Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.os.os2.setup.storage > #64 > unrolled thread

Multiple "Visible" Primaries

Back to article view | Back to comp.os.os2.setup.storage

Contents

  Multiple "Visible" Primaries Felix Miata <UgaddaBkidding.due2UCE@dev.nul> - 2012-11-26 15:22 -0500
    Re: Multiple "Visible" Primaries Paul Ratcliffe <abuse@orac12.clara34.co56.uk78> - 2012-11-26 21:30 +0000
      Re: Multiple "Visible" Primaries Felix Miata <UgaddaBkidding.due2UCE@dev.nul> - 2012-11-26 23:16 -0500
        Re: Multiple "Visible" Primaries "ivan" <ivanjt@free.fr> - 2012-11-27 16:19 +0000
          Re: Multiple "Visible" Primaries Felix Miata <UgaddaBkidding.due2UCE@dev.nul> - 2012-11-27 13:53 -0500
            Re: Multiple "Visible" Primaries "Peter J. Seymour" <Newsgroups@pjsey.demon.co.uk> - 2012-11-28 09:51 +0000
              Re: Multiple "Visible" Primaries Barry Landy <bl10nospam@cam.ac.uk> - 2012-11-28 10:41 +0000
              Re: Multiple "Visible" Primaries Paul Ratcliffe <abuse@orac12.clara34.co56.uk78> - 2012-12-16 23:58 +0000
                Re: Multiple "Visible" Primaries "Peter J. Seymour" <Newsgroups@pjsey.demon.co.uk> - 2012-12-17 08:18 +0000
          Re: Multiple "Visible" Primaries Paul Ratcliffe <abuse@orac12.clara34.co56.uk78> - 2012-12-16 23:55 +0000
        Re: Multiple "Visible" Primaries Paul Ratcliffe <abuse@orac12.clara34.co56.uk78> - 2012-12-16 23:53 +0000

#64 — Multiple "Visible" Primaries

It has recently been brought to my attention that under Win98 (and its DOS 
7.1 version) and later Windows versions, and possibly as early as Win95, that 
multiple primary partitions lacking the hidden attribute are acceptable. For 
many years, more than a decade at least, maybe two, I have been under the 
impression, because of IBM BM behavior, and having read in newsgroups, email 
lists, OS/2 print media, and on the WWW, that this is not acceptable to DOS, 
and presumably to OS/2 and eCS.

I'm interested in knowing if members of the OS2/eCS community have been aware 
of this, for how long if so, and what if any ramifications this ability has 
had or may have on those who multiboot with Windows.

As proof to myself, I used DFSee to partition a 30G laptop HD as follows:

7331M 0x1C primary
6681M 0x1C primary
  251M 0x0C logical
3344M 0x83 logical
7003M 0x0C logical
4001M 0x0C primary/startable

I then booted a M$ DOS 7.1 floppy to:
1-FORMAT C: /S
2-FORMAT D:
3-FORMAT E:

after which I booted from C:, the last/4001M partition.

Next I changed the partitions with DFSee thus:
7331M 0x1C primary
6681M 0x0C primary/startable
  251M 0x0C logical
3344M 0x83 logical
7003M 0x0C logical
4001M 0x12 primary

Again I booted the same floppy and did FORMAT C: /S, after which I booted C:, 
the #2/6681M partition. After that I installed WinXP SP3 from CD onto 
#2/6681M C:. Windows assigned D: to #3/251M, E: to #5/7003M, and, as 
expected, nothing to #1/7331M, #3/251M or #6/4001M.

Next I changed the partitions with DFSee thus:
7331M 0x0C primary/startable
6681M 0x1C primary
  251M 0x0C logical
3344M 0x83 logical
7003M 0x0C logical
4001M 0x12 primary

Again I booted the same floppy and did FORMAT C: /S, after which I booted C:, 
the #1/7331M partition.

The real confirmation tests came next, where I alternately changed the active 
partition while having set both the #1/7331M & #2/6681M partitions to type 
0x0C, and booted successfully to both WinXP and DOS 7.1. As before, the 
active/boot partition became C:, the 0x0C logicals D: & E:, and the 
non-active 0x0C became F: (after moving the OM device from F: to R:). These 
assignments were confirmed by last setting #1/7331M active and booting the 
WinXP installation CD again, whereupon letters were assigned thus:

C: 7331M 0x0C primary/startable
F: 6681M 0x1C primary
D:  251M 0x0C logical
H: 3344M 0x83 logical
E: 7003M 0x0C logical
-- 4001M 0x12 primary

Now I'm contemplating an upgrade to eCS-latest, mostly to see how it would 
work out to replace the 0x12/4001M with eCS and use AiR-Boot instead of BM.
-- 
"The wise are known for their understanding, and pleasant
words are persuasive." Proverbs 16:21 (New Living Translation)

  Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!

