From owner-freebsd-scsi  Wed Mar  6 09:16:16 1996
Return-Path: owner-freebsd-scsi
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id JAA05689
          for freebsd-scsi-outgoing; Wed, 6 Mar 1996 09:16:16 -0800 (PST)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id JAA05683
          for <freebsd-scsi@FreeBSD.org>; Wed, 6 Mar 1996 09:16:13 -0800 (PST)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.12/8.6.9) id EAA30849; Thu, 7 Mar 1996 04:13:42 +1100
Date: Thu, 7 Mar 1996 04:13:42 +1100
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199603061713.EAA30849@godzilla.zeta.org.au>
To: jkh@time.cdrom.com, joerg_wunsch@uriah.heep.sax.de
Subject: Re: HP 4080i as both CD and WORM?
Cc: freebsd-scsi@FreeBSD.org
Sender: owner-freebsd-scsi@FreeBSD.org
Precedence: bulk

>> Well, it still suffers from a few idiosyncracsies and design flaws.
>> The most annoying one is the blatant use of UNIT ATTENTIONs in these
>> drives, which causes four messages at boot time, and still required me

>Yes, those are kind of annoying, I agree.  Wouldn't it be permissable

Why does the SCSI driver log them?  I don't want it to tell me that
I've just changed a removable disk.  For a zip drive, after a media
change, the following messages are logged:
 
	sd1(ncr0:5:0): UNIT ATTENTION asc:28,0
	sd1(ncr0:5:0):  Not ready to ready transition, medium may have changed
	sd1(ncr0:5:0): ILLEGAL REQUEST asc:24,0 Invalid field in CDB
	sd1 could not mode sense (4). Using ficticious geometry

Then if the drive is accessed while no media is present:

	sd1(ncr0:5:0): NOT READY asc:3a,0 Medium not present
	sd1(ncr0:5:0): NOT READY asc:3a,0 Medium not present

(one message for the attempted access to the MBR to determine the slices,
and one to attempt to read the disk).

Then if the drive is accessed again while no media is present:

	sd1(ncr0:5:0): NOT READY asc:3a,0 Medium not present

and the Debugger("sdopen: no slices") is called.  Then after continuing
in the debugger, the message for the attempt to read the disk is logged:

	sd1(ncr0:5:0): NOT READY asc:3a,0 Medium not present

Then if the drive is accessed with media present:

	sd1(ncr0:5:0): UNIT ATTENTION asc:28,0
	sd1(ncr0:5:0):  Not ready to ready transition, medium may have changed
	[Debugger("sdopen: no slices") is called]
	sd1(ncr0:5:0): ILLEGAL REQUEST asc:24,0 Invalid field in CDB
	sd1 could not mode sense (4). Using ficticious geometry

There would be more messages from the slice code if the disk didn't have
a valid DOS partition table on it :-].

>to simply loop, say, 5 times with a slight delay each time and if the
>drive is still returning UNIT ATTENTION (reported as "Target Busy", I
>assume?) simply gobble it and try again, or does the fool thing assert
>UA continuously until something else happens?

Zip drives seem to return just one unit attention.  It doesn't matter
if the new media has been present for a long time.  This makes sense.
The driver should be able to handle it easily if it occurs in open().
Handling media changes that occur while the device is open is much
more difficult.

Bruce