pidfile_open() usage in "mount"

[J David]

On Wed, Aug 26, 2020 at 5:15 PM Rick Macklem <rmacklem at uoguelph.ca> wrote:
> The reason is that, if the local file system being mounted happens to be in
> /etc/exports, then /etc/exports has to be re-loaded for the update.

That makes sense.

> Ideally, all local file systems would be mounted and then one HUP would
> be sent to mountd, but that isn't the way it has worked for a long time
> and changing it would be a POLA violation, imho.

Currently, it happens once per invocation of /sbin/mount, at the very
end, so if you do something that causes multiple mounts (i.e. the -a
flag), it only happens once.  That's a good compromise, and I agree
that decoupling it to happen somewhere else would not be a good idea,
both for the reason you identify, and also because it just doesn't
seem like local filesystems follow the fixed, from-boot-to-shutdown
model that they used to.

Thanks!