From owner-freebsd-scsi  Sun Jan 14 16:32:08 1996
Return-Path: owner-freebsd-scsi
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id QAA03610
          for freebsd-scsi-outgoing; Sun, 14 Jan 1996 16:32:08 -0800 (PST)
Received: from ref.tfs.com (ref.tfs.com [140.145.254.251])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id QAA03604
          Sun, 14 Jan 1996 16:32:03 -0800 (PST)
Received: (from julian@localhost) by ref.tfs.com (8.6.12/8.6.9) id QAA19339; Sun, 14 Jan 1996 16:31:46 -0800
From: Julian Elischer <julian@ref.tfs.com>
Message-Id: <199601150031.QAA19339@ref.tfs.com>
Subject: Re: Calling st_unmount for nrst* devices
To: gibbs@freefall.freebsd.org (Justin T. Gibbs)
Date: Sun, 14 Jan 1996 16:31:46 -0800 (PST)
Cc: freebsd-scsi@freefall.freebsd.org
In-Reply-To: <199601140308.TAA07484@freefall.freebsd.org> from "Justin T. Gibbs" at Jan 13, 96 07:08:48 pm
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Sender: owner-freebsd-scsi@FreeBSD.ORG
Precedence: bulk

> 
> 
> I don't like the fact that after closing the nrst* devices, we
> don't reenable the eject button.  In fact, although I haven't
> tracked down exactly why, even if you rewind the tape via an "mt
> rewind" after doing some operations on the nrst* devices, the drive
> is still locked.  I think  these patches fix the problem, but I'll
> have to do some more testing tomorrow to make sure.  Perhaps some
> of you would be interested in testing these as well?
hmmmmm
I think this was a deliberate act on my part but probably at a time.
To me it makes perfect sense to have it as it is but I bow to 
public opinion.

however I object to the way you've done it.. on the following grounds..

An "unmount" Ends a session and I want to only call the unmount call in
this case.
if you use the "nrst" device you are deliberatly NOT ending the session
If you had wanted to end the session you should have used "rst"
which is defined as ending a session..

If you want to enable the button between whenever ther device os closed,
then you need simply
execute the scsi_prevent() call from st_close(). To call "st_unmount()
breaks my mental model of what is going on and is misleading..
If you DO enable the button, then pushing the button will
eventually result in a "Unit Attention" which will end the session,
(or should.. will check)
but I don't want to call "st_unmount(), because you are NOT
changing the mount session.. you are NOT reloading
device parameters and driver settings.. you ARE leaving the present
session active.

leave it all the same and call scsi_prevent() from st_close()
directly, (and possibly from st_open() to ensure it's locked when open)

thanks
julian