From owner-freebsd-hackers  Wed May 15 10:15:18 1996
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id KAA05671
          for hackers-outgoing; Wed, 15 May 1996 10:15:18 -0700 (PDT)
Received: from dyson.iquest.net (dyson.iquest.net [198.70.144.127])
          by freefall.freebsd.org (8.7.3/8.7.3) with ESMTP id KAA05664;
          Wed, 15 May 1996 10:15:14 -0700 (PDT)
Received: (from root@localhost) by dyson.iquest.net (8.7.5/8.6.9) id MAA22723; Wed, 15 May 1996 12:14:47 -0500 (EST)
From: "John S. Dyson" <toor@dyson.iquest.net>
Message-Id: <199605151714.MAA22723@dyson.iquest.net>
Subject: Re: A question for the VM gurus..!
To: jgreco@brasil.moneng.mei.com (Joe Greco)
Date: Wed, 15 May 1996 12:14:47 -0500 (EST)
Cc: dyson@freebsd.org, hackers@freebsd.org
In-Reply-To: <199605151445.JAA12093@brasil.moneng.mei.com> from "Joe Greco" at May 15, 96 09:45:08 am
X-Mailer: ELM [version 2.4 PL24 ME8]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-hackers@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

> 
> Not portable.  Solaris, at least, implements the madvise() stuff..
> 
Okay..

>
> What it seems to do:
> 
> If you set MADV_SEQUENTIAL on a region, if it has to fault a page "n" in, it
> looks like it discards all pages from 0 to n-1 in that region..  notably it
> does NOT seem to do anything if it doesn't have to fault a page in.
> 
Is there only one process accessing the region?  What happens if there
are other processes using that region also?  Otherwise, that is easy to
do in our VM system.

>
> MADV_RANDOM seems to cause a lot more faults if you are doing sequential
> accesses, as far as I can tell it just tells Solaris not to read ahead.
> 
That is easy also.

>
> MADV_DONTNEED appears to junk pages (asynchronously, from what I can tell)
> 
Easy.

>
> MADV_WILLNEED appears to fault pages (again, asynchrnously, from what I can
> tell).
> 
Hmmm...  time to implement kernel threads.

> 
> I guess in the _short_ term I am most interested in support for
> MADV_DONTNEED, it is the most generally useful change for my application.
> However, it would be nice to have a full suite of this stuff in the future.
> 
Okay, I'll work on it this weekend.  Probably implement everything except
WILLNEED, and will give WILLNEED lip-service by making sure all of the
pages are activated (and perhaps mapped), if they are already in memory.

I will work on it this weekend or soon thereafter.

John