Description of problem:

Since recently, chromium chrashes when run with Ozone/Wayland on an Intel UHD Graphics 630.

Version-Release number of selected component (if applicable):

chromium-90.0.4430.212-1.fc34.x86_64

How reproducible:

Always

Steps to Reproduce:
1. chromium-browser --enable-features=UseOzonePlatform --ozone-platform=wayland
2.
3.

Actual results:

It crashes and prints:

ATTENTION: default value of option allow_rgb10_configs overridden by environment.
ATTENTION: default value of option allow_rgb10_configs overridden by environment.
[8131:8131:0607/082002.996422:ERROR:gpu_init.cc(426)] Passthrough is not supported, GL is egl
[8131:8131:0607/082003.001470:ERROR:sandbox_linux.cc(374)] InitializeSandbox() called with multiple threads in process gpu-process.
[8131:8131:0607/082003.049265:ERROR:gl_surface_egl.cc(778)] EGL Driver message (Error) eglCreateImageKHR: dri2_create_image_khr_texture
[8131:8131:0607/082003.049383:ERROR:gl_image_egl.cc(38)] Error creating EGLImage: EGL_BAD_MATCH
[8131:8131:0607/082003.049472:ERROR:gpu_memory_buffer_factory_native_pixmap.cc(264)] Failed to create GLImage 3840x2096, BGRA_8888, usage SCANOUT
[8131:8131:0607/082003.049601:ERROR:gl_surface_egl.cc(778)] EGL Driver message (Error) eglCreateImageKHR: dri2_create_image_khr_texture
[8131:8131:0607/082003.049659:ERROR:gl_image_egl.cc(38)] Error creating EGLImage: EGL_BAD_MATCH
[8131:8131:0607/082003.049730:ERROR:gpu_memory_buffer_factory_native_pixmap.cc(264)] Failed to create GLImage 3840x2096, BGRA_8888, usage GPU_READ
[8131:8131:0607/082003.049786:ERROR:shared_image_backing_factory_gl_texture.cc(536)] CreateSharedImage: Failed to create bindable image
[8131:8131:0607/082003.049834:ERROR:shared_image_factory.cc(582)] CreateSharedImage: could not create backing.
[8131:8131:0607/082003.050055:ERROR:skia_output_surface_impl_on_gpu.cc(1495)] Failed to make current.
[8131:8131:0607/082003.050337:ERROR:skia_output_surface_impl_on_gpu.cc(1495)] Failed to make current.
[8131:8131:0607/082003.050442:ERROR:skia_output_surface_impl_on_gpu.cc(1495)] Failed to make current.
[8131:8131:0607/082003.052939:ERROR:raster_decoder.cc(1064)]   RasterDecoderImpl: Context lost during MakeCurrent.
[8131:8131:0607/082003.053856:ERROR:raster_decoder.cc(1064)]   RasterDecoderImpl: Context lost during MakeCurrent.
[8131:8131:0607/082003.054087:ERROR:shared_image_stub.cc(452)] SharedImageStub: context already lost
[8131:8131:0607/082003.054671:ERROR:shared_image_stub.cc(452)] SharedImageStub: context already lost
[8090:8090:0607/082003.055428:ERROR:wayland_event_watcher.cc(138)] Fatal Wayland protocol error 7 on interface zwp_linux_buffer_params_v1 (object 32). Shutting down..

Expected results:

It shouldn't crash.

Additional info:

lspci says:
…
00:02.0 VGA compatible controller: Intel Corporation CoffeeLake-H GT2 [UHD Graphics 630]
…
01:00.0 3D controller: NVIDIA Corporation TU117M [GeForce GTX 1650 Mobile / Max-Q] (rev a1)
…

When setting DRI_PRIME=1 it runs just fine.

This issue first appeared recently, around the time, when the 5.12 kernel and mesa 21.1 were introduced to Fedora 34. I cannot tell, though, which, if any, of these updates caused it.

The same issue shows when trying to run some electron based applications with a similar command configuration, such as Visual Studio Code.

The issue does *not* show when running Google's official chrome (version 91.0.4472.77) with a similar command line.