| Age | Commit message (Collapse) | Author |
|---|
|
Later commits will start using _SRGB image views again,
where appropriate.
|
|
|
|
When we cleared the pending backend damage when the output committed,
we would not take into account the output transform. It's easiest to fix
this by just changing pending_commit_damage to always have transformed
coordinates.
|
|
We would apply it twice if we failed direct scanout. Once when we attempt
direct scanout and again when we composite.
|
|
Direct scanout damage will just accumulate on the damage ring while in
direct scanout and properly damage when we exit anyway. On the flip side
since we now manage backend damage submission ourselves, this won't break
that either.
|
|
Compositors who use _build_state are no longer required to call
damage_ring_rotate themselves. This is a minor breaking change.
|
|
The damage ring is only supposed to keep track of damage within
the swapchain
|
|
|
|
|
|
Sadly, the new API is not backwards compatible with the old API. Since
we have already switched all users in wlroots to the new API compositors
are already practically mandated to implement the new API. Let's get rid
of the old one since there is no point.
|
|
|
|
|
|
|
|
|
|
|
|
Also get rid of gles2_texture.owns_texture. We only use the tex/fbo
|
|
|
|
|
|
|
|
This scales better when more render pass options are added.
|
|
|
|
This is useful for e.g. lazily blitting a texture for readback
purposes while rendering.
|
|
|
|
It turns out we forgot about the function declaration in the header.
Also some docs were still referring to that function.
Move the wlr_output_attach_render() docs to
wlr_output_begin_render_pass().
|
|
These comments were a bit misleading:
- "GL_TEXTURE_2D == mutable": not really, imported non-external-only
DMA-BUFs would also use this target, but are not mutable.
- "Only affects target == GL_TEXTURE_2D": same here.
- "If imported from a wlr_buffer": not really, would be NULL if
imported from a shm wlr_buffer.
Adjust these comments to better reflect reality and adjust the check
in gles2_texture_update_from_buffer().
|
|
Let's us share more code with the other code path
|
|
Since wlr_gles2_buffer is now managing importing for us, there is
no reason for us to continue doing this.
|
|
We can get it from the buffer
|
|
We can double import a dmabuf if we use it as a texture target and
a render target. Instead, let's unify render targets and texture dmabuf
imports to use wlr_gles2_buffer which manages the EGLImageKHR
|
|
Since imported textures will be based off of gles2_buffer we have
to destroy textures first or else they will have an invalid reference
to the buffers they are imported from.
|
|
We can't do it while we're creating them because we'll want to use
gles2 buffers for textures soon.
|
|
|
|
This is unused.
|
|
This function touches queued_fb/current_fb, which the renderer has
nothing to do with.
|
|
|
|
This is always true now that we can only render with a buffer.
|
|
|
|
|
|
|
|
This adds an alternate way to use wlr_damage_ring without the
concept of buffer age. Buffer age is a concept inherited from EGL
but there is no real reason why we should continue to use that in
wlroots. Instead, use wlr_buffer pointers.
Eventually, we should be able to remove the buffer age based
functions.
|
|
If the external-only flag is set, then the EGLImage is only
supported for use with GL_TEXTURE_EXTERNAL_OES texture targets.
In particular, the EGLImage cannot be bound to a RBO.
|
|
Why track the alpha here when we can already get that information
elsewhere?
|
|
RMFB implicitly performs a modeset to turn off any CRTC which is
using the FB. This prevents seamless transitions between two DRM
masters from working.
Use the new CLOSEFB IOCTL which doesn't turn off anything and leave
it up to the compositor to turn off outputs on shutdown if it wants
to.
|
|
It's possible that we don't have an EGLDevice if we created the
EGL context from a GBM device. Let's ensure all GPU-accelerated
renderers always have a DRM FD to return by falling back to GBM's
FD.
|
|
This was used by the legacy rendering API. Since begin()/end()
only need to set current_render_buffer and nothing else, we can
drop all of these bits.
|
|
Instead, move the wlr_xwayland_surface_set_withdrawn() and
wlr_xwayland_surface_restack() calls to the MapNotify handler with an
override_redirect check, as they are done too early. This mirrors the logic in
the UnmapNotify handler and fixes a bug where wlr_xwayland_surface_restack()
would be called on an o-r window after the following sequence of requests:
- CreateWindow with override_redirect=True
- ChangeWindowAttributes with override_redirect=False
- MapWindow
Fixes: https://gitlab.freedesktop.org/wlroots/wlroots/-/issues/3770
|
|
Under X11, ConfigureNotify means that the window has already been resized.
Sending ConfigureRequest with the received size is not only useless, but also
can confuse the window manager, which will probably reply with the current
(i.e. *old*) size causing a configure loop.
Fixes: https://gitlab.freedesktop.org/wlroots/wlroots/-/issues/3769
|
|
|
|
The surface scale and transform are applied before the viewport.
Closes: https://gitlab.freedesktop.org/wlroots/wlroots/-/issues/3766
|
|
The commit event fires too late: we're interested in checking the
values at wl_surface.commit time, not after.
|
