18 Jul
2014
18 Jul
'14
9:27 p.m.
I'm looking for a program that runs under DOS 3.1 and will tell me CHS for
all unreadable hard disk sectors. Anybody know where one is I can download?
I can't find search keywords that filters out the modern stuff. A read
write test would also be useful.
18 Jul
18 Jul
11:51 p.m.
On Fri, 18 Jul 2014, David Gesswein wrote:
I'm looking for a program that runs under DOS
3.1 and will tell me CHS for
all unreadable hard disk sectors. Anybody know where one is I can download?
I can't find search keywords that filters out the modern stuff. A read
write test would also be useful.
Depends on the type of drive.
Are you talking about "MFM" (ST506/412), or are you talking about
SATA/PATA/USB, multi-gig, multi-terabytes, etc?
PC-DOS/MS-DOS 3.10 (three point ten) had a limit of 32M for a regular hard
disk.
BUT, it was the first version with the network redirector, and it was
possible (albeit hassle) to access larger drives through MSCDEX or
equivalent. DOS 3.31 was the first one to support hard drives larger than
32M.
Back in those days, the usual programs for testing hard drives, and
finding bad tracks were SpeedStor and SpinRite. DO NOT use old versions
of SpinRite with a [default] setting of "return to service anything that
tests good" - if the manufacturer of the drive says that a given track
can't be trusted, then I don't think that a single successful test by the
program over-rides that!
--
Grumpy Ol' Fred cisin at xenosoft.com
19 Jul
19 Jul
8 a.m.
On Fri, Jul 18, 2014 at 08:51:55PM -0700, Fred Cisin wrote:
On Fri, 18 Jul 2014, David Gesswein wrote:
MFM. Machine is a Compaq portable plus.
I'm looking for a program that runs under
DOS 3.1 and will tell me
CHS for all unreadable hard disk sectors.
Anybody know where one is I can download? I can't find search keywords
that filters out the modern stuff. A read write test would also be useful.
Depends on the type of drive.
Are you talking about "MFM" (ST506/412), or are you talking about
SATA/PATA/USB, multi-gig, multi-terabytes, etc?
Back in those days, the usual programs for testing
hard drives, and
finding bad tracks were SpeedStor and SpinRite. DO NOT use old versions
of SpinRite with a [default] setting of "return to service anything that
tests good" - if the manufacturer of the drive says that a given track
can't be trusted, then I don't think that a single successful test by the
program over-rides that!
SpinRite quick surface scan is just telling me which were flagged bad but
not which currently can't be read. The full test may but if it doesn't
the information I'm looking for will be changed.
I'll try speedstor.
10:41 a.m.
On 07/18/2014 10:51 PM, Fred Cisin wrote:
PC-DOS/MS-DOS 3.10 (three point ten) had a limit of
32M for a regular hard
disk.
I think the disk could be larger, could it not (although practical maximum
for the consumer drive technology of the era was probably around 120MB)? It
was the initial FAT16 implementation that topped out at 32MB, so a drive
would have to be partitioned (and I don't recall what the limit there would
be - although I think the CHS addressing scheme topped out at 8GB,
presumably limits on the number of partitions possible would come into play
first).
I don't think there was anything in 3.10 which stopped it from addressing a
partition beyond the 32M boundary, but maybe I'm wrong. I'd assume that the
drive test program would work at the drive level, not at the partition level.
FAT16b came in with DOS 3.31, and I think its individual partition size was
then capped at 2GB (at least assuming 512-byte sector sizes).
cheers
Jules
20 Jul
20 Jul
7:19 p.m.
> PC-DOS/MS-DOS 3.10 (three point ten) had a limit
of 32M for a regular hard
> disk.
On 07/18/2014 10:51 PM, Fred Cisin wrote:
I think the disk could be larger, could it not (although practical maximum
for the consumer drive technology of the era was probably around 120MB)? It
was the initial FAT16 implementation that topped out at 32MB, so a drive
would have to be partitioned (and I don't recall what the limit there would
be - although I think the CHS addressing scheme topped out at 8GB,
presumably limits on the number of partitions possible would come into play
first).
Good catch; thank you.
Yes, the limit for a logical DOS Drive in 3.10 was 32M, but the physical
disk drive could be substantially larger. It was not at all uncommon to
use a 40M drive, split into 2 (or more) DOS drives.
I don't think there was anything in 3.10 which
stopped it from addressing a
partition beyond the 32M boundary, but maybe I'm wrong. I'd assume that the
drive test program would work at the drive level, not at the partition level.
