| 18 Jan 2024 |
 K900 | So something is fucking with the permissions on the device node possibly | 08:09:57 |
| K900 | This should normally be handled by logind | 08:10:05 |
| K900 | What does ls -la /dev/dri/* output? | 08:10:19 |
 l0b0 | I have enabled libvirtd, if that's relevant. | 08:10:26 |
| l0b0 | ❯ ls -la /dev/dri/*
crw-rw----+ 1 root video 226, 0 Jan 18 21:00 /dev/dri/card0
crw-rw----+ 1 root video 226, 1 Jan 18 21:00 /dev/dri/card1
crw-rw-rw- 1 root render 226, 128 Jan 18 21:00 /dev/dri/renderD128
crw-rw-rw- 1 root render 226, 129 Jan 18 21:00 /dev/dri/renderD129
/dev/dri/by-path:
total 0
drwxr-xr-x 2 root root 120 Jan 18 21:00 .
drwxr-xr-x 3 root root 140 Jan 18 21:00 ..
lrwxrwxrwx 1 root root 8 Jan 18 21:00 pci-0000:03:00.0-card -> ../card0
lrwxrwxrwx 1 root root 13 Jan 18 21:00 pci-0000:03:00.0-render -> ../renderD128
lrwxrwxrwx 1 root root 8 Jan 18 21:00 pci-0000:17:00.0-card -> ../card1
lrwxrwxrwx 1 root root 13 Jan 18 21:00 pci-0000:17:00.0-render -> ../renderD129
| 08:11:13 |
| K900 | OK, how about getfacl /dev/dri/card*? | 08:11:23 |
| l0b0 | ❯ getfacl /dev/dri/card*
getfacl: Removing leading '/' from absolute path names
# file: dev/dri/card0
# owner: root
# group: video
user::rw-
user:victor:rw-
group::rw-
mask::rw-
other::---
# file: dev/dri/card1
# owner: root
# group: video
user::rw-
user:victor:rw-
group::rw-
mask::rw-
other::---
| 08:11:43 |
| l0b0 | (I'm a member of the video group, if that's relevant.) | 08:12:02 |
| K900 | You shouldn't need to be | 08:12:07 |
| K900 | Are you on X or on Wayland? | 08:13:16 |
| l0b0 | Wayland, right now. | 08:13:52 |
| l0b0 | I've also set virtualbox.host.enable = true, although I don't use that much. | 08:14:47 |
| K900 | Try eglinfo | grep "core profile renderer" from a TTY | 08:14:57 |
| l0b0 | And enabled rootless Docker | 08:15:01 |
| K900 | None of the VM stuff should matter for this | 08:15:06 |
| l0b0 | No output from that command. | 08:16:08 |
| l0b0 | Exit code 1 | 08:16:10 |
| K900 | How about just eglinfo? | 08:16:40 |
| K900 | Also, you might need to close your Wayland compositor first | 08:16:49 |
| l0b0 | OIC, will try | 08:17:24 |
| l0b0 | Download eglinfo.log | 08:18:59 |
| l0b0 | That's in a VT | 08:19:16 |
| K900 | Well that looks like it's probing something at least | 08:20:09 |
 @elvishjerricco:matrix.org | In reply to @k900:0upti.me
You should be getting better than half-decent out of Mesa FWIW, at 4K resolution, I only get like 80fps with a 7900 XT | 08:20:34 |
| @elvishjerricco:matrix.org | dunno if that's better on windows or not | 08:20:42 |
| @elvishjerricco:matrix.org | But I also get the same message from eglinfo:
$ eglinfo | grep "core profile rendere"
amdgpu_device_initialize: amdgpu_query_info(ACCEL_WORKING) failed (-13)
amdgpu: amdgpu_device_initialize failed.
| 08:20:56 |
| @elvishjerricco:matrix.org | In reply to @k900:0upti.me
You should be getting better than half-decent out of Mesa * FWIW, at 4K resolution, I only get like 80fps with a 7900 XT in BG3 | 08:21:29 |
| @elvishjerricco:matrix.org | l0b0: are you also running 4K resolution? | 08:21:57 |
| l0b0 | In reply to @elvishjerricco:matrix.org
l0b0: are you also running 4K resolution? Yes | 08:22:09 |
| @elvishjerricco:matrix.org | I'll see how much better or worse it is on windows here in a few minutes after I finish something up | 08:22:33 |
