffers. When we need to render, we
pick a command buffer from the pool which has completed its
operations. If we don't find one, try to allocate a new command
buffer. If we don't have slots in the pool anymore, block like we
did before.
[1]: https://gitlab.freedesktop.org/wlroots/wlroots/-/merge_requests/3574
Up until now we were using a VkFence for GPU-to-CPU
synchronization. This has some limitations that become a blocker
when trying to have multiple command buffers in flight at once
(e.g. for multi-output). It's desirable to implement a command
buffer pool [1], but VkFence cannot be used to track command buffer
completion for individual subpasses.
Let's just switch to timeline semaphores [2], which fix this issue,
make synchronization a lot more ergonomic and are a core Vulkan 1.2
feature.
[1]: https://gitlab.freedesktop.org/wlroots/wlroots/-/merge_requests/3802
[2]: https://www.khronos.org/blog/vulkan-timeline-semaphores
If NULL is returned by vkGetDeviceProcAddr(), either the driver
is buggy, either the wlroots code is buggy. For a valid device and
command name, drivers are not allowed to return NULL per the spec.
This mirrors what the GLES2 renderer does in load_gl_proc().
Added wlr_vk_renderer_get_* functions to allow get the VkInstance,
VkPhysicalDevice, VkDevice, queue family of a wlr_vk_renderer.
Added wlr_vk_renderer_get_current_image_attribs function to allow get
the VkImage of current renderer buffer to use on compositors.
Added wlr_renderer_is_vk function, it's like the wlr_renderer_is_gles2,
returns true if the wlr_renderer is a wlr_vk_renderer.
Added wlr_vk_image_get_attribs function to get a VkImage and it's
extras information (e.g. a VkImageLayout and VkImageFormat of the
VkImage) from a wlr_texture.
Instead of filling the fields one by one, use a struct initializer.
This avoids repeating the name of the variable and is more
consistent with the wlroots code style.
find_extensions() is clunky to use when checking only a single
extension, and it's surprising that it returns NULL on success.
Simplify by replacing it with a check_extension() function which
just checks whether a single extension is in the list.
Instead of having a C file with strings for each shader, move each
shader into its own file. Use a small POSIX shell script to convert
the files into C strings (can't wait for C23 #embed...).
The benefits from this are:
- Improved readability and syntax highlighting.
- Line numbers in shader compiler errors are easier to make sense of.
- Consistency with the Vulkan renderer.
- Shaders will become more complicated as we add color management
features.
In the CREATE_DUMB error code-path, we'd only free() the buffer,
however it's already inserted in the alloc->buffers list at this
point.
Instead, make sure finish_buffer() is safe to call (by populating
drm_fd) and call that function.
INVALID means that the modifier is implicit. However dumb buffers
are guaranteed to be LINEAR, so let's just advertise this. Fixes
cursors with the DRM backend: cursor planes usually only support
LINEAR.
When running with the DRM backend, the Pixman renderer needs to
render the cursor buffer. However, DRM drivers only support linear
buffers for these in general, they don't support implicit modifiers
(aka. INVALID).
Advertise support for LINEAR in the Pixman renderer to fix this.
If we have a render node, it means there is a GPU which could be
used. We probably failed GL because of a kernel or Mesa issue.
Instead of automatically falling back to Pixman, error out.
This makes it more obvious to users when something goes wrong,
instead of silently exposing a slow unaccelerated desktop.
References: https://github.com/swaywm/sway/issues/7194
The Vulkan renderer is still experimental. If GL fails, we don't
want to automatically fall back to it by default.
Fixes: 8bd7170fd9 ("Use env helpers")
See the spec at [1]. tl;dr EGL has terrible defaults: eglTerminate()
may have side-effects on completely unrelated EGLDisplay objects.
This extension allows us to opt-in to get the sane behavior:
eglTerminate() only free's our own EGLDisplay without affecting
others.
