Function io_max_device_available 

pub fn io_max_device_available(path: &Path) -> bool

Whether the filesystem behind path has a block device an io.max ceiling can name.

Test support beside cgroup_delegation_available, for the half of the io.max scenario that needs a ceiling to actually apply: a checkout on tmpfs, on overlayfs or on a FUSE mount has no device for the block layer to bind a limit to, and enforced is unreachable there however well the host delegates cgroups. Answered through [io_max_device], the same resolution a launch performs, so a test cannot stand down on a host the launch would have satisfied. Off Linux there is no io.max at all.