librc: don't create /run/openrc itself on demand.

[navi-desu]

this is a temporary fix, in a system that boots with / as rw, /run unmounted, and has rc_parallel on, openrc would create a dirfd to /run/openrc before mounting /run, thus keeping an open fd to the old /run mountpoint, and creating 'exclusive' lockfiles under it, that subsequent services can't see nor lock on.

Fixes: #859
Fixes: 07d6f4d

[N-R-K]

I don't think I fully follow what's going on. Is it:

1. openrc creates stuff under /run/openrc but /run isn't mounted yet.
2. /run gets mounted and the older stuff can no longer be seen.

Or:

1. openrc opens /run but doesn't create anything.
2. /run is mounted
3. Now it creates stuff, but the dirfd is pointing to the old /run.

If it's the former then I don't think this fixes it, if it's the latter then this patch makes sense. But why call it "temporary fix"?

[navi-desu]

I don't think I fully follow what's going on. Is it:

1. openrc creates stuff under /run/openrc but /run isn't mounted yet.

2. /run gets mounted and the older stuff can no longer be seen.

Or:

1. openrc _opens_ `/run` but doesn't create anything.

2. /run is mounted

3. Now it creates stuff, but the dirfd is pointing to the old `/run`.

If it's the former then I don't think this fixes it, if it's the latter then this patch makes sense. But why call it "temporary fix"?

1. openrc creates /run/openrc and opens a dirfd to it
2. /run gets mounted, so /run/openrc can no longer be seen by new processes
3. then openrc creates /run/openrc/exclusive under the dirfd for the old /run
4. new processes try to check for exclusive in the new /run, find nothing

so it's both, at the same time, and this fixes the former case because it (now, i had forgotten to) restores old behaviour of only creating /run/openrc when rc_deptree_update_needed is called, which is always after do_sysinit

this is also why it's temporary, in my view, i don't like this "solution" of creating the runtime dir as a side-effect of this one "check" function