[1]: https://registry.khronos.org/EGL/extensions/KHR/EGL_KHR_display_reference.txt
We were checking whether any of the features was supported. We need
to check if all of them are.
This makes the check consistent with query_modifier_support() above.
I think the second parameter of the function should be void* instead of
void **, because we use it as a right value in the function.
Signed-off-by: fakechen <chenzigui@kylinos.cn>
Signed-off-by: sunzg <sunzhigang1@kylinos.cn>
From a comment by emersion:
> There is a logic error here: we pass 0xFFFFFFFF to vulkan_find_mem_type, which
> returns an index, and then we logical-and that with mem_reqs.memoryTypeBits.
> Instead we should pass mem_reqs.memoryTypeBits to vulkan_find_mem_type and use
> the result for the memoryTypeIndex field directly. Ideally checking for errors
> (-1 return value) in case no memory type is suitable.
Closes: #3470
This lets the renderer handle the wlr_buffer directly, just like it
does in texture_from_buffer. This also allows the renderer to batch
the rectangle updates, and update more than the damage region if
desirable (e.g. too many rects), so can be more efficient.
GL_ALPHA_BITS is the number of bits of the alpha channel of the
currently bound frame buffer's color buffer -- which is precisely
renderer->current_buffer->rbo . Thus, instead of binding the color
buffer and checking its properties, we can query the already bound
frame buffer.
Note that GL_IMPLEMENTATION_COLOR_READ_{FORMAT,TYPE} are also
properties of frame buffer's color buffer.
Instead of checking whether the wlr_egl dependencies are available
in the GLES2 code, introduce internal_features['egl'] and check
that field.
When updating the EGL dependency list, we no longer need to update
the GLES2 logic.
Whether a texture is opaque or not doesn't depend on the renderer
at all, it just depends on the source buffer. Instead of forcing
all renderers to implement wlr_texture_impl.is_opaque, let's move
this in common code and use the wlr_buffer format to know whether
a texture will be opaque.
`vulkan_format_props_query` calls `query_modifier_support` which
initializes fields of `props` with allocated memory. this memory is
leaked if `query_modifier_support` does not find a supported format
and shmtex is not supported, as in this case `add_fmt_props` ends
up being false and the allocated fields of `props` are never freed.
Replace them with wlr_signal_emit_safe, which correctly handles
cases where a listener removes another listener.
Reported-by: Isaac Freund <ifreund@ifreund.xyz>
Now that the DRM backend no longer depends on GBM, we can make it
optional. The GLES2 renderer still depends on it because of our EGL
device selection.
This is useful for compositors with their own renderers, and for
compositors using the Vulkan renderer.
These formats require EXT_texture_norm16, which in turn needs OpenGL
ES 3.1. The EXT_texture_norm16 extension does not support passing
gl_internalformat = GL_RGBA to glTexImage2D, as can be done for
formats available in OpenGL ES 2.0, so this commit adds a field to
wlr_gles2_pixel_format to provide a more specific internalformat
parameter to glTexImage2D.
This removes an artificial limitation in form of an assert that disallowed the
creation of textures while the renderer is rendering.
A consumer might run its own rendering pipeline and after start of the renderer
still want to create textures for internal usage.
commit 44e8451cd9 ("render/gles2: hide shm formats without GL
support") added the is_gles2_pixel_format_supported() function to
render/gles2/pixel_format.c, whose stated purpose is to "check whether
the renderer has the needed GL extensions to read a given pixel format."
It then used that function to filter the pixel formats returned by
get_gles2_shm_formats().
The result of this change is that RGB formats are no longer reported for
GL drivers that don't implement EXT_read_format_bgra, even when those
formats are supported for rendering (which they have to be for
wlr_gles2_renderer_create() to succeed). This is a pretty clear
regression, since wlr_renderer_init_wl_shm() fails when either of
WL_SHM_FORMAT_ARGB8888 or WL_SHM_FORMAT_XRGB8888 are missing.
