Re: sysctl() bugs

["Jason V. Miller" <jmiller%securityfocus.com@localhost>]

On Fri, Sep 12, 2008 at 06:10:54PM -0400, der Mouse wrote:
> I just had occasion to learn how sysctl hw.iostats works on 4.0.  This
> has pointed up a bug in sysctl(), though whether the code or the doc is
> not clear to me.  The manpage says the return value on success is the
> number of bytes copied out, but I'm seeing zero on success.  (The byte
> count is being stored through the data-length pointer correctly.)
> 
> Which is right?  (I can generate a PR, but want to know whether it's
> sw-bug or doc-bug.)

Someone brought this up last month [1], and the thread concluded with the
suggestion that it was a documentation bug. I'm not sure if it was actually
fixed or not, so it might be worth opening a PR.

FWIW, it looks like FreeBSD [2] returns 0 on success, or -1 on failure, and
OpenBSD [3] returns -1 on failure with no mention of any specific value on
success. This seems to support the conclusion that it's a doc-bug.

[1] http://mail-index.netbsd.org/tech-kern/2008/08/19/msg002473.html
[2] 
http://www.freebsd.org/cgi/man.cgi?query=sysctl&sektion=3&apropos=0&manpath=FreeBSD+7.0-RELEASE
[3] 
http://www.openbsd.org/cgi-bin/man.cgi?query=sysctl&sektion=3&arch=&apropos=0&manpath=OpenBSD+Current

-- 
Jason V. Miller