From owner-freebsd-hackers  Tue Oct 29 15:43:16 1996
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id PAA10179
          for hackers-outgoing; Tue, 29 Oct 1996 15:43:16 -0800 (PST)
Received: from gatekeeper.ray.com (gatekeeper.ray.com [138.125.162.1])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id PAA10161
          for <freebsd-hackers@FreeBSD.org>; Tue, 29 Oct 1996 15:42:59 -0800 (PST)
From: Gregory_D_Moncreaff@ccmail.ed.ray.com
Received: (mailer@localhost) by gatekeeper.ray.com (8.7.5/8.7.3) id SAA27797 for <freebsd-hackers@FreeBSD.org>; Tue, 29 Oct 1996 18:41:49 -0500
Received: from zeus.ed.ray.com by gatekeeper.ray.com; Tue Oct 29 18:41:09 1996
Received: from ccmail.ed.ray.com by ZEUS.ED.RAY.COM (PMDF V4.2-10 #4335) id
 <01IB84DD7JV4001TG9@ZEUS.ED.RAY.COM>; Tue, 29 Oct 1996 18:43:15 EST
Date: Tue, 29 Oct 1996 18:24 -0500 (EST)
Subject: A better kernel random() ??
To: freebsd-hackers@FreeBSD.org
Message-id: <01IB84DEEOQM001TG9@ZEUS.ED.RAY.COM>
MIME-version: 1.0
Content-transfer-encoding: 7BIT
Sender: owner-hackers@FreeBSD.org
X-Loop: FreeBSD.org
Precedence: bulk

        I need a random (psuedo or otherwise) number generator
        for use in kernel network protocols (collision backoff,
        jitter, etc).  I looked, at libkern/random.c and the
        seed is constant.  

        I am looking for suggestions/algorithims for getting a
        more random sequence of numbers.  The two machines are
        running the same kernel and are timed synced.

        I am thinking that a utility program could be written
        to change this seed value via a kernel memory write
        based on some combination of information, such as
                process id,
                address of a network interface,
                ???

        Should the result of such a computaion have any numeric
        properties:
                number of bits set, oddness, not divisible by a
                single digit prime number, etc?


        Any other, better hopefully, ideas?

thanks,
g