This can happen when switching to another tty. A test is to run glxgears in a drm session with xwayland, then switch to another tty and run top. The wayfire cpu usage is 100+%. However with this patch, when the session is inactive, the cpu usage is negligible.
One note is that patches were pushed to seatd recently which make the session->active state reliable. At the time of this writing, seatd does not have a release containing these patches, so for this to work reliably, seatd git master is recommended.
This can happen when switching to another tty. A test is to run glxgears in a drm session with xwayland, then switch to another tty and run top. The wayfire cpu usage is 100+%. However with this patch, when the session is inactive, the cpu usage is negligible.
One note is that patches were pushed to seatd recently which make the session->active state reliable. At the time of this writing, seatd does not have a release containing these patches, so for this to work reliably, seatd git master is recommended.