Immediate crash on Winit wayland backend and wgpu Vulkan renderer

[Jaakkonen]

Is there an existing issue for this?

• [X] I have searched the existing issues.

Is this issue related to iced?

• [X] My hardware is compatible and my graphics drivers are up-to-date.

What happened?

All Iced programs (in the panic sample the game of life example) immediately crash when having vulkan & wayland backends chosen.

Environment: Sway 1.7 Intel alder lake graphics with Mesa 22.3.1

What is the expected behavior?

The application doesn't crash or errors that vulkan & wayland is an unsupported combination (-> Use x11 & vulkan for actually working experience or wayland & gl for experience where most things excluding images work)

Version

master

Operative System

Linux

Do you have any log output?

RUST_BACKTRACE=1 WINIT_UNIX_BACKEND=wayland WGPU_BACKEND=vulkan cargo run --features iced/wgpu --package game_of_life
    Finished dev [unoptimized + debuginfo] target(s) in 0.09s
     Running `target/debug/game_of_life`
thread 'main' panicked at 'Error in Surface::configure: parent device is lost', /home/jaakko/.local/share/cargo/registry/src/github.com-1ecc6299db9ec823/wgpu-0.14.2/src/backend/direct.rs:274:9
stack backtrace:
   0: rust_begin_unwind
             at /rustc/b569c9dc57ee22b6ff94563af856909763dfa24b/library/std/src/panicking.rs:575:5
   1: core::panicking::panic_fmt
             at /rustc/b569c9dc57ee22b6ff94563af856909763dfa24b/library/core/src/panicking.rs:64:14
   2: wgpu::backend::direct::Context::handle_error_fatal
             at /home/jaakko/.local/share/cargo/registry/src/github.com-1ecc6299db9ec823/wgpu-0.14.2/src/backend/direct.rs:274:9
   3: <wgpu::backend::direct::Context as wgpu::Context>::surface_configure
             at /home/jaakko/.local/share/cargo/registry/src/github.com-1ecc6299db9ec823/wgpu-0.14.2/src/backend/direct.rs:1017:13
   4: wgpu::Surface::configure
             at /home/jaakko/.local/share/cargo/registry/src/github.com-1ecc6299db9ec823/wgpu-0.14.2/src/lib.rs:3715:9
   5: <iced_wgpu::window::compositor::Compositor<Theme> as iced_graphics::window::compositor::Compositor>::configure_surface
             at ./wgpu/src/window/compositor.rs:157:9
   6: iced_winit::application::run_instance::{{closure}}
             at ./winit/src/application.rs:258:5
   7: iced_winit::application::run::{{closure}}
             at ./winit/src/application.rs:218:24
   8: winit::platform_impl::platform::wayland::event_loop::EventLoop<T>::run_return
             at /home/jaakko/.local/share/cargo/git/checkouts/winit-57d3141eaf559308/9404575/src/platform_impl/linux/wayland/event_loop/mod.rs:233:9
   9: winit::platform_impl::platform::EventLoop<T>::run_return
             at /home/jaakko/.local/share/cargo/git/checkouts/winit-57d3141eaf559308/9404575/src/platform_impl/linux/mod.rs:748:56
  10: <winit::event_loop::EventLoop<T> as winit::platform::run_return::EventLoopExtRunReturn>::run_return
             at /home/jaakko/.local/share/cargo/git/checkouts/winit-57d3141eaf559308/9404575/src/platform/run_return.rs:62:9
  11: iced_winit::application::platform::run
             at ./winit/src/application.rs:749:17
  12: iced_winit::application::run
             at ./winit/src/application.rs:193:5
  13: iced::application::Application::run
             at ./src/application.rs:216:12
  14: game_of_life::main
             at ./examples/game_of_life/src/main.rs:23:5
  15: core::ops::function::FnOnce::call_once
             at /rustc/b569c9dc57ee22b6ff94563af856909763dfa24b/library/core/src/ops/function.rs:507:5
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.

[KentaTheBugMaker]

no problem on kabylake GT2 with mesa 22.3.1 kernel 5.15 + sway-im-git 1.9

[KentaTheBugMaker]

if you can run with LIBGL_ALWAYS_SOFTWARE=1. there is problem in sway or mesa .

[kaimast]

I had this problem for a few months on my GNOME desktop. I assumed it was related to my hybrid graphics card setup (Intel+Nvidia), but you only seem to have one card and the same problem.

