Mail Archives: opendos/2000/10/27/10:41:01
Hello Alain,
Your message is a little confusing. You seem to have inadvertently
interchanged primary and extended in your discussion in some areas. I will
use the ^^^ to indicate these in your original message and make my comments
after each portion.
----- Original Message -----
From: "Alain" <alainm AT pobox DOT com>
To: <opendos AT delorie DOT com>
Sent: Tuesday, September 26, 2000 5:33 PM
Subject: Re: DRDOS FDISK
> I have successfully added a DR-DOS partition to a MS-DOS 7.1 (win98)
> system, but I didn't use DR-DOS's fdisk ;-)
I do not have a copy of 98 to play with, but have used DRDOS sucessfully
with WIN 3.xx, 95, and ME. So 98 should be the same. I have used DRDOS FDISK
for all of these as well as NT 4.01. I USUALLY use Linux FDISK when
installing Linux, however the newer DRDOS 7.03 FDISK /X will now allow
partitioning many other file systems. I have used it with NTFS. It laso has
OS/2 and many, many other filesystem types. I have not yet used DRDOS FDISK
for Linux ext2 and swap, but see no reason why it should not work.
I am really baffled by all of the talk about DRDOS FDISK. The only thing any
FDISK does is make entries into the last 66 bytes of the very first sector
of track zero on a hdd. Actually the last 2 bytes are the signature bytes
that IBM uses which is 55 AA or AA 55 (depending on if you are reading as in
a hex editor or as the BIOS itself reads it. BIOS will read these two bytes
backwards.)
There are 4 entries of 16 bytes each. One byte indicates if this is the
active partition or not. One byte is for the filesystem type, i.e. DOS,
BIGDOS, NTFS, ext2, FAT 32, DOS extended, etc. Several bytes are used to
show the starting sector and length on partition. This is just simple hex
data entry.
The rest of this sector contains a small crude binary file which is read by
the BIOS and executed for reading the partition table and selecting the
sector to actually boot from. In fact this first sevtor could be any length
you choose, the BIOS will just read the first 512 bytes of whatever is
there. It can find this location simply by detecting the appropriate sync
bits needed to identify this sector on the drive. Most drives that are
manufactured today are mostly compatible with IBM PC and use a HARD SECTOR
FORMAT. However, it would appear that soft sector formatting could also be
used. I ran across a problem that a person was having with his HDD after he
sent it in to Seagate to have it referbished. They sent it back with some
weird sector size that was larger than 512 bytes per sector. This was the
format designed for that particular model number, but for some reason there
were some made with 512 byte sectors and Seagate seemed to be unaware of
this. After this person communicated with Seagate several times, it was
decided to send the drive back and they reformatted it for 512 byte sectors.
I do not see how they could have done that with a HARD sector format unless
they had to actually replace the platters in the drive.
>
> BTW, I lost the message with the information about what should be in
> the OEM signature to be MS-DOS compatible. Can someone please
> resend it to me? :)
What are you referring to here? Is this the OEM of the BOOT SECTOR? NOTE:
The BOOT SECTOR is a different sector than the partition table sector. I
think this is where people are very confused. The BOOT SECTOR is located at
the very first sector of a PARTITION. The Master Boot Record (MBR) is where
the Partition Table is located, not the BOOT RECORD/SECTOR of a partition.
For a hard drive that uses a DOS FDISK the MBR/Partition Table is located at
physical track 0 which is Cylinder 0, Head 0, Sector 0. Normally (without
using some utility to bypass bad tracks, etc.) the first partition will
start on track 1, which is Cylinder 0, Head 1, Sector 0. Note: Stupid
MS/Gates does not use the industry standard of Sector 0 as being the first
sector of a track, instead they use sector 1 as the first sector. This can
get somewhat comfusing when using some DISK EDITORs and switching between
physical and logical interegating of the drive.
What is really confusing in this thread of messages is that people are
talking about more than one thing. My original message that started this
thread was about problems that people claim to have with DRDOS FDISK. I was
not referring to the BOOT SECTOR. I think everything I have read thus far in
this thread about FDISK problems (a couple of exceptions) has been about
problems with the BOOT RECORD/SECTOR. However, I can appreciate this to some
extent, because DRDOS FDISK does format the drive when running FDISK, thus
it will put the DRDOS info in the BOOT RECORD when it does this. However,
you can use whatever format command you wish to change this and format the
partition(s). This is really two different topics and now I seem to
understand that it is not FDISK where the problem is occuring, but in the
format of the BOOT RECORD.
