Having looked at the patch at https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/3304 it seems like a pretty reasonable thing to get put into a point release and as you indicated in the video, point releases in this case are not guarantees of the inclusion of only "bug fixes".

It was implemented in a matter that did not break ABI or backwards compatibility for libmutter ABI 14, so GNOME would still be in the right even if the argument centered around SemVer 2.0.0.

The code that enables the behavior, at least at a quick glance, seems like it would be trivial for Canonical to patch should they decide to be weird about keeping NVIDIA users on an LTS with what could be argued as a sub-optimal experience.

Basically two lines of code: https://gitlab.gnome.org/GNOME/mutter/-/blob/gnome-46/src/backends/native/meta-renderer-native.c?ref_type=heads#L1882

At the end of the day though, I would not be surprised if Canonical just spent some extra engineering time and let it in once it has been validated.

Nvidia secondary GPU copy acceleration (!3304) · Merge requests · GNOME / mutter · GitLab

This fixes three issues that were preventing GPU copies on the Nvidia proprietary driver: gbm_surface_create fails with ENOSYS (Function not...

^GitLab