AFAIK, the Mach microkernel was version 3.0 onwards, and used in projects such as IBM Workplace OS, MKLinux, and GNU HURD. These operating systems rarely progressed past the experimental or research project phase.
AFAIK, the Mach microkernel was version 3.0 onwards, and used in projects such as IBM Workplace OS, MKLinux, and GNU HURD. These operating systems rarely progressed past the experimental or research project phase.
MacOS uses a BSD personality colocated in the same address space as the Mach part, so it is no pure microkernel system.
You can check the last (OpenStep-based) Rhapsody kernel sources from Darwin 0.3 at
https://github.com/evolver56k/Darwin-0.3/tree/master/kernel-...
and the first archived version of the MacOS X xnu kernel at
Their long term roadmap being every time a new framework for userspace drivers gets introduced, the developer community is given one year transition period to migrate their kernel extensions to userspace, before that capability is removed from the following macOS version.
End goal being macOS will provide a microkernel like experience for everyone outside Apple's HQ.
GNU Hurd was based on CMU Mach 3 for a while in lieu of Mach 4, which was an oft cited reasons for delays with the Hurd development due to missing critical features anticipated to arrive in Mach 4. Then CMU dropped Mach, the University of Utah picked it up and developed for some time, and finally the active development fizzled out altogether, which added the further delays with GNU Hurd. GNU Mach was branched off Mach 4 at some point and has been (was?) developed as a separate endeavour.
In reality, the CPU's at the time were not powerful enough to operate message passing, microkernel based operating systems so, whilst the idea itself was great, it was not a viable one. It could have been a different story today, but with the large scale operating system reasearch and experimentation all but having ceased in the last couple of decades, it does not seem likely there will be a large scale microkernel (or something else) success story any time soon, which is kinda sad.
Still very few developers behind it though, compared to other Unix/-like projects.