Date: Thu, 11 May 2000 15:57:39 -0400 (EDT)
Message-Id: <200005111957.PAA13569@indy.delorie.com>
From: Eli Zaretskii <eliz AT delorie DOT com>
To: Martin Stromberg <Martin DOT Stromberg AT lu DOT erisoft DOT se>
CC: djgpp-workers AT delorie DOT com
In-reply-to: <200005111438.QAA05671@lws256.lu.erisoft.se> (message from Martin
	Stromberg on Thu, 11 May 2000 16:38:51 +0200 (MET DST))
Subject: Re: Perfomance of gc-simple
References:  <200005111438 DOT QAA05671 AT lws256 DOT lu DOT erisoft DOT se>
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

> From: Martin Stromberg <Martin DOT Stromberg AT lu DOT erisoft DOT se>
> Date: Thu, 11 May 2000 16:38:51 +0200 (MET DST)
> 
> But Solaris' man page says:
> "The use of MAP_FIXED is discouraged, as it may prevent an
> implementation from making the most effective use of system
> resources."
> 
> I also saw a similar statement on Linux.
> 
> So it's strange that a test for mmap() uses a feature that is
> discouraged!

It uses MAP_FIXED because Grep uses it.  And Grep uses it because it
does an mmap on each file it processes, and it doesn't want to
relocate the working buffer each time, I guess.

Anyway, Grep only uses mmap when given a special command-line option
now, since using mmap can lose in some rare cases where the normal I/O
wins.  I guess that covers the discouragement as well ;-).