[ids1024]

There seem to be a few issues reported on winit with the error message: https://github.com/gfx-rs/wgpu/issues?q=is%3Aissue+parent+device+is+lost+is%3Aopen

Likely using the Iced glow feature (to use OpenGL instead of wgpu) will work. It may be worth testing if wgpus own examples have the same issue, and reporting there with any relevant details.

Presumably it's a driver or wgpu issue rather than something in iced. (Though if other things using wgpu seem to work reliably, that assumption may be wrong.)

[AAlieZZ]

same problem on KDE Plasma Arch Linux

[kaimast]

same problem on KDE Plasma Arch Linux

What GPU(s) do you have?

[AAlieZZ]

same problem on KDE Plasma Arch Linux

What GPU(s) do you have?

NVIDIA GeForce RTX 3060 Laptop GPU

[genusistimelord]

this sounds like either the screen format is wrong or not supported, its attempting to use wgpu Opengl and missing features, a bad driver for the nvidia, or its using the processors built in GPU over the dedicated GPU.

[kaimast]

same problem on KDE Plasma Arch Linux

What GPU(s) do you have?

NVIDIA GeForce RTX 3060 Laptop GPU

Do you also have an integrated GPU? If so, you might be encountering this bug https://github.com/gfx-rs/wgpu/issues/2519.

My current "workaround" is to use the integrated GPU.

[probablySophie]

My current "workaround" is to use the integrated GPU.

Sorry to dredge up an old thread, but how do you use the integrated GPU? Is there a RUSTFLAGS or compile target?

[kaimast]

When you set up your wgpu adapter you can set the power preference to "low" which should pick the integrated GPU.

You can also make your application check for the WGPU_POWER_PREF environment variable if you initialize your adapter like here: https://github.com/gfx-rs/wgpu/blob/trunk/wgpu/src/util/init.rs#L71

[AmmarAbouZor]

Setting the following environment variable made it work for my machine (Sway on Hybrid laptop with Intel + NVIDIA GPU.

VK_ICD_FILENAMES=/usr/share/vulkan/icd.d/intel_icd.x86_64.json. For using Intel graphics Or VK_ICD_FILENAMES=/usr/share/vulkan/icd.d/nvidia_icd.x86_64.json. for Nvidia

I got to this workaround from #1185

I think the problem is in trying to figure out which file to use on hybrid systems with sway

[Dunckleosteus]

