erh@swapsimple.com wrote:

> 	and nothing in those says that I should grovel through the 
> (rather long) _C_FUNCTION_API_ in sysctl(3) even though I'm not trying to 
> write a C program.

Both sysctl(3) and sysctl(8) describe the values that can be used by..
well, sysctl. Since neither veriexecctl(8) or veriexec(4) is used to
set any of those values, like ~every other sysctl knob, these are
documented in the sysctl man-pages.

> 	ah, of course.  Why bother to document the details of veriexec
> under the veriexec man page when it can go in a general sysctl man page.

...together with the details of every other sysctl knob? :)

> Well, whatever.  I'll agree that duplication of information is bad, but the
> veriexec.4 page at least deserves to have an explicit mention of what can be
> found in the other man pages and why I should spend time looking through
> docs that don't at first glace seem to be what I need.

veriexec(4) is to Veriexec what pf(4) is to pf.

> "For details on the values for kern.veriexec.strict see VERIEXEC_STRICT in
> sysctl(3).  For a more in depth discussion of veriexec see the NetBSD guide
> at http://netbsd.org/guide/en/chap-veriexec.html."

That is already there in a shorter, more consistent way:

SEE ALSO
     sysctl(3), sysctl(8), veriexecctl(8)
      ^^^^

> or maybe even an entire veriexec.8 page.

That is a solution I like -- doing a man-page similar to the Veriexec
chapter in the NetBSD guide. You can mail me the man-page and I'll have
it commited after review. :)

Thanks,

-e.

-- 
Elad Efrat
PGP Key ID: 0x666EB914