Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 4 Jun 1998 17:43:51 -0500 (EST)
From:      "John S. Dyson" <dyson@FreeBSD.ORG>
To:        nate@mt.sri.com (Nate Williams)
Cc:        mike@smith.net.au, nate@mt.sri.com, dyson@FreeBSD.ORG, hackers@FreeBSD.ORG
Subject:   Re: kernfs/procfs questions...
Message-ID:  <199806042243.RAA00404@dyson.iquest.net>
In-Reply-To: <199806042233.QAA04941@mt.sri.com> from Nate Williams at "Jun 4, 98 04:33:03 pm"

next in thread | previous in thread | raw e-mail | index | archive | help
Nate Williams said:
> > > I argue shouldn't be exposed to the users then.  If it's exposed, it
> > > should be documented.  In other words, sysctl should go away since very
> > > few (if any) of it's knobs are documented except accidentally.
> > 
> > You aren't clear *which* sysctl should go away.  If you mean sysctl(8), I
> > hope you will be removing gdb, nm, hexdump, etc. as well, as these are 
> > all user-unfriendly tools designed for studying and/or adjusting the 
> > state of complex, undocumented things.
> 
> But users aren't expected to use gdb/nm/hexdump, but sysctl is.  Many of
> these parameters *should* be tweaked to get better performance, avoid
> errors, etc...
>
Only some of them, if any.  Most of the manipulation should be in 
the rc files.  If you want to tweak your system, then tweak it.  There
is no equality of expertise here, and just stay away from what you
don't understand.

> 
> > sysctl(8) is a tool for writing scripts, and performing tasks which 
> > have been previously documented.  It is not a user-friendly interface, 
> > however it still exists to serve a purpose.
> 
> What tasks have been documented that are used by sysctl?
> 
> Tell me where 'sysctl -w net.inet.tcp.rfc13230=0' is documented.  (And,
> no I don't consider it's presence in rc.network documentation. :()
> 
You are hereby appointed to document what you know, to get it started.
Of course, /kernfs is exactly the same thing, except doesn't do enough
for what we need.  I am NOT for implementing a:

echo "0" >/kernfs/net/inet/tcp/rfc1323

It is entirely bogus for the kernel to parse strings like this.  We
probably do it, but that doesn't mean it's right... Even worse would be:

echo "NO" >/kernfs/net/inet/tcp/rfc1323

but in Russia or China, what would one enter????  Parsing of language
or strings in the kernel is really short sighted.  Of course, this
is a degenerate case, but shows that an exposed kernel interface needs
a userland program to be nice.  The sysctl program is such a minimal 
interface.

Try this one:

cat /kernfs/time

Which locale should it use???
Tell me that the above isn't in the domain of hackery anyway???

-- 
John                  | Never try to teach a pig to sing,
dyson@freebsd.org     | it just makes you look stupid,
jdyson@nc.com         | and it irritates the pig.

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199806042243.RAA00404>