From owner-freebsd-current  Tue Mar 12 13:18:34 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id NAA17465
          for current-outgoing; Tue, 12 Mar 1996 13:18:34 -0800 (PST)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id NAA17453
          for <current@freebsd.org>; Tue, 12 Mar 1996 13:18:28 -0800 (PST)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.12/8.6.9) id IAA05823; Wed, 13 Mar 1996 08:11:45 +1100
Date: Wed, 13 Mar 1996 08:11:45 +1100
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199603122111.IAA05823@godzilla.zeta.org.au>
To: bde@zeta.org.au, wollman@lcs.mit.edu
Subject: Re: cvs commit:  src/sys/sys conf.h
Cc: current@freebsd.org
Sender: owner-current@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

>> [using u_longs where u_ints are sufficient]
>> It doesn't help at runtime, and may waste space and time.

>For any sane architecture, that seems unlikely.  What might happen on
>insane architectures doesn't bear thinking about.

Every 32-bit system should have relatively slow 64-bit longs implemented
in software and no long longs.

The i386 architecture provides many examples where less commonly used
types are inefficient.  You have to think about it even though it is
insane :-).

Bruce