Same problem here, running anything on my Nvidia GPU. It seem to run on the integrated GPU. Here is the error message: thread 'main' panicked at 'Failed to call XMapRaised: XError { description: "BadMatch (invalid parameter attributes)", error_code: 8, request_code: 149, minor_code: 4 }', /home/throgg/.cargo/registry/src/github.com-1ecc6299db9ec823/winit-0.27.5/src/platform_impl/linux/x11/window.rs:983:18 stack backtrace: 0: rust_begin_unwind at /rustc/84c898d65adf2f39a5a98507f1fe0ce10a2b8dbc/library/std/src/panicking.rs:579:5 1: core::panicking::panic_fmt at /rustc/84c898d65adf2f39a5a98507f1fe0ce10a2b8dbc/library/core/src/panicking.rs:64:14 2: core::result::unwrap_failed at /rustc/84c898d65adf2f39a5a98507f1fe0ce10a2b8dbc/library/core/src/result.rs:1750:5 3: core::result::Result<T,E>::expect at /rustc/84c898d65adf2f39a5a98507f1fe0ce10a2b8dbc/library/core/src/result.rs:1047:23 4: winit::platform_impl::platform::x11::window::UnownedWindow::set_visible at /home/throgg/.cargo/registry/src/github.com-1ecc6299db9ec823/winit-0.27.5/src/platform_impl/linux/x11/window.rs:981:13 5: winit::platform_impl::platform::Window::set_visible at /home/throgg/.cargo/registry/src/github.com-1ecc6299db9ec823/winit-0.27.5/src/platform_impl/linux/mod.rs:343:50 6: winit::window::Window::set_visible at /home/throgg/.cargo/registry/src/github.com-1ecc6299db9ec823/winit-0.27.5/src/window.rs:631:9 7: iced_winit::application::run_instance::{{closure}} at /home/throgg/.cargo/registry/src/github.com-1ecc6299db9ec823/iced_winit-0.9.1/src/application.rs:309:9 8: iced_winit::application::run::{{closure}} at /home/throgg/.cargo/registry/src/github.com-1ecc6299db9ec823/iced_winit-0.9.1/src/application.rs:251:24 9: winit::platform_impl::platform::sticky_exit_callback at /home/throgg/.cargo/registry/src/github.com-1ecc6299db9ec823/winit-0.27.5/src/platform_impl/linux/mod.rs:849:9 10: winit::platform_impl::platform::x11::EventLoop::run_return::single_iteration at /home/throgg/.cargo/registry/src/github.com-1ecc6299db9ec823/winit-0.27.5/src/platform_impl/linux/x11/mod.rs:329:13 11: winit::platform_impl::platform::x11::EventLoop::run_return at /home/throgg/.cargo/registry/src/github.com-1ecc6299db9ec823/winit-0.27.5/src/platform_impl/linux/x11/mod.rs:448:31 12: winit::platform_impl::platform::EventLoop::run_return at /home/throgg/.cargo/registry/src/github.com-1ecc6299db9ec823/winit-0.27.5/src/platform_impl/linux/mod.rs:748:56 13: <winit::event_loop::EventLoop as winit::platform::run_return::EventLoopExtRunReturn>::run_return at /home/throgg/.cargo/registry/src/github.com-1ecc6299db9ec823/winit-0.27.5/src/platform/run_return.rs:62:9 14: iced_winit::application::platform::run at /home/throgg/.cargo/registry/src/github.com-1ecc6299db9ec823/iced_winit-0.9.1/src/application.rs:870:17 15: iced_winit::application::run at /home/throgg/.cargo/registry/src/github.com-1ecc6299db9ec823/iced_winit-0.9.1/src/application.rs:226:5 16: iced::application::Application::run at /home/throgg/.cargo/registry/src/github.com-1ecc6299db9ec823/iced-0.9.0/src/application.rs:209:12 17: iced::sandbox::Sandbox::run at /home/throgg/.cargo/registry/src/github.com-1ecc6299db9ec823/iced-0.9.0/src/sandbox.rs:153:9 18: yew::main at /home/throgg/Documents/Code/Rust/yew/src/main.rs:5:5 19: core::ops::function::FnOnce::call_once at /rustc/84c898d65adf2f39a5a98507f1fe0ce10a2b8dbc/library/core/src/ops/function.rs:250:5 note: Some details are omitted, run with RUST_BACKTRACE=full for a verbose backtrace.

And here is my system information (in case it can help solve the problem in the future):

OS: Pop!_OS 22.04 LTS x86_64 Host: TUXEDO Stellaris/Polaris AMD Gen4 Standard Kernel: 6.2.6-76060206-generic Uptime: 1 hour, 39 mins Resolution: 1920x1080 DE: Unity WM: Mutter WM Theme: Adwaita CPU: AMD Ryzen 7 6800H with Radeon Graphics (16) @ 3.200GHz GPU: AMD ATI 05:00.0 Rembrandt GPU: NVIDIA GeForce RTX 3060 Mobile / Max-Q Memory: 7866MiB / 15218MiB

[Dunckleosteus]

Hi, I fixed it with the following command:

export VK_ICD_FILENAMES=/usr/share/vulkan/icd.d/nvidia_icd.json

Cheers @AmmarAbouZor

[kaimast]

There also is a potential fix in an upcoming NVIDIA driver (545). See here.

[i509VCB]

Hi, I fixed it with the following command:

export VK_ICD_FILENAMES=/usr/share/vulkan/icd.d/nvidia_icd.json

@Dunckleosteus this is likely an issue with your vulkan and or driver installation. You shouldn't be needing to tell libvulkan where to find icds.

[Dunckleosteus]

Hi, I fixed it with the following command:

export VK_ICD_FILENAMES=/usr/share/vulkan/icd.d/nvidia_icd.json

@Dunckleosteus this is likely an issue with your vulkan and or driver installation. You shouldn't be needing to tell libvulkan where to find icds.

Hi, thank you for your reply. How should I go about fixing the problem ? This may a be silly question but I am a complete linux beginner. Cheers !

[kaimast]

There also is a potential fix in an upcoming NVIDIA driver (545). See here.

This update did indeed fix my problem. I have another unrelated issue now (https://github.com/gfx-rs/wgpu/issues/4627) though.

Maybe this issue can be closed unless other folks can still reproduce it?