*BSD News Article 42165


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!bunyip.cc.uq.oz.au!munnari.oz.au!news.hawaii.edu!ames!hookup!olivea!gossip.pyramid.com!sun.digidesign.com!netcomsv!netapp.com!netapp.com!not-for-mail
From: guy@netapp.com (Guy Harris)
Newsgroups: comp.unix.bsd
Subject: Re: Why select() returns ``exceptional'' for files?
Date: 28 Jan 1995 21:06:35 -0800
Organization: Network Appliance Corporation
Lines: 36
Message-ID: <3gf7or$3k2@nova.netapp.com>
References: <3fois1$5d5@shore.shore.net> <bakulD2wLz4.5Gn@netcom.com> <3g65hg$2ns@nova.netapp.com> <bakulD33vxv.KIt@netcom.com>
NNTP-Posting-Host: 192.9.200.13

Bakul Shah <bakul@netcom.com> wrote:
>guy@netapp.com (Guy Harris) writes:
>>That way, the issue of select-on-exceptional-condition on files, and of
>>changes to the NFS protocol to do stronger cache consistency (which is
>>what eliminating polling for this case really implies, as far as I can
>>tell), are decoupled - which, as far as I'm concerned, is as it should
>>be, as there's no reason to deny people any benefits that might accrue
>>from select-on-exceptional-condition on files merely because we don't
>>have strong(er) cache consistency for NFS.
>
>I don't know.  Seems to me that if there was a protocol that did
>cache consistency, select-on-exceptional-condition-on-files
>(SOECOF) would be almost for free.

That's not a reason to deny people any benefits that might accrue from
select-on-exceptional-condition on files merely because we don't have
strong(er) cache consistency for NFS.

SOECOF on NFS is probably easier to implement than real cache
consistency for NFS, and doesn't require you to design a protocol and
get all the NFS client and server vendors to implement it.

I.e., the fact that SOECOF would come for free *if* you had real cache
consistency isn't really relevant here, given that:

	1) it wouldn't really come for free, given that the vast
	   majority of NFS servers aren't likely to implement a real
	   cache consistency protocol tomorrow, so, unless you want to
	   render applications that use SOECOF unusable over NFS (which
	   may keep people from using SOECOF in their applications),
	   you'll have to implement it over Boring Old NFS V2 and Boring
	   Old NFS V3 anyway;

	2) there's probably enough work involved in designing a real
	   cache consistency protocol for NFS *and* getting it adopted
	   that "for free" is somewhat misleading.