To fix the regression, change is_gles2_pixel_format_supported() to
accept all pixel formats that support rendering, regardless of whether
we can read them or not, and move the check for EXT_read_format_bgra
back into gles2_read_pixels(). (There's already a check for this
extension in gles2_preferred_read_format(), so we're not breaking any
abstraction that wasn't already broken.)
Tested on the NVIDIA 495.46 proprietary driver, which doesn't support
EXT_read_format_bgra.
Fixes: 44e8451cd9 ("render/gles2: hide shm formats without GL support")
This intersects two DRM format sets. This is useful for implementing
DMA-BUF feedback in compositors, see e.g. the Sway PR [1].
[1]: https://github.com/swaywm/sway/pull/6313
Support for EXT_image_dma_buf_import_modifiers doesn't necessarily
indicate support for modifiers. For instance, Mesa will advertise
EXT_image_dma_buf_import_modifiers for all drivers. This is a trick
to allow EGL clients to enumerate supported formats (something
EXT_image_dma_buf_import is missing). For more information, see [1].
Add a new wlr_egl.has_modifiers flag which indicates whether
modifiers are supported. It's set to true if any
eglQueryDmaBufModifiersEXT query returned a non-empty list.
Use that flag to figure out whether the buffer modifier should be
passed to the EGL implementation on import.
[1]: https://github.com/KhronosGroup/EGL-Registry/issues/142
The backends and allocators use INVALID, but the renderer uses
LINEAR. Running a compositor with WLR_RENDERER=pixman results in:
00:00:00.744 [types/output/render.c:59] Failed to pick primary buffer format for output 'WL-1'
This allows creating a wlr_egl from an already-existing EGL display
and context. This is useful to allow compositors to choose the exact
EGL initialization parameters.
If the backend doesn't have a DRM FD, fallback to the renderer's.
This accomodates for the situation where the headless backend hasn't
picked a DRM FD in particular, but the renderer has picked one.
If the backend hasn't picked a DRM FD but supports DMA-BUF, pick
an arbitrary render node. This will allow removing the DRM device
selection logic from the headless backend.
They are never used in practice, which makes all of our flag
handling effectively dead code. Also, APIs such as KMS don't
provide a good way to deal with the flags. Let's just fail the
DMA-BUF import when clients provide flags.
For `required` to disable search the value needs to be of `feature` type.
Checking `gles2` via `in` keyword returns a `bool` but `required: false`
makes the dependency optional instead of disabled.
This new renderer is implemented with the existing wlr_renderer API
(which is known to be sub-optimal for some operations). It's not
used by default, but users can opt-in by setting WLR_RENDERER=vulkan.
The renderer depends on VK_EXT_image_drm_format_modifier and
VK_EXT_physical_device_drm.
Co-authored-by: Simon Ser <contact@emersion.fr>
Co-authored-by: Jan Beich <jbeich@FreeBSD.org>
If we aren't trying to create a dumb buffer allocator, and if the
DRM device has a render node (ie, not a split render/display SoC),
then we can use the render node instead of the primary node. This
should allow wlroots to run under seatd when the current user
doesn't have the permission to open primary nodes (logind has a
quirk to allow physically logged in users to open primary nodes).
This allows callers to specify the operations they'll perform on
the returned data pointer. The motivations for this are:
- The upcoming Linux MAP_NOSIGBUS flag may only be usable on
read-only mappings.
- gbm_bo_map with GBM_BO_TRANSFER_READ hurts performance.
The half-float formats depend on GL_OES_texture_half_float_linear,
not just the GL_OES_texture_half_float extension, because the latter
does not include support for linear magni/minification filters.
The new 2101010 and 16161616F formats are only available on little-
endian builds, since their gl_types are larger than a byte and thus
endianness dependent.
Uses the EXT_device_query extension to get the EGL device matching the
requested DRM file descriptor. If the extension is not supported or no device
is found, the EGL device will be retrieved using GBM.
