Doesn't run out of the box on a TTY on PinePhone Pro

[YaLTeR]

For whenever I do more TTY backend changes. But I'm not sure if I can really fix this without having a device to test.

2024-04-30T03:19:47.120722Z  INFO niri: starting version 0.1.5 (v0.1.5-21-gc0d5001)
2024-04-30T03:19:48.319465Z DEBUG niri_config: loaded config from "/home/alarm/.config/niri/config.kdl"
2024-04-30T03:19:48.518433Z  WARN niri::backend::tty: error getting the render node for the primary GPU; proceeding anyway
2024-04-30T03:19:48.519744Z  INFO niri::backend::tty: using as the render node: "/dev/dri/card0"
2024-04-30T03:19:48.689554Z DEBUG niri::backend::tty: device added: 57857 "/dev/dri/card1"
2024-04-30T03:19:48.692392Z  WARN niri::backend::tty: error adding device: DRM access error: Error loading resource handles on device `Some("/dev/dri/card1")` (Operation not supported (os error 95))

Caused by:
    Operation not supported (os error 95)
2024-04-30T03:19:48.692608Z DEBUG niri::backend::tty: device added: 57856 "/dev/dri/card0"
2024-04-30T03:19:48.942066Z DEBUG niri::backend::tty: this is the primary node
2024-04-30T03:19:49.007396Z DEBUG niri::backend::tty: device changed: 57856
2024-04-30T03:19:49.008252Z DEBUG niri::backend::tty: connecting connector: DSI-1
2024-04-30T03:19:49.008436Z DEBUG niri::backend::tty: picking mode: Mode { name: "720x1440", clock: 74250, size: (720, 1440), hsync: (760, 806, 846), vsync: (1449, 1456, 1463), hskew: 0, vscan: 0, vrefresh: 60, mode_type: ModeTypeFlags(PREFERRED | DRIVER) }
2024-04-30T03:19:49.008590Z DEBUG niri::backend::tty: error setting max bpc: couldn't find max bpc property
2024-04-30T03:19:49.019291Z DEBUG niri::niri: putting output DSI-1 at x=0 y=0
2024-04-30T03:19:49.019828Z  INFO niri: listening on Wayland socket: wayland-1
2024-04-30T03:19:49.019891Z  INFO niri: IPC listening on: /run/user/1000/niri.wayland-1.23319.sock
2024-04-30T03:19:49.026634Z  WARN niri::backend::tty: error creating renderer for primary GPU: Error::NoDevice(DrmNode { dev: 57856, ty: Primary })

[samueldr]

This seems to be an RK3399 issue outright, if not that particular mali "vintage".

On a Pinebook Pro, I dipped in the debug options, and used this to get niri to render:

debug {
  render-drm-device "/dev/dri/card1"
}

I'm not (yet) enough familiar with DRM/DRI stuff to diagnose, but if that helps:

/etc/nixos $ ls -l /dev/dri/by-path/
total 0
lrwxrwxrwx 1 root root  8 May  5 01:02 platform-display-subsystem-card -> ../card0
lrwxrwxrwx 1 root root  8 May  5 01:02 platform-ff9a0000.gpu-card -> ../card1
lrwxrwxrwx 1 root root 13 May  5 01:02 platform-ff9a0000.gpu-render -> ../renderD128

This might end-up breaking in the same way on MT8183. I will likely check that soon enough. Both different vintages of mali though...

[YaLTeR]

Yeah, it's these kmsro devices and the way they interact with various mesa API calls. This is kind of a "poke API on a real device and see what it does" situation.