From owner-cvs-sys  Tue Dec 10 23:04:51 1996
Return-Path: <owner-cvs-sys>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.4/8.8.4) id XAA22823
          for cvs-sys-outgoing; Tue, 10 Dec 1996 23:04:51 -0800 (PST)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.8.4/8.8.4) with ESMTP id XAA22815;
          Tue, 10 Dec 1996 23:04:15 -0800 (PST)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.8.3/8.6.9) id NAA27773; Wed, 11 Dec 1996 13:16:28 +1100
Date: Wed, 11 Dec 1996 13:16:28 +1100
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199612110216.NAA27773@godzilla.zeta.org.au>
To: jkh@time.cdrom.com, toor@dyson.iquest.net
Subject: Re: cvs commit: src/sys/ufs/ufs ufs_readwrite.c
Cc: cvs-all@freefall.freebsd.org, CVS-committers@freefall.freebsd.org,
        cvs-sys@freefall.freebsd.org, dyson@freefall.freebsd.org
Sender: owner-cvs-sys@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

>> #ifdef MMAP_ACTUALLY_BOUGHT_US_PERFORMANCE
>> ..
>> #endif
>> 
>> Be more appropriate?
>> 
>For us, since 2.1.X, I think that your suggested name is more accurate.
>Also, I do think that we'll generally see marginally improved perf now
>since the recent change.  Note that if you enable it in your private
>sources, you will likely want to do the madvise -- if you don't, the
>perf will be just as lousy as it was before.

There are a number of clones of the cmp and cp code in the tree.  E.g.,
xinstall.c has a version of cmp with an arbitrary 8MB limit like cp.
It's not clear that the mmap() version is faster than the read() version.
The read() version uses the semi-arbitrary buffer size of MAXBSIZE.  This
may be too big for starting up (it is too big if the files compare unequal
near the start, but they are likely to compare equal).  `install' sometimes
installs the source, so it should attempt to keep the source cached.

Bruce