LGTM. You may want to add a manual page changes as well.

Looks fine to me, however please let it be further reviewed by and obtain approvals from ports committers.

Move the return value description above usage details.

• Remove "the snapshot of"
• grammar fix

In D36943#839227, @markj wrote:

BTW, a number of refcount functions are not documented here: acquiren, releasen, acquire_if_gt, acquire_if_not_gt.

I will commit this.

In D30050#793887, @pauamma_gundo.com wrote:

Doesn't explain why the manpages group reviewer.

Works on the 10.0.20348.524 Hyper-V hypervisor, with AMD Epyc 7443 CPU.

Superseded by https://reviews.freebsd.org/D34444.

Hi. Here is my most recent snapshot: https://github.com/khng300/freebsd-src/commit/d05aa814f33f003ac12c1631bd69a2e90ddb0401 .

Is anyone going to merge this?

In D34097#771358, @bojan.novkovic_fer.hr wrote:

In D34097#771212, @khng wrote:

Do you think fattach binding to a file in the file system namespace is a necessary characteristic that we need to replicate as well? If not, I think another less-compatible approach is to separate the namespace all together, just like the posixshm facilities.

I don't think that it is a necessary part for the IPC mechanism itself and that it is mostly needed for compatibility.
Separating the namespace sounds like a good idea, and I would be more than glad to update the differential using this approach if you think it is the way to go.

It sounds like a better idea to me to not interact with file system namespace, thus there will be no hassle dealing with file system events like mount/unmount. I am not sure if others oppose this approach though.

Do you think fattach binding to a file in the file system namespace is a necessary characteristic that we need to replicate as well? If not, I think another less-compatible approach is to separate the namespace all together, just like the posixshm facilities.

Oops.

Hmm, osdep/PortMapper.cpp is not being compiled into -one. I need to investigate it further.