In some email I received from Michael Graff, sie wrote:
> 
> Darren Reed <darrenr@reed.wattle.id.au> writes:
> 
> > > Well, yeah.  My point is that for /kern/kernel to be useful, it has to
> > > provide a symbol table for nlist() to use, and for that, it needs to
> > > have the symbols available to it in the first place.
> > 
> > Not to forget that the symbol table should be dynamic in nature.
> 
> I have long thought that grubbing around in the kernel using nlist
> was just wrong.  kernfs might be something we don't want to require
> but sysctl access isn't optional last time I checked.

Oh, the above is for ddb/crash, etc, as much as it is other tools such
as nm or top, etc, that use nlist.