[Gllug] Win&Lin accessible filesystems

Pete Ryland pdr at pdr.cx
Wed Dec 12 11:19:29 UTC 2001


On Wed, Dec 12, 2001 at 10:32:32AM -0000, Paul Brazier wrote:
> > > I also have a partition under Windows which fdisk says is "Win Ext'd
> > > (LBA)" (type f) - can I mount this and what switch do I need (I've
> > > looked through "man mount" but can't find it).
> > 
> > I would not trust what the DOS/Windows fdisk told me. Windows 
> > 95 fdisk 
> > used to report NTFS as HPFS (OS/2's filesystem, which NTFS is 
> > based on) 
> > and refuse to show or delete non-FAT extended partitions.
> 
> This was from the linux version of fdisk, it says the Windows partitions
> are:
> HPFS/NTFS		7
> Win Ext'd (LBA)	f
> 
> > > What would be the best filesystem to use that works r/w 
> > with both OSs if
> > > I re-format the partition? VFAT?
> > 
> > Both Windows 2000 and Linux can safely read and write FAT32, and many 
> > more systems can do FAT 16 (Windows 95, NT4). Only problem is 
> > that FAT does not understand file ownership or permissions, so you not
> 
> > be able to use them. 
> 
> I don't really need file ownership / permissions so I think I'll go with
> creating a FAT32 partition in Windows. I'll look into whether the "Win
> Ext'd (LBA)" is really FAT32 in disguise or something.
> 
> VFAT partitions can only be created in Linux but are mountable by
> Windows, is this right?
> Do they offer any advantages over FAT32?

Don't go doing anything to the Extended partition!!

The short story:  It actually contains other drives!!  They will disappear
too if you remove the extended partition!!

ok, the long story is that the partition table is only capable of holding 4
entries.  So, since around the transisition of MSDOS 3 or so, which made us
reformat our drives anyway, MS introduced a scheme to allow as many
partitions as possible. [1] What they did was only ever to put two entries
into any partition table.  The second partition, however, didn't contain a
filesystem from the first sector, but was basically treated as a logical
disk and started with another partition table.  After that followed a
"logical" partition with a filesystem, and possibly another extended
partition.  So by this process of "chaining" partition tables, you can have
as many partitions as you want.

Anyway, by convention, fdisk authors (of all varieties) have seemed to
notate this chaining by calling (in our notation) /dev/hdx1 the "primary"
partition, and all others (/dev/hdx5 up) "logical" partitions of the
"extended" partition (/dev/hda2), despite in reality this is actually a
*chain* of extended partitions.

hth,
Pete

[1] Q. "Why go to all the trouble of this highly-concocted scheme when they
could have just changed the partition table format to allow more drives?!" I
hear you ask!  The answer is that the mbr format was created and used by the
BIOS manufacturers, not the OS, and some BIOSes wouldn't have handled such a
change!
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 240 bytes
Desc: not available
URL: <http://mailman.lug.org.uk/pipermail/gllug/attachments/20011212/fccd9e9d/attachment.pgp>


More information about the GLLUG mailing list