Depends on the EGL_EXT_device_enumeration to get the list of EGL devices.
This EGL extension has been added in [1]. The upsides are:
- We directly get a render node, instead of having to convert the
primary node name to a render node name.
- If EGL_DRM_RENDER_NODE_FILE_EXT returns NULL, that means there is
no render node being used by the driver.
[1]: https://github.com/KhronosGroup/EGL-Registry/pull/127
Without setting this the EGL implementation is allowed to perform
destructive actions on the buffer when imported: its contents
become undefined.
This is mostly a pedantic change, because Mesa processes the attrib
and does absolutely nothing with it.
Now that we have our own wl_drm implementation, there's no reason
to provide custom renderer hooks to init a wl_display in the
interface. We can just initialize the wl_display generically,
depending on the renderer capabilities.
Khronos refers to extensions with their namespace as a prefix in
uppercase. Change our naming to align with Khronos conventions.
This also makes grepping easier.
Khronos refers to extensions with their namespace as a prefix in
uppercase. Change our naming to align with Khronos conventions.
This also makes grepping easier.
Everything needs to go through the unified wlr_buffer interface
now.
If necessary, there are two ways support for
EGL_WL_bind_wayland_display could be restored by compositors:
- Either by using GBM to convert back EGL Wayland buffers to
DMA-BUFs, then wrap the DMA-BUF into a wlr_buffer.
- Or by wrapping the EGL Wayland buffer into a special wlr_buffer
that doesn't implement any wlr_buffer_impl hook, and special-case
that buffer type in the renderer.
Custom backends and renderers need to implement
wlr_backend_impl.get_buffer_caps and
wlr_renderer_impl.get_render_buffer_caps. They can't if enum
wlr_buffer_cap isn't made public.
We never create an EGL context with the platform set to something
other than EGL_PLATFORM_GBM_KHR. Let's simplify wlr_egl_create by
taking a DRM FD instead of a (platform, remote_display) tuple.
This hides the internal details of creating an EGL context for a
specific device. This will allow us to transparently use the device
platform [1] when the time comes.
[1]: https://github.com/swaywm/wlroots/pull/2671
The wlr_egl functions are mostly used internally by the GLES2
renderer. Let's reduce our API surface a bit by hiding them. If
there are good use-cases for one of these, we can always make them
public again.
The functions mutating the current EGL context are not made private
because e.g. Wayfire uses them.
Add wlr_pixman_buffer_get_current_image for wlr_pixman_renderer.
Add wlr_gles2_buffer_get_current_fbo for wlr_gles2_renderer.
Allow get the FBO/pixman_image_t, the compositor can be add some
action for FBO(for eg, attach a depth buffer), or without pixman
render to pixman_image_t(for eg, use QPainter of Qt instead of pixman).
The types of buffers supported by the renderer might depend on the
renderer's instance. For instance, a renderer might only support
DMA-BUFs if the necessary EGL extensions are available.
Pass the wlr_renderer to get_buffer_caps so that the renderer can
perform such checks.
Fixes: 982498fab3 ("render: introduce renderer_get_render_buffer_caps")
This new API allows buffer implementations to know when a user is
actively accessing the buffer's underlying storage. This is
important for the upcoming client-backed wlr_buffer implementation.
This allows compositors to choose a wlr_buffer to render to. This
is a less awkward interface than having to call bind_buffer() before
and after begin() and end().
Closes: https://github.com/swaywm/wlroots/issues/2618
This allows users to know the capabilities of the buffers that
will be allocated. The buffer capability is important to
know when negotiating buffer formats.
When importing a DMA-BUF wlr_buffer as a wlr_texture, the GLES2
renderer caches the result, in case the buffer is used for texturing
again in the future. When the wlr_texture is destroyed by the caller,
the wlr_buffer is unref'ed, but the wlr_gles2_texture is kept around.
