FWIW, I tend to think eventually everything will. But I am more worried about Wayland than systemd for the *BSDs. AFAIK, all systemd does is start items KDE needs, so with some work I think workarounds can be created.

But, once Firefox changes to Wayland only, that will lock out NetBSD and OpenBSD. I heard FreeBSD has some kind of support for Wayland, but that probably forced FreeBSD to import a lot of Linuxisms.

I get the impression OpenBSD will not bring in those Linux specific items and NetBSD already posted Wayland is to big a project for them to import at this point.

Wayland needs... I think some EGL stuff to create OpenGL or Vulkan contexts in Wayland apps and a kernel API for graphics modesetting (resolution and refresh rate + misc related stuff). Also input support in libinput. Modesetting is perhaps the biggest piece of work, but I dare say not a huge one compared to having drivers for modern GPUs in the first place.

What's going to suck is that every major compositor +wlroots for the non-major ones is also going to need patches to use the BSD modesetting API - something that the BSD guys don't have direct control over and which is maybe a little out of their traditionalist(?) comfort zone.

Or are the BSDs simply going for a copy of the Linux KMS (kernel modesetting) API? Wouldn't be the worst idea, Linux finally seems to have something robust there after earlier attempts that were considered problematic.

OpenBSD is working on Wayland support. They aren’t doctrinally opposed to it, they just want to implement it properly according to their standards. They are having to implement their own equivalents to udev etc to make it work.

FreeBSDs Wayland support is pretty good, it supports Plasma perfectly and most wlroots based compositors. OpenBSD is more experimental, I know Sway (and presumably other simple wlroots based compositors) work alright but I wouldn't hold my breath for KDE support. I haven't heard anything from NetBSD or DragonflyBSD

I agree, Wayland is a much larger problem.