You are absolutely right.
FAT16b came in with DOS 3.31, and I think its
individual partition size was
then capped at 2GB (at least assuming 512-byte sector sizes).
and almost all of the 2G limits should have been 4G, since they were a
limit of 32 bits. But, they ended up at 2G instead of 4G due to
programmers who never expected to come near such enormous limits and used
a signed, instead of an unsigned number. Because of that, most of DOS
was willing to accept the existence of a NEGATIVE 2G file or drive!
(Note: copying a negative sized file to an almost full drive fails to
increase the free space.)
21 Jul
21 Jul
8:54 a.m.
On 07/20/2014 06:19 PM, Fred Cisin wrote:
I can't remember now, did 3.10 support logical drives? I have a feeling it
perhaps didn't, so the limit (from a DOS point of view) would have been
four primary partitions of 32MB each.
Interesting... I didn't realize that it was possible to go beyond the 2GB
limit and for things to still work - I always assumed that normal file ops
would break at that boundary, even if a >2GB filesystem was hand-crafted.
>
PC-DOS/MS-DOS 3.10 (three point ten) had a limit of 32M for a regular hard
> disk.
On 07/18/2014 10:51 PM, Fred Cisin wrote:
I think the disk could be larger, could it not (although practical maximum
for the consumer drive technology of the era was probably around 120MB)? It
was the initial FAT16 implementation that topped out at 32MB, so a drive
would have to be partitioned (and I don't recall what the limit there would
be - although I think the CHS addressing scheme topped out at 8GB,
presumably limits on the number of partitions possible would come into play
first).
Good catch; thank you.
Yes, the limit for a logical DOS Drive in 3.10 was 32M, but the physical
disk drive could be substantially larger. It was not at all uncommon to
use a 40M drive, split into 2 (or more) DOS drives. FAT16b came in
with DOS 3.31, and I think its individual partition size was
then capped at 2GB (at least assuming 512-byte sector sizes).
and almost all of the 2G limits should have been 4G, since they were a
limit of 32 bits. But, they ended up at 2G instead of 4G due to
programmers who never expected to come near such enormous limits and used
a signed, instead of an unsigned number. Because of that, most of DOS
was willing to accept the existence of a NEGATIVE 2G file or drive! 
10:09 a.m.
On 21 July 2014 14:54, Jules Richardson <jules.richardson99 at gmail.com> wrote:
I can't remember now, did 3.10 support logical
drives? I have a feeling it
perhaps didn't, so the limit (from a DOS point of view) would have been four
primary partitions of 32MB each.
I *think* it went:
3.0: 1 hard disk
3.1: 2 hard disks
3.2: 2 partitions per drive, max 32MB
3.3: 1 primary + multiple logical partitions per drive, max 32MB
Compaq 3.31: FAT16 partitions up to 2GB (adopted in PC DOS 4.0 and MS-DOS 4.0)
MS-DOS 7.01 (Windows 95OSR2) - FAT32 partitions
--
Liam Proven ? Profile: http://lproven.livejournal.com/profile
Email: lproven at cix.co.uk ? GMail/G+/Twitter/Flickr/Facebook: lproven
MSN: lproven at hotmail.com ? Skype/AIM/Yahoo/LinkedIn: liamproven
Cell/Mobiles: +44 7939-087884 (UK) ? +420 702 829 053 (?R)
12:36 p.m.
On Mon, 21 Jul 2014, Liam Proven wrote:
I *think* it went:
3.0: 1 hard disk
With my current error rate (unrefreshed dynamic wetware RAM), I am
hesitant to say that anybody is "wrong", but the XT/5160 controller very
definitely supported 2 hard^H^H^H^H fixed drives, and I often used 2 hard
drives on XT; and I think that that was with 2.00, 2.10, 2.11, 3.00.
BTW, using other than 10M called for soldering jumpers (or a 4 position
DIP switch). There were two bits for each drive, and I think that they
may have been 5M, 10M, 16M, 26M.
5170 stored har drive parameters in CMOS, which is another thread.
I heard that the controllers could be modified to support more than one
controller, but I never tried.
3.1: 2 hard disks
3.10 added support for the
"Network Redirector", and enabling MSCDEX to
present a CD-ROM (>650M!) as if it were a resource on a network.
Trying to CHKDSK a CD-ROM: "YOU CAN NOT CHKDSK A NETWORK DRIVE"
I heard that the same technique could be used for big hard drives, with a
lotta work, but I never tried.