This is fine because wlr_gles2_texture listens for wlr_buffer's destroy
event to avoid any use-after-free.
However, with this logic wlr_texture_destroy doesn't "really" destroy
the wlr_gles2_texture. It just decrements the wlr_buffer ref'count.
Each wlr_texture_destroy call must have a matching prior
wlr_texture_create_from_buffer call or the ref'counting will go south.
Wehn destroying the renderer, we don't want to decrement any wlr_buffer
ref'count. Instead, we want to go through any cached wlr_gles2_texture
and destroy our GL state. So instead of calling wlr_texture_destroy, we
need to call our internal gles2_texture_destroy function.
Closes: https://github.com/swaywm/wlroots/issues/2941
Make it so wlr_gles2_texture is ref'counted (via wlr_buffer). This
is similar to wlr_gles2_buffer or wlr_drm_fb work.
When creating a wlr_texture from a wlr_buffer, first check if we
already have a texture for the buffer. If so, increase the
wlr_buffer ref'count and make sure any changes made by an external
process are made visible (by invalidating the texture).
When destroying a wlr_texture created from a wlr_buffer, decrease
the ref'count, but keep the wlr_texture around in case the caller
uses it again. When the wlr_buffer is destroyed, cleanup the
wlr_texture.
This adds a a function to create a wlr_texture from a wlr_buffer.
The main motivation for this is to allow the renderer to create a
single wlr_texture per wlr_buffer. This can avoid needless imports
by re-using existing textures.
GL_RENDERER typically displays a human-readable string for the name
of the GPU, and EGL_VENDOR typically displays a human-readable string
for the GPU manufacturer. EGL_DRIVER_NAME_EXT should give the name of
the driver in use.
References: e8baa0bf39
This function is only required because the DRM backend still needs
to perform multi-GPU magic under-the-hood. Remove the wlr_ prefix
to make it clear it's not a candidate for being made public.
Prior to this commit, WLR_RENDERER was only looked up when the
backend returned a DRM FD.
Make it so WLR_RENDERER is always looked up, so that running wlroots
on a system without a GPU and with WLR_RENDERER=gles2 fails as
expected instead of falling back to the Pixman renderer.
Make it clear GLES2 is being used. Before this commit, various
GL-related information was printed, but not an easy-to-find line
about which renderer is being picked up.
This env var forces the creation of a specific renderer. If no renderer
is specified, the function will try to create all of the renderers one
by one until one is created successfuly.
It allocates in local main memory via shm_open, and provides a FD
to allow sharing with other processes.
This is suitable for software rendering under the Wayland and X11
backends.
The compositor shouldn't write to client buffers if the client
attaches a DMA-BUF to a wl_surface, then attaches a shm buffer.
Make gles2_texture_write_pixels return an error to prevent this
from happening.
The get_drm_fd was made available in an internal header with a53ab146f. Move it
now to the public header so consumers opting in to the unstable interfaces can
make use of it.
PRIME support for buffer sharing has become mandatory since the renderer
rewrite. Make sure we check for the appropriate capabilities in backend,
allocator and renderer.
See also #2819.
All backends use the GBM platform. We can't use it to figure out
whether the DRM backend is used anymore.
Let's just try to always request a high-priority EGL context. Failing
to do so is not fatal.
Split render/display setups have two separate devices: one display-only
with a primary node, and one render-only with a render node. However
in these cases the EGL implementation and the Wayland compositor will
advertise the display device instead of the render device [1]. The EGL
implementation will magically open the render device when the display
device is passed in.
So just pass the display device as if it were a render device. Maybe in
the future Mesa will advertise the render device instead and we'll be
able to remove this workaround.
[1]: https://gitlab.freedesktop.org/mesa/mesa/-/issues/4178
Compute only the transform matrix in the output. The projection matrix
will be calculated inside the gles2 renderer when we start rendering.
The goal is to help the pixman rendering process.
