richter@randvax.UUCP (Susan Richter) (10/01/85)
------ ------ ------ ------ ------ ------ ------ ------ ------ ------
I have an AT&T 7300, and have been having some problems using the floppy.
There are two main problems, one with formatting the disk and the other
with doing file transfers between my system and others (IBM people, you
may want to skip to problem 2).
1) I formatted a diskette using the ua Floppy window stuff, and then
used "iv -t" to see what it thought the format was. OK, it says 40
cylinders, single density. So I replaced "singledensity" with
"doubledensity" in the description (/usr/lib/iv/FDnl, I think).
It didn't like that, so I just took it out altogether, and was rewarded
with a "doubledensity" formatted floppy (according to iv).
Unfortunately, the numbers of blocks, tracks, etc, remained exactly
the same, as did the amount of data that would fit on it (I thought iv
might have just been fooling, so I tried a backup that normally takes
about 1.3 diskettes. It still did. . .).
Then I decided to follow the *example given in the iv man page* (!)
and set cylinders=80. When I formatted the disk this time, there were
a lot of grinding, crunchy noises, but no errors were reported.
When I did "iv -t", voila! 80 cylinders, twice as many
blocks, sectors, etc. I did my 1.3-diskette cpio backup and (it
crunched and ground, but) it all fit!! "1000 blocks saved."
Uh-oh! When I tried to "cpio -i" it, I got this ominous message
about "can't open device /dev/rfp021". Iv NOW says it can't read
the Volume Header Block (VHB).
The Hotline guy said that the floppy drive only does 40 cylinders,
so when I formatted it with 80, it just rewrote the last cylinder
40 extra times. (Question: why did iv report all the info OK, if
that's the case?) Also, he said that even when iv says a diskette is
singledensity, it's really doubledensity (does this make sense to
anyone else? Why would they even bother having that keyword?).
QUESTION: does anybody know the real scoop on the floppy format?
If so, couldja please let me in on it?
2) Armed with a standard "doubledensity" 40-cylinder diskette, I have been
trying to use cpio to read/write files between my machine and an IBM
PC/XT running PC/IX. Neither system can read (cpio) the floppy written
by the other one (which I guess is not all that unexpected), but the
IBM can at least open the device and do raw I/O (like, dd) to dump
the contents of the disk into a file. When I stick an IBM-written
diskette in my machine, I get "can't open device /dev/rfp021", no matter
*what* I try to do (cpio, dd, cat). What's going on? How come the
IBM, which apparently writes its VHB or equivalent somewhere else, can
still function in its absence, while the 7300 clams up?
Has anyone else tried to do this? Did you succeed? Would writing
an alternate device driver for the floppy drive help, or is the
VHB location really set in concrete (please forgive my ignorance; I
don't do hardware)? Or am I making this harder than it really is? . .
If you can help or have had similar problems, please mail to me; if
there's lots of response, I'll bring the topic back to the net for
more discussion. Thanks lots!
- Susan
richter@rand-unix.arpa
..!decvax!ihnp4!trwrb!randvax!richter