RVGPU Running in Nvidia environment Problem.

[nmhn]

We followed the steps in the Readme to set up Weston on Ubuntu 22.04 X11. After initiating RVGPU render and proxy, we launched another Weston instance on the newly created GPU and ran glmark2-es2. However, it seems to be using the local GPU instead of the expected virtio-gpu. Is this normal behavior? The commands we used are as follows:

render

weston --width 2200 --height 1200 &
rvgpu-renderer -b 1280x720@0,0 -p 55667

proxy

sudo -i
modprobe virtio-gpu
modprobe virtio-lo
rvgpu-proxy -s 1280x720@0,0 -n 127.0.0.1:55667

backend

export XDG_RUNTIME_DIR=/tmp
weston --backend drm-backend.so --tty=2 --seat=seat_virtual -i 0

The input and output in the Weston command line are:

glmark2-es2

The render log and lo log are as follows: output_render.log output_lo.log

[FumiyaKohzu]

Thank you for your post. RVGPU requires Mesa to function properly. Please specify Mesa with the LD_LIBRARY_PATH environment variable. Example: export LD_LIBRARY_PATH=/usr/lib/x86_64-linux-gnu We have confirmed stable operation with Mesa version 20.3.5.

We plan to include this information in our README.md. Thank you.