In retrospec of this paragraph that I am replying to now, I am at a loss as
to WHAT signature you want. The last two bytes of the MBR AND each BOOT
RECORD will have the two byte I metioned above for it's signature. This is
not an MS signature, but a requiremant originally demanded by IBM to
recognize various types disk formats. Originally IBM was going to be able to
read all types of formats and even use various formats for the IBM. i.e it
would be able to detect an IBM diskette from an APPLE diskette from a UNIX
diskette. However, many companies using various stupid copy protection
schemes, screamed about this capability becuase it would allow anyone to
copy thier diskettes. So instead of changing these stupid (easy to defeat)
to more sophisticated methods of copy protection, they wanted IBM to change
this. So now we have to have special programs to read these other formats
and some such as the APPLE DOS 3.2 and 3.3 were never developed. (This
really ticks me off.) (I have the Central Point Deluxe Option Board, which
can duplicate just about any type of diskette, including APPLE ][ DOS
3.2/3.3/ProDOS/Mac. However, it cannot duplicate Commordore C-64/128
diskettes because of a different modulation method used.)
> I found a problem in DR-DOS about the dive letters assigment. Let me
> describe it: if I have a drive (only one) with _two_ primary partitions
and
> an extended partition, MS-DOS uses the active partition as C:, the
> extended partition as D: and the other extended partition as E:. This is a
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This is what I found to be confusing. You talk about two primary and one
extended, then you talk about one primary and two extended.
Drive D: will be the other PRIMARY partition and E: and above will be
assigned as logical drives in the EXTENDED partition. MS DOS does the same
thing. NT does not. (At least through MS DOS 5.0, I have not used 6.0 and
later very much to check out the differences.
Primary partitions are assigned first for all drives containing primary
partitions. Then LOGICAL drives are assigned their letters.
Whatever drive is assigned to be the boot drive and whatever partition on
that drive is marked as ACTIVE for that drive will be assigned DRIVE C: This
can included a CDROM/RAM drive, a ZIP drive, Jazz drive, Syquest drive,
LS-120 drive, or any HDD using an IDE interface. With SCSI drives you have
an even greater selection to some degree. Either ID-0 or ID-1 can boot and
you can have several SCSI controllers in one system. However, if a system
has both IDE and SCSI (without a special BIOS) drives, then it will boot
from whatever IDE drive is the assigned boot drive. Ultimately what this
means is that whatever drive is assigned to location 80h will be the boot
device (except of course the floppy drive(s).)
> general rule: extra primary partitions go at the end. With DR-DOS, the
> first extended partition is C:, and the second is D:, leaving the extended
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> partition as E:. The biggest problem is that if you put DR-DOS in the
^^^^^^^^^^^^^^
Again, this is confusing. Here you are talking about three extended
partitions, which cannot be done. YOU MAY ONLY have ONE DOS EXTENDED
partition. You may have many logical drives in a DOS EXTENDED partition. I
suggest that maybe you should use the terms PRIMARY, LOGICAL and PHYSICAL to
designate the drives you are referring to.
> second partition it _cannot_ boot as it changes from C: to D: in the
middle
> of the boot process. So if you want to have more than one partition
> to select which is active, DR-DOS HAS TO BE in the first one. I don't
> know for sure, but this may also afect DR-DOS+Linux in the same drive...
This statement is also very confusing. So rather than disagree with what you
have stated, I will make a statement of my own which relates to this.
I have used multiple operating systems on a single drive and multiple
drives. I have used several different boot mangers. A few years ago I had a
small partition with Novell DOS 7 installed on the first partition of the
drive which was a primary partition. I has OS/2 Warp 3.0 installed on a
logical drive in the extended partition on this drive. I had a small
partition installed for Linux in a primary partition. I had the OS/2 boot
manager installed in a very small primary partition. All of these were on
one drive. I had my 256MB Syquest removable cartridge SCSI drive. I could
format these cartridges for FAT 16, OS/2 floppy, Linux ext2, Linux swap.
These cartridges could be partitioned into several partitions if needed. So
for addition programs and storage space I could select a cartridge formatted
and partitioned for whatever purpose I had for whichever OS I was running.
I did have some problems when switching (unloading and loading/ mounting and
unmounting) between DOS and OS/2 floppy formats and getting the system to
recognize the new cartridges. OS/2 did not like FAT 16 and DOS did not like
the OS/2 floppy format. As long as I did not change the cartridge with a
different format than what it contained when I booted I was okay. This was
only a problem with automatically seeing that a cartridge had been changed.
I could manually unmount and mount the drive and be okay.
At first I used the OS/2 boot manager to select the OS to boot. Later I
installed LILO Linux boot manager to boot whichever OS I wanted. LILO was
installed in the BOOT RECORD of the Linux ext2 partition. I could use the
DOS FDISK and select whichever primary partition I wanted to boot from, i.e.
MBR (to go directly into DOS if I had some other boot problems) OS/2 primary
BOOT SECTOR (not the extended partition/logical drive where OS/2 resided) or
the BOOT RECORD in the ext2 partition for LINUX. NOTE: DOS is incapable of
reading OS/2 HPFS, OS/2 boot manager partition, Linux ext2 or Linux swap
partitions. However, the BIOS can read the first Sector of any partition.
Thus I could boot from the MBR, OS/2 Boot Manager, or Linux LILO Boot
Manager simply by running FDISK and changing the active partition.
So I do not see any reason why DRDOS has to be on the first primary
partition. There may be some restrictions when using other OSes as to where
they can reside and where they must boot from. Also utilities like Stacker,
Double Drive, Double Space, etc. may have stupid requirments. Whichever
partition is the active partition will be drive C: This partition will be
assigned to location 80h and the BIOS will boot from 80h (or floppy.)
However, let's take a look at the entire boot process before I proceed. Well
I'll just start with the boot process for the drive/partition assigned to
80h instead of going through the whole nine yards. First the BOOT RECORD is
read. The Boot Record contains information about the partition, such as:
sector size, number of sides, sectors per track, cylinders, etc. It also
conatains a boot program called the boot loader. This boot loader program
will tell the BIOS what program(s) is to be loaded next. This may be DOS,
Linux, NT, OS/2, UNIX, Solaris, etc or a boot manager such as OS/2, LILO.
DRDOS LOADER, NT LOADER, etc. To make things a little easier to understand,
lets just assume we have one drive with a single primary partition and DOS
installed. For MS users the files names will be IO.SYS and MSDOS.SYS. For
DRDOS and PCDOS and others they will be IBMBIO.COM and IBMDOS.COM
respectively.
1. IBMBIO.COM will be loaded
2. IBMDOS.COM will be loaded
3. CONFIG.SYS will be read and appropriate modifications will be invoked.
4. COMMAND.COM will be loaded. (Note; the name and location of this file can
be modified in the CONFIG.SYS file.)
5. AUTOEXEC.BAT file will be run. (Note: the name and location of this file
can be modified in the CONFIG.SYS file.)
Now if drive designations are changed during the boot process, like Stacker,
SuperStore, Double Drive, Double Space, and others do, then you may have to
modify the COMSPEC line in the CONFIG.SYS file and make two copies of the
AUTOEXEC.BAT file and put on copy on each of the affected drives. So even if
there is some reason for drive letters to change when booting to DOS, the
fix is easy and is done in the above mentioned drive compression program. NT
does not seem to use any such scheme for it's drive compression, nor does
WIN 9.x FAT 16 compression.)
I have not played around with installing DRDOS on any other partition other
than the first primary of the first physical IDE drive since Novell DOS 7.0,
but see no reason why that should have changed. I have helped people install
DR/MS DOS on ZIP and LS-120 IDE drives and they could boot to DOS this way
instead of FDISKing and reformatting their HDD. Most manufactures ship their
systems with a single primary partition with the OS installed.(Really
stupid!) I can't blame people from not wanting to have to go through all of
this so they can have an independent DOS by totally wiping everything out on
their HDD and installing WINDOZE and everything from scratch. MS seems to
think that people should just run DOS MODE 7.x and leave it at that.
(Idiots!)
As far as your statement about Linux goes, I'll address that here.
In Linux, the primary partitions are assigned this way (assuming one IDE
drive in the system for now.)
hda1
hda2
hda3
hda4
Drives in the extended partition are assigned as follows:
hda5
hda6
hda7
hda8
hda9
.
,
,
hda24
SCSI would be the same as above except they would be sda instead of hda and
will end at sda16 intead of sda24.
The second IDE drive would be hdb instead of hda. etc.