Felix Miata  ***  http://fm.no-ip.com/

[toc] | [next] | [standalone]

#65

On Mon, 26 Nov 2012 15:22:27 -0500, Felix Miata
<UgaddaBkidding.due2UCE@dev.nul> wrote:

> It has recently been brought to my attention that under Win98 (and its DOS 
> 7.1 version) and later Windows versions, and possibly as early as Win95, that 
> multiple primary partitions lacking the hidden attribute are acceptable. For 
> many years, more than a decade at least, maybe two, I have been under the 
> impression, because of IBM BM behavior, and having read in newsgroups, email 
> lists, OS/2 print media, and on the WWW, that this is not acceptable to DOS, 
> and presumably to OS/2 and eCS.

Back in the old days (I started with DOS 3.0 in Jan. 1986, quickly followed
by DOS 3.1 to fix all the bugs), multiple primary partitions were ALL that
you had. The letters were assigned in disk order for disk 0 and then for
disk 1 (if you had such luxury).
I think extended partitions and logical drives came in with DOS 4.0
This was before OS/2 1.0 came out in Dec. 1987 (IIRC).

> I'm interested in knowing if members of the OS2/eCS community have been aware 
> of this, for how long if so,

Er, about 25 years!

> and what if any ramifications this ability has had or may have on those
> who multiboot with Windows.

No idea. I gave up using multiple primaries when the drive letter
assignments were done differently depending on what OS you booted.
I forget the details now.

[toc] | [prev] | [next] | [standalone]

#66

On 2012-11-26 21:30 (GMT) Paul Ratcliffe composed:

> On Mon, 26 Nov 2012 15:22:27 -0500, Felix Miata wrote:

>> multiple primary partitions lacking the hidden attribute are acceptable.

> I think extended partitions and logical drives came in with DOS 4.0
> This was before OS/2 1.0 came out in Dec. 1987 (IIRC).

I just dug out 3.2 & 3.3 boot floppies. On system with a motherboard dating 
to around 1997 with a K6/2 CPU, and partitioning as follows:

  251M 0x16 primary
  7.8M 0x01 primary/startable
   79M 0x83 primary
  384M 0x82 logical
5428M 0x83 logical

FDISK from 3.2 reports
  32M non-DOS
   1M FAT active
  10M non-DOS
741M non-DOS

FDISK from 3.3 reports
  32M non-DOS
   1M FAT active
  10M non-DOS
741M Ext DOS
No logicals drives defined.

Obviously these antiques don't understand LBA or even FAT16B, but 3.3 does 
recognize 0x05 as extended.

>> I'm interested in knowing if members of the OS2/eCS community have been aware
>> of this, for how long if so,

> Er, about 25 years!

Did you notice the phrase "lacking the hidden attribute" (key component of 
the subject)?
-- 
"The wise are known for their understanding, and pleasant
words are persuasive." Proverbs 16:21 (New Living Translation)

  Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!

Felix Miata  ***  http://fm.no-ip.com/

[toc] | [prev] | [next] | [standalone]

#67

On Tue, 27 Nov 2012 04:16:04 UTC, Felix Miata 
<UgaddaBkidding.due2UCE@dev.nul> wrote:

> On 2012-11-26 21:30 (GMT) Paul Ratcliffe composed:
> 
> > On Mon, 26 Nov 2012 15:22:27 -0500, Felix Miata wrote:
> 
> >> multiple primary partitions lacking the hidden attribute are acceptable.
> 
> > I think extended partitions and logical drives came in with DOS 4.0
> > This was before OS/2 1.0 came out in Dec. 1987 (IIRC).
> 
> I just dug out 3.2 & 3.3 boot floppies. On system with a motherboard dating 
> to around 1997 with a K6/2 CPU, and partitioning as follows:
> 
>   251M 0x16 primary
>   7.8M 0x01 primary/startable
>    79M 0x83 primary
>   384M 0x82 logical
> 5428M 0x83 logical
> 
> FDISK from 3.2 reports
>   32M non-DOS
>    1M FAT active
>   10M non-DOS
> 741M non-DOS
> 
> FDISK from 3.3 reports
>   32M non-DOS
>    1M FAT active
>   10M non-DOS
> 741M Ext DOS
> No logicals drives defined.
> 
> Obviously these antiques don't understand LBA or even FAT16B, but 3.3 does 
> recognize 0x05 as extended.
> 
> >> I'm interested in knowing if members of the OS2/eCS community have been aware
> >> of this, for how long if so,
> 
> > Er, about 25 years!
> 
> Did you notice the phrase "lacking the hidden attribute" (key component of 
> the subject)?

