From owner-freebsd-questions  Tue Mar 12 09:44:11 1996
Return-Path: owner-questions
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id JAA00890
          for questions-outgoing; Tue, 12 Mar 1996 09:44:11 -0800 (PST)
Received: from alpha.xerox.com (alpha.Xerox.COM [13.1.64.93])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id JAA00884
          for <questions@freebsd.org>; Tue, 12 Mar 1996 09:44:09 -0800 (PST)
Received: from gemini.sdsp.mc.xerox.com ([13.231.132.20]) by alpha.xerox.com with SMTP id <15041(15)>; Tue, 12 Mar 1996 09:43:29 PST
Received: from gnu.mc.xerox.com (gnu.sdsp.mc.xerox.com) by gemini.sdsp.mc.xerox.com (4.1/SMI-4.1)
	id AA02593; Tue, 12 Mar 96 12:43:30 EST
Received: by gnu.mc.xerox.com (4.1/SMI-4.1)
	id AA08403; Tue, 12 Mar 96 12:43:29 EST
Message-Id: <9603121743.AA08403@gnu.mc.xerox.com>
To: Luigi Rizzo <luigi@labinfo.iet.unipi.it>
Cc: msmith@atrad.adelaide.edu.au, questions@freebsd.org
Subject: Re: non-blocking read ? 
In-Reply-To: Your message of "Tue, 12 Mar 1996 09:33:14 PST."
             <199603121733.SAA20865@labinfo.iet.unipi.it> 
Date: Tue, 12 Mar 1996 09:43:28 PST
From: "Marty Leisner" <leisner@sdsp.mc.xerox.com>
Sender: owner-questions@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk


>
>That's my problem... some programs (noticeably the "harvest" cache)
>work as a single-threaded process which serves concurrent
>requests by trying to avoid blocking calls. It appears to do a nice
>job, but for some things (say reads/seeks etc. from disk) it appears
>that there is no way to avoid delays -- at least to my knowledge.
>
>	Thanks
>	Luigi
>====================================================================
>Luigi Rizzo                     Dip. di Ingegneria dell'Informazione
>email: luigi@iet.unipi.it       Universita' di Pisa
>tel: +39-50-568533              via Diotisalvi 2, 56126 PISA (Italy)
>fax: +39-50-568522              http://www.iet.unipi.it/~luigi/
>====================================================================

How long a "delay" are you having?

You may want to look at the mit pthreads package...

If you start forking I don't see how you can save time...

Maybe an architecture with another process which is a filing daemon, connecting to the
other process with pipes (where each pipe maps to a file).

Select works fine on pipes...

marty