| 14 Dec 2025 |
 niklaskorz | 
Download grafik.png | 14:12:03 |
 K900 | Interesting | 14:13:29 |
| K900 | Very interestinf | 14:13:49 |
| K900 | * Very interesting | 14:13:54 |
| K900 | ls /dev/dri/*? | 14:14:39 |
| niklaskorz | ❯ ls /dev/dri/*
/dev/dri/card0
/dev/dri/by-path:
pci-0000:01:00.0-platform-simple-framebuffer.0-card
| 14:15:50 |
| niklaskorz | vulkaninfo shows llvmpipe as GPU id 1 and nvidia as GPU id 0 | 14:16:30 |
| niklaskorz | hm but also vulkaninfo prints this warning: | 14:17:44 |
| niklaskorz | WARNING: [Loader Message] Code 0 : Path to given binary /nix/store/ymzqq38wncg2c0b1qjxhf646xd61i1pw-nvidia-x11-580.119.02-6.18/lib/libGLX_nvidia.so.580.119.02 was found to differ from OS loaded path /nix/store/ymzqq38wncg2c0b1qjxhf646xd61i1pw-nvidia-x11-580.119.02-6.18/lib/libGLX_nvidia.so.0
| 14:17:50 |
| K900 | That's fine | 14:27:06 |
| niklaskorz | kwin6 logs:
Dez 14 15:04:58 nk-desktop kwin_wayland[2787]: No backend specified, automatically choosing drm
Dez 14 15:04:58 nk-desktop kwin_wayland[2787]: Could not find edid for connector DrmConnector(id=37, gpu="/dev/dri/card0", name="Unknown-1", connection="Connected", countMode=1)
Dez 14 15:04:58 nk-desktop kwin_wayland[2787]: Failed to open drm node: ""
Dez 14 15:04:58 nk-desktop kwin_wayland[2787]: couldn't find dev node for drm device
Dez 14 15:04:58 nk-desktop kwin_wayland[2787]: No render nodes have been found, falling back to primary node
| 14:27:08 |
| K900 | I wonder if the race is just the Nvidia driver not loading in time | 14:27:29 |
| K900 | So it's picking up the simplefb device | 14:27:29 |
| K900 | Because if that's the case, removing Mesa won't help | 14:27:56 |
| niklaskorz | in that case I should not get anything to render at all though, right? | 14:34:28 |
| niklaskorz | when mesa is disabled, that is | 14:34:37 |
| K900 | There's an even softwarer fallback, technically | 14:35:05 |
| niklaskorz | I'll boot into the generation that removes mesa from graphics.package to see what those logs say | 14:35:24 |
| K900 | But it might also crash and just keep restart looping until the Nvidia device shows up | 14:35:35 |
| niklaskorz | on a successful boot, it looks like this: | 14:41:46 |
| niklaskorz | ❯ ls /dev/dri/*
/dev/dri/card1 /dev/dri/renderD128
/dev/dri/by-path:
pci-0000:01:00.0-card pci-0000:01:00.0-render
| 14:41:50 |
| niklaskorz | Which makes me wonder if that card0 is the ryzen integrated graphics after all | 14:44:31 |
| K900 | No, it's simpledrm | 14:45:36 |
| K900 | Which is a placeholder not real driver | 14:45:41 |
| K900 | That runs on top of your EFI framebuffer | 14:45:47 |
| K900 | And yeah this is definitely not affected by Mesa | 14:45:57 |
| K900 | So my guess is that it's actually just racing, but not where you think it is | 14:46:13 |
| K900 | Maybe try early loading nvidia-drm | 14:51:42 |
| niklaskorz | out of interest... what would be the place to look up the timing of the nvidia driver being loaded? | 14:53:54 |
| niklaskorz | to compare it against the kwin logs | 14:54:03 |
