SUMMARY: Problem with odd blocksizes on 9-track tape drives

From: Ken Rossman (
Date: Mon Jul 08 1991 - 18:06:29 CDT

Just as an FYI addendum, for those who have the same or similar setup as I
do (namely 9-track tape drives with Xylogics 472 tape controllers):

Our original problem was that the Xylogics was apparently writing nulls at
the end of odd size tape blocks (overwriting the last character in that
block), and my question related to whether the kernel driver, the
controller, or the drive was doing this, and whether it could be fixed.

The answer is that it is (was) a Xylogics controller firmware problem, and
that it has already been fixed in the current rev. In fact, as things
turned out, we had one bad drive (old rev) and one good drive, and I had
just never tested this problem on drive #1!!!

Anyway, the short story is, if you have these controllers and you want to
avoid this problem yourself, make sure that you have PROM revision level
"957H" or later. Anything that is "957G" or earlier will have this problem
with odd size tape blocks. I've verified this is the case by having Sun
check our controllers and them telling us that indeed one is in rev and the
other out (I then tested both drives to verify).

The PROM you want to check is in the upper righthand corner area of the
board (with the bus pins pointing down, external cable connector edge up),
at about position E-10.

Ken Rossman
Computer Center
Columbia University ken@cumin.bitnet
212-854-4876 uunet!columbia!ken

"You can't teach a new dog Ultrix"

This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:21 CDT