Alexey Dobriyan <adobriyan@sw.ru> writes:

> I found why first version should be rejected, and, no, it is not
> documentation updates. Here is version 2:

I care about the documentation because it is the best thing
we have for enforcing sane sysctl table usage.

After looking at the problem.  While I can't remove  ctl_name
from the sysctl tables easily.  What I can do is check at
registration time if a sysctl table is sane, and print
an error message and fail to register that table if it has problems.

Once that is in place I will have no problems killing CTL_UNNUMBERED.

Eric