[jackpot51]

Namespaces on redox are probably what you are looking for. They are incredibly powerful, but difficult to describe. Processes can offer isolated filesystems called "schemes". These are usually at the default namespace (1). Processes can move to the null namespace (0) which disallows them from opening file descriptors from most schemes. Moving to a new namespace allows a process to control the view of the system for all its children. This can be used, for example, to implement chroot.

Syscalls are mostly sync, but with some syscalls for doing blocking operations asynchronously, like on Linux.