-
Notifications
You must be signed in to change notification settings - Fork 12
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
No windows when switched to eGPU #29
Comments
@olivergs You can use both method 2 and method 3 at the same time and this is my preferred way for my personal system using Fedora 40 as method 3 adds the specific variables used by the GNOME desktop. |
Thanks @ewagner12 for the fast response. My setup is a LG Gram 15 laptop with intel Irix XE iGPU, connected via USB-C to an LG display. Only one cable is connected to the laptop. That cable is most of the time connected to the USB-C connector in the monitor, but I was hotplugging the eGPU to use it as a CUDA accelerator with no display. The eGPU is a TH3P4G3 with an RTX3060Ti 8GB card on it. Before using your script I was able to connect the laptop to the egpu, and from the same egpu mount I can connect to the USB-C monitor, so I can have the eGPU connected for CUDA work, but still use the USB-C display. After using it, I was able to boot in the card. No display on USB-C. neither in laptop screen, but in the NVIDIA DP I can see the gray screen, then login to desktop and see the background. The problem is that neither mouse, nor keyboard are responsive after booting in the eGPU mode. The new problem is that even reconfiguring everything to defaults and removing the script using the uninstall option, I can't plug the eGPU anymore, because if I do, then the USB-C display freezes, no display on NVIDIA DP so I've to power off the laptop. I can't see anything so I do not know if it is a kernel oops or what it is. I will try as you said and set this to use mode 2 and 3 at the same time and let's see what happens. Thanks again for your help. |
After some extra testing, it seems that if I connect the eGPU to the monitor through DP, and try to use the eGPU secondary thunderbolt port to downlink to the monitor, wayland gets crazy and is not able to understand where the display should go and stops working completely. Now I've removed the DP cable from the RTX3060 it works as before: I can connect the eGPU and from the eGPU rack directly to the USB-C in the monitor and worked as before. Anyway I still have the problems to connect the eGPU as primary display, but I will continue experimenting. I suppose it is just a small configuration quirk that is keeping me away of getting the windows displayed. |
I've set up the mode 2 and was able to boot and get display on the eGPU DP out.
The problem is that GDM just shows a gray screen and if I press enter and then my user password it logs in into my normal Wayland but it only shows the background image. No GNOME shell, no windows. Just the background.
If I go to other consoles by using Ctrl+Alt+F3 they do not show anything, and if I come back to my wayland session, some times it shows the background like if I pressed the windows button in the GNOME session (the desktop backgound is zoomed out in the center of the screen) but there is no favorite bar neither the top bar.
Maybe Wayland is not aware of where to send the display for the windows? Do not know how to debug this. Can you point me to the right direction please?
My system: Fedora 40
The text was updated successfully, but these errors were encountered: