Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

odd graphical behaviour when using the experimental mesa driver #221

Open
GardenerAether opened this issue Jul 22, 2024 · 2 comments
Open

Comments

@GardenerAether
Copy link

GardenerAether commented Jul 22, 2024

approximately a month ago (24th of june) since i updated my system. im using hyprland but i tested on sway and i3 and the behavior is reproducible there as well (i didnt assume it would be different but i was kinda losing my mind while trying to figure this out)

first noticed some weird graphical behavior in hotel 09 on the day i updated, but i recently (today) had the time to properly look into the issue and found an assortment of shaders that are entirely busted on my end

hotel09 screenshots hi09 example 1 hi09 example 2
lover 2 lover 2 example
nested octahedrons 2 nested octahedrons 2 example
selfie girl selfie girl example

i did try to look at some of the shader code to see if i could pinpoint what exact part of it gets borked, and here's some notes on what i found:

  • i was able to entirely fix the nested octahedrons shader and the selfie girl by setting the clamp (at Buffer A:60 and Image:48 respectively) to some absurdly low non-zero minima, even if that minima is negative, so clamping is broken specifically for exactly zero
  • the way in which clamping breaks it is that it seems to set it to some extremely small value very close to zero. i initially thought it was zero because at first it just looked like a black screen but i realized you could recover those values back (even if they are extremely crusty) by scaling the clamped value by some gargantuan factor
  • i could not use this to fix the lover 2 shader, so theres either some secret additional thing that is broken or theres something else that im missing regarding exactly what way in which the clamping breaks

none of these issues were present prior the 24th of june

i do not know if this is something extremely specific to my machine or how to fix it. i do not actually know if this is a bug with mesa or something in my configuration, but im not sure what could happen in my local configuration that would cause clamping to break specifically

i will emphasize that all of these are all webgl, as i dont really have a setup for testing arbitrary opengl code outside of shadertoy right now (though i do plan to set something up later), but i also dont see how my webgl could differ from anybody else's which leads me to believe its a problem with mesa

none of my desktop environments function properly without the experimental driver (hyprland doesnt even launch) so i havent been able to test if these are the same without it

if you think this is the wrong place for me to post this issue please do let me know

@GardenerAether
Copy link
Author

is this literally not reproducible by a single soul

@rowanG077
Copy link
Contributor

I don't see the same output as you but it's still broken. You'd best report that upstream: AsahiLinux/linux#72

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants