Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Fdformat questions.

0 views
Skip to first unread message

John Henders

unread,
Jul 28, 1993, 8:25:36 AM7/28/93
to

In /etc/ there is a file called fdprm, which seems to have settings
for accessing extended floppy formats. I can't figure out how to get
them to work, though, and grepping faqs for fdformat doesn't yeild much
either.
Can I format floppys at >standard format, like 10x80 for a 720k
disk? How?
Also, can anyone tell me how to get 30 character filenames from the
minix mkfs?


--
John Henders GO/MU/E d* -p+ c+++ l++ t- m--- s/++ g+ w+++ -x+

Helmut Geyer

unread,
Jul 28, 1993, 9:44:24 AM7/28/93
to
In article <1993Jul28....@jonh.wimsey.bc.ca>, jhen...@jonh.wimsey.bc.ca (John Henders) writes:
|>
|> In /etc/ there is a file called fdprm, which seems to have settings
|> for accessing extended floppy formats. I can't figure out how to get
|> them to work, though, and grepping faqs for fdformat doesn't yeild much
|> either.
|> Can I format floppys at >standard format, like 10x80 for a 720k
|> disk? How?
|> Also, can anyone tell me how to get 30 character filenames from the
|> minix mkfs?
|>

You have to apply a patch to the kernel to get non-standard floppy formats.
This patch is on sunsite in kernel/misc-patches/fdpatch.tar.Z. You have to
fiddle a bit with it to get it working, because it is against an older kernel
version. If you can't get it right, I can mail you an exact patch for 0.99pl11.
(I do not have it at this site, otherwise I would do that now).
But I do not use setfdprm, I always use non-standard formats with the exact
device names.

|>
|> --
|> John Henders GO/MU/E d* -p+ c+++ l++ t- m--- s/++ g+ w+++ -x+

Helmut Geyer ge...@kalliope.iwr.uni-heidelberg.de

Kerry Hoath

unread,
Aug 5, 1993, 12:50:30 PM8/5/93
to
In article <1993Jul29.1...@embl-heidelberg.de> ho...@embl-heidelberg.de writes:
>>|> In /etc/ there is a file called fdprm, which seems to have settings
>
>> You have to apply a patch to the kernel to get non-standard floppy formats.
>[...]

>> But I do not use setfdprm, I always use non-standard formats with the exact
>> device names.
>
>And there is your problem: you should use setfdprm. If you do, you
>do not need to change anything to the stock-Linus-supplied-kernel. I
>happily use 820kB floppies for some unimportant archiving.
>
How do you format the 820k disks then? Do you do that under dos?
I'd like to use single-stepped 800k 5.25 inch disks but can not format/read
them under linux.
TYFYA, Kerry.
--
ke...@gotss1.apana.org.au
ke...@scs.apana.org.au
ke...@lsupoz.apana.org.au

John Henders

unread,
Aug 5, 1993, 5:06:38 PM8/5/93
to
ke...@gotss1.apana.org.au (Kerry Hoath) writes:
>>
>>And there is your problem: you should use setfdprm. If you do, you
>>do not need to change anything to the stock-Linus-supplied-kernel. I
>>happily use 820kB floppies for some unimportant archiving.
>>
>How do you format the 820k disks then? Do you do that under dos?
>I'd like to use single-stepped 800k 5.25 inch disks but can not format/read
>them under linux.

Formatting is not the problem, making a filesystem on the disk
afterwards is, however.
From the man page for setfdprm.

Without any options, setfdprm loads the device (usually
/dev/fd0 or /dev/fd1) with a new parameter set with the
name entry found in /etc/fdprm (usually named 360/360,
etc.). These parameters stay in effect until the media is
changed.

However, after formatting, doing mkfs /dev/fd1 <number of blocks>
with number of blocks set to the previously formatted non-standard
format just gives me errors and a whole bunch of floppy controller
resets printed to the console.
Fdformat, I discovered by accident, will also happily attempt to
format a low density 3.5 floppy to 14400. I don't know if it actually
writes the format to the floppy (doesn't the lack of a HD hoole stop the
drive mechanism from actually writing to the floppy, like the write
protect tab?), but it certainly will not take a filesystem afterwards.
I have the kernel patches, but they are for 99pl10, which I haven't
installed, so when I do, I'll report whether this fixes the problem, if
there's any interest.

Jason Neudorf

unread,
Aug 6, 1993, 2:38:55 PM8/6/93
to
Jason Neudorf (neu...@cs.usask.ca) wrote:

(absolutely nothing, quoting the last article--I'm sorry, grovel, grovel,
please don't flame. . .)

I managed to apply the patch to .99p11 by adding a typecast or two.

This allowed me to format AND mke2fs a couple of disks at higher than
normal densities (1.44meg on 5.25", 1.68meg on 3.5"), but I'm not sure how
reliable this is. When I tried to use more than 80 tracks I got errors.

