SiFuh has quit [Remote host closed the connection]
SiFuh has joined #crux-devel
ardo has quit [Ping timeout: 240 seconds]
ardo has joined #crux-devel
r0ni has quit [Server closed connection]
r0ni has joined #crux-devel
<jaeger>
hrmm... having a weird issue with vulkan on this system... ERROR: [Loader Message] Code 0 : vkCreateInstance: Found no drivers!
<jaeger>
using an AMD GPU (RX580) that has worked in the past with the amdgpu kernel DRI/DRM driver
<jaeger>
I've rebuilt all the vulkan packages with no luck
<jaeger>
Wonder if maybe it's wayland/xwayland not working, guess I should try it with a normal X session
<jaeger>
ok, same in plain X
<ukky>
jaeger: Can you strace the loading process? Any chance that your system has extra libc.so or libstdc++.so on the lib path?
<jaeger>
I can't imagine where an extra libc or libstdc++ would come from but yeah, going to try strace next
<beerman>
which kernel are you using? I am hitting a weird bug with with everything 6.1.z upwards
<beerman>
not this though, i barely get a working session and if i get i am easy to trigger a fault in amdgpu and it just crashes
<beerman>
with my rx580..
<jaeger>
6.1.61
<jaeger>
sway by itself seems to behave just fine
<jaeger>
Though I did have one odd firefox crash when opening a file save dialog
<ukky>
jaeger: Just a few days ago I couldn't load /usr/lib64/dri/nouveau_dri.so because an SDK from 2020 places ancient libstdc++.so on the lib path when SDK loads.
<jaeger>
What SDK was that?
<ukky>
Xilinx Vitis 2020.2
<jaeger>
definitely not using that here, never heard of it
<ukky>
It's being used for FPGA chips
<jaeger>
Yeah, I looked it up
<jaeger>
Ah ha! Found the problem, radeon_icd.x86_64.json was corrupted somehow, had garbage in it
<jaeger>
rebuilt mesa ports to recreate it, works now in X at least
<jaeger>
and seems to work in wayland too, good
<beerman>
ah, weird
<ukky>
Are you loading it as root?
<jaeger>
nope
<jaeger>
and of course my user is in the video group, etc.
<jaeger>
It was just the file from mesa being messed up
<ukky>
ok
<jaeger>
Always something new, heh. Hadn't seen that happen before.
<beerman>
just checked, i get this beauty after turning the monitors off and then turning them back on. same is triggered when the screen gets locked and it sends the monitors to sleep
<beerman>
6.0.whatever works flawlessly. 6.1 upwards produces this
<beerman>
ukky thanks for checking in. while i researched a bit, it seemed like its not the first time that amdgpu suffered such a bad bug for way too long. But I don't have a hub attached, and I haven't tried unplugging anything to see if this unfucks it. Even if it would, I dunno what that would mean for me
<beerman>
actually one thing. and while we are talking about, i'll try throwing that out
<beerman>
well that and a bios update I guess..
<ukky>
Power management related bugs are nasty. Sometimes the reason for the issue is the bug in ACPI tables in the System BIOS.
<beerman>
yeah, the one thing (supposedly bad pcie card) doesn't seem to do any difference. i'll try the bios update next
<beerman>
It's not the BIOS either
<ukky>
Have anybody seen, or reported similar issue with non-AMD graphics?
<beerman>
I don't think that I can remember anything for other vendors