[comp.unix.xenix.sco] tape drive errors, is there a listing

danielw@wyn386.mi.org (Daniel Wynalda) (04/19/91)

I have a bunch of 60 MB tape drives in house.  They vary from Archive
to Wangtek to Caliper.  All of the drives are similar in design and take
the same type of tape cartridge.  I have a machine here that doesn't
seem to like my tape drives and I can't find anywhere that explains the
error messages I receive from the drivers.

All of the drives appear to have this characteristic on this machine:

If I TURN OFF the computer, and re-boot SCO Xenix 386 2.2.3 (the
machine is a 33Mhz 80386/33 from DTK), the tape drive will work fine --
for a while.  I've noticed I can cause the error in 33Mhz mode, but if I
turn off the cache (turbo mode), things appear to work normally again (after
a reset of the tape drive).

I get one of the following from the kernel on the console when the drive 
freezes up.  

--
Wed Apr 17 13:42:58
ct cmd error(2,40)

Wed Apr 17 13:43:38
ct cmd error(2,C0)

Wed Apr 17 13:52:32
ct cmd error(1,60)
--

I've looked everywhere in the manual and can't find any reference to
these error messages.  I'm sure they mean something to someone at SCO
but not to me....

Could someone explain where these come for and what it is I have a problem 
with?   I see the (1,60) error more than the rest and I imagine its
a speed related problem, but I'm not sure... I get a similar problem with
the Caliper drive on an 80386/25 occasionally -- that machine is running
Xenix 386 2.3.2.  


Any hints?



-- 
Daniel Wynalda       | (616) 866-1561 X22 Ham:N8KUD Net:danielw@wyn386.mi.org
Wynalda Litho Inc.   | 8221 Graphic Industrial Pk. | Rockford, MI  49341 

paulz@sco.COM (W. Paul Zola) (05/04/91)

In article <892@wyn386.mi.org> danielw@wyn386.mi.org (Daniel Wynalda) writes:
}I have a bunch of 60 MB tape drives in house.  They vary from Archive
}to Wangtek to Caliper.  All of the drives are similar in design and take
}the same type of tape cartridge.  I have a machine here that doesn't
}seem to like my tape drives and I can't find anywhere that explains the
}error messages I receive from the drivers.
}
[ questions about the meaning of ct cmd errors deleted ]

The following article from the SCO Information Tools database should help 
with your problem.   I'm posting because I think this may be of general
interest.

################## cut here ################ cut here ##################

Generic QIC-02 tape drive: "ct cmd error(%d, %b)".
        
RELEASE:    SCO XENIX System V Operating System Generic
            SCO UNIX System V/386 Operating System Generic

PROBLEM:    The message "ct cmd error (%d,%b)", where %d is 0, 1 or 2 
	    and %b is a a hexadecimal number, appears quite frequently 
	    when the tape drive is in use.

CAUSE:      The interrupt handler contains a loop which polls the tape, 
	    waiting for the status to say Ready or Exception.  If neither 
	    bit gets set in a reasonable amount of time, the loop runs out 
	    and the status read may not be accurate.  As machines have become 
	    faster, this time period may have become too short for some 
	    configurations, and hence the error messages.

SOLUTION:   The error codes %d and %b in the message "ct cmd error (%d, %b)" 
	    indicate the following, depending on the values of %d and %b:

	    when %d is:	 	it means the tape driver is:
	    ---------------------------------------------------------

		0		trying to sleep at interrupt time
		1		waiting for READY or EXEC has failed
		2		waiting for READY, active has failed		



	   when	%b is:		the command issued to controller is:
	   ----------------------------------------------------------

		0x20		position tape
		0x40		write data
		0x60		write file mark
		0x80		read data
		0xa0		read file mark
		0xc0		read status
		0xff		impossible command value

################## cut here ################ cut here ##################

I hope this helps you with your problem.

-
Paul Zola			Support Technical Lead, Operating System
				paulz@sco.COM 
Gotta tend the earth if you want a rose.  - Emily Saliers
    DISCLAIMER: I speak for myself, and not for SCO.


}
}-- 
}Daniel Wynalda       | (616) 866-1561 X22 Ham:N8KUD Net:danielw@wyn386.mi.org
}Wynalda Litho Inc.   | 8221 Graphic Industrial Pk. | Rockford, MI  49341 

mcs@fps.com (Mike Sanfratello) (05/10/91)

In article <16615@scorn.sco.COM> paulz@sco.COM (W. Paul Zola) writes:
>
>PROBLEM:    The message "ct cmd error (%d,%b)", where %d is 0, 1 or 2 
>	    and %b is a a hexadecimal number, appears quite frequently 
>	    when the tape drive is in use.
>
>CAUSE:      The interrupt handler contains a loop which polls the tape, 
>	    waiting for the status to say Ready or Exception.  If neither 
>	    bit gets set in a reasonable amount of time, the loop runs out 
>	    and the status read may not be accurate.  As machines have become 
>	    faster, this time period may have become too short for some 
>	    configurations, and hence the error messages.
                                                                            
This is very poor driver construction; blind spin loops should be outlawed!
I've seen the ct driver source, and I'm sure you'll agree with me that this
is a very poor example of a tape driver and should be redone.