Maybe I should have used setfdprm instead, but I wasn't sure about the
parameters (I lost the man page).

ERIK VASAASEN

unread,
Aug 7, 1993, 12:08:04 PM8/7/93
to
In article <23u8fv$l...@tribune.usask.ca> neu...@cs.usask.ca (Jason Neudorf) writes:
>From: neu...@cs.usask.ca (Jason Neudorf)
>Subject: Re: Fdformat questions.
>Date: 6 Aug 1993 18:38:55 GMT

>This allowed me to format AND mke2fs a couple of disks at higher than
>normal densities (1.44meg on 5.25", 1.68meg on 3.5"), but I'm not sure how
>reliable this is. When I tried to use more than 80 tracks I got errors.

I'm sitting here right now and formatting 3.5"DD disks as 829440 bytes
disks (1620 sectors, 10 sectors a track, makes 81 tracks I guess). The
number of tracks you can put on the disk are dependent on the drive, you can
sometimes get 83 (On _some_ machines this works well..), but if you want to
be able to read them on an _amiga_, you have to use only 81 tracks or so. (I
was amazed when I found out that PC disks formatted with 81 tracks, more
sectors a track, and smarter interleaving (gains some 20% in speed) could be
read under Amiga Dos 3.0.)

It would be a really good idea if someone found a rawrite like utility that
enabled you to write 1.44 mb to 5.25 disks, as this would make distribution
a hole lot simpler for people with only 5.25 drives.

BTW: I'm using fdfroat version 1.8, and I think it's availabe as fdform18.
lzh or something like that. (Under ms-loss. sigh)

I've been using this program for a couple of years, and it isn't less
reliable than the speed / size gain.. Problem is: Don't lend it to someone
who will think something is wrong (you have to run a TSR under DOS) and run
fixdisk.

Erik

______
/ __ \ _|| Erik Vasaasen | email: er...@dhhalden.no
| / \ |_/ oo Enerbakken 9 | (finger er...@odin.dhhalden.no to see if
|\____/ / _\_) 1768 Halden | I'm in, or er...@fenris.dhhalden.no
\______/__/ Norway | to check on unix....)
----------------------------------/
...."If there isn't a silicon heaven, where does all the calculators go?...

Risto Kankkunen

unread,
Aug 7, 1993, 6:43:10 PM8/7/93
to
> However, after formatting, doing mkfs /dev/fd1 <number of blocks>
>with number of blocks set to the previously formatted non-standard
>format just gives me errors and a whole bunch of floppy controller
>resets printed to the console.

For some reason, fdformat resets the disk parameters set by setfdparm,
even if you had used the -p (permanent) switch. Just run setfdparm again
with the right params before making the fs.

--
Other people like awk too, although in these days of kitchen sink
languages with chainsaw syntax and dishwasher semantics [see perl],
it sure looks light-weight. --Ozan S. Yigit

Henri Faber

unread,
Aug 9, 1993, 4:56:40 AM8/9/93
to
In article <241b5u$n...@klaava.Helsinki.FI> kank...@klaava.Helsinki.FI (Risto Kankkunen) writes:
> However, after formatting, doing mkfs /dev/fd1 <number of blocks>
>with number of blocks set to the previously formatted non-standard
>format just gives me errors and a whole bunch of floppy controller
>resets printed to the console.

For some reason, fdformat resets the disk parameters set by setfdparm,
even if you had used the -p (permanent) switch. Just run setfdparm again
with the right params before making the fs.

It even resets the disk parameters between the formatting of the disk and the
checking for bad sectors, so using fdformat without the -n switch will result
in errors with a non-standard disk geometry.

I have formatted some disks at a capcity of 1722K, but the result is extremely
slow. But when I do the low-level format the same disks at the same capacity
with the DOS based fdformat, disk access is much faster. The DOS based
fdformat interleaves the sectors on the disk, so that an entire track can be
read within two rotations of the disk. When a disk without interleaving is
read, the drive just misses the next sector so that the disk has to spin 21
times to read a track.

Now for my question: Is it possible to tell fdformat or setfdprm to interleave
a disk?


Henri Faber
hen...@es.ele.tue.nl

Heiko Schroeder

unread,
Aug 9, 1993, 9:16:08 AM8/9/93
to
hen...@python.es.ele.tue.nl (Henri Faber) writes:

>Now for my question: Is it possible to tell fdformat or setfdprm to interleave
>a disk?

The fdpatch I wrote includes a change to the format routine which addresses
this problem. (BTW, this is the patch that was mentioned by some people
before. It was written for an older version. I'll post a version which will
patch cleanly into pl11 as sone as I have the time (i.e. the exams are over))

CU
Heiko

--
Email: | Snail-mail: Heiko Schroeder
| Lerchenweg 120
he...@pool.informatik.rwth-aachen.de | 52223 Stolberg
| GERMANY

0 new messages