Like Paul I have known about this for about 25 years and, in fact, 
still have an old DOS computer with a 10 MB hard disk that has 4 non 
hidden primary partitions on it.

Where the idea of primary partitions having to be hidden originates I 
don't know.  It might come from trying to set up a multi boot system 
booting from a primary partition when each boot partition required the
same letter therefore you had to hide the one(s) you weren't booting 
from.  This is only applicable to windows systems because OS/2 quite 
happily boots from a logical partition.

ivan
-- 

[toc] | [prev] | [next] | [standalone]

#68

On 2012-11-27 16:19 (GMT) ivan composed:

> Where the idea of primary partitions having to be hidden originates I
> don't know.  It might come from trying to set up a multi boot system
> booting from a primary partition when each boot partition required the
> same letter therefore you had to hide the one(s) you weren't booting
> from.  This is only applicable to windows systems because OS/2 quite
> happily boots from a logical partition.

It's hard for me to imagine applicability only to Windows, as it is IBM's own 
Boot Manager that insists on hiding when more than one supported primary is 
present on the same HD. Newer boot managers provide configurability of 
whether and which to hide.
-- 
"The wise are known for their understanding, and pleasant
words are persuasive." Proverbs 16:21 (New Living Translation)

  Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!

Felix Miata  ***  http://fm.no-ip.com/

[toc] | [prev] | [next] | [standalone]

#69

On 2012-11-27 18:53, Felix Miata wrote:
> On 2012-11-27 16:19 (GMT) ivan composed:
>
>> Where the idea of primary partitions having to be hidden originates I
>> don't know.  It might come from trying to set up a multi boot system
>> booting from a primary partition when each boot partition required the
>> same letter therefore you had to hide the one(s) you weren't booting
>> from.  This is only applicable to windows systems because OS/2 quite
>> happily boots from a logical partition.
>
> It's hard for me to imagine applicability only to Windows, as it is
> IBM's own Boot Manager that insists on hiding when more than one
> supported primary is present on the same HD. Newer boot managers provide
> configurability of whether and which to hide.

Are you confusing primary partitions with the same drive letter with 
primary drives in general? Certainly on my machines, multiple primaries 
are not hidden except when for instance there are two 'C' drives and one 
of them is booted, then the other 'C' drive is marked hidden. It is 
hardly an issue. If the two OS's are instead installed on for instance 
'C' and 'D' drives, none of the primaries are ever hidden.

[toc] | [prev] | [next] | [standalone]

#70

On Wed, 28 Nov 2012, Peter J. Seymour wrote:

:>On 2012-11-27 18:53, Felix Miata wrote:
:>> On 2012-11-27 16:19 (GMT) ivan composed:
:>> 
:>> > Where the idea of primary partitions having to be hidden originates I
:>> > don't know.  It might come from trying to set up a multi boot system
:>> > booting from a primary partition when each boot partition required the
:>> > same letter therefore you had to hide the one(s) you weren't booting
:>> > from.  This is only applicable to windows systems because OS/2 quite
:>> > happily boots from a logical partition.
:>> 
:>> It's hard for me to imagine applicability only to Windows, as it is
:>> IBM's own Boot Manager that insists on hiding when more than one
:>> supported primary is present on the same HD. Newer boot managers provide
:>> configurability of whether and which to hide.
:>
:>Are you confusing primary partitions with the same drive letter with primary
:>drives in general? Certainly on my machines, multiple primaries are not hidden
:>except when for instance there are two 'C' drives and one of them is booted,
:>then the other 'C' drive is marked hidden. It is hardly an issue. If the two
:>OS's are instead installed on for instance 'C' and 'D' drives, none of the
:>primaries are ever hidden.
:>

and this of course is how things always were in the "old days" (before 
logical partitions).

-- 
Barry Landy                        Email: Remove nospam in from address
192, Gilbert Road, Cambridge CB4 3PB

[toc] | [prev] | [next] | [standalone]

#73

On Wed, 28 Nov 2012 09:51:52 +0000, Peter J. Seymour <Newsgroups@pjsey.demon.co.uk> wrote:

>>> Where the idea of primary partitions having to be hidden originates I
>>> don't know.  It might come from trying to set up a multi boot system
>>> booting from a primary partition when each boot partition required the
>>> same letter therefore you had to hide the one(s) you weren't booting
>>> from.  This is only applicable to windows systems because OS/2 quite
>>> happily boots from a logical partition.
>>
>> It's hard for me to imagine applicability only to Windows, as it is
>> IBM's own Boot Manager that insists on hiding when more than one
>> supported primary is present on the same HD. Newer boot managers provide
>> configurability of whether and which to hide.
>
> Are you confusing primary partitions with the same drive letter with 
> primary drives in general? Certainly on my machines, multiple primaries 
> are not hidden except when for instance there are two 'C' drives and one 
> of them is booted, then the other 'C' drive is marked hidden. It is 
> hardly an issue. If the two OS's are instead installed on for instance 
> 'C' and 'D' drives, none of the primaries are ever hidden.

