Nvidia fixes for scanout gbm allocation

[vaxerski]

Nvidia cannot render to linear buffers part 2 the sequel

ref #26
ref hyprwm/Hyprland#7005

needs testing

[Agent00Ming]

hyprwm/Hyprland@72bce7e can't build against this branch because FormatUtils.hpp isn't included with the other headers and it isn't used with relative path(?)

aq (fix-nv) patch used to get Hyprland to compile: include.txt

crash on some keybind bt.txt disregard. hyprwm/Hyprland#7025

[vaxerski]

uhhh ci succeeds...

Also I still need confirmation whether this works

[Agent00Ming]

The monitor connected through thunderbolt/usb-c to displayport does indeed get video output on this branch (where it doesn't on main 4c72cd4)

[Agent00Ming]

I don't think CI checks for building Hyprland off of this branch of aq

[61/185] Building CXX object CMakeFiles/Hyprland.dir/src/protocols/DRMLease.cpp.o
FAILED: CMakeFiles/Hyprland.dir/src/protocols/DRMLease.cpp.o 
/usr/bin/c++ -DHAS_EXECINFO -DHYPRLAND_DEBUG -DHyprland_EXPORTS -DUSES_SYSTEMD -I/home/ming/documents/git/hyprland/. -I/home/ming/documents/git/hyprland/src -I/home/ming/documents/git/hyprland/subprojects/udis86 -I/home/ming/documents/git/hyprland/protocols -I/home/ming/documents/git/hyprland/subprojects/udis86/libudis86 -isystem /usr/include/uuid -isystem /usr/include/pango-1.0 -isystem /usr/include/cairo -isystem /usr/include/glib-2.0 -isystem /usr/lib/glib-2.0/include -isystem /usr/include/harfbuzz -isystem /usr/include/freetype2 -isystem /usr/include/pixman-1 -isystem /usr/include/libdrm -isystem /usr/include/libpng16 -isystem /usr/include/libmount -isystem /usr/include/blkid -isystem /usr/include/fribidi -isystem /usr/include/sysprof-6 -g -std=gnu++23 -Wall -Wextra -Wno-unused-parameter -Wno-unused-value -Wno-missing-field-initializers -Wno-narrowing -Wno-pointer-arith -fmacro-prefix-map=/home/ming/documents/git/hyprland/= -pthread -Winvalid-pch -include /home/ming/documents/git/hyprland/build/CMakeFiles/Hyprland.dir/cmake_pch.hxx -MD -MT CMakeFiles/Hyprland.dir/src/protocols/DRMLease.cpp.o -MF CMakeFiles/Hyprland.dir/src/protocols/DRMLease.cpp.o.d -o CMakeFiles/Hyprland.dir/src/protocols/DRMLease.cpp.o -c /home/ming/documents/git/hyprland/src/protocols/DRMLease.cpp
In file included from /home/ming/documents/git/hyprland/src/protocols/DRMLease.cpp:3:
/usr/include/aquamarine/backend/DRM.hpp:10:10: fatal error: FormatUtils.hpp: No such file or directory
   10 | #include "FormatUtils.hpp"
      |          ^~~~~~~~~~~~~~~~~
compilation terminated.

[vaxerski]

@fufexan fix

[vaxerski]

fixed

[vaxerski]

@fufexan fix ci

[fufexan]

not failing tho?

[vaxerski]

#27 (comment)

[fufexan]

So you want me to create a new workflow that builds hl main with current-rev aq?

[vaxerski]

on pr yes just liuke hl

no wait

[vaxerski]

the issue is the CI passed even tho the code from this branch wasnt compiling, so its not building this PR

[fufexan]

CI isn't lying, I've just built this on WSL.

[nixos@nixos:~]$ nix build github:hyprwm/aquamarine/fix-nv -L --extra-experimental-features 'nix-command flakes' --print-out-paths
/nix/store/xiaclrn5cggrmqbpb14afybi5ci7mi8r-aquamarine-0.1.0+date=2024-07-24_c63080e

[vaxerski]

jfc

last commit is fixed, you wont build dd8f312 though and it has a tick

[Agent00Ming]

this works for single nvidia gpu outputting through thunderbolt/usb-c to displayport, multigpu also works but this revealed that hotplug doesn't work for multigpu intel+nvidia (not a regression in this branch, it exists in main too)

[vaxerski]

ok mergging thennnn

[darkwater]

This breaks rendering on AMD graphics, I get a black screen

[vaxerski]

trace log please

[vaxerski]

@darkwater should be fixed in latest please check

[darkwater]

Nope, still broken

hypr.log

edit: have a works.log too, you can diff the two and see some interesting differences

[vaxerski]

they aren't trace, but thanks.

[vaxerski]

lets move to #28