AvdanOS New Push: Rendering Backend

[Sammy99jsp]

Hello, everyone! Since Avdan has released a new video, we're trying to do a new push for development.
As part of this, we need to integrate with updated versions of our deps, as well as new Wayland protocols.
Anyway, since we have to rewrite everything, we might as well write things properly this time.

This PR includes rewriting the backend from scratch to use more performant tools like Sync Files to synchronise things between {Vulkan, OpenGL, CPU} and DRM {frame is ready to scanout; scanout finished, frame ready to write over}.

To-Do:

• General:
  • Implement double triple-buffered rendering through a backend-agnostic API.
  • Remove the hard-coded XRGB8888 format and pick the best for the display (HDR ftw!)
• Vulkan
• OpenGL
• CPU
• Skia:
  • Vulkan
  • OpenGL
  • CPU

This works on my machine ™️ with nvidia_drm, but it'd be nice to test with other graphics cards/integrated too.

[SunightMC]

i can eventually properly test on all GPUs, intel igpu, intel arc, amd, whatever you need (like next week)

[MrSerge01]

holy moly

[Sammy99jsp]

Rewritten the description of this PR to add a few more to-dos.

Trying to think of a good API for the format negotiation. Maybe:

impl Negotiate<Format> for Vulkan {
    fn applicable(&self, formats: &[Format]) -> Vec<Format> {
        todo!("Actual impementation here!")
    }
}

Same thing for the Skia stuff. Perhaps an extension trait:

pub trait SkiaBackend: Backend {
    fn make_context(&mut self) -> skia_safe::gpu::DirectContext<'_>{ todo!() }  // or whatever;
}

[Sammy99jsp]

I have resorted to a nasty hack (modified from Smithay) whilst we wait for Smithay/drm-rs#224 to close.

Apart from that, relatively simple to implement the CPU backend.