Date: Tue, 5 Jan 2010 18:01:48 +1100 (EST) From: Ian Smith <smithi@nimnet.asn.au> To: Polytropon <freebsd@edvax.de> Cc: Jerry McAllister <jerrymc@msu.edu>, freebsd-questions@freebsd.org, Fbsd1 <fbsd1@a1poweruser.com> Subject: Re: re-write is this booting info correct? Message-ID: <20100105163423.L50666@sola.nimnet.asn.au> In-Reply-To: <20091230170958.f5014fef.freebsd@edvax.de> References: <20091228230618.7FB7310656F1@hub.freebsd.org> <20091229134453.G81420@sola.nimnet.asn.au> <20091230170958.f5014fef.freebsd@edvax.de>
next in thread | previous in thread | raw e-mail | index | archive | help
This message is in MIME format. The first part should be readable text,
while the remaining parts are likely unreadable without MIME-aware tools.
--0-214674698-1262674908=:50666
Content-Type: TEXT/PLAIN; charset=ISO-8859-1
Content-Transfer-Encoding: 8BIT
On Wed, 30 Dec 2009, Polytropon wrote:
> On Tue, 29 Dec 2009 16:29:56 +1100 (EST), Ian Smith <smithi@nimnet.asn.au> wrote:
> > In freebsd-questions Digest, Vol 291, Issue 3, Message: 1
> > On Mon, 28 Dec 2009 21:04:57 +0800 Fbsd1 <fbsd1@a1poweruser.com> wrote:
[..]
> > All of these, at least from DOS 3 (c. '86?) use the same MBR setup, a
> > maximum of 4 Primary Partitions, one (and only one) of which may be an
> > Extended DOS Partition, containing as many Logical Drives as you like;
> > they're formed as a linked list, though I never used past Drive J: with
> > OS/2 (HPFS). (I'm using caps here to refer to the DOS nomenclature)
>
> The number is de-facto limited to 26 maximum for ALL drive
> letters - keyword is LETTER: A up to Z. A: and B: are
> reserved for floppy disk drives, C: is the booting partition
> (usually a primary DOS partition), D: up to Z: can be:
> - other primary partitions
> - optical drives
> - fake drives refering to directories (SUBST command)
> - external drives (INTERLNK / INTERSVR commands)
> The order of the drives is somewhat arbitrary, so you
> can't always predict drive letter behaviour.
All true. Plus perhaps virtual drives provided by a Domain Controller
(eg Samba) pointing to various network resources users can access.
> > In all of these, you can't access more than one Primary Partition from
> > any DOS-based OS; if you wish to have drives D:, E:, F: (etc) then these
> > _must_ be in the single Extended Partition - so your statement above is
> > not correct in that respect.
>
> I'm not sure about this. It's long time ago, so my brain isn't
> up to date anymore. :-) When I try to remember, I have the
> idea in mind that it WAS possible to partition a drive with
> primary partitions (max. 4).
Oh you can partition it that way, but DOSes can only see one Primary
Partition (PP) at a time, the Active one, on any one disk; eg you could
have say DOS 6 and Win2k in separate PPs; booting either would call that
one its Drive C: and any other PPs are then not visible to that OS.
FreeBSD of course can mount any of the Primary or Extended Partitions as
slices, as can Linux AFAIK, so this is really just a DOS/win limitation,
rather than being any consequence of the MBR-based system itself.
> I'll check this - and I actually CAN, because I still have
> a DOS machine (6.22) running well; it's mostly used for
> programming mobile radios and for disk operations in a
> museal content (robotron resurrection). :-)
Goodo :) I think my ancient OS/2 tower is past booting these days.
I've since dusted off (which took a while :) my User's Guide to OS/2
Warp, which has very detailed info on all this. I was talking before
about single-disk systems, as was fbsd1. Strange things happen to what
any DOS-based OS sees if there are also Primary Partitions on HD#2 ..
DOS(etc) sees the PP marked active on HD#1 as C:, always, and DOS 3-6 at
least, and I suspect DOS 7 (win9x through XP) can only boot from HD #1.
Further, DOS <= 3.3 required that PP to be within the first 32MB, and
all to 6.x need the bootable PP to be within the first 1024 cylinders.
However, DOS allocates any active PP on the second disk as Drive D:, so
even if there is an Extended Partition on HD #1, its Drive Letters will
be allocated AFTER the D: drive on HD#2, as first E:, F: etc on HD#1
then any more on HD#2 as G: etc. This used to provide much 'fun' for
folks later adding another HD who had hardcoded links to other drives.
Partition Magic used to understand (and display) all these intricacies,
and gparted and friends likely do also.
> > > An alternate method is to allocate an “extended dos partition” and then
> > > sub-divide it into logical dos drives lettered C, D, E, F. One of these
> >
> > Not limited to F: as above (adding the DOS colon as Polytropon suggests)
>
> My suggestion comes from documentation where "C:" is preferred
> to "C" (in context of drive letters), like "The C: drive is
> the booting drive", or "On floppy A: you'll find no files".
Sure; when in Rome speak Latin, as it were. OK, Italian these days :)
> > I'm not sure about NT, but certainly DOS 3 to 7
> > cannot boot from other than drive C: - though DOS Drive C: need not be
> > the first physical disk partition, indeed there can be several, though
> > only the first one marked Active is called C: by DOS on any one boot.
>
> DOS doesn't provide a native means for boot selection, so
> this statement appears to be correct in relation to my
> memories.
I'm not sure if the DOSes that can multiboot (NT, W2k, XP) can do so
from another PP on HD#1 or not; certainly they have to start from C:
> > > Microsoft/Windows partition and the FreeBSD slice is where the operating
> > > system software is installed. Microsoft/Windows operating system creates
> > > default folders that share the space in the partition. The FreeBSD
> >
> > It's not clear what you mean here by 'folders that share the space'?
>
> It seems to refer to the fact that the functional separation
> in "Windows" is done through directories ("folders"), instead
> of partitions.
Oh, OK I guess, while that's also true of FreeBSD directories within a
partition within a slice, but it's more files than directories that are
sharing the space anyway; directories are generally very small files.
Fbsd1, I hope the more gory details have helped rather than confused :)
O&O, Ian
--0-214674698-1262674908=:50666--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20100105163423.L50666>