[comp.sys.amiga.tech] Kronos tape drive

crash@ckctpa.UUCP (Frank J. Edwards) (04/08/90)

Thanks, Larry, for the info about your tape driver (I *thought* you were
the one :-), and I think you're right -- that info may be useless for
the Kronos package.

I have located, however, some header files that deal with using
SCSI-Link.library, which is a run-time library interface to the controller
card.  Now the $64,000 question:

Does anyone have the "SCSI-Link.lib" linker library, i.e. stub routines,
for using this library?  The SCSIDos3.0 docs say to "contact CLtd"...
Well, I think we know what the result of that will be! :-)  Did anyone
else get those files before CLtd closed up?  (Or, does anyone know what
their *new* name will be [smirk] ;-)

Larry:  the CLtd docs imply that QuarterBack can be used with a regular
SCSI tape drive; just configure the tape drive as a harddisk.  Since
QB talks directly to the .device it will access the tape in a sequential
manner (is there any other way?)...  Any opinions? comments?

Thanks all!
-----
Frank J. Edwards		ComputerKnowledge Corp
2677 Arjay Court		12740 Hillcrest, Suite 212
Palm Harbor, FL  34684-4505	Dallas, TX  75230
Phone:  (813) 786-3675		(214) 385-9700 / (800) 227-9700

lphillips@lpami.wimsey.bc.ca (Larry Phillips) (04/09/90)

In <90.04.08.15.44.14.537@ckctpa.UUCP>, crash@ckctpa.UUCP (Frank J. Edwards) writes:
>
>Larry:  the CLtd docs imply that QuarterBack can be used with a regular
>SCSI tape drive; just configure the tape drive as a harddisk.  Since
>QB talks directly to the .device it will access the tape in a sequential
>manner (is there any other way?)...  Any opinions? comments?

An early C Ltd controller had documentation with it that told how you could use
a tape as a destination for a DiskCopy. The idea was to configure both the
source disk and the destination tape, via mountlist entries, as disks having
the same geometry. I never did try it, since I only had the C Ltd controller in
the machine for a very short time.

How you go about making tape work is going to depend entirely on the driver
implementation for the C Ltd. controller. If it supports SCSIDirect in the
manner of scsidisk.h, then that's probably the way to go. Otherwise, you will
have to hope that their library has the hooks for sending commands to generic
SCSI devices. I think it probably does, snce that same driver could be used for
a printer, if I remember correctly.

In more general terms, once you figure out how you can communicate with the
tape, you have a number of ways you can approach the problem. The quickest, and
probably the easiest way is to write individual programs that use the tape. The
advantage here is that you don't need to learn how to write and install
handlers, libraries, and so on. The disadvantage is that only your program(s)
will be able to access the tape.

A more general solution would be to write a handler that implements a file
system specifically tailored to tape. It would allow you to mount the tape as a
device, and to write to or read from tape, albeit with the limitations
inherent to a serial medium. The handler will have to recognize Amigados
packets and in turn communicate with the driver itself. The handler could
implement a file system for general copying, and could include extra goodies to
make the tape into a 'raw' device for such things as custom backup formats,
tar, etc.

I haven't looked into it deeply at all yet, so I don't know what probolems you
might run into. For example, there may be a need for packets that are not
standard Amigados packets, in which case, any operations requiring those
packets would be accessable only to custom-written programs that knew how to
build the right packets.

For further information of writing a handler, I recommend Transactor for the
Amiga, Vol 2, issues 2, 3, and 4, in which a couple of 'Software Distillery'
folks show how it's done.

Good luck.

-larry

--
Entomology bugs me.
+-----------------------------------------------------------------------+ 
|   //   Larry Phillips                                                 |
| \X/    lphillips@lpami.wimsey.bc.ca -or- uunet!van-bc!lpami!lphillips |
|        COMPUSERVE: 76703,4322  -or-  76703.4322@compuserve.com        |
+-----------------------------------------------------------------------+

new@udel.EDU (Darren New) (04/10/90)

>>Larry:  the CLtd docs imply that QuarterBack can be used with a regular
>>SCSI tape drive; just configure the tape drive as a harddisk.  Since
>>QB talks directly to the .device it will access the tape in a sequential
>>manner (is there any other way?)...  Any opinions? comments?

Well, my comment is that it doesn't work.  I've tried it and Diskcopy won't
hack it.  I've written my own program to read the first block from the
tape drive and I always get error 20. This program reads the hard disk
without any problem.  Soooo..... I suspect that the tape stuff is not
up to snuff unless there is somebody out there who REALLY DOES have it
working.  Maybe it's my tape drive, but I can't tell without somebody
having gotten it working.          -- Darren

P.S., Is there anybody at UDel or nearby with a copy of quarterback I
can borrow?    -- Darren

bbaker@cbmvax.commodore.com (Bob Baker) (04/11/90)

I tried using an Archive tape with the C Ltd interface and drivers
several years ago with no success. I even shipped my drive to them
in order for them to try and provide support, again with no success.
The bottom line at that time was that they seemed incapable of 
understanding soft errors and retries on tape drives and their software
would always abort when either was encountered. 

For their own tape driver, they played games in software to keep the 
tape streaming... like writing four copies of each block on tape to
keep the tape from stopping (thus cutting capacity to 25%).

I've worked with tape drivers and related code before and it shouldn't
be all that hard to implement something on the Amiga. In my case, we 
were implementing a very simple file system that would allow booting
the operating system and/or diagnostic facilities from tape. We simply
added a header block at the front of each file that contained the
filename and other required info (load address, length, etc.) and 
wrote a filemake at the end of each file. Searching for a file simple 
meant reading the header block and comparing the filemane. If not 
correct, perform an forward space file to position at the next file. 
Very simple but effective and easy to implement. 

I'd really love to get my Archive working on the Amiga. If anyone has
code already started they need help with let me know.

-Bob