From owner-freebsd-hackers  Thu Feb 15 06:22:01 1996
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id GAA01703
          for hackers-outgoing; Thu, 15 Feb 1996 06:22:01 -0800 (PST)
Received: from terra.Sarnoff.COM (terra.sarnoff.com [130.33.11.203])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id GAA01698
          for <hackers@FreeBSD.org>; Thu, 15 Feb 1996 06:21:58 -0800 (PST)
Received: (from rminnich@localhost) by terra.Sarnoff.COM (8.6.12/8.6.12) id JAA14407; Thu, 15 Feb 1996 09:17:31 -0500
Date: Thu, 15 Feb 1996 09:17:31 -0500 (EST)
From: "Ron G. Minnich" <rminnich@Sarnoff.COM>
To: Luigi Rizzo <luigi@labinfo.iet.unipi.it>
cc: hackers@FreeBSD.org
Subject: Re: Paging questions
In-Reply-To: <199602151302.OAA07491@labinfo.iet.unipi.it>
Message-ID: <Pine.SUN.3.91.960215090347.14306B-100000@terra>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-hackers@FreeBSD.org
Precedence: bulk

> More questions/thoughts on paging support for small-memory system.
> First, a preliminary question:
>     How expensive is it to get a page fault which immediately
>     returns without doing anything (lets call this a Quick Fault)

OK, i did a test of this on a p90. If you short stop the trap code in 
i386/i386/trap.c and just return on a user fault from a page fault it 
happens in about 4 microseconds on a p90, fbsd 2.05R, neptune chipset. 

IF you do much more, and in particular if pmap_unuse_pt gets called, it 
goes to more like 100. 

ron