Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 17 Mar 1999 19:16:56 -0800
From:      Josh MacDonald <jmacd@CS.Berkeley.EDU>
To:        mjacob@feral.com
Cc:        freebsd-scsi@FreeBSD.ORG
Subject:   Re: SCSI Tape (HP SureStore T20) trouble
Message-ID:  <19990317191656.43093@helen.CS.Berkeley.EDU>
In-Reply-To: <Pine.LNX.4.04.9903170419340.17718-100000@feral-gw>; from Matthew Jacob on Wed, Mar 17, 1999 at 04:31:02AM -0800
References:  <19990316235547.14559@helen.CS.Berkeley.EDU> <Pine.LNX.4.04.9903170419340.17718-100000@feral-gw>

next in thread | previous in thread | raw e-mail | index | archive | help
Quoting Matthew Jacob (mjacob@feral.com):
> > 
> > (sa0:ncr0:0:5:0): SPACE. CDB: 11 1 ff ff ff 0 
> > (sa0:ncr0:0:5:0): ILLEGAL REQUEST asc:2c,0
> > (sa0:ncr0:0:5:0): Command sequence error
> > (sa0:ncr0:0:5:0): unable to backspace over one of double filemarks at EOD- opting for safety
> > 
> > and it rewinds the tape even though I opened the nrsa0 device (which
> > supposedly does NOT rewind).  I am able to restore from the dump, but
> > it seems broken that it does not rewind.  Also, the drive says it supports
> (I'm sure you mean "It is broken that it rewinds")
> > hardware compression, but 'mt status' says it is unsupported.  Does 
> > anyone know whats the trouble with either of these problems?
> 
> There are actually three problems here:
> 
> Problem #1: The driver apparently has a problem backspacing over a filemark.
> This indicates to me that this should be treated as if it were a QIC drive
> that only has one filemark at EOD. Frankly, I would have loved to enforce
> only one filemark at EOD for all except 1/2" reel drives, but folks shot that
> down months ago.
> 
> Fix: Quirk the drive to have the SA_QUIRK_1FM.

How does one go about that?

> Problem #2: When this failed, the driver tried to eject the tape.
> Well, that doesn't work for a lot of drives. This has been discussed on
> this list quite a bit recently. 
> 
> Fix: I proposed a change that will not have the drive rewind, but will
> disallow further writes until either a rewind, eject, or a space to EOD
> is performed. This fix is in progress. If I didn't have a 100 degree
> fever, it'd probably be done some time in the next day or so.

Cool.

> Problem #3: Compression is being reported as the driver tried to use some
> 'standard' means to enable compression. Unfortunately, the method chosen
> is for a mode page that's in SCSI-3, not for SCSI-2.
> 
> Fix: use the Device Configuration page first and then try the Date
> Compression page). However, this may not always guarantee compression if
> compression is enabled via some drive specific mechanism, e.g. a density
> code. A manual for this device may have this information. Passing that
> information back to someone to fold into the driver will be appreciated.

The manual has very little information.  Only the following, much of which
doesn't seem of interest to you.  The entire "Tape Format" section
from the appendix:

Interface: SCSI-2

Tape Format:

Format: QIC-3220
Media: HP 20GB/Travan TR-5 (QIC-3220)
Number of tracks: 108
Bit density: 106400 bpi
Encoding method RLL1,7
Error Correction: Level-10 Reed-Solomon

So I dug for information and found some information at:

	http://www.qic.org/html/qicstan.html

There is a complete specification for the QIC-3220 tape format.  It
says that it uses the ADLC-1 compression algorithm.  In a seperate
document at the same site, QIC-154 discusses the compression algorithm
and QIC-157 is the latest standard available that discusses the SCSI
command set (dated 1995).

-josh


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-scsi" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990317191656.43093>