Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watnot!watmath!clyde!cbatt!ihnp4!ptsfa!lll-lcc!seismo!mimsy!chris
From: chris@mimsy.UUCP
Newsgroups: comp.unix.questions
Subject: Re: input editing at interrupt time (was Re: tty driver advice)
Message-ID: <5632@mimsy.UUCP>
Date: Sun, 1-Mar-87 07:22:21 EST
Article-I.D.: mimsy.5632
Posted: Sun Mar  1 07:22:21 1987
Date-Received: Mon, 2-Mar-87 05:44:06 EST
References: <75@pixdoc.UUCP> <5623@mimsy.UUCP> <14304@sun.uucp>
Organization: U of Maryland, Dept. of Computer Science, Coll. Pk., MD 20742
Lines: 15

In article <14304@sun.uucp> guy%gorodish@Sun.COM (Guy Harris) writes:
>A streams-based driver could do the canonicalization in a service
>procedure; they are not run at interrupt priority level.

Actually, I had considered that.  A rather standard scheme for
interrupt processing is to drop the event into a queue, then schedule
a special process to run it.  That process is, in essence, an
interrupt (and is often implemented using software interrupts).

4.2 and 4.3BSD use this scheme for network input.  That is the
difference between splnet() and splimp(): the former blocks software
input interrupts, while the latter blocks hardware input interrupts.
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7690)
UUCP:	seismo!mimsy!chris	ARPA/CSNet:	chris@mimsy.umd.edu