LASTDRIVE was needed if you wanted to go past 5? drives (4 floppies, 2
hard drives, and a CD-ROM)
3.2: 2 partitions per drive, max 32MB
also 720K
floppies (also present in some OEM versions of 2.11)
Early 5170s needed DRIVER.SYS or DRIVPARM for 720K
DRIVPARM was present in MS-DOS 3.20, and present but undocumented in
PC-DOS 3.20. BUT, my real OEM 5170 BIOS's would give "UNRECOGNIZED
CONFIG.SYS COMMAND" when using either MS-DOS OR PC-DOS DRIVPARM, but
aftermarket BIOS worked with either! That never made sense to me.
Are you sure that multiple partitions weren't present in prior versions?
Are you sure that it wasn't FOUR?
3.3: 1 primary + multiple logical partitions per
drive, max 32MB
Compaq 3.31: FAT16 partitions up to 2GB (adopted in PC DOS 4.0 and MS-DOS 4.0)
While probably the most POPULAR, Compaq was not unique for 3.31.
23 Jul
23 Jul
7:09 p.m.
On 21 Jul 2014, at 16:09, Liam Proven <lproven at gmail.com> wrote:
[...]
I *think* it went:
3.0: 1 hard disk
3.1: 2 hard disks
3.2: 2 partitions per drive, max 32MB
3.3: 1 primary + multiple logical partitions per drive, max 32MB
Compaq 3.31: FAT16 partitions up to 2GB (adopted in PC DOS 4.0 and MS-DOS 4.0)
MS-DOS 7.01 (Windows 95OSR2) - FAT32 partitions
The Wikipedia entry on FAT and related topics are very extensively OCD on such topics even
by the usual Wikipedia standards. I recommend them as you'll probably find them a good
read.
I waded through most of it as part of my research into writing filesystem drivers, but
haven't yet used the information in anger. So I can only confirm that it is definitely
extensive but not that it is definitely correct :)
21 Jul
21 Jul
11:34 a.m.
On Mon, 21 Jul 2014, Jules Richardson wrote:
Interesting... I didn't realize that it was
possible to go beyond the 2GB
limit and for things to still work - I always assumed that normal file ops
would break at that boundary, even if a >2GB filesystem was hand-crafted.
In general, with few exceptions, you can not go bast 2G, because MOST of
the programs also used a 32 bit SIGNED number. The 2G limit applies not
only to drives, but also to the FILE SIZE in FAT16.
However, in NT4?, FAT16 CAN go up to 4G (4G minus one byte for
nit-pickers). That can, of course result in nasty incompatabilities
between NT4 and other products for any drives (or files) between 2G and
4G.
To waste a little time: step on the DIRectory sectors of a floppy. Take
the last 4 bytes of one of the FPDEs and change it to FF FF FF FF, and see
which versions of DOS and programs say that the file size is -1 byte.
Change another one to 80 00 00 00 to get -2147483648. Copying a file of
negative size to a disk fails to increase the free space.
BTW, I am using "logical" "drive" to refer to MS-DOS's
presentation of a
partition as if it were a disk drive. I am not referring to the various
types of drives and partitions. Thus, I am referring to splitting a 4096
into two 32M and one 16M as creating three "logical drives". I don't
remember when Microsoft began support for durther divisions, and
again altered the terminology.
4:35 a.m.
On Fri, 18 Jul 2014, Fred Cisin wrote:
PC-DOS/MS-DOS 3.10 (three point ten) had a limit of
32M for a regular hard
disk.
You're wrong, this was the FAT partition size limit, not the disk size
limit. You could well have a much larger drive in your system and
partition it accordingly, as long as the disk BIOS could address
everything properly.
Christian
11:16 a.m.
> PC-DOS/MS-DOS 3.10 (three point ten) had a limit
of 32M for a regular hard
> disk.
On Mon, 21 Jul 2014, Christian Corti wrote:
You're wrong, this was the FAT partition size
limit, not the disk size
limit. You could well have a much larger drive in your system and
partition it accordingly, as long as the disk BIOS could address
everything properly.
You're right. As I said in th4e previous retraction, the 32M limit is for
a DOS [before 3.31] logival "drive". That limit is NOT a limit to the
size of the physical drive, and it was quite common to use larger drives,
such as 40M or 80M, broken up into multiple DOS logical drives.
3616
days inactive
3620
days old
11 comments
6 participants
participants (6)
-
abuse＠cabal.org.uk -
cc＠informatik.uni-stuttgart.de -
cisin＠xenosoft.com -
djg＠pdp8online.com -
jules.richardson99＠gmail.com -
lproven＠gmail.com