*BSD News Article 72491


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!nntp.coast.net!news-res.gsl.net!news.gsl.net!usenet.eel.ufl.edu!news-feed-1.peachnet.edu!paperboy.wellfleet.com!newsmaster
From: Robert Withrow <bwithrow@baynetworks.com>
Newsgroups: comp.unix.bsd.freebsd.misc
Subject: Re: DAT Hell
Date: Mon, 01 Jul 1996 09:48:19 -0400
Organization: Bay Networks, Inc.
Lines: 24
Message-ID: <31D7D723.794BDF32@baynetworks.com>
References: <4pvmjj$flb@mercury.mcs.com> <4q0j0s$h0@uriah.heep.sax.de>
	  <31C8A0C4.15FB7483@rwwa.com> <4qgo0o$q31@uriah.heep.sax.de>
	  <31D027F0.41C67EA6@baynetworks.com> <4quof1$9p@uriah.heep.sax.de>
NNTP-Posting-Host: dino.wellfleet.com
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Mailer: Mozilla 2.01 (X11; I; FreeBSD 2.1.0-RELEASE i386)

J Wunsch wrote:
> The problem is that there are far too many different timeout values
> for the different SCSI commands. 

It seems like one could make many of them factors of some
small set of ``primary'' timeouts, though.

> but still then, you have to assign
> them on a per-device basis (not per-driver), since e.g. a Sony DAT
> will require far more time to rewind than the good ol' Archive Viper
> you happen to have in the same machine (using the same `st' driver).

Given the designof the system variable interfaces this should,
in theory, be possible.  It is simply a matter of selecting
the correct ``instance ID''.  A uniform instid tagging for
device things seems like a good idea anyway.

I'd suggest that the architects decide what the correct instance
ID for device variables should be, even though I know actually
getting code to do it would have to await a suitable volunteer...

-- 
Robert Withrow -- (+1 508 436 8256)
BWithrow@BayNetworks.com