>[... sysctl architecture]
>
>All sounds good to me.
>
>I dont want to tell tales, but when adding the bpf buffersize sysctl,
>we found out the hard way that the raw sysctl_createv() API is
>unecessarily error-prone. :-/.

well...that's only because i tossed off some code into a piece of
email without really thinking it through.  the error was found very
easily, would not have occurred if i hadn't used cut-n-paste, and
might veyr well have occurred for the same reason even if i was using
shorter "create" macros.

fwiw, we could have "tested it" as an lkm first, to see how it would
do.  :)

-- 
|-----< "CODE WARRIOR" >-----|
codewarrior@daemon.org             * "ah!  i see you have the internet
twofsonet@graffiti.com (Andrew Brown)                that goes *ping*!"
werdna@squooshy.com       * "information is power -- share the wealth."