Partitions don't (or didn't pre-LVM) have assigned drive letters.
You could not have two 'C' drives. It was a meaningless concept.

[toc] | [prev] | [next] | [standalone]

#74

On 2012-12-16 23:58, Paul Ratcliffe wrote:
> On Wed, 28 Nov 2012 09:51:52 +0000, Peter J. Seymour <Newsgroups@pjsey.demon.co.uk> wrote:
>
>>>> Where the idea of primary partitions having to be hidden originates I
>>>> don't know.  It might come from trying to set up a multi boot system
>>>> booting from a primary partition when each boot partition required the
>>>> same letter therefore you had to hide the one(s) you weren't booting
>>>> from.  This is only applicable to windows systems because OS/2 quite
>>>> happily boots from a logical partition.
>>>
>>> It's hard for me to imagine applicability only to Windows, as it is
>>> IBM's own Boot Manager that insists on hiding when more than one
>>> supported primary is present on the same HD. Newer boot managers provide
>>> configurability of whether and which to hide.
>>
>> Are you confusing primary partitions with the same drive letter with
>> primary drives in general? Certainly on my machines, multiple primaries
>> are not hidden except when for instance there are two 'C' drives and one
>> of them is booted, then the other 'C' drive is marked hidden. It is
>> hardly an issue. If the two OS's are instead installed on for instance
>> 'C' and 'D' drives, none of the primaries are ever hidden.
>
> Partitions don't (or didn't pre-LVM) have assigned drive letters.
> You could not have two 'C' drives. It was a meaningless concept.
>
Let's put it this way. Using the configuration in question, I boot up to 
say XP and the next time I come to BM the XP partition shows as C and 
the OS2 partition is hidden. If I then boot up instead to OS2, the next 
time I come to BM, the OS2 partition shows as C and the XP partition is 
hidden. So strictly speaking there is ever only one C partition, but 
because of the switching involving it is convenient to refer to the two 
of them as being the set of possible C partitions (The 2 'C' 
partitions). I suppose the same sort of scenario could occur with three 
or more partitions, but I never tried that.

[toc] | [prev] | [next] | [standalone]

#72

On 27 Nov 2012 16:19:56 GMT, ivan <ivanjt@free.fr> wrote:

> Like Paul I have known about this for about 25 years and, in fact, 
> still have an old DOS computer with a 10 MB hard disk that has 4 non 
> hidden primary partitions on it.
>
> Where the idea of primary partitions having to be hidden originates I 
> don't know.  It might come from trying to set up a multi boot system 
> booting from a primary partition when each boot partition required the
> same letter therefore you had to hide the one(s) you weren't booting 
> from.  This is only applicable to windows systems because OS/2 quite 
> happily boots from a logical partition.

Yes, it was another bit of brilliance from those geniuses in Redmond.

[toc] | [prev] | [next] | [standalone]

#71

On Mon, 26 Nov 2012 23:16:04 -0500, Felix Miata <UgaddaBkidding.due2UCE@dev.nul> wrote:

> On 2012-11-26 21:30 (GMT) Paul Ratcliffe composed:
>
>> On Mon, 26 Nov 2012 15:22:27 -0500, Felix Miata wrote:
>
>>> multiple primary partitions lacking the hidden attribute are acceptable.
>
>> I think extended partitions and logical drives came in with DOS 4.0
>> This was before OS/2 1.0 came out in Dec. 1987 (IIRC).
>
> I just dug out 3.2 & 3.3 boot floppies. On system with a motherboard dating 
> to around 1997 with a K6/2 CPU, and partitioning as follows:
>
>   251M 0x16 primary
>   7.8M 0x01 primary/startable
>    79M 0x83 primary
>   384M 0x82 logical
> 5428M 0x83 logical
>
> FDISK from 3.2 reports
>   32M non-DOS
>    1M FAT active
>   10M non-DOS
> 741M non-DOS
>
> FDISK from 3.3 reports
>   32M non-DOS
>    1M FAT active
>   10M non-DOS
> 741M Ext DOS
> No logicals drives defined.
>
> Obviously these antiques don't understand LBA or even FAT16B, but 3.3 does 
> recognize 0x05 as extended.

Fair enough. I meant 3.2 instead of 4.0 then. Been a long time.

> Did you notice the phrase "lacking the hidden attribute" (key component of 
> the subject)?

Of course I noticed. This hidden attribute wasn't even designed and
multiple primary partitions were recognised perfectly well.

[toc] | [prev] | [standalone]

Back to top | Article view | comp.os.os2.setup.storage

csiph-web