<stratofall>
I recently bought a new GPU, RTX 5080 and checking my xorg logs etc. Everything seems to be loaded but ultimately the xorg session fails to load.
<stratofall>
I am not currently in Crux, but was looking for any ideas to try on my next attempt.
<jaeger>
stratofall: I might be misremembering but I think the fix was adding 'nvidia_dgm.fbdev=1' to the boot cmdline
<jaeger>
er... 'nvidia_drm.fbdev=1', typo in the first one
<jaeger>
In my case it didn't fail to load the session, just had to wait for the delay
<stratofall>
Alright, thank you, I will definately give this a shot.
<stratofall>
I had installed the driver. Reloaded dracut, made sure the modules were loaded by checking lsmod. The xorg log file showed it loading the modules. But ultimately failing and closing back to the command line.
<stratofall>
Just as a quick overview.
<jaeger>
What error(s) did it eventually throw?
<stratofall>
I wish I wrote it down. But it was something like, no compatible card found. I think. Although it shows and seems to be loaded during the dmesg.
<jaeger>
If it happens again, I suggest pasting the log somewhere
<stratofall>
Yeah, will do.
stratofall has quit [Quit: Lost terminal]
tsundokuboi has quit [Ping timeout: 260 seconds]
tsundokuboi has joined #crux
leah2 has quit [Server closed connection]
leah2 has joined #crux
tsundokuboi has quit [Read error: Connection reset by peer]
tsundokuboi has joined #crux
_whitelogger has joined #crux
tsundokuboi has quit [Ping timeout: 260 seconds]
tsundokuboi has joined #crux
tsundokuboi has quit [Ping timeout: 260 seconds]
tsundokuboi has joined #crux
stratofall has joined #crux
<stratofall>
I tried with the above suggestion. I am still getting the same result. This is a fresh install of Crux. I am using startx, and my .xinitrd only has exec i3 in it.
<cruxbridge>
<pitillo (@pitillo:crux.nu)> so isn't there a nvdia_uvm module?
<stratofall>
I can modprobe nvidia_uvm but with it loaded same issue. It's not sticking after reboots.
<cruxbridge>
<pitillo (@pitillo:crux.nu)> you can force manual load of a module on rc.local, or check archwiki udev rule to see if it's persistent after a reboot... but if it doesn't solve the xorg start issue... doesn't matter really
<stratofall>
Yeah, I can load it and test, but same Xorg.log error.
<cruxbridge>
<pitillo (@pitillo:crux.nu)> is your user in the video group? just to confirm
<stratofall>
audio::16:stratofall
<stratofall>
video::17:stratofall
<stratofall>
Now I'm starting to wonder if I need some of the noveau or any other random kernal modules enabled...
<cruxbridge>
<pitillo (@pitillo:crux.nu)> ummm I would say that you shouldn't be using any nouveu module if you are using proprietary driver... but may be any other missing kernel module
<stratofall>
Yeah, I can't say I've had this issue before with nvidia cards. Guess I jumped the gun possibly on this one. My previous cards work without issue.
z812 has quit [Server closed connection]
z812 has joined #crux
<stratofall>
I was testing three cards. Gigabyte 4080 Super, Asrock 7900xtx and this Asus 5080 Astral.
<stratofall>
Anyways, I'll watch chat in the logs at least. But I'm heading out. I really do appreciate the suggestions and someone else taking a look at this.
<cruxbridge>
<pitillo (@pitillo:crux.nu)> you are welcome, I hope someone with a similar card can give a real hand on this
flabba has joined #crux
lavaball has quit [Remote host closed the connection]
lavaball has joined #crux
tsundokuboi has joined #crux
tsundokuboi has quit [Ping timeout: 260 seconds]
bleb has quit [Server closed connection]
bleb has joined #crux
flabba has quit [Ping timeout: 245 seconds]
flabba has joined #crux
<jaeger>
Very odd, haven't ever seen that error before. Can you also paste an 'lspci' output at some point?
farkuhar has quit [Quit: nyaa~]
tsundokuboi has joined #crux
bleb has quit [Ping timeout: 258 seconds]
remiliascarlet has quit [Ping timeout: 258 seconds]
<jaeger>
Note that I only have 3000-series cards to test
stratofall has joined #crux
<stratofall>
I just spent a while, getting kernel 6.16.3 installed. Testing what I could. I ended up removing the Crux nvidia package. I tried the nvidia driver from their website. Same issue... I uninstalled that and then manually did the open driver MIT licence. And... I am now in i3.
<jaeger>
Interesting, maybe my open port would have fixed it, too, then. Very odd
<stratofall>
I'll remove this and try the package you built in just a moment. I guess I didn't realize the 5xxx series cards required the open driver. I thought they could still use the proprietary one.
<jaeger>
Maybe that's nvidia's way of forcing the move
<stratofall>
bbiaf, gonna try and test before I head to work.
stratofall has quit [Quit: leaving]
stratofall has joined #crux
<stratofall>
I must be doing it wrong. I thought pkgadd nvidia-open-test should work. I get invalid package name.
<cruxbridge>
<pitillo (@pitillo:crux.nu)> unpack that file, inside probably you will find two packages.... nvidia#.... and nvidia-open#.... so you can then add the nvidia-open one
<stratofall>
ahh yeah, completely my bad.
<stratofall>
I'll have to try once I get home from work. Thank you all. I will report if that package works tonight.
<jaeger>
Sounds good
DaViruz has quit [Ping timeout: 260 seconds]
DaViruz has joined #crux
<r0ni>
i see complaints over many broken things with 6.16... tbh I'd stay away from it.. usb3, i915, nvidia, diff places, diff issues... i'll stick to 6.12 for now ;)