One of the problems with DRDOS FDISK and I do not remember for certain which
versions did this and some MSDOS versions also do this and that is:
When you use Linux FDISK,CFDISK, etc. to partition your drive and later use
DOS FDISK and rewrite the partition table for any reason including changing
the active partition, it will play musical chairs with the position in which
Linux assigned the partitions. Thus hda2 may become hda3, etc. This does not
bother DOS itself, but NT and other OSes will also see these changes. DRDOS
7.03 FDISK has eliminated this problem by always assigning the DOS EXTENDED
partition to the last of the four partition entries in the MBR. I know that
DRDOS 7.01 and 7.02 (Open and Caldera DOS.) I do not recall if Novell DOS 7
did this. I did have it installed when I first started playing around with
Linux, but cannot remember if it did this. I know that some MSDOS versions
also did this, but do not recall which versions and I have not fooled around
with any versions later than 6.0.
Also for some strange reasons that I do not understand, DOS FDISK cannot
delete non standard partitions. DRDOS FDISK 7.03 can. I have seen some other
FDISK programs that can delete non standard DOS partitions. In the past, I
had a short little program that could be entered through DEBUG and wipe out
the partition table. Most people do not have a disk editor or peograms like
Disk Manager, etc. The first MS fiasco that I ran across about this was when
people tried to upgrade from MS DOS 3.x to MSDOS 5.0. This may also have
been true with some early DRDOSes, but do not remeber having any such
problem with my system, but I never did have any DRDOS 3.xx on my systems
and those that I used at work that did have DRDOS 3.41, I did not upgrade to
DRDOS 5.0. What happened is that people that had MSDOS 3.x installed could
not re-partition their drives with MSDOS 5.0 FDISK. They could not delete
the MSDOS 3.x partitions. The only way you could do this was by booting to
DOS 3.x from floppy and using it's FDISK and remove the partitions. Then
boot from MSDOS 5.0 floppy and running it's FDISK and partition the drive
that way. (Really, really stupid.) Of course you could use a Disk Editor or
a third party program such as Disk Manager. (I made a lot of money on this
fiasco!)
I am not certain how multiple drives with multiple primary partitions and
each with an extended partitions work with IDE drives. I can check out the
SCSI stuff sometime. I just installed another IDE drive in my bother's
computer from his old computer. It is a 120MB Conner drive and made it the
master and changed his WD 850MB drive to a slave. He is in the process of
backing up his files so that we can put DRDOS 7.03 and WINDOZE 3.1 on the
120MB and NT on the 850MB drive. I will experiment with multiple primary
partitions on each drive and a logical drive on each and see what happens
since I will be FDISKing and reformatting both drives anyway.
He currently has Novell DOS 7 on the 120 and DRDOS 7.03 on the 850 and
WINDOZE 3.1 on the 850. So I'll play with it some before I install NT. I
have to Install 95 first on the 120MB in order to install NT. For some
reason we can no longer install NT from WIN 3.1. I believe he did once a
long time ago and am certain he had Novell DOS 7.0 installed. Also the 7.03
that is currently on his system may be the upgrade from 7.02 instead of the
7.03 full release. I don't remember when the last time he trashed the system
and re-installed DOS, and not certain which 7.03 he now has on floppy. I
will make certain he has the current 7.03. There were some files that were
different in the upgrade and full versions. Some of the files were different
sizes. I do not recall which ones but some of them were amoung the system fi
les.
The NT version he has is the student version that came with the MSCE course
and does not have the standard files I need to make to NT boot and install
diskettes. I have to install it off a CD. I will use the WINDOZE 95a that
comes with those CDs and not use the 95B, that way I can use a lot less
space to install 95 on the 120MB drive and use the complete 850 for the NTFS
filesystem. He like to play those stupid old arcade games, like DOOM,
Wizardry, etc. He cannot get half of them to run in DOS and uses WINDOZE 3.1
for thos and a ton of them are WINDOZE only games. Maybe some or all of
these games will run under NT. Something will have to go if he wants to run
them all under DRDOS/WINDOZE 3.1. There will not be enough space. Since I do
not play these games, I do not have a clue as to what will work and what
will not work with NT.
In closing, I think you may be mistaken about how the primary partiton and
extended partitions are assigned their drive letters in DRDOS. Howver, since
your message was very confusing, I'll check DRDOS 7.03 on my brother's
computer before I install NT. I may even check out 7.05, but do not see why
there would be any difference as I would still be using the 7.03 FDISK. I
may also check out the Novell DOS 7.0 FDISK as well. If he will let me play
with it for several days, I may also check out MSDOS 6.22xxxx whatever it
was!
Also, Does anyone know what variation of MSDOS is used with the NT Command
Prompt? All I show when I run ver is WINNT 4.0.
Pat
__________________________________________________
Do You Yahoo!?
Talk to your friends online with Yahoo! Messenger.
http://im.yahoo.com
- Raw text -